VOLUM 3 ISSUE 2 - University of Iowa...Corresponding Author: Jessica E. Goetz, PhD; Phone: (319)...

VOLUM 39 ISSUE 2

Jocelyn T. Compton, MSc, MD Nathan R. Hendrickson, MD, MS

Brandon G. Wilkinson, MD

2019 Electronic Publication

Volume 39 Issue 2

BASIC SCIENCEDeep Learning for Chondrocyte Identification in Automated Histological Analysis of Articular Cartilage Linjun Yang, MS; Mitchell C. Coleman, PhD; Madeline R. Hines, BS; Paige N. Kluz, PhD; Jessica E. Goetz, PhD ...................................................................1The Effect of Pulsed Electromagnetic Field and Combined Magnetic Field Exposure Time on Healing of a Rabbit Tibial Osteotomy Douglas C. Fredericks, BS; Emily B. Petersen, DVM; Madison Rhodes, BS; Grace A. Larew, BS; James V. Nepola, MD ...................................................9 Electrospun PLGA and β-TCP (ReBOSSIS-85) in a Lapine Posterolateral Fusion Model J. Christopher Nepola, Emily B. Petersen, DVM; Nicole DeVries-Watson, PhD; Nicole Grosland, PhD; Douglas C. Fredericks, BS ..................................................................................................................................................................................................................................................................................... 19

SHOULDER AND SPORTSLow Serum Albumin Levels Are Associated with Increased 30-Day Cardiopulmonary Complications, Reoperation, and Readmission Rates Following Total Shoulder Arthroplasty Danny Lee, BS; Ryan Lee, MBA; Megan T. Cross, BS; Andrew Tran, MD; Jason Kappa, MD; Sam Moghtaderi, MD ..................................................... 27The Effect of Turf Toes Injuries on Player Performance in the National Football League Andre Tran, MD; Jason Kappa, MD; Evan Smith, MD; Micahel Hoy, BS; Jacob Farrar, BS; Avinash Chandran, PhD; Rajeev Pandarinath, MD......................................................................................................................................................................................................................................................................................... 35Outcomes of Surgical Management of Borderline Hip Dysplasia: A Systematic Review Cameron Barto,MD; Elizabeth Scott, MD; Zain M. Khazi, BS; Michael Willey, MD; Robert Westermann, MD ................................................................... 40

HANDDescriptive Epidemiology and Return to Sport after Hand Fractures in NCAA Athletes Christopher N. Carender, MD; Joseph A. Buckwalter V, MD, PhD; Natalie A. Glass, PhD; Robert W. Westermann, MD........................................... 49Variable-Angle Locking Compression Plate Fixation of Distal Radius Volar Rim Fractures Mengcun Chen, MD; Daniel J. Gittings, MD; Shuhua Yang, MD; Guohui Liu, MD, PhD; Tian Xia, MD, PhD ......................................................................... 55

SPINEEvaluation of Predictors and Outcomes of Bracing with Emphasis on the Immediate Effects of In-Brace Correction in Adolescent Idiopathic Scoliosis Tzu Chuan Yen, BS; Stuart L. Weinstein, MD ..................................................................................................................................................................................................................................... 62Impact of Surgery on the Quality of Life of Adolescent Idiopathic Scoliosis Pedro Fernandes, MD, PhD, FEBOT; Joaquim Soares do Brito, MD, FEBOT; Isabel Flores, PhD; Jacinto Monteiro, MD, PhD ............................ 66

TUMOR, FOOT AND ANKLE, AND OTHER TOPICSPVNS Masquerading as a Metastasis: Imaging Features and Biopsy Technique Ramanan Rajakulasingam, FRCR; Jennifer Murphy, FRCR; Idris Badreddine, MB, ChB; Steven James, FRCR; Rajesh Botchu, FRCR................................................................................................................................................................................................................................................................................................ 73Evaluation of Asymptomatic Contralateral Foot Deformities Using the Tripod Index Courtney Carlson, MD; Craig Akoh, MD; Chamnanni Rungprai, MD; Phinit Phisitkul, MD ................................................................................................................ 76Automated Mobile Phone Messaging Utilizing a Cognitive Behavioral Intervention: A Pilot Investigation Chris A. Anthony, MD; Edward O. Rojas, BS; Jill Kain, RN, Natalie Glass, PhD; Apurva S. Shah, MD, MBA; Tammy Smith, MS; Benjamin J. Miller, MD, MS .................................................................................................................................................................................................................................................................................. 85Oswestry Disability Index: Is Telephone Administration Valid? Christopher T. Martin, MD; Alexandra K. Yaszemski, MBA; Charles G. T. Ledonio, MD; Tara C. Barrack, PA-C; David W. Polly, Jr., MD ............................................................................................................................................................................................................................................................................................. 92

THE IOWA ORTHOPEDIC JOURNAL2019 Volume 39 Issue 2

EDITORS Jocelyn T. Compton, MSc, MD Nathan R. Hendrickson, MD, MS

STAFF ADVISERS J. Lawrence Marsh, MD Jose A. Morcuende, MD

Volume 39 Issue 2 1

AbstractBackground: Histology-based methods are com-

monly used in osteoarthritis (OA) research because they provide detailed information about cartilage health at the cellular and tissue level. Computer-based cartilage scoring systems have previously been developed using standard image analysis techniques to give more objective and reliable evaluations of OA severity. The goal of this work was to develop a deep learning-based method to segment chondrocytes from histological images of cartilage and validate the resulting method via comparison with human segmentation.

Methods: The U-Net approach was adapted for the task of chondrocyte segmentation. A training dataset consisting of 235 images and a validation set consisting of 25 images in which individual chondrocytes had been manually segmented, were used for training the U-Net. Chondrocyte count, detection accuracy, and boundary segmentation of the trained U-Net was evaluated by comparing its results with those of human observers.

Results: The U-Net chondrocyte counts were not significantly different (p = 0.361 in a paired t-test) than the algorithm trainer counts (Pearson correla-tion coefficient = 0.92). The five expert observers had good agreement on chondrocyte counts (intra-class correlation coefficient = 0.868), however the resulting U-Net counted a significantly fewer chondrocytes than the average of those expert ob-servers (p < 0.001 in a paired t-test). Chondrocytes were accurately detected by the U-Net (F1 scores

= 0.86, 0.90, with respect to the selected expert observer and algorithm trainer). Segmentation ac-curacy was also high (IOU = 0.828) relative to the algorithm trainer.

Conclusions: This work developed a method for chondrocyte segmentation from histological images of arthritic cartilage using a deep learning approach. The resulting method detected chon-drocytes and delineated them with high accuracy. The method will continue to be improved through expansion to detect more complex cellular features representative of OA such as cell cloning.

Clinical Relevance: The imaging tool developed in this work can be integrated into an automated cartilage health scoring system and helps provide a robust, objective and reliable assessment of OA severity in cartilage.

Keywords: u-net, deep learning, chondrocyte segmentation, cartilage, osteoarthritis

INTRODUCTIONOsteoarthritis (OA) is a joint disease associated with

painful degenerative changes in articular cartilage and the underlying subchondral bone. Patients typically pres-ent to physicians with end-stage OA after the cartilage has completely eroded and significant bony changes have taken place. Therefore, a large amount of OA-related research is focused on intervening earlier in the natural history of the disease, while cartilage is still present. De-tailed information about cartilage health at the cellular and tissue level for this research is often obtained via histology-based methods. The histological characteristics of early OA include loss of the integrity of the articular surface and loss of proteoglycan (PG) content, together which can result in the disruption of the extra cellular matrix (ECM). In this environment, chondrocytes initially become more active, replicating and producing more PG in an attempt to stabilize the ECM, but later dying within the compromised ECM.1 The severity of early OA is often characterized by scoring schemes based on these histological features.

One of the earliest and most widely implemented scor-ing schemes is the Mankin score, which independently evaluates articular surface integrity, cellularity, PG deple-tion and tidemark integrity, and then sums the sub-scores

DEEP LEARNING FOR CHONDROCYTE IDENTIFICATION IN AUTOMATED HISTOLOGICAL ANALYSIS OF ARTICULAR CARTILAGE

Linjun Yang, MS1,2; Mitchell C. Coleman, PhD1,3; Madeline R. Hines, BS1,3; Paige N. Kluz, PhD1; Marc J. Brouillette, PhD1; Jessica E. Goetz, PhD1.2

1Department of Orthopedics and Rehabilitation, University of Iowa, Iowa City, IA, USA2Department of Biomedical Engineering, University of Iowa, Iowa City, IA, USA3Department of Radiation Oncology, University of Iowa, Iowa City, IA, USACorresponding Author: Jessica E. Goetz, PhD; Phone: (319) 384-4275; Fax (319) 335-7530; Email: [email protected]: The authors report no potential conflicts of interestrelated to this study.Sources of Funding: This work was funded in part from grants from the NIH/NIAMS (AR055533 and AR070914) and the Department of Defense (W81XWH1810658).

L. Yang, M.C. Coleman, M.R. Hines, P.N. Kluz, M.J. Brouillette, J.E. Goetz

2 The Iowa Orthopedic Journal

of each feature to obtain a single value to represent OA severity.2 There have been a number of other similar scor-ing systems patterned on the Mankin score, including the OARSI, O’Driscoll, ICRS, and Knutsen scoring systems,3-5 and further species-specific adaptations of several of these individual scoring schemes. Many of these systems were developed to formalize the extent to which the tissue is affected by the severity of degeneration, and they utilize a “stage” and “grade” approach then multiply these numbers to obtain a degeneration “score”.6 Despite the variation in scoring schemes, a common feature to all histological scoring schemes is the necessity to assess cartilage cel-lularity and deviations from normal in order to obtain this OA severity score.

Conventional application of any scoring scheme and thus the assessment of cellularity is performed by an experienced individual observing the specimen under a microscope and subjectively assigning a score. This ap-proach usually results in high intra- and inter-observer variability.7 Therefore, more objective and reliable computer-based scoring systems are highly desirable. To develop such algorithms, methodologies to automati-cally detect chondrocytes, articular cartilage boundaries, etc. using image analysis techniques must be developed. Identifying an object in a histological image requires it to visually appear different from the surrounding structures so that programs can be developed to extract it from the surrounding image. There are many different techniques to achieve this identification, including thresholding, in which an object is identified by having a brightness or a color pattern that meets a specific cutoff value; edge detection using a variety of gradient-based filters to identify abrupt changes in color or brightness; watershed-based methods which separate each object based on local minimums on object boundaries;8 and blob detection which utilizes a shape-detection method to identify objects of interest.9 The challenge of using any of these techniques to identify chondrocytes in histological images of arthritic cartilage is that there is a huge variety in the appear-ance of cartilage associated with the disease as well as the histological processing and staining itself (Figure 1).

Modern artificial intelligence-type approaches have been brought to bear on challenging image analysis tasks similar to identifying individual cells within images. “Deep learning” is one such technique that has achieved enormous success in advanced image analysis applica-tions. The use of convolutional neural networks (CNN or ConvNets), a typical deep learning model designed to process image data, has performed excellently in several computer vision applications including object detection and semantic segmentation.10,11 In contrast to conventional segmentation methods, in which the computer is explic-itly instructed to execute a series of programmer-selected

steps to perform a segmentation task, a CNN model can “learn” to do the task, essentially developing an internal series of steps needed to perform the segmentation from a set of raw input images and the associated “correct” seg-mentations. To do this, the CNN iteratively outputs a seg-mentation image (prediction), which is computationally compared with the gold standard human segmentation for the image (ground truth). The model is programmed to learn from this comparison by optimizing the internal routine in-use to output a prediction that is progressively closer to the ground truth. This process is referred to as training. It has been shown that a CNN can not only learn to identify lower level image features like color and edges, but also higher level image features such as histograms of oriented gradients.12 This ability to identify complicated features in image representation allows the CNN to outperform traditional image processing methods. We hypothesized that an appropriately trained CNN model could achieve excellent chondrocyte segmentation accuracy. The goal of this work was to adapt and train the U-Net, a CNN model with providing excellent cell segmentation in other contexts,13 to perform chondrocyte segmentation and then to compare the automated results to the gold standard of human segmentations.

METHODSGeneral Approach and U-Net Architecture

The U-Net is a symmetrical CNN model consisting of a contracting path followed by an expansive path (Figure 2).13 Beginning at the contracting path, the histology

Figure 1. Common variations in chondrocyte appearance in normal and osteoarthritic cartilage. OA results in loss of red PG staining (upper right and lower left) and cloning (lower left). Thick or torn sections result in deformed or indistinct chondrocytes (lower right). High PG intensities can also partially obscure cells present only in the slide-side half of the tissue section (lower left). Composing a single analysis algorithm to address this level of variation has proven challenging.

Volume 39 Issue 2 3

Deep Learning for Chondrocyte Identification in Automated Histological Analysis of Articular Cartilage

A

image is convoluted using different learnable filters to extract different features (such as colors and edges) at each location in the image. This operation results in an intermediate, multi-channel image with each channel representing the detection of a certain feature using the associated learning filter. Next, the “max-pooling” layer down-samples the intermediate image with respect to its height and width in order to obtain the most repre-sentative feature combination in each local region. This process (convolution and max-pooling) is repeated to extract higher level features, e.g. different appearances of nuclei. At the deepest level, the intermediary image containing the highest level features enters the expansive path. It becomes up-sampled and concatenated with the corresponding intermediate image that was generated along the contracting path. The objective is to localize critical information from high-level features to a known location in the histology image space. The expansive path terminates with an image of the same size and resolution as the input histology image. A loss function is then used to quantify the difference between the model prediction segmentation image and the ground truth segmentation image. Therefore, selection of an appropriate loss function for comparison with the ground truth image will depend on the nature of the image being analyzed.

Training the algorithm also requires two different input datasets, both of which are selected to encompass the range of appearance of chondrocytes in the cartilage. Training is conducted by providing the algorithm with each pair of images/ground truth segmentations that are in the designated “training set”. The corresponding loss is then minimized over multiple iterations of the algorithm in which the convolution filters are modified. To prevent overfitting to the training set, after every training itera-tion (called an epoch), the model is required to predict on a fully separate set of images/ground truth segmentations, termed the validation set. This ensures that the model will remain accurate for yet unseen images. The model is saved when the lowest validation loss is achieved.

U-Net Adaption and Training for Chondrocyte Segmentation

In this study, the U-Net framework was adapted for analysis of 256 x 256 color (RGB) histology images. Each convolutional layer of the U-Net utilized 3 x 3 filters fol-lowed by a batch normalization layer and rectified linear unit activation function.14 A batch normalization layer was also added after the concatenation layer (bottom of the U) along the expansive path. At the final output layer, sigmoid activation was applied to result in an output im-age with values between 0 and 1, which represents the probability value range of a given pixel being a cell (1) or not in a cell (0). A binary cross entropy loss function was utilized to compute the probability difference between the U-Net prediction image and the ground truth segmenta-tion image.15 For each pixel, the difference between the prediction value and the ground truth value (loss) at that pixel location is computed, weighted, and normalized by the sum of weights over the entire image to form the final loss function to be minimized by optimization. The weight is taken from a weight map, which is a precomputed constant from the ground truth segmentation (Figure 3). Higher weight values force the associated feature to be learned with greater emphasis. In this work, the weight value for each pixel in the weight map was the inverse of the frequency of the pixel’s ground truth class (cell or not cell) appearing in the full ground truth map. This approach forces the model to preferentially learn cells. Similarly, the values of the pixels along the borders of very closely adjacent cells were weighted higher to im-prove border detection the border between very closely adjacent cells (Figure 3).

To train the algorithm, safranin-o/fast green-stained histological sections of rabbit cartilage with varying degrees of arthritic changes induced by ACL transec-tion16 and medial meniscus destabilization17 were used. Histological sections were digitized using a stage scan-ner microscope (Olympus VS110, Olympus America Inc., Center Valley, PA, USA) at a resolution of 322.25 nm/

Figure 2. U-Net architecture:13 the left side of the “U” shape represents the contracting path, while the right side represents the expansive path.



A B C

pixel. Training and validation image (256 x 256 pixels) sets of 235 and 25 images, respectively, were selected to encompass chondrocytes with different sizes, shapes, ap-pearances, and zones of origin. The chondrocytes in each image in both sets were manually segmented by a single individual leading algorithmic development (hereafter referred to as the “algorithm trainer”) using MATLAB (R2018b, The Mathworks, Natick MA). To increase the number of images in the training set in order to reduce model overfitting, data augmentation was utilized to increase the number of images in the training data set without requiring additional manual segmentation. The data augmentation included image rotation, mirroring, and brightness adjustment to reproduce the variability of cartilage histology in practice.

The U-Net model was implemented using the open source deep learning library Keras with TensorFlow backend (https://keras.io/). An NVIDIA Tesla K80 GPU was used for model training. The model was randomly initialized according to the work by Glorot and Bengio,18 and dropout layers were applied at the expansive path to further prevent overfitting.19 Due to the GPU memory limitation, a batch size of 1 pair of training images at a time was used. The model was trained for 60 epochs (train-ing iteration + validation check) using an Adam optimizer (learning rate = 5*10-4, momentum = 0.99).20

Model EvaluationThe trained U-Net was evaluated in terms of perfor-

mance in the following three related tasks: chondro-cyte counting, chondrocyte detection, and chondrocyte boundary segmentation. As the goal of this work was to improve upon manual segmentation and assessment of chondrocytes, data from the U-Net were compared with manual cell segmentations performed by five expert observers with experience identifying chondrocytes in histological, immunohistochemical and fluorescent confo-cal microscopic images. This comparison was performed on a totally separate set of 30 images (testing set). Five of the 30 images in the set were repeated to assess intra-observer variability in the segmentation process. Image selection, format and segmentation were performed as described above. Manual segmentation of this additional testing set was also performed by the algorithm trainer and compared with U-Net’s performance. Intra-class cor-relation coefficient (ICC) among the expert observers was calculated to quantify inter-observer variability in chondrocyte counting. An ICC value greater than 0.8 was considered good agreement between the observers.21 For each of the five observers, a Pearson’s correlation coeffi-cient was computed between the replicate segmentations of the repeated images in the data set. Similarly, Pearson’s correlation coefficients were computed between the U-Net predictions and the average of the expert observers and between the U-Net predictions and the algorithm trainer.

Chondrocyte detection accuracy was evaluated using an F1 score.22 Given a pair of segmentations (prediction and manual segmentation), the F1 score is calculated as the harmonic mean of precision (P) and recall (R) using the following equations:

where Ntp, Nfp, and Nfn denote the number of true posi-tives (TP), false positives (FP), and false negatives (FN), respectively.22 A U-Net predicted cell covering more than 40 percent of the manually segmented cell was considered a true positive; otherwise it was considered a false posi-tive. A false negative was defined as less than 40 percent of a manually segmented cell being intersected by a U-Net-predicted cell. Precision or recall values close to 1.0 indicate very few FP or FN, respectively. A F1 score of 1.0 thus represents perfect detection. The F1 score was computed relative to the algorithm trainer and relative to a representative expert observer.

To evaluate the accuracy of the U-Net chondrocyte segmentation, an intersection over union (IOU) metric (Figure 4) was computed between the U-Net segmenta-tions and the algorithm trainer segmentations.13 A perfect U-Net segmentation is identical to the ground truth seg-

Figure 3. Example histology image of chondrocytes (top left) with the associated ground truth segmentation (bottom left). The U-Net is trained using a series of these image/segmentation pairs. The segmentation mask is overlaid on the histology image on the top right. A weight map (bottom right) assigning loss weights for each pixel is precomputed from the ground truth segmentation. Warmer colors represent higher weight value assigned to the loss of the cor-responding pixel, which forces the U-Net to learn preferentially in those higher-weighted regions.

Volume 39 Issue 2 5


mentation, which gives an IOU value of 1.0. No overlap of cells between the two segmentations yields a value of zero. Object-wise IOU was calculated for each individual chondrocyte in each of the 35 test images, and a total IOU value was computed by summing the chondrocyte size-weighted IOU of all chondrocytes.

RESULTSU-Net training with the parameters specified above

required approximately 90 minutes. Both the training loss and validation loss started from infinity and decreased below 0.2 after 20 epochs. The model was saved when it achieved the lowest validation loss (0.152) (Figure 5).

There was good inter-observer agreement (ICC = 0.868) in cell counting among the five expert observers. Intra-rater agreement was excellent with Pearson’s correlation coefficients of 0.984, 0.979, 0.994, 0.922, and 0.974. Despite reasonable correlation between the U-Net and the expert observer chondrocyte counts (0.786, 0.771, 0.75, 0.85, and 0.858), the U-Net predictions of chondrocyte counts were on average 24% (±15%) lower than the average of the expert observer counts (Table 1). This difference was statistically significant (p < 0.001 in a paired t-test). In contrast, the Pearson’s correlation coefficient between the trained U-Net model and the algorithm trainer was higher (0.92), and there were no statistically significant differences in the number of chondrocytes counted (p = 0.361 in a paired t-test). Chondrocytes were accurately detected, with a majority of true positives and very few false detection (both FP and FN) (Figure 6). F1 scores (for accuracy of cell detection) of 0.86 and 0.90 were achieved for the selected expert observer and algorithm trainer, respectively. Segmentation accuracy relative to the al-gorithm trainer was also high (IOU = 0.828). Generally, chondrocytes with an elongated shape (as would be found

Figure 4. A schematic illustration of intersection over union (IOU) is shown on the left. IOU is calculated based on the overlapping (intersection) and the total (union) area of a given cell pair. On the right are two example segmentation comparisons overlaid on the original image yellow contours represent the manual segmentation by the algorithm trainer, and transparent blue mask represents U-Net segmentation. An overall IOU > 0.95 is achieved in these two examples, indicating excellent chondrocyte segmentation.

Figure 5. The output from the training process is plotted, illustrating the minimization of both training loss and validation loss. Training loss decreases steadily through iterations while validation loss gener-ally decreases in a fluctuating manner. The minimum validation loss occurred at the 44th epoch and the model was saved.

Figure 6. Chondrocyte detection by U-Net com-pared with the manual segmentations of two dif-ferent human observers. The images analyzed with the trained U-Net and evaluated by the human observers are shown in the top row. The resulting

U-Net segmentations are compared with one of the expert observ-ers (middle row) and with the algorithm trainer (bottom row). The middle and bottom rows are colorized manual segmentation maps with colors indicating detection accuracy.

Figure 7. An example of relatively poor performance of our trained U-Net is shown. Several chondrocytes from superficial zone are missed, as indicated by their blue color in the middle image. Seg-mentation accuracy was also reduced in the superficial zone as indicated with a lower IOU shown in the right image (zoomed in view of red boxed area).



Table 1. Cell Counts for Each Individual Observer and the Associated U-Net Counts

Expert Observer Expert Algorithm

1 2 3 4 5 Observer Average U-Net Trainer

Image 1 66 72 65 73 70 69 58 59

Image 2 21 25 22 22 23 23 18 20

Image 3 35 44 35 48 35 39 27 27

Image 4 45 41 40 48 44 44 41 37

Image 5 42 42 41 44 43 42 26 30

Image 6 28 42 22 29 27 30 27 22

Image 7 45 51 38 47 40 44 32 31

Image 8 57 60 54 59 64 59 56 53

Image 9 40 43 39 37 40 40 31 32

Image 10 67 70 65 37 68 61 40 46

Image 11 32 44 29 33 32 34 33 31

Image 12 51 48 46 39 39 45 23 26

Image 13 36 45 35 36 33 37 34 32

Image 14 34 34 29 30 28 31 19 25

Image 15 42 52 33 32 36 39 25 25

Image 16 41 45 37 45 43 42 36 36

Image 17 67 69 60 65 67 66 55 53

Image 18 63 55 51 60 59 58 53 47

Image 19 19 30 18 22 21 22 19 18

Image 20 76 75 72 71 68 72 50 52

Image 21 52 48 49 42 45 47 30 35

Image 22 26 40 26 25 26 29 24 21

Image 23 44 50 43 46 42 45 35 36

Image 24 54 56 53 46 49 52 37 46

Image 25 37 40 37 37 41 38 32 31

Image 26 35 55 30 35 37 38 32 23

Image 27 34 58 34 36 36 40 30 30

Image 28 56 58 48 58 49 54 48 43

Image 29 41 51 35 37 36 40 27 28

Image 30 42 45 43 35 38 41 12 25

Image 31 43 49 41 44 41 44 26 33

Image 32 27 46 25 27 28 31 27 22

Image 33 60 67 57 60 57 60 56 54

Image 34 95 102 93 87 83 92 58 73

Image 35 48 51 47 49 43 48 23 26

Volume 39 Issue 2 7


in superficial zone) were less accurately detected (Figure 7). The U-Net was also less accurate in identification of small cells, out of plane cells, or in torn cartilage with a “fuzzy” appearance.

DISCUSSIONHistopathological features encode important informa-

tion about cartilage pathology and are often interpreted by human observers to evaluate cartilage health. Com-puter-based, automated evaluation of cartilage has been limited by challenges associated with using standard im-age analysis routines on the wide variety of chondrocyte appearances in osteoarthritic cartilage.23 As cellularity (cell density, organization, etc), is a key feature of dif-ferent stages of OA severity, inaccurate cell counts limit the utility of some computational analysis algorithms. In this study, a state-of-the-art deep learning method to improve automated chondrocyte segmentation was developed and evaluated. A specific CNN architecture termed the U-Net was trained using normal and arthritic rabbit cartilage samples and compared to gold-standard chondrocyte segmentations. The trained U-Net model was able to automatically and objectively identify cells accurately compared to several human observers. Fur-thermore, the chondrocyte boundaries identified by the trained U-Net model were very similar to those traced by human observers, especially for chondrocytes that were not immediately adjacent to the articular surface.

In contrast to traditional image processing methods which require explicit computer programming to teach the computer to detect cells from low-level features, the U-Net can detect and segment cells by learning multiple high-level features such as cell components (cytoplasm and nucleus) and cellular organization. This approach as-sists with separating cells that are immediately adjacent to or even overlapping each other, while remaining robust to many artifacts of histological processing. The U-Net has previously been adapted to other cell segmentation tasks where different cell types are the objects of the interest with excellent results;24 however, to the best of our knowledge, this work is among the first to use the U-Net approach to identify chondrocytes with a goal of automatically and objectively assessing OA severity.

There are several limitations to this work that should be considered. First, the results of the correlation analysis indicated that our trained U-Net correlated better with the segmentations of the algorithm trainer than with those of the expert observers. This makes intuitive sense given that the U-Net was trained on the segmentation pattern of that algorithm trainer. Training the U-Net on a set of images representing expert consensus segmenta-tion may result in a slightly differently trained model that is able to more accurately replicate expert observer cell counts. The U-Net was trained and tested on rabbit

cartilage, and it is not yet clear if a single model will be accurate for evaluating cartilage of different species, or if separate algorithms will need to be trained for the variety species used to model OA. It is still challenging for the model to detect and segment chondrocytes that reside in the superficial zone. Some of those cells are very small, elongated in shape, and have homogeneous color patterns, which is a combination of features that in deeper zones is related to artifact or out of plane cells. Further training of the U-Net that includes more images of cells from the superficial zone with this appearance and more images of true non-cells from the deeper zones could improve performance of the U-Net prediction in a wider variety of cartilage locations. Finally, only a single probability threshold was utilized in this work, and it is possible that a systematic evaluation of threshold values could be performed to identify an optimal value for cell retention in the U-Net images.

In this project, state-of-the-art deep learning has applied a version of the U-Net method to detect and segment chon-drocytes from histologic images of arthritic cartilage. This is a first attempt to bring machine learning to the critical histological data analysis nodes within the field of orthopedic research. Future work will include improving the U-Net’s performance to agree with expert opinion, as well as expanding the application to detect more complex cellularity features representative of OA such as cell cloning and more complex histological artifacts such as those associated with fraying or torn tissue. The result-ing chondrocyte segmentation tool will be integrated into an automated Mankin scoring system23 and used more generally as a robust, objective technique for chondrocyte identification in image processing.

ACKNOWLEDGMENTThe authors wish to thank Dr. Stephen Baek from the

Department of Industrial Engineering at the University of Iowa for helpful discussions of deep learning method-ology.

REFERENCES1. Blanco, F.J., et al., Osteoarthritis chondrocytes die

by apoptosis. A possible pathway for osteoarthritis pathology. Arthritis Rheum, 1998. 41(2): p. 284-9.

2. Mankin, H., Biochemical and metabolic aspects of osteoarthritis. The Orthopedic clinics of North America, 1971. 2(1): p. 19-31.

3. O’Driscoll, S.W., F.W. Keeley, and R.B. Salter, The chondrogenic potential of free autogenous periosteal grafts for biological resurfacing of major full-thickness defects in joint surfaces under the influ-ence of continuous passive motion. An experimental investigation in the rabbit. J Bone Joint Surg Am, 1986. 68(7): p. 1017-35.



4. Mainil-Varlet, P., et al., Histological assessment of cartilage repair: a report by the Histology Endpoint Committee of the International Cartilage Repair So-ciety (ICRS). J Bone Joint Surg Am, 2003. 85-A Suppl 2: p. 45-57.

5. Knutsen, G., et al., Autologous chondrocyte im-plantation compared with microfracture in the knee. A randomized trial. J Bone Joint Surg Am, 2004. 86-a(3): p. 455-64.

6. Pritzker, K.P., et al., Osteoarthritis cartilage histopathology: grading and staging. Osteoarthritis Cartilage, 2006. 14(1): p. 13-29.

7. Pauli, C., et al., Comparison of cartilage histopathol-ogy assessment systems on human knee joints at all stages of osteoarthritis development. Osteoarthritis Cartilage, 2012. 20(6): p. 476-85.

8. Lin, G., et al., A hybrid 3D watershed algorithm incorporating gradient cues and object models for automatic segmentation of nuclei in confocal image stacks. Cytometry A, 2003. 56(1): p. 23-36.

9. Byun, J., et al., Automated tool for the detection of cell nuclei in digital microscopic images: application to retinal images. Mol Vis, 2006. 12: p. 949-60.

10. Redmon, J., et al. You only look once: Unified, real-time object detection. in Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.

11. Long, J., E. Shelhamer, and T. Darrell. Fully convolutional networks for semantic segmentation. in Proceedings of the IEEE conference on computer vision and pattern recognition. 2015.

12. Dalal, N. and B. Triggs. Histograms of oriented gradients for human detection. in Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on. 2005. IEEE.

13. Ronneberger, O., P. Fischer, and T. Brox. U-net: Convolutional networks for biomedical image segmen-tation. in International Conference on Medical image computing and computer-assisted intervention. 2015. Springer.

14. Ioffe, S. and C.J.a.p.a. Szegedy, Batch normaliza-tion: Accelerating deep network training by reducing internal covariate shift. 2015.

15. Goodfellow, I., et al., Deep learning. Vol. 1. 2016: MIT press Cambridge.

16. Tochigi, Y., et al., Instability dependency of osteo-arthritis development in a rabbit model of graded anterior cruciate ligament transection. J Bone Joint Surg Am, 2011. 93(7): p. 640-7.

17. Goetz, J.E., et al., Time-dependent loss of mito-chondrial function precedes progressive histologic cartilage degeneration in a rabbit meniscal destabili-zation model. J Orthop Res, 2017. 35(3): p. 590-599.

18. Glorot, X. and Y. Bengio. Understanding the diffi-culty of training deep feedforward neural networks. in Proceedings of the thirteenth international conference on artificial intelligence and statistics. 2010.

19. Srivastava, N., et al., Dropout: a simple way to pre-vent neural networks from overfitting. 2014. 15(1): p. 1929-1958.

20. Kingma, D.P. and J.J.a.p.a. Ba, Adam: A method for stochastic optimization. 2014.

21. Koo, T.K. and M.Y. Li, A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J Chiropr Med, 2016. 15(2): p. 155-63.

22. Chen, H., et al., DCAN: Deep contour-aware net-works for object instance segmentation from histology images. Med Image Anal, 2017. 36: p. 135-146.

23. Moussavi‐Harami, S.F., et al., Automated ob-jective scoring of histologically apparent cartilage degeneration using a custom image analysis program. 2009. 27(4): p. 522-528.

24. Zhang, M., et al., Image Segmentation and Clas-sification for Sickle Cell Disease using Deformable U-Net. 2017.

Volume 39 Issue 2 9

ABSTRACTBackground: Calcium phosphate materials have

been employed clinically as bone void fillers for several decades. These materials are most often provided in the form of small, porous granules that can be packed to fill the wide variety of size and shape of bony defects encountered. ReBOSSIS-85 (RB-85) is a synthetic bioresorbable bone void filler for the repair of bone defects with handling characteristics of glass wool-like (or cotton ball-like). The objective of this study is to evaluate the in vivo performance of RB-85 (test material), compared to a commercially available bone void filler, Mastergraft Putty (predicate material), when combined with bone marrow aspirate and iliac crest autograft, in an established posterolateral spine fusion rabbit model.

Methods: One hundred fifty skeletally mature rabbits had a single level posterolateral fusion per-formed. Rabbits were implanted with iliac crest bone graft (ICBG), Mastergraft Putty™ plus ICBG, or one of 4 masses of ReBOSSIS-85 (0.2, 0.3, 0.45, or 0.6 g) plus ICBG. Plain films were taken weekly until euthanasia. Following euthanasia at 4, 8, and 12 weeks, the lumbar spine were tested by manual palpation. Spinal columns in the 12 week group were also subjected to non-destructive flexibility testing. MicroCT and histology were performed on a subset of each implant group at each euthanasia period.

Results: Radiographic scoring of the fusion sites indicated a normal healing response in all test groups. Bilateral radiographic fusion rates for all test groups were 0% at 4 weeks; ICBG 43%, Mas-

tergraft Putty 50%, RB-85-0.2g 0%, RB-85-0.3g 13%, RB-85-0.45g 38%, and RB-85-0.6g 63% at 8 weeks; and ICBG 50%, Mastergraft Putty 50%, RB-85-0.2g 0%, RB-85-0.3g 25%, RB-85-0.45g 36%, and RB-85-0.6g 50% at 12 weeks.

Spine fusion was assessed by manual palpation of the treated motion segments. At 12 weeks, ICBG, MGP, and RB-85-0.6g were fused mechani-cally in at least 50% of the rabbits. All groups demonstrated significantly less range of motion in both flexion/extension, lateral bending, and axial rotation compared to normal unfused controls.

Histopathology analysis of the fusion masses, in all test groups, indicated an expected normal response of mild inflammation with macrophage and multinucleated giant cell response to the graft material at 4 weeks and resolving by 12 weeks. Regardless of test article, new bone formation and graft resorption increased from 4 to 12 weeks post-op.

Conclusions: This animal study has demon-strated the biocompatibility and normal healing features associated with the ReBOSSIS-85 bone graft (test material) when combined with autograft as an extender. ReBOSSIS-85 was more effective when a larger mass of test article was used in this study.

Clinical Relevance: ReBOSSIS-85 can be used as an extender negating the need for large amounts of local or iliac crest bone in posterolateral fusions.

Keywords: synthetic bone graft substitute, pos-terolateral fusion, animal model

INTRODUCTIONIliac crest autograft (ICBG) has long been consid-

ered the gold standard in achieving long-term spinal arthrodesis,1-11 despite the morbidity rate associated with graft harvest.12,13 In response, a number of commercially available synthetic bone graft extenders and autograft alternatives have been developed.14-20 Most common sub-stitutes and extenders are allografts, demineralized bone matrices, collagen matrices with bone morphogenetic proteins (BMPs), and ceramics.1 Ceramic bone grafts are porous and resorbable making them optimal for bone ingrowth.21-29 Ceramic bone grafts are osteoconductive

ELECTROSPUN PLGA AND β-TCP (REBOSSIS-85) IN A LAPINE POSTEROLATERAL FUSION MODEL

J. Christopher Nepola1, Emily B. Petersen1, Nicole DeVries-Watson2, Nicole Grosland2, Douglas C. Fredericks1

1Department of Orthopedics and Rehabilitation, The University of Iowa, Iowa City, IA USA2Department of Biomedical Engineering, The University of Iowa, Iowa City, IA USACorresponding author: Douglas Fredericks, Department of Orthopedics and Rehabilitation, Phone: (319) 335-4377, Fax: (319) 335-4378, Email: [email protected]: The authors report no potential conflicts of interest related to this study.Sources of Funding: This study was funded by OrthoRebirth through a grant to the Institution.

J. C. Nepola, E. B. Petersen, N. DeVries-Watson, N. Grosland, D.C. Fredericks


and can be combined with osteoinductive factors such as bone marrow aspirate to maximize bone growth.30

Calcium phosphate materials have been employed clinically as bone void fillers for several decades. These materials are most often provided in the form of small, porous granules that can be packed to fill the wide variety of size and shape of bony defects encoun-tered. Recently, ReBOSSIS85 was developed and is a synthetic bioresorbable bone void filler for the repair of bone defects with handling characteristics of glass wool-like (or cotton ball-like) material. ReBOSSIS85 is a synthetic, resorbable bone void filler. It is composite material consisting of (by weight) 40% beta-tricalcium phosphate (β-TCP), 30% siloxane-containing vaterite (a form of calcium carbonate, CaCO3), and 30% poly(L-lactide-co-glycolide). Siloxane-containing vaterite has been synthesized by doping calcium carbonate with 2-3% of siloxane with an amino group.34 Fibrous composites were prepared by blending siloxane-containing vaterite and β-TCP with PLGA using kneading and subsequent electrospinning. The electrospinning process used in manufacturing ReBOSSIS-85 results in a glass wool-like physical form. The diameter of the fibers in the PLGA scaffold ranges from 50 μm to 100 μm. The intercon-nected structure allows for formation of new bone throughout the network of interconnecting pores.31-33 The silicon content of the final ReBOSSIS-85 material is 0.5% - 1% by weight. A trace amount of silicon spe-cies has been reported to stimulate osteogenic cells to mineralize and enhance bone formation.34-37 The product form allows for versatility in handling and ease of use and can be hydrated with bone marrow aspirate (BMA). ReBOSSIS-85 has demonstrated to be osteoconductive, providing a scaffold for cell attachment and supporting formation of osseous tissue across bony defects.

This purpose of this study was to evaluate and com-pare the spine fusion performance of the ReBOSSIS-85 to a FDA approved bone graft substitute Mastergraft Putty in a spine fusion animal model, using iliac crest autograft as a control. Both products are designed to be hydrated with bone marrow aspirate and can be used when combined with autologous bone.

METHODSOne hundred fifty male skeletally mature New Zea-

land White rabbits weighing 4.5–5.5 kg were entered into the study (Table 1). Female rabbits were not used as they may undergo instantaneous ovulation and pseudopregnancy due to extraneous stimulation, and subsequent fluctuations in hormones may impact bone metabolism. To avoid potential impact on study outcomes, male animals were selected for this study. The number of animals per treatment was determined

by a power analysis to show a 20% difference in flexion/extension via biomechanical testing (alpha of 0.05 and a power of 90). All procedures were approved by the Institutional Animal Care Use Committee (#6031701) and conducted at The University of Iowa Department of Orthopaedics Animal Research Surgicenter, Bone Healing Research Lab, and the Iowa Spine Research Lab, USA. Throughout the study, animals were individually caged and monitored twice daily for signs of pain and discomfort.

Surgical procedureAll operative procedures were performed in a dedicat-

ed surgical suite using inhalation anesthesia and aseptic techniques. Surgical anesthesia was maintained with 1–5% isoflurane delivered in oxygen. Cardiorespiratory monitoring was conducted throughout the procedure.

All rabbits were prepared for surgical procedures using standard operating procedures. After the pre-anesthetic had taken effect, rabbits were clipped free of fur over the surgical area and scrubbed with Chlorhexi-dine soap and wiped with isopropyl alcohol (alternated three times each). Chlorhexidine solution was applied just prior to surgical incision. Rabbits were placed prone on the operating table. The surgical approach to the spine was identical in all rabbits. A dorsal midline skin incision, approximately 15 centimeters long, was made from L1 to the sacrum, and then the fascia and muscle were incised over the L5-L6 transverse processes (TPs). The TPs were then decorticated with a high-speed burr to ensure bleeding host bone. Vertebral bodies and lamina were avoided and not decorticated.

Table 1.Number of Animals per Treatment Group

Treatment Group

#Animals/ Time Point4

weeks8

weeks12

weeks Total

Iliac Crest Bone Graft (ICBG) 5 8 12 25

Mastergraft Putty + ICBG 5 8 12 25

ReBOSSIS-85 (0.2g)/BMA + ICBG (1:1) 5 8 12 25

ReBOSSIS-85 (0.3 g)/BMA + ICBG 5 8 12 25



Total #Animals 30 48 72 150

Volume 39 Issue 2 11

Electrospun PLGA and β-TCP (ReBOSSIS-85) in a Lapine Posterolateral Fusion Model

A

All animals had approximately 2.5–3.0 cc of corti-cocancellous bone graft obtained bilaterally from the iliac crest (total of 5-6 cc). This volume of graft is the maximum amount which can be harvested from the rab-bit iliac crest without significant animal morbidity.31,38-42

In the Mastergraft Putty™ animals, 1.5 cc of implant was hydrated with bone marrow aspirate (BMA, 1.5 cc) + ICBG (1.5cc) was placed in the paraspinal bed between the transverse processes. The ReBossis-85™ (RB-85), regardless of amount, was hydrated with 1.5 cc of BMA and 1.5 cc of ICBG (Figure 1) was added and then placed in the paraspinal bed. ICBG only rabbits received approximately 3.0 cc of ICBG per side. Fascia and skin were closed with 3-0 Vicryl and the skin stapled.

Radiographic assessmentVentral/dorsal radiographs were obtained with a

Simon DR (Quantum) RAD-X High Frequency Radio-graphic Imaging System, (model: E7242X), and stored using Whitecap PACs system. Radiographs were examined to confirm graft placement and assess graft migration, osteolysis, fracture, and/or any other adverse events. In addition, radiographs were graded for bilateral fusion. Three reviewers blinded to treatment group and time point determined bilateral fusion rates for each study group at each time point. Trabecular bone pat-terns, change in implant appearance, and graft density of term radiographs were compared to immediate post-op radiographs to assess fusion. Final fusion results were determined by agreement of at least two of the three observers.

Five specimens (histology animals) were scanned us-ing a SkyScan 1176 Micro-CT. Internal structures were reconstructed as a series of 2D cross sections which were then used to analyze morphological parameters

of the specimen. The SkyScan 1176 was calibrated ac-cording to the manufacturers recommended schedule for contrast, positioning and density.

A rectangular ROI of 250 mm2 (10 x 25 mm) was placed across the fusion site inclusive of the transverse processes sagittal plane, bilateral, center of defect) and areas of bone, implant, and soft tissue were calculated.

Manual palpation analysisThe primary outcome used to determine fusion was

manual palpation56 and flexibility analysis.24,57After remov-ing the spine, fusion was graded by three independent observers (blinded to treatment group and time point) as ‘fused’ if no detectable motion was present at the treated segment when tested in flexion and extension. The fusion was graded as ‘not fused’ if motion was present. Final results were determined by agreement of at least two of the three observers.

Mechanical testingBiomechanical non-destructive stiffness testing was

performed following manual palpation in the 12 week group to quantify stiffness and differences between groups. Testing consisted of flexion/extension, lateral bending, and torsion to a pre-determined, sub-failure load.

The vertebral bodies cranial and caudal to the fused motion segment were embedded in Bondo/Fiberglass material using 2 inch PVC piping. The specimens were mounted in a biaxial servo-hydraulic materials testing machine (858 Bionix II, MTS Corporation, Eden Prairie, MN, USA) retrofitted with two spine gimbals and a pas-sive XZ table. Custom-made rigid body markers compris-ing infrared light emitting diodes affixed between two small aluminum plates were placed on each vertebral body and the two gimbals to track the segmental mo-tions. Nondestructive flexibility tests were performed about each axis of rotation (i.e., flexion-extension, right-left lateral bending, and right-left axial rotation) by ap-plying an isolated ±0.27 Nm (0 Nm, 0.09 Nm, 0.18 Nm, and 0.27 Nm) moment about each of the primary axes. Each test initiated and concluded in the neutral position with zero load. Three loading and unloading cycles were performed with motion data collected on the third cycle (the first two cycles served as preconditioning). The displacement of each vertebrae was measured us-ing an optoelectronic motion capture system (Optotrak, Northern Digital, Waterloo, Ontario, Canada); the output of which was synchronized with that of the MTS. Dur-ing testing the specimens were kept moist with saline solution spray. Stiffness was determined and compared to normal controls (10 normal rabbit lumbar columns, historic internal laboratory controls).

Figure 1. RB-85 hydrated with BMA and filled with ICBG.



A B C

HistologyFive specimens from each group were randomly se-

lected for histological evaluation. Fusion sites of each animal were processed for histology and sectioned in the sagittal plane to obtain a total of six sections per animal (3 per side of each fusion mass). For each side, sections were created adjacent to the vertebral body, through the center of the fusion mass, and through the lateral aspect of the fusion mass, spaced approximately 3 mm apart. Sections were stained with H&E.

All six sections from each fusion mass (3 per side) were subject to semiquantitative histopathology analysis. Specimens from each slide were evaluated by the PI and a board certified veterinary pathologist and assigned scoring based on a modified ISO 10993-6, Annex E (Table 2). An overall assessment was recorded for each fusion mass (one score per side).

Statistical analysisStatistical analysis was performed on the microCT and

histomorphometry data, as well as the flexion-extension, lateral bending and axial rotation data from the flexibility analysis. All data was analyzed to a 95% confidence level (p<0.05) using a 2-tailed t-test assuming unequal variance in Microsoft Excel. Unless otherwise noted, all data are reported as the mean and one standard deviation.

RESULTSAll test articles were administered successfully. Rab-

bits were alert and eating within 3 hours of surgery. Daily observations of each rabbit were conducted and general overall health was determined. The rabbits, regardless of treatment, were in good health at time of

Table 2. Histopathology Analysis Scoring System

ResponseScore (*phf = per high powered field)

0 1 2 3 4

Polymorphonuclear cells 0 Rare, 1-5/phf* 5-10/phf Heavy infiltrate Packed

Lymphocytes 0 Rare, 1-5/phf* 5-10/phf Heavy infiltrate Packed

Plasma cells 0 Rare, 1-5/phf* 5-10/phf Heavy infiltrate Packed

Macrophages 0 Rare, 1-5/phf* 5-10/phf Heavy infiltrate Packed

Giant cells 0 Rare, 1-5/phf* 3-5/phf Heavy infiltrate Packed

Necrosis 0 Slight Moderate Marked Severe

Infection 0 Slight Moderate Marked Severe

Fibrinous exudates 0 Slight Moderate Marked Severe

Neovascularisation 0 Minimal capillary prolif-eration focal, 1-3 buds

Groups of 4-7 capillaries with supporting fibroblas-

tic structures

Broad band of capil-laries with supporting

structures

Extensive band of capil-laries with supporting fibroblastic structures

Fibrocytes/ fibro-connective tissue, fibrosis 0 Narrow band Moderately Thick band Extensive band

Material degradation / Graft resorption 0 Slight Moderate Marked Severe (100% degraded)

Bone Formation 0 Slight Moderate Marked Extensive

Figure 2. Post-op and twelve week plain radiographs.



euthanasia. There were no complications related to the test articles in-vivo. Three (3) rabbits were euthanized prior to end of study due to complications associated with ICBG harvest (1 ICBG 8w, 1 RB-85-0.45g 12w, and 1 RB-85-0.6g 4w). These complications have been noted in previous studies by the PI and others.12,14-20,34,44-47

Necropsy of the animals was unremarkable regardless of test group. Macroscopic analysis of the implant sites demonstrated healthy tissue with no apparent adverse effects such as inflamed, necrotic, or devascularized tissue surrounding the defect sites.

Radiographic Assessment: Radiographs at 4, 8 and 12 weeks showed a normal healing response over time with no adverse reactions for all test groups (Figure 2). The loss of graft distinction at the host bone margins indicates a progression in host integration, implant remodeling and new bone formation over time. No frac-tures, osteolysis, or other adverse reactions were evident during radiographic examination for all test groups. Graft position migration from immediate post-op radiographs to radiographs greater than 1-week post-op is typically observed as the forces of overlaying muscle compresses the graft. These migration observations are typical of synthetic bone grafts applied in this animal model and are not a confounding variable. The radiographic fusion analysis is summarized in Table 3.

The MicroCT scans supported the radiographic findings, showing a normal osteoconductive healing response over time with no adverse reactions for all test groups. The 4 week microCT scans showed host integration and new bone formation originating from the transverse process margins in all animals. The 8 and 12 week scans showed a progression of new bone formation and graft remodeling across the paraspinal bed from the 4 week time point (Figure 3).

MicroCT morphometry analysis was performed us-ing a rectangular region of interest (ROI) of 250 mm2 placed across the fusion site inclusive of the transverse

Table 3. Bilateral Radiographic Fusion Rate4 weeks 8 weeks 12 weeks

ICBG 0% 43% 50%

MGP 0% 50% 50%

RB-85 (0.2) 0% 0% 0%

RB-85 (0.3) 0% 13% 25%

RB-85 (0.45) 0% 38% 36%

RB-85 (0.6) 0% 63% 50%

Figrure 3. Sagittal slices of fusion masses at twelve weeks.

Table 4. MicroCT Tissue Ares

4 weeks 8 weeks 12 week

Bone Area mm2

Implant Area mm2

Soft Tissue Area mm2

Bone Area mm2

Implant Area mm2


Bone Are

mm2

Implant Area mm2


ICBG 55.99 N/A 194.01 59.56 N/A 190.44 53.93 N/A 196.07

Mastergraft Putty 39.72 41.14 169.15 59.12 11.33 179.55 42.13 4.17 203.70

ReBOSSIS-85 (0.2 g) 45.70 N/A 204.30 37.52 N/A 212.48 47.75 N/A 202.25






processes (sagittal plane, bilateral, center of defect) and areas of bone, residual implant (Mastergraft Putty, MGP) and soft tissue were calculated. ReBOSSIS-85, regard-less of group, was not distinguishable from background noise and/or soft tissue and therefore was not measured. Mean tissue areas are presented in Table 4.

Flexibility Assessment of Stiffness: Spine fusion was assessed by manual palpation of the treated motion seg-ments. No implanted motion segments were fused at 4 weeks, regardless of implant type. At 12 weeks, ICBG, MGP, and RB-85-0.6g were fused mechanically in at least

50% of the rabbits. No other test groups reached the 50% threshold for fusion in this model. Fusion results are listed in Table 5.

Motion analysis of the test groups in this study were compared to historical internal laboratory data of normal unfused rabbit spines. Normal motion of the rabbit lumbar spine was determined by testing 10 nor-mal (untreated/ unfused) rabbit lumbar spinal columns using the same testing methods. Motion in the flexion/extension plane was used as the primary measurement to demonstrate fusion (Figure 4) as manual palpation testing is based off of motion detected in flexion/exten-sion. General findings from the biomechanics indicate that the 6 test article groups had significantly less motion than the unfused normals. Mastergraft Putty, RB-85-0.45g, and RB-85-0.6g were not significantly different from ICBG positive control. No significant differences were seen among Mastergraft Putty and RB-85-0.3g, RB-85-0.45g, or RB-85-0.6g treated fusions. Both RB-85-0.45g and RB-85-0.6g had significantly less motion than RB-85-0.2g motion segments.

Histologic Assessment: The majority of specimens, regardless of implant type or time of implantation, had

Figure 4. Range of motion at twelve weeks.

Table 5. Manual Palpation Rates4 weeks 8 weeks 12 weeks

ICBG 0% 43% 50%

MGP 0% 50% 50%

RB-85 (0.2) 0% 0% 8%

RB-85 (0.3) 0% 13% 25%

RB-85 (0.45) 0% 38% 36%

RB-85 (0.6) 0% 63% 50%



Table 6. Hisopathology ScoringIm

plan

t

Term

Pol

ymor

pho-

nucl

ear

cells

Lym

phoc

ytes

Pla

sma

cells

Mac

roph

ages

Gia

nt c

ells

Nec

rosi

s

Infe

ctio

n

Fibr

inou

s ex

udat

es

Neo

vasc

ular

-is

atio

n

Fibr

osis

Gra

ft re

sorp

tion

New

Bon

e Fo

rmat

ion

ICBG 4 Mean 1.00 1.00 1.00 1.00 1.30 0.00 0.00 0.00 2.40 3.00 1.40 1.20

St.Dev. 0.00 0.00 0.00 0.00 0.48 0.00 0.00 0.00 0.52 0.00 0.52 0.42

MGP 4 Mean 1.00 2.00 2.00 1.70 1.80 0.00 0.00 0.00 2.00 3.00 1.30 1.00

St.Dev. 0.00 0.00 0.00 0.48 0.42 0.00 0.00 0.00 0.00 0.00 0.48 0.00RB-85-0.2g 4 Mean 0.80 0.70 0.30 1.80 1.78 0.00 0.00 0.00 1.10 3.00 0.00 1.00

St.Dev. 0.42 0.48 0.48 0.42 0.44 0.00 0.00 0.00 0.32 0.00 0.00 0.00RB-85-0.3g 4 Mean 0.90 0.80 0.20 2.10 1.70 0.00 0.00 0.00 1.50 3.00 0.00 1.00

St.Dev. 0.57 0.42 0.42 0.74 0.67 0.00 0.00 0.00 0.53 0.00 0.00 0.00RB-85-0.45g 4 Mean 0.80 0.70 0.10 2.10 1.80 0.00 0.00 0.00 1.60 3.00 0.00 1.00

St.Dev. 0.42 0.48 0.32 0.74 0.79 0.00 0.00 0.00 0.52 0.00 0.00 0.00RB-85-0.6g 4 Mean 0.88 1.00 0.13 2.00 1.88 0.00 0.00 0.00 1.63 3.00 0.00 1.00

St.Dev. 0.35 0.00 0.35 0.00 0.64 0.00 0.00 0.00 0.52 0.00 0.00 0.00

ICBG 8 Mean 0.50 1.00 1.00 1.20 1.90 0.00 0.00 0.00 2.40 3.00 1.80 2.20

St.Dev. 0.53 0.00 0.00 0.42 0.32 0.00 0.00 0.00 0.52 0.00 0.42 0.63

MGP 8 Mean 1.00 2.90 2.00 2.00 2.00 0.00 0.00 0.00 1.10 3.00 2.00 2.30

St.Dev. 0.00 0.32 0.00 0.00 0.47 0.00 0.00 0.00 0.32 0.00 0.47 0.48RB-85-0.2g 8 Mean 0.80 0.60 0.30 2.10 1.50 0.00 0.00 0.00 1.40 2.60 1.00 1.60

St.Dev. 0.63 0.52 0.48 1.10 0.85 0.00 0.00 0.00 0.52 0.52 0.00 0.84RB-85-0.3g 8 Mean 0.90 0.20 0.30 2.20 1.20 0.00 0.00 0.00 1.30 2.70 1.30 1.90

St.Dev. 0.57 0.42 0.48 0.79 0.42 0.00 0.00 0.00 0.48 0.48 0.48 0.74RB-85-0.45g 8 Mean 0.90 0.70 0.40 2.00 1.70 0.00 0.00 0.00 1.50 2.70 1.40 2.70

St.Dev. 0.32 0.48 0.52 0.82 0.95 0.00 0.00 0.00 0.71 0.67 0.52 0.95RB-85-0.6g 8 Mean 0.80 0.50 0.30 1.90 1.30 0.00 0.00 0.00 1.50 2.00 1.40 2.80

St.Dev. 0.42 0.53 0.48 0.57 0.67 0.00 0.00 0.00 0.53 1.05 0.52 0.92

ICBG 12 Mean 0.50 1.00 1.00 0.90 0.90 0.00 0.00 0.00 2.80 3.00 2.00 2.70

St.Dev. 0.53 0.00 0.00 0.32 0.74 0.00 0.00 0.00 0.42 0.00 0.00 0.48

MGP 12 Mean 0.40 1.10 0.70 0.80 0.90 0.00 0.00 0.00 1.00 0.80 3.20 3.20

St.Dev. 0.52 1.45 0.95 1.03 1.20 0.00 0.00 0.00 0.00 1.03 1.03 1.03RB-85-0.2g 12 Mean 0.80 0.20 0.30 1.60 1.20 0.00 0.00 0.00 1.20 2.00 2.00 2.00

St.Dev. 0.42 0.42 0.48 0.52 0.42 0.00 0.00 0.00 0.79 0.00 0.00 0.67RB-85-0.3g 12 Mean 0.80 0.50 0.30 1.40 1.20 0.00 0.00 0.00 1.40 1.60 2.20 2.40

St.Dev. 0.42 0.53 0.48 0.97 0.63 0.00 0.00 0.00 0.70 0.70 0.63 1.26RB-85-0.45g 12 Mean 0.60 0.20 0.00 0.90 0.80 0.00 0.00 0.00 1.50 1.50 2.30 2.10

St.Dev. 0.52 0.42 0.00 0.99 0.42 0.00 0.00 0.00 0.71 0.53 0.48 0.88RB-85-0.6g 12 Mean 0.30 0.70 0.00 0.60 0.80 0.00 0.00 0.00 0.90 1.10 2.70 3.30

St.Dev. 0.48 0.48 0.00 0.52 0.42 0.00 0.00 0.00 0.57 0.57 0.48 0.82



mild inflammation. Most often there were very low numbers of macrophages and multinucleated giant cells with some scattered, often erythrocytes, and plasma cells in the early time points. In all sections, there is moder-ate neovascularization and fibrous tissue interspersed throughout the sections (Table 6). Graft resorption and new bone formation increased with time from 4 to 12 weeks (Figure 5).

DISCUSSIONIn this study we evaluated the in vivo performance

of an electrospun PLGA fabric combined with calcium carbonate (vaterite) and β-TCP, (ReBOSSIS-85) and compared it to Mastergraft Putty. Both materials were combined with bone marrow aspirate and iliac crest auto-graft, in an established posterolateral spine fusion rabbit model. The study also included a control study group using only iliac crest autograft (ICBG). The test groups were evaluated for spine fusion rate, new bone forma-tion, graft resorption and inflammatory response using radiographic, microCT, biomechanical and histological endpoints at 4, 8 and 12 weeks following implantation.

Radiographic scoring of the fusion sites indicated a normal healing response in all test groups, with no adverse reactions and similar progressions of new bone formation over time. Fusion rates for Mastergraft Putty and ReBOSSIS-85 (0.6g) were equivalent to ICBG at 8 and 12 weeks. The remaining ReBOSSIS-85 groups (0.2, 0.3, 0.45 g) were fused 0%, 25%, and 36% respectively at 12 weeks. MicroCT scans supported the radiographic observations, showing a normal osteoconductive healing response in all groups with remodelling of the fusion site over time.

Manual palpation results, at 4 weeks, demonstrated no fusions. At 12 weeks, ICBG, MGP, and RB-85-0.6g were fused mechanically in at least 50% of the rabbits. No other test groups reached the 50% threshold for fu-sion in this model. All groups demonstrated significantly less range of motion in both flexion/extension, lateral bending, and axial rotation compared to normal unfused controls.

Biodegradable polymer scaf folds are commonly used in bone reconstruction, tissue engineering, drug delivery systems, and three-dimensional biomaterials. When combined with ceramic composite bone fillers like calcium carbonate, these hybrids demonstrate excellent bone-forming and mechanical properties.50-54 Electrospin-ning of biodegradable polymer scaffolds can create me-chanically strong polymer fiber meshes with even higher porosity and flexibility, allowing for easy handling, while control over pore sizing in turn can control soft tissue intrusion at the regeneration site.55-58 These electrospun PLA and PLGA fabrics combined with calcium carbonate (vaterite) show excellent hydroxyapatite (HA)-forming Figure 5. Histology slides at twelve weeks.



ability, cellular compatibility, and biocompatibility.59,60 The bone forming ability of these hybrid materials can be increased by doping the calcium carbonate particles with siloxane.61,62 The siloxane hybrids (SiPVHs) allow a controlled release of soluble silica species and calcium ions.48 Trace amounts of soluble silica and calcium have been suggested to enhance the mineralization and bone forming abilities of osteogenic cells and the soluble silica species and calcium ions released by the SiPVHs reportedly stimulate the proliferation and differentiation of murine osteoblast-like cells.63-65 This indicates a possi-bility of siloxane being used to create new bone-repairing and regenerating materials.

Histopathology analysis of the fusion masses, from each test article and time point, indicated an expected normal response for resorbable calcium phosphates and collagen graft material when present. Mild inflammation with macrophage and multinucleated giant cell response to the graft material was evident in all test groups. Re-gardless of test article, new bone formation and graft resorption increased from 4 to 12 weeks post-op.

This animal study has demonstrated the biocompat-ibility and normal healing features associated with the ReBOSSIS-85 bone graft (test material) when combined with bone marrow aspirate and autograft as an extender. ReBOSSIS-85 was more effective when a larger mass of test article was used in this study.

REFERENCES1. Pugely AJ, Petersen EB, Devries-Watson N,

Fredericks DC. Influence of 45S5 Bioactive Glass in A Standard Calcium Phosphate Collagen Bone Graft Substitute on the Posterolateral Fusion of Rabbit Spine. Iowa Orthop J. 2017;37:193-198.

2. Arrington ED, Smith WJ, Chambers HG, et al. Complications of iliac crest bone graft harvesting. Clin.Orthop.Relat Res. 1996:300–9.

3. Xu R, Ebraheim NA, Yeasting RA, Jackson WT. Anatomic considerations for posterior iliac bone harvesting. Spine. 1996;21(9):1017-20.

4. Heppenstall B. Bone grafting. In: McCollster Evarts C, ed. Surgery of the Musculoskeletal System. Vol 1. New York: Churchill Livingston, 1983:89-106.

5. Chase SW, Herndon CH: The fate of autogenous and homogenous bone grafts: A historical review. J Bone Joint Surg 37A:809-841, 1955.

6. Friedlaender GE: Current concepts review: Bone grafts. J Bone Joint Surg 69A:786-790, 1987.

7. Goldberg VM, Stevenson S: Natural history of autografts and allografts. Clin Orthop 225:7-16, 1987.

8. Hall MB, Vallerand WP, Thompson D, Hartley G: Comparative anatomic study of anterior and pos-terior iliac crests as donor sites. J Oral Maxillofac Surg 49:560-563, 1991.

9. Hsu K, Zucherman JF, White AH: Bone Grafts and Implants in Spine Surgery. In White AH, Rothman RH, Ray CD (eds). Lumbar Spine Surgery. St Louis, CV Mosby 434-458, 1987.

10. Kurz LT, Garfin SR, Booth RE: Harvesting autog-enous iliac bone grafts: A review of complications and techniques. Spine 14:1324-1331, 1989.

11. Fredericks DC, Smucker J, Petersen EB, et al. Effects of direct current electrical stimulation on gene expression of osteopromotive factors in a posterolat-eral spinal fusion model. Spine. 2007;32(2):174-81.

12. Silber JS, Anderson DG, Daffner SD, et al. Donor site morbidity after anterior iliac crest bone harvest for single-level anterior cervical discectomy and fu-sion. Spine. 2003;28(2):134-9.

13. Schnee CL, Freese A, Weil RJ, Marcotte PJ. Analysis of harvest morbidity and radiographic outcome using autograft for anterior cervical fusion. Spine. 1997;22(19):2222–2227.

14. Smucker JD, Petersen EB, Al-hili A, Nepola JV, Fredericks DC. Assessment of SiCaP-30 in a Rabbit Posterolateral Fusion Model with Concurrent Chemotherapy. Iowa Orthop J. 2015;35:140-6.

15. Smucker JD, Petersen EB, Nepola JV, Fred-ericks DC. Assessment of Mastergraft(®) strip with bone marrow aspirate as a graft extender in a rabbit posterolateral fusion model. Iowa Orthop J. 2012;32:61-8.

16. Smucker JD, Fredericks DC. Assessment of Progenix(®) DBM putty bone substitute in a rab-bit posterolateral fusion model. Iowa Orthop J. 2012;32:54-60.

17. Smucker JD, Petersen EB, Fredericks DC. Assessment of MASTERGRAFT PUTTY as a graft extender in a rabbit posterolateral fusion model. Spine. 2012;37(12):1017-21.

18. Smucker JD, Bobst JA, Petersen EB, Nepola JV, Fredericks DC. B2A peptide on ceramic gran-ules enhance posterolateral spinal fusion in rabbits compared with autograft. Spine. 2008;33(12):1324-9.

19. Fredericks D, Petersen EB, Watson N, Gro-sland N, Gibson-Corley K, Smucker J. Com-parison of Two Synthetic Bone Graft Products in a Rabbit Posterolateral Fusion Model. Iowa Orthop J. 2016;36:167-73.

20. Fredericks DC, Petersen EB, Sahai N, et al. Evaluation of a novel silicate substituted hydroxy-apatite bone graft substitute in a rabbit posterolateral fusion model. Iowa Orthop J. 2013;33:25-32.

21. Schliephake H, Neukam FW, Klosa D. Influence of pore dimensions on bone ingrowth into porous hydroxylapatite blocks used as bone graft substi-tutes. A histometric study. Int J Oral Maxillofac Surg. 1991;20:53–8. doi: 10.1016/S0901-5027(05)80698-8.



22. Holmes RE, Wardrop RW, Wolford LM. Hydrox-ylapatite as a bone graft substitute in orthognathic surgery. J Oral Maxillofac Surg 1988: 46: 661-71.

23. Holmes RE. Bone regeneration within a coralline hydroxyapatite implant. Plast Reconstr Surg 1979: 63: 626-33.

24. Holmes RE, Buchholz RW, Mooney Porous hy-droxyapatite as a bone-graft substitute in metaphyseal defects. J Bone Joint Surg 1986:68A:904-11.

25. Klawitter JJ, Hulbert SF. Application of porous ceramics for the attachment of load bearing internal orthopedic applications. J Biomed Mater Res Sym-posium 1971: 2: 161-229.

26. Yuan H, Yang Z, Li Y, Zhang X, De Bruijn JD, De Groot K. Osteoinduction by calcium phosphate biomaterials.J Mater Sci Mater Med. 1998;9:723–6. doi: 10.1023/A:1008950902047.

27. Baas J, Svaneby D, Jensen TB, Elmengaard B, Bechtold J, Soballe K. Coralline hydroxyapatite granules inferior to morselized allograft around uncemented porous Ti implants: unchanged fixation by addition of concentrated autologous bone marrow aspirate. J Biomed Mater Res A. 2011;99:9–15. doi: 10.1002/jbm.a.33156.

28. van Haaren EH, Smit TH, Phipps K, Wuisman PI, Blunn G, Heyligers IC. Tricalcium-phosphate and hydroxyapatite bone-graft extender for use in impaction grafting revision surgery. An in vitro study on human femora. J Bone Joint Surg Br 2005; 87: 267–271.

29. Johnson KD, Frierson KE, Keller TS, et al. Porous ceramics as bone graft substitutes in long bone defects: a biomechanical, histological, and radiographic analysis. J Orthop Res 1996;14:351–69.

30. Odri G-A, Revert R, Delecrin J, et al. Effect of adding bone marrow to ceramic graft materials with different interconnectivities in lumbar arthrodesis. Journal of Orthopaedic Science. 2013;18(2):321-330. doi:10.1007/s00776-012-0336-3.

31. Boden SD, Martin GJ, Jr., Morone M et al. The use of coralline hydroxyapatite with bone marrow, autogenous bone graft, or osteoinductive bone pro-tein extract for posterolateral lumbar spine fusion. Spine 1999;24:320–7.

32. Hutmacher, D.W. Scaffolds in tissue engineering bone and cartilage. Biomaterials 21, 2529, 2000.

33. Reece GP, Patrick Jr CW. Tissue engineered con-struct design principles. In: Patrick Jr CW, Mikos AG, McIntire LV, editors. Frontiers in tissue engineering. New York, USA: Elsevier Science, 1998. p. 166-96.

34. Obata A, Hotta T, Wakita T, Ota Y, Kasuga T. Electrospun microfiber meshes of silicon-doped vaterite/poly(lactic acid) hybrid for guided bone regeneration. Acta Biomater. 2010;6(4):1248-57.

35. Xynos ID, Edgar AJ, Buttery LD, Hench LL, Polak JM. Ionic products of bioactive glass dissolu-tion increase proliferation of human osteoblasts and induce insulin-like growth factor II mRNA expression and protein synthesis. Biochem Biophys Res Com-mun. 2000;276(2):461-5.

36. Reffitt, D.; Ogston, N.; Jugdaohsingh, R.; Cheung, H.; Evans, B.A.J.; Thompson, R.; Pow-ell, J.; Hampson, G. Orthosilicic acid stimulates collagen type 1 synthesis and osteoblastic differen-tiation in human osteoblast-like cells in vitro. Bone 2003, 32, 127–135.

37. Thian ES, Huang J, Best SM, et al. The re-sponse of osteoblasts to nanocrystalline silicon-substituted hydroxyapatite thin films. Biomaterials. 2006;27(13):2692-8.

38. Boden SD, Schimandle JH, Hutton WC. An experimental lumbar intertransverse process spinal fusion model. Radiographic, histologic, and biome-chanical healing characteristics. Spine 1995;412–20.

39. Grauer JN, Erulkar JS, Patel TC et al. Biome-chanical evaluation of the New Zealand white rabbit lumbar spine: a physiologic characterization. Eur-SpineJ 2000;9:250–5.

40. Grauer JN, Vaccaro AR, Kato M et al. Develop-ment of a New Zealand white rabbit model of spinal pseudarthrosis repair and evaluation of the potential role of OP-1 to overcome pseudarthrosis. Spine 2004;29:1405–12.

41. Boden SD, Schimandle JH, Hutton WC. Lumbar intertransverse-process spinal arthrodesis with use of a bovine bone-derived osteoinductive protein. A preliminary report. JournalofBone&JointSurgery 1995;77:1404–17.

42. Boden SD, Schimandle JH, Hutton WC et al. In vivo evaluation of a resorbable osteoinductive com-posite as a graft substitute for lumbar spinal fusion. JournalofSpinalDisorders 1997;10:1–11.

43. Valdes M, Palumbo M, Appel AJ, Mcallister S, Ehrlich M. Posterolateral intertransverse lumbar arthrodesis in the New Zealand White rabbit model: II. Operative technique. Spine J. 2004;4(3):293-9.

44. Gupta AR, Shah NR, Patel TC, Grauer JN. Peri-operative and long-term complications of iliac crest bone graft harvesting for spinal surgery: a quantita-tive review of the literature. Int Med J. 2001;8:163–6.

45. Magit DP, Maak T, Trioano N, et al. Healos/re-combinant human growth and differentiation factor-5 induces posterolateral lumbar fusion in a New Zea-land white rabbit model. Spine. 2006;31(19):2180-8.

46. Cockin J: Autologous bone grafting-complications at the donor site. J Bone Joint Surg 53B:153, 1971.



47. Shiels SM, Raut VP, Patterson PB, Barnes BR, Wenke JC. Antibiotic-loaded bone graft for reduc-tion of surgical site infection in spinal fusion. Spine J. 2017;17(12):1917-1925.

48. V.W. Bramwell, J.E. Eyles, H. Oya Alpar Particu-late delivery systems for biodefence subunit vaccines Adv Drug Deliv Rev, 57 (2005), pp. 1247-1265

49. J.P. Santerre, K. Woodhouse, G. Laroche, R.S. Labow Understanding the biodegradation of polyure-thanes: from classical implants to tissue engineering materials Biomaterials, 26 (2005), pp. 7457-7470

50. S.S. Kim, M.S. Park, O. Jeon, C.Y. Choi, B.S. Kim Poly(lactide-co-glycolide)/hydroxyapatite composted scaffolds for bone tissue engineering Biomaterials, 27 (2006), pp. 1399-1409

51. K. Zang, Y. Wang, M.A. Hillmyer, L.F. Francis Processing and properties of porous poly(l-lactide)/bioactive glass composites Biomaterials, 25 (2004), pp. 2489-2500

52. T. Kasuga, H. Maeda, K. Kato, M. Nogami, K. Hata, M. Ueda Preparation of poly(lactic acid) composites containing calcium carbonate (vaterite) Biomaterials, 24 (2003), pp. 3247-3253

53. H. Liao, H. Mutvei, M. Sjöström, L. Hammar-ström, J. Li Tissue responses to natural aragonite (Margaritifera shell) implants in vivo Biomaterials, 21 (2000), pp. 457-468

54. R. Vago, D. Plotquin, A. Bunin, I. Sinelnikov, D. Atar, D. Itzhak Hard tissue remodeling using biofabricated coralline biomaterials J Biochem Bio-phys Methods, 50 (2002), pp. 253-259

55. T.J. Still, H.A. von Recum Electrospinning: ap-plications in drug delivery and tissue engineering Biomaterials, 29 (2008), pp. 1989-2006

56. W.J. Li, C.T. Laurencin, E.J. Caterson, R.S. Tuan, F.K. Ko Electrospun nanofibrous structure: a novel scaffold for tissue engineering J Biomed Mater Res, 60 (2002), pp. 613-621

57. X. Xin, M. Hussain, J.J. Mao Continuing dif-ferentiation of human mesenchymal stem cells and induced chondrogenic and osteogenic lineages in electrospun PLGA nanofiber scaffold Biomaterials, 28 (2007), pp. 316-325

58. T. Wakita, A. Obata, G. Poologasundarampillai, J.R. Jones, T. Kasuga, Preparation of electrospun siloxane-poly(lactic acid)-vaterite hybrid fibrous membranes for guided bone regeneration, Compos. Sci. Technol. 70 (13) (2010) 1889–1893.

59. H. Maeda, T. Kasuga, L.L. Hench Preparation of poly(l-lactic acid)–polysiloxane–calcium carbonate hybrid membranes for guided bone regeneration Biomaterials, 27 (2006), pp. 1216-1222

60. H. Maeda, T. Kasuga, M. Nogami Bonelike apatite coating on skeleton of poly(lactic acid) composite sponge Mater Trans, 45 (2004), pp. 989-993

61. A. Obata, S. Tokuda, T. Kasuga Enhanced in vitro cell activity on silicon-doped vaterite/poly(lactic acid) composites Acta Biomater, 5 (2009), pp. 57-62\

62. Obata A, Hotta T, Wakita T, Ota Y, Kasuga T. Electrospun microfiber meshes of silicon-doped vaterite/poly(lactic acid) hybrid for guided bone regeneration. Acta Biomater. 2010;6(4):1248-57.

63. I.D. Xynos, A.J. Edgar, L.D.K. Buttery, L.L. Hench, J.M. Polak Ionic products of bioactive glass dissolution increase proliferation of human osteo-blasts and induce insulin-like growth factor II mRNA expression and protein synthesis Biochem Biophys Res Commun, 276 (2000), pp. 461-465

64. J.E. Gough, J.R. Jones, L.L. Hench Nodule forma-tion and mineralisation of human primary osteoblasts cultured on a porous bioactive glass scaffold Bioma-terials, 25 (2004), pp. 2039-2046

65. J.R. Jones, O. Tsigkou, E.E. Coates, M.M. Ste-vens, J.M. Polak, L.L. Hench Extracellular matrix formation and mineralization on a phosphate-free porous bioactive glass scaffold using primary human osteoblast (HOB) cells Biomaterials, 28 (2007), pp. 1653-1663


ABSTRACTBackground: This study compares effectiveness

of two commercially available signals, Pulsed Elec-tromagnetic Field (PEMF) and Combined Magnetic Field (CMF) clinical signals, to stimulate bone healing in rabbit tibial osteotomies.

Methods: One millimeter osteotomies in New Zealand White rabbits, stabilized with external fixators, were exposed daily to either signal for 30 minutes, three or six hours. Osteotomized sham controls received no signal exposure. Analyses of torsional strength, periosteal callus area and frac-ture healing stage demonstrated dose responses to increasing daily exposures to both signals.

Results: By 14 days torsional strength increased over shams in the three and six hour-treated groups, significant only for the six hour groups (p<0.05). By 21 days both three and six hour-treated groups were significantly stronger than shams (p<0.05, p<0.005) and the PEMF 30 minute treated group also showed significance (p<0.05). PEMF versus CMF-treated groups were not different at any exposure time.

Conclusions: Both CMF and PEMF signals were most effective in this model when used for six hours per day.

Clinical Relevance: In this model we dem-onstrate that though both PEMF and CMF are “bioactive” and promote healing at shorter and longer exposure dosages, there exists an “optimal” threshold effect of 6 hours/day electromagnetic wave stimulation for bone healing.

Keywords: animal model, fracture healing, Combined Magnetic Field, Pulsed Electromagnetic Field

INTRODUCTIONCombined electromagnetic field and pulsed electro-

magnetic field technologies commercially available for use to promote osteogenesis are prescribed according to directives provided by manufacturer’s in keeping with indications as permitted by the FDA. Current treatment protocols are derived from submissions made commensu-rate with the clinical model testing provided. However, in-adequate dosing literature exists to provide clinicians with guidance with respect to optimal application duration. Considerable variation in prescribed time of both PEMF and CMF signals is recommended by manufacturer’s.

Three different electromagnetic devices are currently available for use in the treatment of nonunion bone frac-tures.1-3 The recommended daily exposure times for these devices vary from 30 minutes to ten hours per day. All three signals are reported to stimulate healing of non-union fractures.4-8 A retrospective analysis of the PEMF clinical data indicated that the time it took to heal the nonunions showed a dose response related to the length of daily exposure.9 The median time to heal was reduced by 60% if the PEMF device was used for ten hours per day as compared to use for one hour per day. Garland et al.7 reported that a similar PEMF device when used fewer than three hours per day in a prospective clinical trial was less effective in healing non-union fractures (35.7%, 5/14) than when used three hours or more per day (80%, 108/135). No similar analysis of clinical dose response has been published for the CMF clinical data.

Therefore, this study was designed to quantitatively investigate whether the PEMF and CMF signals demon-strate different optimal exposure times in stimulating the repair of rabbit tibial osteotomies.

METHODSSkeletally mature male New Zealand White rabbits

weighing 3-4 kg were obtained from Myrtle Rabbitry (Thompson Station, TN, USA). A one millimeter trans-verse osteotomy was created in the left tibia of each rabbit with a Gigli wire as previously described.10 The perios-teum was preserved and the osteotomy stabilized with an external fixator (M100, Orthofix, Verona, Italy). All rabbits were permitted immediate unrestricted weight bearing in standard rabbit cages following surgery. Ra-diographs were taken post-surgery and weekly thereafter for the duration of the study.

THE EFFECT OF PULSE ELECTOMAGNETIC FIELD AND COBMINED MAGNETIC FIELD EXPOSURE TIME ON HEALING

OF A RABBIT TIBIAL OSTEOTOMY

Douglas C. Fredericks1; Emily B. Petersen1; Madison Rhodes1; Grace A. Larew1; James V. Nepola1

1Department of Orthopedics and Rehabilitation, The University of Iowa, Iowa City, IA USACorresponding author: Douglas Fredericks, Department of Orthopedics and Rehabilitation, Phone: (319) 335-4377, Fax: (319) 335-4378, Email: [email protected]: The authors report no potential conflicts of interest related to this study.

Sources of Funding: This study was funded by EBI Medical Systems, Inc., through a grant to the Institution.

D.C. Fredericks, E.B. Petersen, M.Rhodes, G.A. Larew, J.V. Nepola.


Experimental DesignEach rabbit was randomly assigned to one of three

treatment groups: unexposed sham controls, PEMF-exposed and CMF-exposed. All rabbits were placed in rabbit restrainers to undergo treatment. Exposures to the PEMF and CMF signals were for 30 minutes, three and six hours daily. Treatment was initiated 24 hours post-surgery and continued until sacrifice at 14 or 21 days at which times the treated tibiae underwent biomechanical testing. Sham controls at 14 days were handled identically with the rabbits placed in restrainers for 30 minutes, three and six hours daily but not exposed to either signal. In this group and in previous studies using this model, the length of time the sham rabbits were in restrainers had no effect on any biomechanical or radiographic healing parameters. For this reason the 21 day sham rabbits were all exposed for one hour per day to reduce the number of rabbits which would need to be euthanized.

The PEMF signal consists of a pulsed magnetic field comprised of 4.5 ms duration bursts of 20 pulses repeated at 15 Hz. During each pulse the magnetic field rises from 0 to 1.6 mT in 200 ms and then decays back to zero in 25 ms.2

The CMF signal consists of the collinear sum of two magnetic fields; a 76.6 Hz sine wave of 20 mT amplitude (40 mT peak to peak) superimposed with a net static mag-netic field of 20 mT.3 To more closely mimic the clinical use of these signals, no attempt was made to offset the Earth’s magnetic field other than in the case of the CMF signal where the net static field parallel to the 76.6 Hz sine wave is normally held at 20 mT as part of the device. For both signals, paired Helmholtz coils were attached to the rabbit restrainers in a vertical position and centered over the location of the osteotomy site.

Investigators were blinded to the knowledge of which treatment each specimen received.

Radiographical StudiesAnteroposterior and lateral radiographs were taken

at weekly intervals to evaluate callus formation and as-sess adequacy of fixation. The area of periosteal callus and diaphyseal bone at the osteotomy site was measured using digitized images of X-rays. Metrics software by Orthographics (Salt Lake City, UT, USA) was used for the analysis. Periosteal callus and diaphyseal bone were outlined along the bone between the two inner-most fix-ator screws in both the lateral and anteroposterior films. Area was calculated from each view and was expressed as a ratio of periosteal callus to diaphyseal bone.

Biomechanical TestingThe osteotomized tibiae were harvested, cleaned of

soft tissues, and immediately prepared for biomechanical testing as previously described.10 Mechanical tests were performed using a custom-made torsion apparatus on an MTS 810 uniaxial testing machine (MTS, Minneapolis, MN, USA). The proximal end of the specimen was firmly anchored on the apparatus. Torque was applied to the dis-tal end of the specimen at a rate of 1.76 degrees per second until failure occurred, and the torque versus angular defor-mation data were recorded. Torque was calculated as the product of the tensile force (measured in the cable with a load cell) and the length of the moment arm. Torsional strength (Nm) and torsional stiffness (Nm per degree) were determined for each specimen by the methods of White et al.11 Torsional strength was defined as the maximum torque before the first abrupt drop in the torque

Table 1. Changes in Strength and Stiffness as a Function of Daily Exposure TimeHours of Exposure/Day 0.5 3 6

n Mean (SD) n Mean (SD) n Mean (SD)

Torque (Nm) 14 DaysSham 6 1.18 (0.35) 6 0.98 (0.51) 4 0.93 (0.17)CMF 8 1.13 (0.66) 8 1.34 (0.43) 8 1.58 (0.33)

PEMF 6 0.89 (0.28) 7 1.41 (0.29) 6 1.77 (0.23)

Torque (Nm) 21 DaysSham 6 2.10 (0.27) 6 2.1 (0.27) 6 2.10 (0.27)CMF 5 2.28 (0.23) 7 2.59 (0.41) 7 2.96 (0.40)

PEMF 4 2.50 (0.26) 7 2.59 (0.39) 7 3.14 (0.38)

Stiffness (Nm/Deg) 14 DaysSham 6 0.12 (0.05) 6 .12 (0.03) 4 0.10 (0.08)CMF 8 0.11 (0.05) 8 0.14 (0.09) 8 0.15 (0.04)

PEMF 6 0.09 (0.06) 7 0.13 (0.09) 6 0.16 (0.07)

Stiffness (Nm/Deg) 21 DaysSham 6 0.18 (0.04) 6 0.18 (0.04) 6 0.18 (0.04)CMF 6 0.18 (0.05) 7 0.18 (0.03) 7 0.20 (0.06)

PEMF 4 0.18 (0.04) 7 0.18 (0.06) 7 0.24 (0.09)


PEMF/CMF Exposure Time Effect on Healing

A

versus angular deformation curve. Stiffness, which is the slope of the rising portion of the curve, was estimated as the ratio of the torsional strength and the angle of failure. A statistical power analysis was performed based on previ-ous experience with rabbit tibia torsion tests. Assuming a variation of 13.25% for torsional strength and a target of 80% confidence, a sample size of six was found to be sufficient to identify a 20% difference between treatment groups. The difference between groups was analyzed by one-way analysis of variance and a Mann-Whitney test.

Biomechanical Stages of Fracture HealingPost torsional testing radiographs of the specimens

were taken to characterize the location of the failure site. The location of the failure fracture was combined with biomechanical stiffness data to assign one of four biomechanical stages of fracture repair to each tibia, as described by White et al.11 The four stages were defined as follows: Stage I - failure through the original fracture site with low stiffness; Stage II - failure through the original fracture site with high stiffness; Stage III - failure partially through the original fracture site and partially through intact bone, with high stiffness; and Stage IV - failure entirely through intact bone, with high stiffness.

RESULTSOf 111 rabbits used in this study, 97 were available for

assessment. Five rabbits died of unknown etiology (one sham, three PEMF and one CMF). Nine suffered second-ary fractures through the proximal pins prior to study completion (one sham, six PEMF and two CMF).

Biomechanical Testing

Torque versus angular deformation curves had a very

consistent shape for all specimens, including a highly linear portion that rose to a peak value and then dropped abruptly at the point of failure. Biomechanical results are presented in Tables 1 and 2 and Figures 1 and 2. At 14 days, osteotomies exposed for 30 minutes per day to either signal showed no increase in mean torsional strength over sham controls. Mean torsional strength values of both groups treated with PEMF or CMF for three hours per day were increased over sham controls, but this difference was not statistically significant. A significant increase in mean torsional strength was found in both signal-exposed groups treated for six hours per day when compared to sham controls. Comparison of torsional strengths between the PEMF and CMF treated groups showed no significant differences at any exposure time. The sham controls also showed no significant differences in mean torsional strength whether restrained for 30 minutes, three or six hours per day.

At 21 days postoperatively, tibial osteotomies treated with PEMF showed a mean torsional strength value significantly higher than sham controls for all three exposure times. Mean torsional strength in the CMF group was significantly higher than sham controls in the three and six hour per day exposure groups but not in the 30 minutes group. No significant differences were found between PEMF and CMF treatment groups at any exposure. Osteotomies treated for six hours per day with either PEMF or CMF were significantly stronger than those treated for ½ hour per day (p=0.016 PEMF, p=0.007 CMF). The PEMF group treated for six hours per day was also stronger than the three hour-treated group (p=0.02).

Increased stiffness over shams was seen in the three and six hour-treated groups at 14 days and in the six hour-treated groups at 21 days, although none reached significance of p<0.05 (Figure 2). Both PEMF and CMF six hour-treated groups at 21 days achieved torsional stiffness not significantly different from normal intact bone whereas shams had not. The stiffness and strength values in the 6 hour-treated groups are rising toward normal intact values regardless of treatment modality.

Radiological StudiesBoth CMF and PEMF signals stimulated increases in

mean periosteal callus area over non-stimulated shams at 14 days when used for 3 and 6 hours per day, but not for 30 minutes per day. Six hour daily exposures produced the largest callus areas. At 21 days, mean callus areas were increased over shams at all three daily exposure times to both signals with maximal differences seen in the groups exposed for three hours per day (Figure 3). There were no significant differences in the mean periosteal callus areas between CMF and PEMF-treated osteotomies under any of the experimental conditions.

Figure 1. Biomechanical testing of osteotomy healing showing tor-sional strength values as a function of treatment with PEMF or CMF signals at three different daily exposure times after 14 days (a) and 21 days (b) post-osteotomy.



A B C

Biomechanical Stages of Fracture HealingResults presented in Table 3 show that exposure to

either the CMF or the PEMF signals for three or six hours per day stimulated earlier appearance of the more mature fracture stages (III and IV) as compared to the sham controls. At 14 days sham controls and groups ex-posed to PEMF or CMF for 30 minutes per day fractured only through the osteotomy site (Stages I and II). At three hour daily exposures to either signal, increasing numbers of tibias reached Stage II, and one CMF and two PEMF fractured through adjacent bone (Stage III). With six hour daily treatment, 50% of the CMF and 67% of the PEMF group reached Stage III. No sham controls reached Stage III by 14 days. At 21 days no improvement over shams was observed for either group treated for 30 minutes per day. At three hours per day, 57% of both CMF and PEMF-treated osteotomies had reached Stage III as compared to 33% of the shams. With six hour daily exposures, 57% of the CMF and the PEMF groups reached Stage III compared to 33% of the shams. 14% of CMF and 29% of PEMF-treated groups fractured entirely through bone adjacent to the osteotomy (Stage IV). No sham controls reached Stage IV by 21 days.

DISCUSSIONThis study has demonstrated a clear dose response

correlating daily exposure time with the rate of healing of rabbit tibial osteotomies. Both the CMF and PEMF signals are “bioactive”, stimulating greater biomechanical

strength when used for six hours per day than when used for three hours or 30 minutes daily exposures. There were no significant differences in mean biomechanical strength between osteotomies treated with PEMF or CMF when exposed for similar time periods (30 minutes, three or six hours per day). Both signals stimulated increases in torsional stiffness but not to the same extent as torsional strength. This may indicate that the signals are chang-ing the composition of the healing bone resulting in an increased capacity to absorb energy prior to fracture (i.e. accelerated remodeling).

Increases in callus area and fracture healing matura-tion stage also showed a dose dependence on daily expo-sure time to both of the signals. As fractures advance from Stage III to IV, the torsion test changes from a measure of fracture gap tissue properties to a measure of the adjacent intact bone properties. This explains why the six hour-treated groups at 21 days achieved stiffness values not dif-ferent from intact bone. The fact that torsional strengths in these groups still fall short of intact bone might be due to changes in bone density associated with disuse.

Other studies using the PEMF or CMF signals have also shown a dose relationship to different daily exposure times. Factors such as signal intensity and pulse number can also have an influence on healing. Using the PEMF signal to enhance bone formation in a model of fracture callus formation in the rat, Aaron and Ciombor12 have shown increased stimulation of cartilage and bone in ani-mals treated for eight hours per day compared to those treated for four or one hour per day. Fitton-Jackson and Bassett13 showed that 45Ca uptake and synthesis of colla-gen by chick embryonic limb rudiments in vitro was also differentially stimulated by an increase in exposure time to the PEMF signal. By manipulating dosage, intensity, and duration of PEMF therapy, Girolamo et al.(Girolamo)

Table 2. Statistical Analysis of Differences in Torsional Strength (One Way Anova)

vs. CMF vs. Sham14 DAYS

0.5 Hr./Day PEMF p=0.436 p=0.153

CMF p=0.870

3 Hrs./Day PEMF p=0.749 p=0.083

CMF p=0.169

6 Hrs./Day PEMF p=0.449 p<0.001*

CMF p<0.004*

21 DAYS

0.5 Hr./Day PEMF p=0.216 p=0.047*

CMF p=0.267

3 Hrs./Day PEMF p=0.795 p=0.027*

CMF p=0.052*

6 Hrs./Day PEMF p=0.312 p<0.001*

CMF p<0.001*

*Statistically significant

Figure 2. Biomechanical testing of osteotomy healing showing stiffness as a function of treatment with PEMF or CMF signals at three different daily exposure times after 14 days (a) and 21 days (b) post-osteotomy.



found that specific conditions were most effective at enhancing cell viability and up-regulating the production of various growth factors in human tendon cells. Using PEMF therapy Li et al.15 discovered that, at 400 kV/m, the proliferation of osteoblasts increased from 0 to 400 pulses, but decreased once pulses reached 2800.

The CMF signal used to stimulate bone healing in a rabbit fibular osteotomy model16 showed that the greatest biomechanical stiffness resulted from 24 hour exposures with three and ½ hour daily exposures showing smaller increases over controls. Smith et al.17 treated eight day chick femoral rudiments in vitro for 30 minutes , one, four or 24 hour daily exposures to the CMF signal. Fe-mur width and diaphyseal collar length was greatest in rudiments treated for four hours per day, and diaphyseal collar was maximally thickened at one hour per day. The authors state that the results, “. . . clearly show that expo-sures of one or four hours per day seem to produce more marked results for most of the parameters than either 24 hours or 30 minutes per day.” There was a clear dose ef-fect with femora treated with minimal exposure time, 30 minutes, demonstrating little improvement over controls. In a castration induced rat model of osteoporosis, Magee et al.18 reported that exposure to the CMF signal for two hours per day (30 minutes four times per day) was more effective in inhibiting bone loss than treatment for 30 minutes per day.

The CMF studies reported by Ryaby et al.19 found that rat fibular fractures exposed to the CMF signal for 30 minutes per day showed no significant differences in torsional strength or stiffness in comparison to untreated controls after either 21 or 30 days. This corroborates find-

ings in the present study which also demonstrated limited stimulation at 30 minutes per day exposure to CMF.

A dose response relationship has also been reported for stimulation of proteoglycan synthesis20 and for osteotomy repair21 using electromagnetic signals other than the PEMF and CMF signals described in this study.

Interestingly, the osteogenic response in this osteoto-my model appears to be similar whether the rabbits were exposed to the PEMF or CMF signal for each of the doses, 30 minutes, three or six hours per day. The peak magnetic fields of the two signals differ from each other quite radi-cally; the CMF signal being about 40X smaller than the PEMF signal. In addition, the CMF signal contains a 76.6 Hz sine wave, whereas the PEMF signal is a pulse burst signal consisting of frequency components from 15 Hz to greater than 10 kHz. Although the frequency content of the two signals are very different, Fourier analysis of the PEMF signal reveals frequency components of less than 100 Hz (including one at 75 Hz) with amplitudes similar to that of the CMF signal. Since there is evidence suggesting that it is the low frequency components of the pulsed electromagnetic fields which are responsible for the osteogenic response,22 this may be the reason why both CMF and PEMF have similar effects on osteogenesis in the rabbit osteotomy model.

Both PEMF and CMF signals have been shown clini-cally to affect healing nonunion and delayed union frac-tures.4,5,8,23-26 Recommended daily use for the PEMF signal is ten hours per day, which has been reported to stimulate the fastest time to heal.20 While treating tibial delayed unions and non-union fractures, it was noted that increased union probability coincided with increased PEMF expo-sure; however, it was not statistically significant due to study limitations.27 Healing will be enhanced with shorter daily exposures to the PEMF signal, but the time to heal is longer. To our knowledge, there are no published clini-

Table 3. Effect of PEMF and CMF Exposure on the Four Biomechanical Stages of Fracture Healing

14 Days 21 Days

Fracture Stage I II III IV I II III IV

0.5 Hour/Day

Sham 2 3 - - - 4 2 -

CMF 4 4 - - - 4 2 -

PEMF 4 2 - - - 3 1 -

3 Hours/Day

Sham 4 2 - - - 4 2 -

CMF 2 5 1 - - 3 4 -

PEMF 1 4 2 - - 3 4 -

6 Hours/Day

Sham 3 1 - - - 4 2 -

CMF 1 3 4 - - 2 4 1

PEMF 1 1 4 - - 1 4 2

Figure 3. Radiographic analysis of the relative area of periosteal callus to diaphyseal bone as a function of daily exposure time to the PEMF or CMF signal 14 days (a) and 21 days (b) post-osteotomy.



cal studies using the CMF signal for more than one-half hour per day, so clinical efficacy with longer daily usage remains unknown.

The PEMF and CMF signals do not stimulate maximal osteogenic responses when used for ½ hour per day in this rabbit tibial osteotomy model or in the other animal models discussed above. Although none of these animal studies are nonunion models, these results suggest that longer exposures to either the PEMF or CMF signal may provide the optimal treatment regimen for obtaining fu-sion of nonunions.

ACKNOWLEDGMENTFunding was provided to the University of Iowa by a

grant from EBI Medical Systems, Inc., Parsippany, New Jersey.

REFERENCES1. American Medical Electronics Physio-Stimä, PMA

No. 850007, February 21, 1986.2. Bi-Osteogenâ System 204, PMA No. 790002, November

6, 1979.3. OrthoLogicâ 1000 Bone Growth Stimulator, PMA No.

910066, March 1994.4. Bassett CA, Mitchell SN, Gaston SR. Pulsing elec-

tromagnetic field treatment in ununited fractures and failed arthrodeses. JAMA. 1982 Feb 5; 247(5): 623-628.

5. Bassett CA, Mitchell SN, Gaston SR. Treatment of ununited tibial diaphyseal fractures with pulsing electromagnetic fields. J Bone Joint Surg Am. 1981 Apr; 63(4): 511-523.

6. Frykman GK, Taleisnik J, Peters G, Kaufman R, Helal B, Wood VE, Unsell RS. Treatment of nonunited scaphoid fractures by pulsed electromag-netic field and cast. J Hand Surg Am. 1986 May; 11(3): 344-349.

7. Garland DE, Moses B, Salyer W. Long-term follow-up of fracture nonunions treated with PEMFs. Contemporary orthopaedics JID - 8219527.

8. Perry CR, Zoltan JD, Gershuni DH, et al. In: Treatment of problem fractures using resonant mag-netic fields. Trans comb orthop mtg 63; ; 1992.

9. Pethica BA, Brownell J. In: The dose-response relationship in PEMF therapy of ununited fractures. Transactions of the eighth annual meeting of the bioelectrical repair and growth society; October 9-12; Washington, D.C. ; 1988.

10. Fredericks DC, Nepola JV, Baker JT, Abbott J, Simon B. Effects of pulsed electromagnetic fields on bone healing in a rabbit tibial osteotomy model. Journal of orthopaedic trauma JID - 8807705.

11. White AA, Panjabi MM, Southwick WO. The four biomechanical stages of fracture repair. J Bone Joint Surg Am. 1977 Mar; 59(2): 188-192.

12. Aaron RK, McCiombor DK. Acceleration of ex-perimental endochondral ossification by biophysical stimulation of the progenitor cell pool. J Orthop Res. 1996 07/01; 2018/12; 14(4): 582-589.

13. Fitton-Jackson S. The response of skeletal tissues to pulsed magnetic fields. Tissue Culture in Medical Research. 1980; 2: 21-28.

14. de Girolamo L, ViganÃ² M, Galliera E, Stanco D, Setti S, Marazzi MG, Thiebat G, Corsi Romanelli MM, Sansone V. In vitro functional response of hu-man tendon cells to different dosages of low-frequency pulsed electromagnetic field. Knee Surgery, Sports Traumatology, Arthroscopy. 2015 11/01; 23(11): 3443-3453.

15. Li K, Ma S, Li Y, Ding G, Teng Z, Liu J, Ren D, Guo Y, Ma L, Guo G. Effects of PEMF exposure at different pulses on osteogenesis of MC3T3-E1 cells. Arch Oral Biol. 2014; 59(9): 921-927.

16. Weinstein AM, McLeod BR, Smith SD, Liboff AR. In: Ion resonance turned electromagnetic fields increase healing rate in osteotomized rabbitts. Trans orthop res soc 36th annual meeting; 1990.

17. Smith SD, McLeod BR, Liboff AR. Effects of reso-nant magnetic fields on chick femoral development in vitro. J Bioelectr. 1991 01/01; 10(1-2): 81-99.

18. Magee FP, Weinstein AM, Fitzsimmons RJ, et. al. The use of low-energy combined AC and DC magnetic fields in the prevention of osteopenia. In: Brighton CT, Pollack SR, editors. Electromagnetics in Medicine and Biology. San Francisco: San Francisco Press; 1991. p. 171--176.

19. Ryaby JT, Huene D, Magee FP, Naser PR. In: Effects of combined AC/DC magnetic fields on heal-ing in a closed femoral fracture model. Trans orthop res soc 39th annual meeting; ; 1993.

20. Bee JA, Liu HX, Clarke N, Abbott J. Modulation of cartilage extracellular matrix turnover by pulsed electromagnetic fields (PEMF). In: Lyall F, El Haj AJ, editors. Biomechanics and Cells. Society for Ex-perimental Biology Seminar Series 54 ed. Cambridge: Cambridge University Press; 1994. p. 224--269.

21. Baker JT, Fredericks DC, Abbott J, Aper R, Brown T, Nepola JV. In: The effects of pulsed electromagnetic field (PEMF) stimulation on fresh fracture healing in the rabbit tibial osteotomy model. Trans orthop res soc 40th annual meeting; ; 1994.

22. McLeod KJ, Rubin CT. The effect of low-frequency electrical fields on osteogenesis. The Journal of bone and joint surgery.American volume JID - 0014030.



23. Sharrard WJ, Sutcliffe ML, Robson MJ, Maceachern AG. The treatment of fibrous non-union of fractures by pulsing electromagnetic stimulation. J Bone Joint Surg Br. 1982; 64(2): 189-193.

24. Sharrard WJ. A double-blind trial of pulsed electro-magnetic fields for delayed union of tibial fractures. J Bone Joint Surg Br. 1990 May; 72(3): 347-355.

25. Streit A, Watson BC, Granata JD, Philbin TM, Lin HN, O’Connor JP, Lin S. Effect on clinical out-come and growth factor synthesis with adjunctive use of pulsed electromagnetic fields for fifth metatarsal nonunion fracture: A double-blind randomized study. Foot Ankle Int. 2016 Sep; 37(9): 919-923.

26. Boopalan PRJVC, Chittaranjan SB, Balamuru-gan R, Nandakumar NS, Sabareeswaran A, Mohanty M. Pulsed electromagnetic field (PEMF) treatment for fracture healing. Ovid. 2009/07; 2009/08; 20(4); 423-428

27. Assiotis A, Sachinis NP, Chalidis BE. Pulsed elec-tromagnetic fields for the treatment of tibial delayed unions and nonunions. A prospective clinical study and review of the literature. Journal of Orthopaedic Surgery and Research. 2012 06/08; 7(1): 24.


ABSTRACTBackground: Hypoalbuminemia has been associ-

ated with several medical complications following surgery in a variety of orthopedic procedures. Hypoalbuminemia has previously been shown to have an increased risk for transfusions, hospital stay longer than three days, and mortality following total shoulder arthroplasty (TSA). This study seeks to further assess the relationship between low se-rum albumin and morbidity to allow surgeons to both preoperatively optimize patients and assess the risk of surgery prior to TSA.

Methods: The American College of Surgeons National Surgical Quality Improvement Program® database was queried to identify 14,494 TSA pa-tients, 6,129 (42.23%) who met inclusion criteria. Patients who had shoulder hemiarthroplasty, revi-sion TSA, or incomplete serum albumin data were excluded. Demographic factors, preoperative co-morbidities, and acute complication rates were as-sessed between hypoalbuminemic (n=485; 7.91%) and a propensity-matched control cohort (n=485), controlling for differences in patient demographics and comorbidities. Multivariate propensity-adjusted logistic regression analyses were used to assess hypoalbuminemia as an independent risk factor for specific postoperative complications.

Results: Hypoalbuminemic patients undergoing TSA demonstrated significantly higher rates of pulmonary complications (p=0.006), unplanned intubation (p=0.014), DVT/PE (p=0.014), cardiac complications (p=0.033), infectious complications (p=0.025), blood transfusions (p<0.001), reopera-tion (p=0.007), extended length of stay (> 4 days) (p=0.036), unplanned readmission (p=0.001), and mortality (p=0.025) in the 30-day postoperative

period when compared to the propensity-matched control cohort. On multivariate regression analy-ses, hypoalbuminemia independently increased the risk for pulmonary complications (OR 9.678, p=0.031), blood transfusions (OR 2.539, p<0.001), reoperation (OR 5.461, p=0.032), and readmission (OR 2.607, p=0.007).

Conclusions: Hypoalbuminemic patients under-going TSA had increased rates of overall cardiac and pulmonary complications, unplanned intuba-tions, DVT/PE’s, overall infectious complications, increased incidence of blood transfusions, reop-eration, extended LOS (> 4 days), readmission, and death. Multivariate analyses demonstrated that low albumin was independently associated with increased risk for pulmonary complications, blood transfusions, reoperation, and readmission. Preoperative albumin levels in patients undergoing TSA may help with preoperative risk stratification and optimization.

Level of evidence: IIIKeywords: reoperation, mortality, pulmonary,

cardiac, albumin, hypoalbuminemia, total shoul-der arthroplasty, readmission, blood transfusion, unplanned intubation, complication

INTRODUCTIONThe use of total shoulder arthroplasty (TSA) for the

treatment of osteoarthritis has increased significantly from an incidence rate of 6.1 per 100,000 patients in 2005 to 13.4 per 100,000 patients in 2013.1-3 As its popularity increases,4-5 surgeons must better be able to anticipate and manage complications associated with TSA.6 Though TSA has comparatively lower overall complication rates than hip and knee arthroplasty,7-8 these complications can drastically lower quality of life for patients and increase healthcare costs. Therefore, it is critical to identify and potentially modify risk factors in order to reduce such complications.

Hypoalbuminemia has previously been reported as a risk factor for various complications in orthopedic surgi-cal patients, including infection and impaired wound heal-ing.9-10 Hypoalbuminemia has numerous etiologies and is most commonly used as a marker of malnutrition, which has been previously associated with poor outcomes in or-

LOW SERUM ALBUMIN LEVELS ARE ASSOCIATED WITH INCREASED 30-DAY CARDIOPULMONARY COMPLICATIONS,

REOPERATION, AND READMISSION RATES FOLLOWING TOTAL SHOULDER ARTHROPLASTY

Danny Lee BS1; Ryan Lee MBA1; Megan T. Cross BS1; Andrew Tran MD2; Jason Kappa MD2; Sam Moghtaderi MD2

1The George Washington University School of Medicine and Health Sciences; The George Washington University2Department of Orthopaedic Surgery; The George Washington UniversityCorresponding author: Danny Lee; Phone (201) 921-0659; [email protected]: The authors report no potential conflicts of interest related to this study.Sources of Funding: No sources of funding declared.

D. Lee, R. Lee, M.T. Cross, A. Tran, J. Kappa, S. Moghtaderi


thopedic patients.11-13 Garcia et al. previously reported the effect of malnutrition in 1,681 TSA patients from 2005 to 2013 using the American College of Surgeons maintained National Surgical Quality Improvement Program (ACS-NSQIP) database, and demonstrated hypoalbuminemia to be an independent risk factor for blood transfusions, extended length of stay beyond 3 days, and increased mortality rate.14 Since then, the number of available TSA cases in the ACS-NSQIP database with complete albumin values has increased approximately four-fold. This expansion of data warrants an update to depict any existing association between hypoalbuminemia and surgi-cal complications following TSA that may have not been previously reported.

The present study uses the ACS-NSQIP database to examine the effect of hypoalbuminemia on short-term perioperative and postoperative complications in patients undergoing primary TSA (both anatomic and reverse arthroplasty). Specifically, this study seeks to determine differences in patient characteristics for those with hypoalbuminemia and normal albumin levels, determine whether low albumin levels were associated with increased rates of certain complications, and further establish hypoalbuminemia as an independent risk factor for specific postoperative complications following TSA.

METHODSPatient Selection

The ACS-NSQIP database was queried for all patients who had undergone primary TSA (Current Procedure Terminology CPT code 23472) from 2005 to 2016.15 The study was exempt from Institutional Review Board ap-proval as it utilized patient data from a publicly available, deidentified database maintained by the ACS. Patients at least 18 years of age who had complete serum albumin concentration levels were included in the study. Hypoalbu-minemia was defined as a serum albumin level of less than 3.5 g/dL. Patients were stratified into two cohorts—those with normal albumin levels and those with low serum albumin. Patients who had undergone hemiarthroplasty procedures and revision shoulder arthroplasty procedures, corresponding to CPT codes 23470 and 23473/23474 re-spectively, were excluded from this study. Patients were also excluded if their serum albumin concentration was missing or unknown.

VariablesDemographic factors such as age, sex, race, and body

mass index (BMI) were included for analysis. Preopera-tive comorbidities that were analyzed include smoking history, diabetes mellitus, dyspnea, chronic obstructive pulmonary disease (COPD), ascites, ventilator dependence, congestive heart failure (CHF), hypertension requiring

medications, acute renal failure, dialysis, disseminated cancer, wound infections, steroid use for chronic condi-tions, weight loss of more than 10% of the patient’s body weight in the 6 months leading up to the surgery, hemato-logic disorders, preoperative blood transfusions, systemic sepsis, and preoperative functional status. Additionally, serum sodium levels, blood urea nitrogen, creatinine, albu-min, white blood cell count, hematocrit level, and platelet count were included.

The following outcome variables were assessed: superficial incisional surgical site infections (SSI), deep incisional SSI, organ/space SSI, pneumonia, unplanned intubation, pulmonary embolism (PE), ventilator de-pendence for greater than 48 hours, progressive renal insufficiency, acute renal failure, urinary tract infection (UTI), stroke, cardiac arrest requiring cardiopulmonary resuscitation (CPR), deep venous thromboembolism (DVT), myocardial infarction (MI), blood transfusion requirements, systemic sepsis, septic shock, return to the operating room, extended LOS (length of stay) (≥ 4 days), unplanned readmission, and death. Variables where more than 10% of patients had missing information were excluded from analysis.

Statistical AnalysisDifferences in demographics, preoperative comorbidi-

ties between hypoalbuminemic patients and those with normal albumin levels were assessed using Pearson’s chi-squared tests, Fischer’s exact tests, and one-way analysis of variance (ANOVA). Further analysis was completed by using propensity score matching without replacement to generate a matched control in a 1:1 man-ner with randomized case-drawing for comparison with the 485 hypoalbuminemic TSA patients. This matched control cohort of 485 patients with normal albumin levels, controlled for significant patient demographic factors and pre-operative comorbidities, was analyzed against hypo-albuminemic patients using Pearson’s Chi-Squared tests for categorical variables. ANOVA was used for continuous variables, such as age and laboratory values, to evaluate for differences in incidence rates and means, respectively, for each variable of interest. Multivariate propensity-adjusted logistic regression models were generated to assess hypoalbuminemia’s role as an independent risk factor for specific 30-day complications following primary TSA. Odds ratios (OR) and 95% confidence intervals (CI) were calculated for each postoperative complication to quantify the risk association. An alpha of 0.05 was used to assess significance in univariate and multivariate analyses. All statistical analyses were performed using the IBM® SPSS® Statistics Version 25 software. (IBM Corp, Armonk, NY, USA)


Low Albumin’s Association with Complications in TSA

A

Table 1. Patient Demographics and Preoperative Comorbidities Stratified by Serum Albumin

Normal Albumin PM-Matched Control Hypoalbuminemia P-Value PM P-Value

DEMOGRAPHICS n=5644 (%) n=485 (%) n=485 (%)

Age (Mean ± SD)a 68.86 ± 9.820 70.26 ± 9.442 71.24 ± 10.388 <0.001 0.125

Sex <0.001 0.005

Female 3102 54.96% 318 65.57% 358 73.81%

Male 2540 45.00% 167 34.43% 127 26.19%

Race <0.001 0.179

American Indian/Alaskan Native 35 0.62% 2 0.41% 1 0.21%

Asian or Pacific Islander 35 0.62% 3 0.62% 0 0.00%

Black or African American 248 4.39% 29 5.98% 37 7.63%

Hispanic 2 0.04% 0 0.00% 1 0.21%

White 4965 87.97% 424 87.42% 429 88.45%

Unknown 359 6.36% 27 5.57% 17 3.51%

PRE-OPERATIVE COMORBIDITIES

Smoking History 599 10.61% 52 10.72% 66 13.61% 0.042 0.169

Diabetes Mellitus 1021 18.09% 93 19.18% 113 23.30% 0.005 0.116

Dyspnea <0.001 0.012

No Dyspnea 5229 92.65% 442 91.13% 413 85.15%

At Rest 19 0.34% 1 0.21% 4 0.82%

Moderate Exertion 396 7.02% 42 8.66% 68 14.02%

COPD 367 6.50% 38 7.84% 70 14.43% <0.001 0.001

Ascites 1 0.02% 0 0.00% 1 0.21% 0.152 1.000

Congestive Heart Failure 31 0.55% 3 0.62% 11 2.27% <0.001 0.031

Hypertension 3840 68.04% 335 69.07% 354 72.99% 0.024 0.179

Acute Renal Failure 3 0.05% 0 0.00% 2 0.41% 0.053 0.499

Disseminated Cancer 16 0.28% 4 0.82% 5 1.03% 0.021 1.000

Wound Infection 38 0.67% 5 1.03% 9 1.86% 0.010 0.282

Steroid Use 317 5.62% 38 7.84% 56 11.55% <0.001 0.051

Bleeding Disorder 158 2.80% 25 5.15% 24 4.95% 0.007 0.883

Blood Transfusions 12 0.21% 8 1.65% 9 1.86% <0.001 0.807

Functional Status <0.001 <0.001

Independent 5456 96.67% 465 95.88% 436 89.90%

Partially Dependent 127 2.25% 13 2.68% 42 8.66%

Totally Dependent 6 0.11% 1 0.21% 5 1.03%

LABORATORY VALUESa

Sodium (mEq/L) 139.32 ± 3.073 139.21 ± 3.092 138.61 ± 3.776 <0.001 0.007

Blood Urea Nitrogen (mg/dL) 18.47 ± 7.165 18.71 ± 7.810 19.15 ± 10.314 0.057 0.459

Creatinine (mg/dL) 0.96 ± 0.584 0.96 ± 0.545 1.06 ± 0.879 <0.001 0.026

Albumin (g/dL) 4.11 ± 0.355 4.12 ± 0.349 3.12 ± 0.367 <0.001 <0.001

White Blood Cells (103 c/mL) 7.20 ± 2.524 7.35 ± 2.207 7.86 ± 2.609 <0.001 0.001

Hematocrit (%) 40.63 ± 4.479 40.02 ± 4.495 36.37 ± 5.298 <0.001 <0.001

Platelets (per mL) 243.59 ± 69.043 243.63 ± 67.250 245.32 ± 96.123 0.614 0.753

PM: Propensity Matched; SD: standard deviation; COPD: chronic obstructive pulmonary disease;aValues expressed as Mean ± Standard Deviation.All other values expressed as (%) and N. Bold indicates statistical significance.



A B C

RESULTSA total of 14,494 patients underwent TSA between

2005 and 2016. 6,129 patients met inclusion criteria for study, of which 485 patients (7.91%) had hypoalbumin-emia as defined by a serum albumin concentration of less than 3.5 g/dL. A significantly higher proportion of females were found to have hypoalbuminemia (73.81%) compared to the normal albumin cohort (54.96%; p<0.001). Univariate analysis also yielded significant differences in race between the two cohorts (p=0.003; Table 1). The hypoalbuminemia cohort consisted of significantly more individuals who identified as “White” (88.45%) and African-American (7.63%) than the control cohort. Patients with hypoalbuminemia had higher rates of all comorbidities analyzed, with the exception of excessive recent weight loss. The mean serum albumin levels be-tween the two cohorts were also significantly different; the low albumin patient cohort had a mean albumin level of 3.12 g/dL, while the normal cohort had a mean of 4.11 g/dL (p<0.001; Table 1).

When comparing the propensity-matched control cohort and the hypoalbuminemic patients, the hypoal-buminemia cohort consisted of a significantly larger proportion of female patients (73.81% vs. 65.57%; p=0.005). The hypoalbuminemic TSA patients also demonstrated significantly higher rates of dyspnea (85.15% vs. 91.13%; p=0.012), COPD (14.43% vs. 7.84%; p=0.001), CHF (2.27% vs. 0.62%; p=0.031), and functional dependence (9.69% vs. 2.89%; p<0.001) in comparison to the matched control cohort (Table 1).

Hypoalbuminemic patients experienced a significantly longer mean LOS from operation to discharge (2.75 days vs. 2.20 days; p<0.001) and a significantly longer total length of hospital stay from admission to discharge (3.46 days vs. 2.26 days; p<0.001) than the matched control cohort. However, the mean number of days between admission and operation were not significantly different (p=0.894). The mean operative time (p=0.848) and time under anesthesia (p=0.921) were not significantly differ-ent between hypoalbuminemic patients and those with normal albumin levels (Table 2).

Patients with hypoalbuminemia had significantly higher rates of numerous 30-day postoperative com-plications. Hypoalbuminemic patients demonstrated higher rates of pulmonary complications (2.06% vs. 0.21%; p=0.006), unplanned intubation (1.24% vs. 0.00%; p=0.014), DVT/PE (1.24% vs. 0.00%; p=0.014), cardiac complications (1.44% vs. 0.21%; p=0.033), and infectious complications (1.03% vs. 0.00%; p=0.025). These patients also demonstrat-ed significantly higher rates intraoperative/postopera-tive blood transfusions (14.43% vs. 5.36%; p<0.001), return to the operating room (2.47% vs. 0.41%; p=0.007), extended LOS (> 4 days) following the procedure (16.91% vs. 12.16%; p=0.036), unplanned readmission (6.80% vs. 2.47%; p=0.001),

and mortality (1.03% vs. 0.00%; p=0.025; Table 3).Multivariate propensity-adjusted logistic regression

analysis showed hypoalbuminemia to be independently as-sociated with increased odds of pulmonary complications (OR 9.678, 95% CI 1.223-76.569, p=0.031), intraoperative/postoperative blood transfusions (OR 2.539, 95% CI 1.562-4.125, p<0.001), reoperation (OR 5.461, 95% CI 1.158-25.745, p=0.032), and unplanned readmission (OR 2.607, 95% CI 1.292-5.259, p=0.007). (Table 4)

DISCUSSIONThis study demonstrated increased rates of overall car-

diac and pulmonary complications, unplanned intubations, DVT/PE’s, overall infectious complications, increased incidence of blood transfusions, reoperation, extended LOS (> 4 days), readmission, and death in patients with hypoalbuminemia. Given these potential complications, a thorough evaluation of patients found to have low serum albumin is warranted. Patients should be appropriately referred to nutritional services for preoperative optimiza-tion if the hypoalbuminemia is secondary to nutritional deficiencies prior to TSA.

Albumin comprises more than 50% of whole blood and has been shown to carry out numerous responsibilities. Re-cent literature has shown albumin’s role to extend beyond that of a carrier protein with its anticoagulant properties. While the mechanism is incompletely understood, low albumin can lead to a hypercoagulable state, potentially causing increased rates of DVT and subsequent PE.16-20 In this study, low albumin was associated with an increased rate of DVT/PE, which has not been demonstrated in previous analyses of TSA..14

This study found increased rates of required blood

Table 2. Perioperative Time Variables Stratified by Serum Albumin Levels

OPERATIVE VARIABLES

PM-Generated Control Cohort

Hypoalbumin-emia

PM P-Value

aTime Under Anesthesia (min) 191.98 ± 62.688 193.39 ± 75.517 0.912

aOperative Time (min) 113.22 ± 44.736 113.75 ± 42.027 0.848

aTotal Length of Hospital Stay (days)

2.26 ± 1.981 3.46 ± 3.668 <0.001

aDays to Operation from Admission 0.81 ± 16.579 0.71 ± 2.112 0.894

aDays to Discharge from Operation

2.20 ± 1.889 2.75 ± 2.211 <0.001

PM: Propensity-matched;aValues expressed as Mean ± Standard DeviationBold indicates statistical significance.



transfusion in TSA patients intraoperatively and post-operatively in patients with low serum albumin. These results support previous studies on transfusion rates in hypoalbuminemic patients undergoing TSA and revision total knee arthroplasty (TKA).10-14 Hypoalbuminemia can serve as a surrogate marker for chronic kidney disease and chronic inflammatory states, both of which have the potential to cause preoperative anemia and subse-quently, higher rates of transfusion.21-23 The need for blood transfusions is likely multifactorial and individual to each patient, necessitating further research to clarify the relationship between anemia requiring transfusions and low albumin levels.

Low albumin has often been used as a surrogate marker for overall poorer health. Similar to Garcia et al., this study also found increased rates of pulmonary complica-tions in hypoalbuminemic patients undergoing primary TSA.14 This study additionally found an increased rate of unplanned intubation in the hypoalbuminemic group

as well.14 This finding is in line with other reports of increased rates of unplanned intubation in hypoalbumin-emic patients undergoing surgical treatment for geriatric hip fractures and revision TKA.10,24 Albumin helps main-tain intravascular volume, prevents fluid extravasation, and supports the vascular endothelial surface.25-26 It is plausible that decreased albumin levels could perturb a tenacious fluid homeostasis, potentially leading to adverse clinical complications.

Reports in the cardiac literature indicate that hy-poalbuminemia is strongly associated with increased rates of coronary disease, MI, and other cardiac events in the general population.27-30 This study found increased rates of cardiac complications in patients with low albumin, an association previously unreported in patients undergoing TSA.14 The exact mechanism behind hypoalbuminemia’s role in causing various cardiac complications, such as MI or heart failure is, as of yet, incompletely understood. However, the results of this study indicate that low albu-

Table 3. Adjusted 30-Day Postoperative Complication Analysis by Albumin Status

POST-OPERATIVE COMPLICATIONPM-Control Cohort Hypoalbuminemia PM P-Value

n=485 n=485

Pulmonary 1 0.21% 10 2.06% 0.006

Pneumonia 1 0.21% 6 1.24% 0.058

Unplanned Intubation 0 0.00% 6 1.24% 0.014

Ventilator Dependence (>48 hours) 0 0.00% 2 0.41% 0.157

Vascular 2 0.41% 7 1.44% 0.094

DVT/PE 0 0.00% 6 1.24% 0.014

CVA/Stroke 2 0.41% 1 0.21% 0.563

Cardiac 1 0.21% 7 1.44% 0.033

Cardiac Arrest 0 0.00% 2 0.41% 0.157

Myocardial Infarction 1 0.21% 5 1.03% 0.101

Infectious 0 0.00% 5 1.03% 0.025

Sepsis 0 0.00% 3 0.62% 0.083

Septic Shock 0 0.00% 1 0.21% 0.317

Surgical Site Infections 0 0.00% 1 0.21% 0.317

Renal 6 1.24% 9 1.86% 0.435

UTI 6 1.24% 8 1.65% 0.590

Renal Insufficiency 0 0.00% 1 0.21% 0.317

Blood Transfusionsa 26 5.36% 70 14.43% <0.001

Reoperation 2 0.41% 12 2.47% 0.007

Extended LOS (≥ 4 days) 59 12.16% 82 16.91% 0.036

Unplanned Readmission 12 2.47% 33 6.80% 0.001

Mortality 0 0.00% 5 1.03% 0.025

PM: Propensity-matched; DVT: deep venous thromboembolism; PE: pulmonary embolism; CVA: cerebral vascular accident; UTI: urinary tract infection; LOS: length of stay;aWithin 72 hours of starting operationBold indicates statistical significance.



min level may be useful as an indicator or screening tool for further work up or preoperative optimization to help mitigate the effects of cardiac postoperative complica-tions.

This current study found an overall increase in infec-tious complications in the low albumin cohort. Hypoalbu-minemia has been reported to be associated with increased infection rates in a broad spectrum of clinical settings.31,32 Similar to this study, a recent meta-analysis analyzed patients with albumin levels <3.5 g/dL and demonstrated a nearly 2.5-fold increased risk of SSI in orthopedic proce-dures.33 This study also found increased rates of extended LOS, reoperation, readmission, and mortality in patients with low albumin. Low albumin levels have been well associated with poor healing reserves in the literature, likely secondary to poor nutritional status.34 This poor healing capability may contribute to the increased infec-tion and reoperation rates. A limitation inherent to this study is that the ACS-NSQIP database does not delineate the reason for readmissions. Albumin levels as a surrogate marker for nutritional status may be overutilized as tools such as the Mini Nutritional Assessment and Subjective Global Assessment have proven to be effective in evaluat-

ing nutritional status.35,36 However, as the results of the current study demonstrate, the importance of low albumin levels as a potential marker for an arduous postoperative course in primary TSA cannot be overlooked. Wilson et al. suggest that obtaining an albumin level at admission may prove beneficial in orthopedic lower extremity trauma patients due to the high rate of readmission, reoperation, and overall postoperative complications.37

Similar to Garcia et al.,14 our study demonstrated that patients with hypoalbuminemia were at increased risk of blood transfusions in the multivariate logistic regression models. Our study did not demonstrate hypoalbuminemia to be an independent risk factor for mortality and extend-ed LOS (> 4 days) in the postoperative period following TSA.14 We also additionally report hypoalbuminemia to be an independent risk factor for reoperation, readmission, and pulmonary complications – relationships previously unidentified in TSA patients.14 Further work investigat-ing the role of hypoalbuminemia and its potential role in increased risk of mortality following TSA is warranted to better risk-stratify patients preoperatively.

The current study has several limitations. Only vari-ables that were recorded for more than 85% of the patients

Table 4. Multivariable Propensity-Adjusted Logistic Regression Analyses Assessing Hypoalbuminemia as a Risk Factor for Postoperative Complications

POST-OPERATIVE COMPLICATION P-Value Odds Ratio 95% CI

Pulmonary 0.031 9.678 1.223 76.569Pneumonia 0.102 5.922 0.704 49.811Unplanned Intubation * 0.993 >999.99 0.000 -Ventilator Dependence (>48 hours) * 0.993 >999.99 0.000 -

Vascular 0.133 3.41 0.687 16.920DVT/PE * 0.992 >999.99 0.000 -CVA/Stroke 0.474 2.582 0.192 34.653

Cardiac 0.081 6.579 0.791 54.714Cardiac Arrest * 0.993 >999.99 0.000 -Myocardial Infarction 0.203 4.159 0.464 37.304

Infectious * 0.992 >999.99 0.000 -Sepsis * 0.993 >999.99 0.000 -Septic Shock* 0.992 >999.99 0.000 -Surgical Site Infections * 0.992 >999.99 0.000 -

Renal 0.683 1.251 0.427 3.662UTI 0.889 1.082 0.358 3.267Renal Insufficiency 0.993 >999.99 0.000 -

Blood Transfusionsa <0.001 2.539 1.562 4.125Reoperation 0.032 5.461 1.158 25.745Extended LOS (≥ 4 days) 0.250 1.246 0.857 1.812Unplanned Readmission 0.007 2.607 1.292 5.259Mortality 0.993 >999.99 0.000

CI: Confidence Interval; DVT: deep venous thromboembolism; PE: pulmonary embolism; CVA: cerebral vascular accident; UTI: urinary tract infec-tion; LOS: length of stay;*Had an incidence rate of 0.00% in the PM-generated control cohort (Table 3)aWithin 72 hours of starting operationBold indicates statistical significance.



were included in this study. A limitation inherent to all clinical registries is the lack of verification of data entries that may influence analyses. The ACS-NSQIP database only reports outcomes within a 30-day surveillance period. Recording adverse events and complications for a longer time can provide further insight into longer term out-comes. Another limitation of this study lies in the exclu-sion of variables that did not have at least a 90% reporting rate. For example, pneumonia and angina were excluded. While the exclusion helps the analyses presented in the study remain as accurate as possible, excluding such variables may lead to unidentified relationships that may be in fact significant. A dependence on CPT/ICD-9 codes is another limitation as these codes were not originally designed for data analysis and could be subject to coding bias with financial incentives or coding errors.38 Despite propensity-score matching and multivariate analysis, the effect of confounding cannot be excluded given the large number of variables that may increase risk of complica-tions in TSA. Finally, all patients who had undergone TSA were included in this retrospective study. The present study did not distinguish patients based on their need for an anatomic total shoulder arthroplasty versus a reverse design as the CPT code is the same for both procedures.

CONCLUSIONHypoalbuminemic TSA patients had increased rates of

overall cardiac and pulmonary complications, unplanned intubations, DVT/PE’s, overall infectious complications, increased incidence of blood transfusions, reoperation, extended LOS (> 4 days), readmission, and death. Multi-variate analyses demonstrated low albumin to indepen-dently increase the risk for pulmonary complications, blood transfusions, reoperation, and readmission. Shoulder surgeons may wish to proceed with caution when deciding on elective TSA for hypoalbuminemic patients and defer treatment until medical or nutritional status are optimi-zated. Future research into identifying a minimum serum albumin level could prove useful in establishing guidelines in pre-operative counseling in this specific population of patients suffering from degenerative shoulder disease requiring TSA.

REFERENCES1. Dillon MT, Chan PH, Inacio MCS, et al. Yearly

Trends in Elective Shoulder Arthroplasty, 2005-2013. Arthritis Care (Hoboken). 2017;69:1574-1581.

2. Triplet JJ, Everding NG, Levy JC, et al. Anatomic and Reverse total shoulder arthroplasty in patients older than 80 years. Orthopedics. 2015;38:e904-10.

3. Levy JC, Everding NG, Gil CC Jr, et al. Speed of recovery after shoulder arthroplasty: a comparison of reverse and anatomic total shoulder arthroplasty. J Shoulder Elbow Surg. 2014;23:1872-1881.

4. Issa K, Pierce CM, Pierce TP, et al. Shoulder Arthroplasty Demographics, Incidence, and Complica-tions-A Nationwide Inpatient Sample Database Study. Surg Technol Int. 2016;XXIX:240-246.

5. Kim SH, Wise BL, Zhang Y, et al. Increasing in-cidence of shoulder arthroplasty in the United States. J Bone Joint Surg Am. 2011;93:2249-54.

6. Aldinger PR, Raiss P, Rickert M, et al. Complica-tions in shoulder arthroplasty: an analysis of 485 cases. Int Orthop. 2010;34:517-524.

7. Shields E, Iannuzzi JC, Thorsness R, et al. Peri-operative complications after hemiarthroplasty and total shoulder arthroplasty are equivalent. J Shoulder Elbow Surg. 2014;23:1449-53.

8. Fegringer EV, Mikuls TR, Michaud KD, et al. Shoulder arthroplasties have fewer complications than hip or knee arthroplasties in US veterans. Clin Orthop Relat Res. 2010;468:717-22.

9. Yuwen P, Chen W, Lv H, et al. Albumin and surgi-cal site infection risk in orthopaedics: a meta-analysis. BMC Surg. 2017;17:7.

10. Kamath AF, Nelson CL, Elkassabany N, et al. Low albumin is a risk factor for complications after revision total knee arthroplasty. J Knee Surg. 2017;30:269-275.

11. Cross MB, Yi PH, Thomas CF, et al. Evaluation of malnutrition in orthopaedic surgery. J Am Acad Orthop Surg. 2014; 22:193–199.

12. Jensen JE, Jensen TG, Smith TK, et al. Nutri-tion in orthopaedic surgery. J Bone Joint Surg Am. 1982;64:1263–1272.

13. Pratt WB, Veitch JM, McRoberts RL. Nutritional status of orthopedic patients with surgical complica-tions. Clin Orthop Relat Res. 1981;155:81–84.

14. Garcia GH, Fu MC, Dines DM, et al. Malnutri-tion: a marker for increased complications, mortality, and length of stay after total shoulder arthroplasty. J Shoulder Elbow Surg. 2016;25:193-200.

15. American College of Surgeons National Surgical Qual-ity Improvement Program. User guide for the 2016 participant use data file. (Available at: https://www.facs.org/~/media/files/quality%20programs/nsqip/nsqip_puf_userguide_2016.ashx); 2018.

16. Galanakis DK. Anticoagulant albumin fragments that bind to fibrinogen/fibrin: possible implications. Semin Thromb Hemost. 1992;18:44-52.

17. Paar M, Rossmann C, Nusshold C, et al. Antico-agulant action of low, physiologic, and high albumin levels in whole blood. PLoS One. 2017;12(8):e0182997.

18. Lam FW, Cruz MA, Leung HC, et al. Histone induced platelet aggregation is inhibited by normal albumin. Thromb Res. 2013;132:69-76.



19. Poteser M, Wakabayashi I. Serum albumin induces iNOS expression and NO production in RAW 267.4 macrophages. Br J Pharmacol. 2004;143:143-151.

20. Maclouf J, Kindahl H, Granström E, et al. Interactions of prostaglandin H2 and thromboxane A2 with human serum albumin. Eur J Biochem. 1980;109(2):561-6.

21. Furth SL, Cole SR, Fadrowski JJ, et al. The association of anemia and hypoalbuminemia with accelerated decline in GFR among adolescents with chronic kidney disease

22. Babitt JL, Lin HY. Mechanisms of Anemia in CKD. J Am Soc Nephrol. 2012;23:1631-1634.

23. Kaysen GA. Biochemistry and Biomarkers of In-flamed Patients: Why Look, What to Assess. Clin J Am Soc Nephrol. 2009;4 Suppl 1:S56-63.

24. Bohl DD, Shen MR, Hannon CP, et al. Serum Albumin Predicts Survival and Postoperative Course Following Surgery for Geriatric Hip Fracture. J Bone Joint Surg Am. 2017;99:2110-2118.

25. Arques S, Ambrosi P. Human serum albumin in the clinical syndrome of heart failure. J Card Fail. 2011;17:451-8.

26. Purvey M, Allen G. Managing acute pulmonary oedema. Aust Prescr. 2017;40(2):59-63.

27. Danesh J, Collins R, Appleby P, et al. Association of fibrinogen, C-reactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies. JAMA. 1998;279:1477-82.

28. Gopal D, Kalogeropoulos AP, Georgiopoulou VV, et al. Serum Albumin Concentration and Heart Failure Risk. The Health, Aging, and Body Composi-tion Study. Am Heart J. 2010;160:279-285.

29. Nelson JJ, Liao D, Sharrett AR, et al. Serum al-bumin level as a predictor of incident coronary heart disease: the Atherosclerosis Risk in Communities (ARIC) study. Am J Epidemiol. 2000;151:468-77.

30. Yang Q, He YM, Cai DP, et al. Risk burdens of mod-ifiable risk factors incorporating lipoprotein (a) and low serum albumin concentrations for first incident acute myocardial infarction. Sci Rep. 2016;6:35463.

31. Minatoguchi S, Nomura A, Imaizumi T, et al. Low serum albumin as a risk factor for infection-related-in-hospital death among hemodialysis patients hos-pitalized on suspicion of infectious disease: a Japanese multicenter retrospective cohort study. Ren Replace Ther. 2018;4:30.

32. Yang C, Liu Z, Tian M, et al. Relationship Between Serum Albumin Levels and Infections in Newborn Late Preterm Infants. Med Sci Monit. 2016;22:92-B.

33. Yuwen P, Chen W, Lv H, et al. Albumin and surgi-cal site infection risk in orthopaedics: a meta-analysis. BMC Surg. 2017;17:7.

34. Guo S, DiPietro LA. Factors Affecting Wound Heal-ing. J Dent Res. 2010;89(3):219-229.

35. Vellas B, Guigoz Y, Garry PJ, et al. The Mini Nutritional Assessment (MNA) and its use in grad-ing the nutritional state of elderly patients. Nutrition. 1999;15:116-22.

36. Fontes D, Generoso Sde V, Toulson Davisson Correia MI. Subjective global assessment: a reliable nutritional assessment tool to predict outcomes in critically ill patients. Clin Nutr. 2014;33:291-5.

37. Wilson J, Lunati M, Grabel Z, et al. Hypoalbu-minemia is an Independent Risk Factor for 30-Day Mortality, Postoperative Complications, Readmission, and Reoperation in the Operative Lower Extremity Orthopedic Trauma Patient. J Orthop Trauma. 2019. doi: 10.1097/BOT.0000000000001448. [Epub ahead of print]

38. Yoshihara H, Yoneoka D. Understanding the statistics and limitations of large database analyses. Spine (Phila Pa 1976). 2014;39:1311-1312.


ABSTRACTBackground: “Turf toe” results from hyperdor-

siflexion of the first metatarsophalangeal joint, injuring the plantar capsuloligamentous complex. We hypothesized that National Football League (NFL) player performance following turf toe injury would decrease in comparison to controls at the same position.

Methods: Demographics, return to play, and season performance data on players sustaining turf toe injuries in the NFL from 2010-2015 were collected. An Offensive Power Rating (OPR=[total yards/10]+[total touchdowns x6]) or Defensive Power Rating (DPR=total tackles+[total sacks x2]+[total interceptions x2]) was calculated for each player. Control data were collected for NFL players in 2013 with no history of turf toe injury. Statistical analysis was performed using Wilcoxon Rank Sum tests.

Results: Twenty-four injured players and 436 controls were included. Nineteen players returned to play within the regular season of injury (mean 36.7 ± 28.9 days). Seventeen players were re-moved from team injury reports for turf toe within the regular season (mean 42.6 ± 26.2 days). Three players required season-ending surgery. Compari-son of 1-year post- versus pre-injury revealed an insignificant median OPR difference (-18.9 IQR -43.4 to 10.3 vs. control -12.2 IQR -46.2 to 47.7, p = 0.328) and median DPR difference (-1.0 IQR -26.0 to 17.0 vs. control 2.0 IQR -15.0 to 18.0, p = NA). Comparison of 2-year data revealed no significant median OPR dif ference (-32.6 IQR -122.2 to 1.0 vs. control -20.7 IQR -72.6 to 44.7, p = 0.327) and median DPR difference (-5.0 IQR -19.0 to 6.0 vs. control -4.5 IQR -22.0 to 12.5, p= NA).

Conclusions: Turf toe results in significant loss of playing time. Despite the long recovery period, NFL players have similar performance following injury compared to controls. The effect of turf toe injuries on performance is variable.

Level of evidence: IVKeywords: football, turf toe, performance, sports

INTRODUCTION“Turf toe” is a common injury affecting high level

athletes, occurring in 10.9% of elite college football play-ers.1 Initially coined in 1976 by Bowers in response to an increasing injury pattern related to new turf, “turf toe” represents a hyperextension injury to the hallux metatarsophalangeal (MTP) joint resulting in a plantar capsuloligamentous sprain or tear, which can ultimately lead to progressive deformity.2 As a result, previous literature has focused on identifying and modifying risk factors for injury prevention.3 Turf toe represents a spectrum of injury, ranging from mild attenuation to complete disruption of plantar tissue.4 Treatment can range from symptomatic management to surgical re-pair or reconstruction. While many of these cases can be managed non-operatively, up to 50% of patients may experience persistent pain and stiffness, with decreased mobility compared to other uninjured toes.5,6 In the setting of persistent dysfunction or instability, surgical intervention may be indicated in the high-level athlete to prevent late deformity.7

Recently, a greater analytical emphasis has been placed on the management of orthopedic injuries in the high-level athlete, focusing on statistical performance after return to play.8,9 Offensive and defensive power ratings have previously been described to rate overall performance in NFL athletes to adjust for production and statistics affected by position played and provide a com-prehensive metric to compare across position groups.9 To date, no data is available that quantitatively compares the change in performance in athletes suffering turf toe injury. Professional American football athletes, particu-larly players at “skill” positions, place unique demands on the MTP joint, with injury potentially detracting from performance even after return to play. Given the persistent symptoms experienced by NFL players with turf toe, we hypothesized that post-injury performance

THE EFFECT OF TURF TOE INJURIES ON PLAYER PERFORMANCE IN THE NATIONAL FOOTBALL LEAGUE

Andrew Tran, MD1, Jason Kappa, MD1, Evan Smith, MD1, Michael Hoy, BS2,Jacob Farrar, BS2, Avinash Chandran, PhD3, Rajeev Pandarinath, MD1

1George Washington University Department of Orthopaedics2George Washington University School of Medicine & Health Sciences3Milken Institute of Public HealthCorresponding Author: Andrew Tran; Phone: (202) 741-3300; Email: [email protected]: The authors report no potential conflicts of interestrelated to this study.Sources of Funding: No sources of funding declared.

A. Tran, J. Kappa, E. Smith, M. Hoy, J. Farrar, A. Chandran, R. Pandarinath


and statistical output in NFL players would decrease in comparison to healthy controls at the same position. Our ultimate goal is to better understand and predict the short and long-term consequences of turf toe injury

METHODSGame summaries, weekly team injury reports, team

press releases, and media reports (ESPN.com, NFL.com, Rotoworld.com) were used to identify players sustaining “turf toe” injuries from 2010-2015 NFL seasons. Players who sustained a turf toe injury and played at positions where an Offensive Power Rating (OPR) or Defensive Power Rating (DPR) could be calculated were included: running back (RB), tight end (TE), wide receiver (WR), defensive lineman (DL), linebacker (LB), and defensive back (DB). Players who did not have a minimum of one regular season of play before and after the year of injury were excluded from the study. A control group was gen-erated from all active roster RBs, TEs, WRs, DLs, LBs, and DBs in the NFL in 2013 who did not sustain a turf toe injury during the study period, which encapsulates performance metrics from 2010-2015. Players who did not participate in the 2012 and 2014 regular seasons were excluded as controls.

Demographics and return to play statistics were col-lected on injured players during the index season. The following variables were recorded: date of injury, player age, number of prior years of NFL experience, last year of NFL play if during the study period, whether the player retired due to turf toe, operative vs. non-operative treatment of turf toe injury, time to return to play, and time off injury report. Time to return to play was de-fined as the number of days after turf toe injury at which the player participated in at least one in-game play. Days until off injury report was defined as the number of days after turf toe injury until the player was no longer listed on the team injury report for a toe injury.

Performance data was collected for a minimum of one year before and up to two years after the season of injury. Offensive performance data was collected using the following variables: games played, yards rushing and receiving, and touchdowns. Defensive performance vari-ables included games played, interceptions and fumbles recovered, tackles, and quarterback sacks. Using a previ-ously validated performance evaluation score,8,9 an Of-fensive Power Rating (OPR = [total rushing and receiving yards/10] + [total TDs x 6]) and Defensive Power Rating (DPR = total tackles + [total sacks x 2] + [total intercep-tions and fumbles recovered x 2]) was generated for each year of data for injured and control subjects. OPR and DPR as metrics for player performance have been shown to have high intra-class correlation coefficients and criterion-oriented validity using Pro Bowl selection

as a benchmark.8 In accordance with this previously described approach for performance analysis, control players with low statistical production from 2010-2015 (sum of OPR < 150 or sum of DPR < 35) were excluded from analysis.6

Group differences (between injured and non-injured groups) in continuous variables were evaluated using Wilcoxon Rank Sum tests. Associations between injury group (injured and non-injured groups) and categorical variables were evaluated using Fisher’s Exact tests. Statistical significance was assessed at the 0.05 level, and all analyses were conducted using SAS 9.4 (SAS Institute, Cary NC).

RESULTSFrom 2010-2015, 50 players sustained a “turf toe”

injury and 24 players (11 offense and 13 defense) met criteria for inclusion in this study [Table-1]. Twenty-six players were excluded for one or more of the following criteria: six were offensive linemen, three were quarter-backs, nine did not play at least one full regular season before the season of injury (index), and 10 did not play at least one full regular season after index. Of the 13 of-fensive players who met criteria for inclusion, six were running backs (RBs), two were tight ends (TEs), and three were wide receivers (WRs). Of the 13 defensive players meeting inclusion criteria, five were defensive

Players sustaining turf

toe injuryControls P value

Offense

Number of players 11 72 --Age (years) 25.45 26.29 0.25Number of games played in 2 seasons prior to index season

23.00 23.72 0.92

Position 0.29Receiver 5 20Rusher 6 52

Defense

Number of players 13 364Age (years) 26.38 26.92 0.24Number of games played in 2 seasons prior to index season

27.00 25.48 NA*

Position 0.83Defensive Line 3 106Linebacker 5 111Defensive Back 5 147

Table 1. Demographics of Players Sustaining Turf Toe Injury and Controls

*indicates inability to estimate probability using Wilcoxon rank sum test.


Effect of Turf Toe Injuries on NFL Player Performance

A

backs (DBs), three were defensive linemen (DLs), and five were linebackers (LBs). Of the 731 RBs, TEs, WRs, DBs, DLs, and LBs who participated in the 2013 NFL season (in this case, index) and did not sustain a turf toe injury from 2010-2015, 72 offensive players and 364 defensive players met criteria for inclusion in this study. The remaining players were excluded for one or more of the following criteria: 273 did not play at least one full regular season before and after index and 292 had low performance statistics (5-year cumulative OPR < 150 or DPR < 35). No differences were found in demographics when comparing players sustaining turf toe injury to their respective control groups [Table -1].

Eighteen players (75%) returned to play within the regular season of injury at a median of 28 (IQR 20.25 to 42) days [Figure-1a]. The remaining six players returned to play during the following regular season. Three play-ers required season-ending surgery for turf toe, and one player who did not meet criteria for inclusion this study retired at the end of the injury season due to persistent pain from turf toe. Fifteen players (62.5%) were eventu-ally removed from team injury reports for toe injury at a median of 42 (IQR 27 to 51) days [Figure-1b].

There were no significant differences in offensive player performance between players with turf toe in-jury and their respective controls. Injured players had a median OPR of 84.0 (IQR 4.7 to 109.3) the season before index, 67.3 (IQR 10.8 to 114.3) the season after index, and 45.3 (4 to 114.3) two seasons after index. In comparison, control players had a median OPR of 89.2 (IQR 53.4 to 154) the season before index, 91.45 (IQR 42.8 to 149.08) the season after index, and 87.35 (IQR 46.1 to 141.02) 2 seasons after index [Figure-2]. From the season before to one season after index, injured players had a median change in OPR of -18.9 (IQR -43.4 to 10.3) vs. -12.2 (IQR -46.2 to 47.7) for controls (p = 0.328). Data using two seasons after index demonstrated a median change in OPR of -32.6 (IQR -122.2 to 1.0) vs. -20.7 (-22.0

to 12.5) for controls (p = 0.327).There were also no significant differences in defen-

sive player performance between injured players and controls. Injured players had a median DPR of 45 (IQR 34 to 79) the season before index, 51 (IQR 36 to 64) the season after index, and 52.5 (IQR 42.25 to 55.5) two seasons after index. Control players had a median DPR of 37 (IQR 20 to 60) the season before index, 40 (IQR 22 to 58.25) the season after index, and 37 (IQR 19 to 25) two seasons after index [Figure-3]. From the season prior to index to one season after, injured players had a median change in DPR of -1.0 (IQR -26.0 to 17.0) vs. 2.0 (IQR -15.0 to 18.0) for controls. Data examining two seasons post-index demonstrated a median change in DPR of -5.0 (IQR -19.0 to 6.0) vs. -4.5 (IQR -22.0 to 12.5) for controls. Exact probabilities for DPR analysis were unable to be estimated using Wilcoxon rank sum test.

DISCUSSIONTurf toe represents a spectrum of injury and ranges

from partial to complete disruption of the plantar tissues on the first MTP joint. First described in athletes in 1976, turf toe has become more extensively scrutinized for long-term sequelae with the goal of determining optimal management. In athletes, injury typically occurs when an axial load is applied to the heel of a plantarflexed foot with the great toe in extension, typically as part of a tackle.10 The risk of this injury increases with age, artificial playing surfaces, and more flexible shoe wear3 and will affect nearly 10% of high-level football athletes,10 and has been reported to occur in up to 45% of NFL play-ers.11 This study found significantly fewer cases of turf toe injury, likely secondary to methodology requiring injury to be severe enough that it was recorded on the team weekly injury report.

This study demonstrates that players who sustained a turf toe injury missed a substantial portion of the season, with a median return to play of 28 days. The wide range

Figure 1. Number of days (a) to return to play and (b) until removed from the injury report for all turf toe injured players. FS denotes players who not return until the following season.

A. Tran, J. Kappa, E. Smith, M. Hoy, J. Farrar, A. Chandran, R. Pandarinath


A B C

in return to play metrics reflects a wide spectrum of in-jury severity for “turf toe,” and underlines the difficulty in studying this injury at a population level. One player was not included in this study because he retired due to injury and did not return to play the following season. Of the players who did return and met inclusion criteria, 25% of them did not return to play until the following season, with three (12.5%) requiring season ending surgery. A study using surveillance data of National Collegiate Athletic Association football players found that turf toe had an incidence of 0.062 per 1000 athlete exposures and comprised 0.83% of all football injuries.3 1.74% of players required operative intervention and the average time lost from injury was 10.1 days. While not directly comparable, the large discrepancy in metrics from those of this study may stem from the inherent differences in surveillance data and media reported data, the latter

skewing towards more severe grades of injury. Though it did not reach statistical significance, of-

fensive players affected by turf toe did show a trend toward lower performance after injury compared with controls. This is in line with previous thought that the effect of turf toe is dependent on the demands placed on the foot. It has also been previously postulated that position played may impact time missed due to injury, with linemen reported to better tolerate the injury than sprinters.2 Not only does turf toe injury occur more frequently in running backs and wide receivers due to the mechanism, it may have a significant impact on performance, particularly considering the long-standing symptoms that can persist after injury.1,5,10

There are several limitations to this study. This study was retrospective which limits its full clinical applicability. This was accounted for by comparing injured patients to

Figure 2. Median offensive power rating and interquartile range of injured players vs controls per season.

Figure 3. Median defensive power rating and interquartile range of injured players vs controls per season.


Effect of Turf Toe Injuries on NFL Player Performance

similar healthy controls to determine if the difference in performance between seasons was different by players with turf toe injuries. Additionally, this study did not stratify patients by grade or injury severity, as this data was not publicly available. Furthermore, the limited number of patients who met criteria for study would not have allowed for subgroup analysis based on grade of injury. This is partially because turf toe is a relatively infrequent injury, occurring in approximately 0.06 per 1000 athlete exposures.3 This may partially explain the high variance in the data. The variability in individual player performance and playing time in a team sport with frequent substitutions may also contribute to this high variance in both the study and control groups. An a priori power analysis was not performed, and a post-hoc power analysis was determined to be non-contributory.13 Regardless, given the different positions affected and the limited overall sample size, these power ratings provide an acceptable proxy by which to compare performance.

Turf toe represents a spectrum of injury that can lead to a significant loss of playing time in NFL athletes depending on the degree of soft tissue rupture. Despite this loss of time, professional football players appear to return to a similar level of play following a turf toe injury compared to healthy controls when measured by composite offensive and defensive metrics. Further examination with a larger series of patients would be useful to understand the degree to which injury severity may affect performance.

REFERENCES1. West, Robin, and James P. Bradley. “Incidence

and Variance of Foot and Ankle Injuries in Elite College Football Players.” The American Journal of Orthopaedics 40.1 (2011): 40–44.

2. Bowers, K. Douglas, and R. Bruce Martin. “Turf-Toe: a Shoe-surface Related Football Injury.” Medicine & Science in Sports & Exercise 8.2 (1976): 81–83.

3. George, Elizabeth et al. “Incidence and Risk Fac-tors for Turf Toe Injuries in Intercollegiate Football: Data from the National Collegiate Athletic Association Injury Surveillance System.” Foot & ankle interna-tional 35.2 (2014): 108–115.

4. Porter, David, and Lew Schon. “Baxter’s The Foot and Ankle in Sport.” 2nd ed. Mosby, 2007. 411–433.

5. Brophy, Robert H. et al. “Effect of Turf Toe on Foot Contact Pressures in Professional American Football Players.” Foot & ankle international 30.5 (2009): 405–409.

6. Mann, R. A. “Adult Hallux Valgus and Complication of Hallux.” Surgery of the Foot 1 (1986): 95–98.

7. Anderson, Robert B. “Turf Toe Injuries of the Hal-lux Metatarsophalangeal Joint.” Techniques in Foot & Ankle Surgery 1.2 (2002): 102–111.

8. Carey, James L. et al. “Outcomes of Anterior Cruciate Ligament Injuries to Running Backs and Wide Receivers in the National Football League.” The American Journal of Sports Medicine 34.12 (2006): 1911–1917.

9. McHale, Kevin J. et al. “Outcomes of Lisfranc Inju-ries in the National Football League.” The American journal of sports medicine 44.7 (2016): 1810–1817.

10. Anderson, Robert B., Kenneth J. Hunt, and Jeremy J. McCormick. “Management of Common Sports-related Injuries About the Foot and Ankle.” JAAOS-Journal of the American Academy of Ortho-paedic Surgeons 18.9 (2010): 546–556.

11. Brophy, Robert et al. “Prevalence of Musculoskel-etal Disorders at the NFL Combine--Trends from 1987 to 2000.” Medicine & Science in Sports & Exercise 39.1 (2007): 22–27.

12. Rodeo, Scott A. et al. “Turf-toe: An Analysis of Metatarsophalangeal Joint Sprains in Professional Football Players.” The American journal of sports medicine 18.3 (1990): 280–285.

13. Lenth, R. V. Post hoc power: tables and commentary. Iowa City: Department of Statistics and Actuarial Sci-ence, University of Iowa 1–13 (2007).

Volume 39 Issue 2 I

ABSTRACTBackground: Multiple reports have detailed

clinical outcomes in surgically treated patients with borderline hip dysplasia. The purpose of this systematic review was to define patient outcomes following these surgical interventions.

Methods: Searches were developed using an iterative process of gathering and evaluating terms. Comprehensive strategies including both index and keyword methods were devised for the following databases: PubMed, Embase, and Cochrane CEN-TRAL. Independent review and data abstraction was performed by two authors. Inclusion criteria were clear delineation of outcomes for patients with borderline hip dysplasia (Lateral center edge angle (LCEA) 18-25°) and outcomes following hip ar-throscopy and/or periacetabular osteotomy (PAO), including patient reported outcomes (PROs), re-vision arthroscopy, and conversion to Total Hip Arthroplasty. Exclusion criteria included alternative surgical procedures including “shelf” or “salvage” osteotomies, studies without patient outcomes or clearly delineated results for patients with border-line dysplasia, inclusion of <5 borderline patients, or inadequate follow-up defined as < 12 months.

Results: Thirteen of 2109 articles met inclusion criteria for full data analysis. 505 patients (mean age 29.6 years, 67.7% female) with borderline dysplasia (mean LCEA 22.3°) were treated with hip arthroscopy. The majority of studies reported outcomes using the modified Harris Hip Score (mHHS) and Hip Outcome Score Sports Specific Subscale (HOS-SSS); all showed post-operative improvement with mean increase of 21.0 and 26.8 points, respectively. Revision arthroscopy rate was 7.5% (31/412), and a total hip arthroplasty conversion rate was 4.0% (18/455). The aver-

age combined total reoperation rate was 13.7% (69/505).

Conclusion: Arthroscopic management of bor-derline hip dysplasia is associated with meaning-ful improvement in PRO scores. However, there appears to be a high reoperation rate including both arthroscopic and open revision procedures. There is a paucity of studies reporting outcomes of PAO in patients with borderline hip dysplasia.

Level of evidence: Systematic review of Level III and Level IV studies.

Keywords: borderline hip dysplasia, outcomes, hip arthroscopy

INTRODUCTIONDevelopmental hip dysplasia (DHD) is a condition char-

acterized by under-coverage of the acetabulum over the femoral head, leading to abnormal biomechanics and joint instability.1,2 DHD can also lead to a myriad of soft tissue abnormalities including acetabular labral hypertrophy or damage, chondral damage and premature osteoarthritis.3-7 DHD is often assessed by measuring the lateral-center edge angle (LCEA) as described by Wiberg, with LCEA <20 degrees denoting classic DHD.8 Borderline hip dys-plasia (BHD) is defined by mild under-coverage with an LCEA of 18-25 degrees.9 Peri-acetabular osteotomy (PAO) is the treatment of choice for skeletally mature patients with DHD, with multiple studies reporting favorable long term outcomes for hips with LCEA <20.10-13 Consensus var-ies, however, regarding the surgical management of BHD.

Hip arthroscopy is used to treat various intraarticular pathologies including femoroacetabular impingement (FAI) and labral tears with favorable long term out-comes.14-16 The incidence of these intraarticular patholo-gies is high in patients with BHD.17-19 Arthroscopy as an adjunct to open hip surgery has therefore become increas-ingly common in the management of BHD,20-22 with some surgeons performing combined arthroscopic and open hip techniques to address intraarticular and extraarticular pathologies such as FAI, labral damage and acetabular under-coverage for BHD with a single anesthesia event. Staged procedures are also commonly used. Several stud-ies have shown clinical benefit of combined management of BHD with favorable early outcomes.21,23,24 However, it is uncertain as to the best overall treatment strategy in

OUTCOMES OF SURGICAL MANAGEMENT OF BORDERLINE HIP DYSPLASIA: A SYSTEMATIC REVIEW

Cameron Barton, MD1; Elizabeth Scott, MD1; Zain M Khazi, BS1; Michael Willey, MD1; Robert Westermann, MD1

1Department of Orthopedics and Rehabilitation, The University of Iowa, Iowa City, IA USACorresponding author: Cameron Barton, MDEmail: [email protected]: The authors report no potential conflicts of interest related to this study.Sources of Funding: No sources of funding declared.

C. Barton, E. Scott, Z. M. Khazi, M. Willey, R. Westermann

II The Iowa Orthopedic Journal

these patients with regards to open versus arthroscopic treatment, their order and timing if both are performed, and their individual and combined clinical outcomes.

As the management of BHD is not well understood, the purpose of this systematic review was to perform a detailed compilation of the literature to elucidate patient-reported outcomes and failure rates after surgi-cal intervention in patients with BHD. Our hypothesis is that arthroscopic treatment alone of BHD will be as-sociated with improvement in clinical outcomes with low complication rates.

METHODSA careful search strategy was developed using an

iterative process of gathering and evaluating terms to identify publications pertaining to outcomes following hip arthroscopy and/or periacetabular osteotomy in the setting of BHD. Comprehensive strategies including both index and keyword methods were devised for the following databases: PubMed, Embase, and Cochrane CENTRAL. Articles were first identified using three main concept terms and associated Medical Subject Head-ing terms (MeSH) of 1) “arthroscopy” or “periacetabular osteotomy” with MeSH terms of “Arthroscopy AND (hip Or hip joint),” “osteotomy AND (hip or hip joint), and “acetabulum/surgery” 2) “hip dysplasia” with MeSH terms of “hip dislocation” and “hip joint/pathology” and 3) “outcomes” with MeSH terms of “treatment outcome,” “radiography,” “pain,” “range of motion,” “recovery of func-tion,” disease progression,” “postoperative complications,” “total hip arthroplasty,” “return to sport,” “recurrence,” and “follow-up studies.” EMTREE search terms included “periacetabular osteotomy,” “arthroscopy AND hip,” “oste-otomy AND hip,” “acetabulum” with surgery subheading, “hip dysplasia,” “hip disease” with surgery subheading, “follow-up.” Articles were independently reviewed by two reviewers (C.B., E.S.). Articles were first chosen by title and abstract for relevance, and then underwent full-text review. Articles without abstracts were chosen for full text review. Levels of evidence were determine by cur-rent Cochrane guidelines, following previously defined categories by Cook et al.25

The following criteria were used to include article in final analysis: 1) study included patients with BHD, either entire cohort or subgroup, defined as LCEA 18-25 degrees, 2) treatment consisted of hip arthroscopy and/or acetabu-lar osteotomy including Ganz, Bernese, periacetabular, or rotational curved osteotomy, and 3) reported patient outcomes including data on survivorship, complications, clinical outcomes, patient reported outcomes (PRO), re-turn to play, revision surgery, and/or conversion to total hip arthroplasty (THA). Exclusion criteria consisted of review articles, hip osteotomies other than previously

mentioned including “shelf” or “salvage” osteotomies, pe-diatric osteotomies for congenital hip dysplasia including Salter, Pemberton and others, results not specific to BHD subgroup, articles without patient outcomes (i.e. purely biomechanical or radiographic data), non-English, follow-up <1 years, and small sample size (<5 patients). We also excluded studies on procedures performed for other hip di-agnoses such as slipped capital femoral epiphysis (SCFE), Legg-Calve-Perthes, and post-infectious deformities.

After full text review and determination of inclusion, data acquisition included extraction of baseline patient demographic variables including sex, age, preoperative LCEA, preoperative Tonnis grade, and preoperative PRO scores including modified Harris Hip Score (mHHS), Hip Outcome Score (HOS for Activities of Daily Living [ADL] and Sports Specific Subscale [SSS]), Nonarthritic Hip Score (NAHS), visual analog scale (VAS), and patient satisfaction score. Procedural data on specific hip ar-throscopy procedures including capsular plication, labral repair versus debridement, femoroplasty, acetabuloplasty, iliopsoas lengthening, and others were also recorded. Postoperative data including postoperative PROs, LCEA, complications, revision surgery, and conversion to THA were also recorded.

Descriptive analysis were used to report study charac-teristics, total rate of complications, revision hip arthros-copy, and conversion to THA. Additionally, mean change in preoperative and postoperative PRO scores were cal-culated. To determine if the mean change for each PRO met clinical significance, established minimal clinically important difference (MCID) were utilized for mHHS,26 HOS-ADL,26 HOS-SSS,26 and VAS.27 Beck et al. determined

Figure 1. PRISMA flowchart for search strategy.

Volume 39 Issue 2 III

Outcomes of Surgical Management of Borderline Hip Dysplasia: A Systematic Review

A

MCID for mHHS, HOS-ADL, and HOS-SSS are 8.5, 9.2, and 13.7 point increase at 2 years after hip arthroscopy for BHD, respectively.26 Similarly, Martin et al. determined that a decrease of 15 mm met MCID for VAS at 1 year after hip arthroscopy.27 A reasonable established MCID was not found for NAHS, therefore, achieving clinically significant change for NAHS was not assessed.

RESULTSIn December 2017, our search identified 2109 articles

from PubMed, Embase, and Cochrane CENTRAL data-bases. 153 articles were chosen for full text review after abstract and title review (Figure 1). Thirteen articles met inclusion and exclusion criteria and underwent full data analysis (Table 1). All thirteen articles included BHD patients undergoing hip arthroscopy only. Articles on acetabular osteotomies that met our inclusion and ex-clusion criteria were not identified in the search. Of the thirteen articles included in the review, five were level III evidence and 8 were level IV evidence articles. In total, 505 patients (mean age 29.6 years, 67.7% female) with borderline dysplasia (mean LCEA 22.3°) were included in the final analysis. Mean follow-up across studies report-ing follow-up specific to borderline sub-groups was 35.6.

A variety of arthroscopic procedures were reported, including labral repair, labral debridement, femoroplasty,

rim decompression, chondroplasty, iliopsoas lengthen-ing, ligamentum teres debridement, loose body removal, microfracture, capsular closure, capsular decompression, trochanteric bursectomy, and iliotibial band release. The use of labral repair or debridement varied, with some performing labral procedures on all patients,28,29 and oth-ers reporting labral procedures in 50% of patients;30,31 the average rate of labral procedures during arthroscopy was 76.9%. Specifically, labral debridement ranged from 0-50% across studies. Rates of osteoplasty and chondroplasty were difficult to analyze, as many studies did not report on these procedures individually. Cam decompression rate ranged widely, from 0-100%. A comparison of outcomes of specific arthroscopy procedures could not be performed, as studies varied greatly in their baseline patient demo-graphics, indications, and reported follow-up.

Regarding clinical outcomes, all but one study reported improvement in PRO scores (Table 2). Mean change in pre and postoperative PRO scores assessed met MCID. Nine studies reported mHHS and HOS-SSS scores with a calculated mean improvement of 21.0 and 26.8 points, respectively, which is greater than the 8.5 and 13.7 point increase required to achieve MCID, respectively.3,28,31-37 Similarly, six studies reported HOS-ADL with mean improvement of 17.3 which is greater than the 9.2 point improvement required to achieve MCID.3,28,32-35 Six stud-

Table 1. Study Characteristics

Study Year Level of Evidence Patients n (% female)

Mean Age, yr (range)

Follow-up, yr (range)

Preoperative LCEA (range)

Domb et al.31 2013 IV- Case Series 22 (82) 20 (14-39) 2.3 (1.4-3.25) 22.2 (15-25)

Chandrasekaran et al.3 2017 IV- Case Series 55 (63) 24.3 (13.2-38.7) 2.1 (2.0-2.53) 22.1 (18-25)

Evans et al.30 2017 IV- Case Series 21 (81) 15.5 (13.1-17.5) 2.19 22.3 (20-25)

Nawabi et al.29 2016 III- Cohort Study 55 hips (48) 29.8 1.30 (0.33-3.38) 22.4 (18.4-24.9)

Fukui et al.25 2015 IV- Case Series 102 hips (50) 35 (18-69) 3.33 (2.0-8.08) 23 (20-25)

Matsuda et al.27 2017 IV- Case Series 8 (50) 49.6+/-7.6 2 19 (16-24)

Hatakeyama et al.26 2017 III- Case Control 45 (66) 31.4 (12-65) 3.54 (2.0-6.05) 23.2 (20-25)

Kalore and Jiranek28 2012 IV- Prognostic Study 50 (88) 35 for repair, 41 for debridement

2.75 23

Grammatopoulos et al.41 2017 III- Case Control 43 40.9 (16-55) for entire cohort

4.5 (0.4-8.3) for entire cohort

20 (15-25)

Chaharbakhshi et al.33 2017 III- Prospective Cohort Study 40 (90) 28.2 (13.3-51.9) 3.87 (2- 6.94) 22.05 (18-25)

Cetanovich et al.32 2017 III- Retrospective Cohort Study 36 (75) 31.5 (SD 11.8) ≥ 2 yrs 23.4 (SD 1.5)

McCarthy et al.44 1998 IV- Retrospective Review 20 (65) 36 (16-54) 24 (27-41) 23 (19-27)

Domb et al.31 2018 IV- Case Series 19 (90) 22.9 (15.5-39.3) 5.73 (5- 6.98) 21.7 (18-24)

HOOS, Hip disability Osteoarthritis Outcome Score; HOS-ADL, Hip Outcome Score–Activities of Daily Living; HOS-SSS, Hip Outcome Score–Sports Specific Subscale; LT- ligamentum teres, LCEA, lateral center edge angle; mHHS, modified Harris Hip Score; NA, not available; NR, not reported; NAHS, Nonarthritic Hip Score; PRO, patient-reported outcome; SF-36, 36-item short form; THA, total hip arthroplasty; VAS, visual analog scale; WOMAC, Western Ontario and McMaster Osteoarthritis Index.


IV The Iowa Orthopedic Journal

A B C

Table 2. Procedures Performed and Patient Reported Outcome Measures Assessed

Study Procedures Performed PROs evaluated

Preoperative PRO Scores

Postoperative PRO Scores

Domb et al.31

capsular plication (100%), labral repair (95%), labral debridement (4.5%), LT debridement (59%), iliopsoas release (68%), osteoplasty/

chondroplasty (41%)

mHHS, HOS-ADL, HOS-SSS,

NAHS, and VAS

mHHS= 69, HOS-ADL= 72.9,

HOS-SSS= 49, NAHS= 68.6,

VAS= 5.8

mHHS= 86.2, HOS-ADL= 89.6,

HOS-SSS= 77, NAHS= 85.9,

VAS= 2.9

Chandrasekaran et al.3

capsular plication (100%), labral repair (67%), labral debridement (31%), iliopsoas lengthen-ing (62%), LT debridement (52%), osteoplasty

(58%), chondroplasty (47%), loose body removal (9%), microfracture (5%)


NAHS, and VAS

mHHS= 63.7, HOS-ADL= 68.8, HOS-SSS = 46.4,

NAHS= 63.7, VAS= 5.64

mHHS= 84.4, HOS-ADL= 86.3, HOS-SSS= 74.8,

NAHS= 83.8, VAS= 2.47

Evans et al.30

capsular plication (100%), labral repair (62%), labral debridement (38%), femoroplasty

(71%), iliopsoas lengthening (71%), LT de-bridement (14%), iliopsoas bursectomy (14%),

synovectomy (14%)


NAHS, and VAS


NAHS= 62.75, VAS= 6.29

mHHS= 91, HOS-ADL= 94.74, HOS-SSS= 81.59,

NAHS= 91.94, VAS= 1.31

Nawabi et al.29

cam osteoplasty (98%) labral repair (69%), capsular closure (100%), labral debridement (31%), LT debridement (24%), subspine de-compression (64%), acetabuloplasty (27%)

mHHS, HOS-ADL,

HOS-SSS, and iHOT-33

mHHS= 61.7, HOS-ADL= 76, HOS-SSS= 54.6, iHOT-33= 44.4


iHOT-33= 80

Fukui et al.25

capsular closure (100%), labral repair (100%), microfracture (18%), cam osteoplasty (78%),

acetabuloplasty (5%), capsular plication (37%), LT debridement (93%)

mHHS, HOS-ADL, HOS-Sport, SF12 PCS, SF12 MCS,

and WOMAC

mHHS= 63.5, HOS-ADL= 70.9, HOS-sport= 51.4 SF12 PCS= 42.5, SF12 MCS= 52.4,

WOMAC= 25.3

mHSS= 84.9, HOS-ADL= 83.7, HOS-Sport= 75.7 SF12 PCS= 50.9

SF12-MCS= 54.1, WOMAC= 9.7

Matsuda et al.27

T-capsulotomy without closure (100%), labral repair (50%), labral debridement (50%),

acetabuloplasty (50%), cam osteoplasty (100%), microfracture (25%)

NAHS NAHS= 52.9 NAHS=69.17 (1 yr); 52.13 (2 yrs)

Hatakeyama et al.26 labral repair (100%), cam osteoplasty (93%), capsular plication (100%)

mHHS and NAHS

Success cohort: mHHS= 72.1 NAHS= 58.8

Reoperation Cohort: mHHS= 68.1 NAHS= 60.0

Success Cohort: mHHS= 100 NAHS= 98.8

Reoperation Cohort: mHHS= 75.9 NAHS= NR

Kalore and Jiranek28 labral repair (50%), labral debridement (50%) mHHS and HOOS

Labral Debridement mHHS= 48.3,

Labral Repair mHHS= 59.43

Labral Debridement mHHS= 65.19,

HOOS-Pain= 73;

Labral Repair mHHS= 75.28,

HOOS-Pain= 78

Grammatopoulos et al.41

limited capsulotomy, labral repair (44% for entire cohort), microfracture (25% for entire cohort), osteochondroplasty (81% for entire

cohort)

NAHS and HOOS

NAHS=46, HOOS=71

NAHS=62, HOOS=71

(continued on next page)

Volume 39 Issue 2 V


ies reported VAS with mean improvement of 3.9, which is greater than the 1.5 improvement required to achieve MCID.3,33-37 Seven studies reported NAHS with mean improvement of 18.8. 3,29,30,33,34,36,37

The average reported rate of revision after arthros-copy was 7.5% (31/412), and the conversion rate to ar-throplasty was 4.0% (18/455), however several studies reported combined reoperation rate which included THA and hip arthroscopy without delineating by type of revi-sion surgery (Table 3). Kalore et al. reported the highest revision rate at 44%, however, they defined revision sur-gery as revision hip arthroscopy, conversion to THA, or PAO.31 The average combined total reoperation rate for any cause was 13.7% (69/505) across all studies.

DISCUSSIONHip dysplasia at skeletal maturity can be managed sur-

gically with open techniques that reorient the acetabulum including PAO, Ganz, and others.8,10,13 The utility of hip arthroscopy, which has become increasingly used for the

management of FAI, remains controversial in hip dyspla-sia.38-40 Many reports describe unsatisfactory outcomes in patients with moderate to severe hip dysplasia (LCEA <18 degrees) treated solely with arthroscopic techniques, with high rates of reoperation.31,41,42 The best treatment course for patients with BHD (LCEA 18-25) remains uncertain. PAO for the borderline dysplastic hip may not provide significant benefit to offset the invasive nature of the procedure; alternatively, some arthroscopic techniques may exacerbate or fail to improve symptoms of mild dysplasia.42,43 The overall purpose of two procedures are vastly different, with the purpose of the PAO to improve bony acetabular coverage, while arthroscopy can address associated conditions including capsular laxity, labral pathology, femoral acetabular impingement, cartilage injury, loose bodies, or soft tissue contracture.10-12,14-16 The purpose of this study was to evaluate the reported outcomes of hip arthroscopy and acetabular reorienting osteotomies in these patients, to determine appropriate treatment strategies and direction of future research.

Table 2. Procedures Performed and Patient Reported Outcome Measures Assessed (continued)

Study Procedures Performed PROs evaluated

Preoperative PRO Scores

Postoperative PRO Scores

Chaharbakhshi et al.33

capsular plication (100%), labral repair (65%), selective debridement (35%), LT tear debride-

ment (50%), internal snapping hip iliopsoas fractional lengthening (45%), femoroplasty

(70%), Loose body removal (2.5%), synovectomy (5%), notchplasty (2.5%), acetabular microfrac-

ture (5%), acetabular decortication (37.5%)

mHHS, NAHS,

HOS-SSS, VAS, and

patient satisfaction

20 patients in each group

LT group: mHHS= 64.1,NAHS= 63.9,

HOS-SSS= 44.1, VAS= 5.3

Non-LT group: mHHS=66.9, NAHS= 67.7,

HOS-SSS= 50.4, VAS= 5.5

20 patients in each group

LT group: mHHS= 81.3, NAHS= 81.7,

HOS-SSS= 68.1, VAS= 2.7

Non-LT group: mHHS= 87.4, NAHS= 88.4,

HOS-SSS= 75.6, VAS= 2.1

Cetanovich et al.32

capsular plication (100%), labral repair (92%), acetabuloplasty (11%), femoral osteoplasty (100%), capsular closure (100%), troch bur-

sectomy (3%), IT band release (3%), cartilage delamination (39%)

mHHS, HOS-ADL,

HOS-SSS, and VAS

mHHS= 57.2, HOS-ADL= 65.4,

HOS-Sports= 44.5, VAS= 7.6

mHHS= 79.9, HOS-ADL= 88.6,

HOS-Sports=73.6, VAS= 1.4

McCarthy et al.44 labral repair or debridement (90%), loose body removal (15%), chondral procedures (35%)

Absence of preoperative

pain

Absence of preoperative pain= 0%

Absence of preoperative pain= 85%

Return to preoperative sporting and

leisure activities= 85%

Domb et al.34

capsular plication (100%), labral repair (95%), selective debridement (5%), minimal rim

decortication only (86%), LT debridement (57%), iliopsoaas fractional lengthening (52%), femoroplasty (52%), limited acetabular chon-

droplasty (29%), loose body removal (9.5%), trochanteric bursectomy (5%)

mHHS, NAHS,

HOS-SSS, VAS, and

patient satisfaction

mHHS= 70.3, HOS-SSS= 52.1,

NAHS= 68.3, VAS= 5.6

mHHS= 85.9, HOS-SSS= 70.8,

NAHS= 87.3, VAS= 1.8

HOOS, Hip disability Osteoarthritis Outcome Score; HOS-ADL, Hip Outcome Score–Activities of Daily Living; HOS-SSS, Hip Outcome Score–Sports Specific Subscale; LT- ligamentum teres, LCEA, lateral center edge angle; mHHS, modified Harris Hip Score; NA, not available; NR, not reported; NAHS, Nonarthritic Hip Score; PRO, patient-reported outcome; SF-36, 36-item short form; THA, total hip arthroplasty; VAS, visual analog scale; WOMAC, Western Ontario and McMaster Osteoarthritis Index.


VI The Iowa Orthopedic Journal

After a comprehensive literature review, we identified 13 studies that met our inclusion and exclusion criteria pertaining to hip arthroscopy for the treatment of BHD. In contrast, we did not identify any studies with acetabular reorienting osteotomies for BHD. This finding likely indi-cates that open procedures for BHD are rarely performed. Additionally, this may also reflect the limited number of hip preservation surgeons trained to perform osteotomy compared to arthroscopy, and/or hesitancy to indicate open procedures for BHD as initial treatment thereby reserving this procedure for those who fail arthroscopic repair. In total, 505 patients (mean age 29.6 years, 67.7% female) with borderline dysplasia (mean LCEA 22.3°) were treated with hip arthroscopy and followed for a mean of 35.6 months. The majority of studies reported outcomes using the modified Harris Hip Score (mHHS) and Hip Outcome Score Sports Specific Subscale (HOS-SSS); and nearly all showed post-operative improvement which met the threshold of established MCID. Revision arthroscopy rate was 7.5% (31/412), and a total hip arthroplasty con-version rate was 4.0% (18/455). The average combined total reoperation rate was 13.7% (69/505).

All of the studies included PROs, however use of in-dividual PRO instruments varied widely, making direct comparison of outcomes challenging and difficult to interpret. It is clear that unified use of PRO instruments would improve hip preservation research. The mHHS and HOS were most frequently cited, whereas, some studies provided qualitative data on return to sport and

resolution of pre-operative pain. All studies except one,30 reported postoperative improvement in PROs. Matsuda et al. reported on 8 patients with an average follow-up of 24 months; preoperative NAHS score was 52.88, and 52.13 at 24 months.30 Their study population had several charac-teristics that likely pre-disposed it to inferior outcomes including average preoperative LCEA of 19, the lowest average of our included studies, as well as the highest average patient age at 49.6 years.30 The mean age across all studies included was 29.6 years. These hips may have had a greater burden of preoperative osteoarthritis and chondrolabral pathology; accordingly, they also had the highest percentage of patients undergoing labral debride-ment (50%) compared to repair (50%).

The substantial variation in type of arthroscopic procedure(s) performed for BHD also implies a continued uncertainty in how to treat this condition. We encountered labral repair, debridement, and reconstruction, bony pro-cedures including femoroplasty, rim decompression, and chondroplasty, iliopsoas lengthening, ligamentum teres debridement, loose body removal, microfracture, trochan-teric bursectomy, iliotibial (IT) band release and a large variety of capsular procedures including decompression, plication, or simple closure. The labrum was consistently repaired in some studies,28,29 and intermittently addressed in others,30,31 for an average rate of repair of 76.9%. The majority of studies did not provide subanalysis of clinical outcomes by type of arthroscopic procedure, and denotes a serious paucity in hip preservation research that needs

Table 3. Revision Hip Arthroscopy or Conversion to Total Hip Arthroplasty

Study

Revision Arthroscopy,

n (%)

Mean Time to Revision Arthroscopy,

months

Total Conversions to Total Hip Arthroplasty,

n (%)

Mean Time to Conversion to Total Hip

Arthroplasty, yr

Domb et al.31 2 (9) 17.5 0 Not Reported

Chandrasekaran et al.3 6 (11) Range: 4.07 to 50.2 0 Not Reported

Evans et al.30 0 Not Reported 0 Not Reported

Nawabi et al.29 2 (4) 16.7 0 Not Reported

Fukui et al.25 7 (7) 24 5 (5) 2 (range: 0.85- 4)

Matsuda et al.27 0 Not Reported 2 (25) Not Reported

Hatakeyama et al.26 2 (4) Not Reported 7 (16) Not Reported

Kalore and Jiranek28 Not Reported Not Reported Not Reported Not Reported

Grammatopoulos et al.41 Not Reported Not Reported 4 (10.4) Not Reported

Chaharbakhshi et al.33 6 (15) 19.4 (range: 1.8-48.1) 3 (7.5) 3.5 (range: 1.4-5.2)

Cetanovich et al.32 1 (3%) Not Reported 0 Not Reported

McCarthy et al.44 1 (5%) Not Reported 2 (10) Not Reported

Domb et al.34 4 (19%) 25.1 (range: 4.1-50.1) 0 Not Reported

Volume 39 Issue 2 VII


to be improved upon in future studies. Biomechanical and clinical reports support the use of

capsular repair or plication for borderline dysplasia; all studies performed capsular plication except for Matsuda et al. and Grammatopoulous et al..30,44 These studies re-ported a 25% (2/8) and 9.3% (4/43) rate of conversion to THA, respectively, both of which were higher than the average conversion to THA in this study (4%).

Several studies specifically evaluated risk factors for failure with arthroscopy in BDH. Hatakeyama et al. found preoperative age >42 years, disruption of Shenton’s line, osteoarthritis, Tonnis angle ≥15, and LCEA ≤17 predictive of failure defined as revision surgery.29 Intraoperative predictors were severe acetabular chondral damage and femoral chondral damage of any severity.29 Kalore et al. reported outcomes in 50 BHD patients, including 25 with labral debridement and 25 with labral repair; labral re-pair was associated with lower revision rates that labral debridement; both cohorts had poorer outcomes, however, than hips without dysplasia (LCEA >25).31 This is con-sistent with prior studies which report inferior outcomes with labral debridement in hips with LCEA <25.41,45

Revision arthroscopy rate was 7.5% (31/412), and conversion rate to arthroplasty was 4.0% (18/455), for a combined total reoperation rate of 13.7% (69/505). This rate is much higher than reported for patients with normal acetabular coverage, with rates ranging from 3.6 to 7.5% in patients undergoing hip arthroscopy for FAI .46 Due to the heterogeneous nature of the arthroscopic procedures reported, this statistic alone provides little insight; instead we recommend evaluation of studies which report outcomes for specific procedures, such as labral repair with capsular closure, which demonstrated comparable results in several studies.34,28

LimitationsThis study had several notable limitations. The body

of literature in BHD patients is limited. We found no stud-ies in BHD patients meeting our inclusion and exclusion criteria reporting on results of open procedures such as PAO. Therefore, a comparison could not be made between open and arthroscopic procedures in these patients. Due to the paucity of studies in BHD patients, we included studies with short follow-up (≥ 1 year), small cohorts (≥5 patients), and retrospective design corresponding to low levels of evidence (all with level III or IV).

The analysis of this systematic review was limited by the inconsistent patient reported outcomes across studies. Further, the type of arthroscopic procedures varied, as well as the patient baseline demographics including age, and length of follow-up. Therefore, determination of fac-tors affecting outcomes could not be adequately assessed across studies.

CONCLUSIONArthroscopic management of BHD is associated with

meaningful improvement in PRO scores, despite signifi-cant variation in specific procedures performed. However, studies report a high rate of revision arthroscopy or subsequent open procedures. No studies on PAO or other open procedures were able to be included within the date range searched. These findings speak to the paucity of outcomes data for patients with BHD, and a need for more robust studies on both specific arthroscopic techniques and open procedures.

REFERENCES1. Cooperman D. What is the evidence to support ac-

etabular dysplasia as a cause of osteoarthritis? Journal of pediatric orthopedics. 2013;33 Suppl 1:S2-7.

2. Henak CR, Abraham CL, Anderson AE, et al. Patient-specific analysis of cartilage and labrum mechanics in human hips with acetabular dysplasia. Osteoarthritis and cartilage. 2014;22(2):210-217.

3. Chandrasekaran S, Darwish N, Martin TJ, Su-arez-Ahedo C, Lodhia P, Domb BG. Arthroscopic Capsular Plication and Labral Seal Restoration in Borderline Hip Dysplasia: 2-Year Clinical Outcomes in 55 Cases. Arthroscopy. 2017;33(7):1332-1340.

4. McCarthy JC, Lee JA. Acetabular dysplasia: a paradigm of arthroscopic examination of chondral injuries. Clinical orthopaedics and related research. 2002(405):122-128.

5. McWilliams DF, Doherty SA, Jenkins WD, et al. Mild acetabular dysplasia and risk of osteoarthritis of the hip: a case-control study. Annals of the rheumatic diseases. 2010;69(10):1774-1778.

6. Murphy SB, Ganz R, Muller ME. The prognosis in untreated dysplasia of the hip. A study of radiographic factors that predict the outcome. J Bone Joint Surg Am. 1995;77(7):985-989.

7. Reijman M, Hazes JM, Pols HA, Koes BW, Bierma-Zeinstra SM. Acetabular dysplasia predicts incident osteoarthritis of the hip: the Rotterdam study. Arthritis and rheumatism. 2005;52(3):787-793.

8. Ganz R, Klaue K, Vinh TS, Mast JW. A new periacetabular osteotomy for the treatment of hip dysplasias. Technique and preliminary results. Clini-cal orthopaedics and related research. 1988(232):26-36.

9. Ricciardi BF, Fields KG, Wentzel C, Kelly BT, Sink EL. Early Functional Outcomes of Periacetabu-lar Osteotomy After Failed Hip Arthroscopic Surgery for Symptomatic Acetabular Dysplasia. Am J Sports Med. 2017;45(11):2460-2467.

10. Troelsen A, Elmengaard B, Soballe K. Medium-term outcome of periacetabular osteotomy and predic-tors of conversion to total hip replacement. J Bone Joint Surg Am. 2009;91(9):2169-2179.


VIII The Iowa Orthopedic Journal

11. Steppacher SD, Tannast M, Ganz R, Siebenrock KA. Mean 20-year followup of Bernese periacetabular osteotomy. Clinical orthopaedics and related research. 2008;466(7):1633-1644.

12. Matheney T, Kim YJ, Zurakowski D, Matero C, Millis M. Intermediate to long-term results fol-lowing the Bernese periacetabular osteotomy and predictors of clinical outcome. J Bone Joint Surg Am. 2009;91(9):2113-2123.

13. Lerch TD, Steppacher SD, Liechti EF, Tan-nast M, Siebenrock KA. One-third of Hips After Periacetabular Osteotomy Survive 30 Years With Good Clinical Results, No Progression of Arthritis, or Conversion to THA. Clinical orthopaedics and related research. 2017;475(4):1154-1168.

14. Brunner A, Horisberger M, Herzog RF. Sports and recreation activity of patients with femoroac-etabular impingement before and after arthroscopic osteoplasty. Am J Sports Med. 2009;37(5):917-922.

15. Philippon MJ, Briggs KK, Yen YM, Kupper-smith DA. Outcomes following hip arthroscopy for femoroacetabular impingement with associated chon-drolabral dysfunction: minimum two-year follow-up. The Journal of bone and joint surgery British volume. 2009;91(1):16-23.

16. Haviv B, Singh PJ, Takla A, O’Donnell J. Arthroscopic femoral osteochondroplasty for cam lesions with isolated acetabular chondral damage. The Journal of bone and joint surgery British volume. 2010;92(5):629-633.

17. Fujii M, Nakashima Y, Jingushi S, et al. In-traarticular findings in symptomatic developmental dysplasia of the hip. Journal of pediatric orthopedics. 2009;29(1):9-13.

18. Domb BG, Lareau JM, Baydoun H, Botser I, Millis MB, Yen YM. Is intraarticular pathology common in patients with hip dysplasia undergoing periacetabular osteotomy? Clinical orthopaedics and related research. 2014;472(2):674-680.

19. Ross JR, Zaltz I, Nepple JJ, Schoenecker PL, Clohisy JC. Arthroscopic disease classification and interventions as an adjunct in the treatment of acetabu-lar dysplasia. Am J Sports Med. 2011;39 Suppl:72S-78S.

20. Domb BG, LaReau J, Redmond JM. Combined hip arthroscopy and periacetabular osteotomy: indi-cations, advantages, technique, and complications. Arthroscopy techniques. 2014;3(1):e95-e100.

21. Ricciardi BF, Mayer SW, Fields KG, Wentzel C, Kelly BT, Sink EL. Patient Characteristics and Early Functional Outcomes of Combined Arthroscopic Labral Refixation and Periacetabular Osteotomy for Symptomatic Acetabular Dysplasia. Am J Sports Med. 2016;44(10):2518-2525.

22. Kim KI, Cho YJ, Ramteke AA, Yoo MC. Peri-acetabular rotational osteotomy with concomitant hip arthroscopy for treatment of hip dysplasia. The Journal of bone and joint surgery British volume. 2011;93(6):732-737.

23. Domb BG, LaReau JM, Hammarstedt JE, Gupta A, Stake CE, Redmond JM. Concomitant Hip Arthroscopy and Periacetabular Osteotomy. Arthros-copy. 2015;31(11):2199-2206.

24. Nassif NA, Schoenecker PL, Thorsness R, Clo-hisy JC. Periacetabular osteotomy and combined femoral head-neck junction osteochondroplasty: a minimum two-year follow-up cohort study. J Bone Joint Surg Am. 2012;94(21):1959-1966.

25. Cook DJ, Guyatt GH, Laupacis A, Sackett DL. Rules of evidence and clinical recommendations on the use of antithrombotic agents. Chest. 1992;102(4 Suppl):305S-311S.

26. Beck EC, Nwachukwu BU, Chahla J, et al. Patients With Borderline Hip Dysplasia Achieve Clinically Significant Outcome After Arthroscopic Femoroacetabular Impingement Surgery: A Case-Control Study With Minimum 2-Year Follow-up. Am J Sports Med. 2019;47(11):2636-2645.

27. Martin RL, Kivlan BR, Christoforetti JJ, et al. Minimal Clinically Important Difference and Substantial Clinical Benefit Values for a Pain Visual Analog Scale After Hip Arthroscopy. Arthroscopy. 2019;35(7):2064-2069.

28. Fukui K, Briggs KK, Trindade CA, Philippon MJ. Outcomes After Labral Repair in Patients With Femoroacetabular Impingement and Borderline Dys-plasia. Arthroscopy. 2015;31(12):2371-2379.

29. Hatakeyama A, Utsunomiya H, Nishikino S, et al. Predictors of Poor Clinical Outcome After Ar-throscopic Labral Preservation, Capsular Plication, and Cam Osteoplasty in the Setting of Borderline Hip Dysplasia. Am J Sports Med. 2018;46(1):135-143.

30. Matsuda DK, Gupta N, Khatod M, et al. Poorer Arthroscopic Outcomes of Mild Dysplasia With Cam Femoroacetabular Impingement Versus Mixed Femo-roacetabular Impingement in Absence of Capsular Repair. Am J Orthop (Belle Mead NJ). 2017;46(1):E47-E53.

31. Kalore NV, Jiranek WA. Save the torn labrum in hips with borderline acetabular coverage. Clinical orthopaedics and related research. 2012;470(12):3406-3413.

32. Nawabi DH, Degen RM, Fields KG, et al. Out-comes After Arthroscopic Treatment of Femoroac-etabular Impingement for Patients With Borderline Hip Dysplasia. Am J Sports Med. 2016;44(4):1017-1023.

Volume 39 Issue 2 IX


33. Evans PT, Redmond JM, Hammarstedt JE, Liu Y, Chaharbakhshi EO, Domb BG. Arthroscopic Treatment of Hip Pain in Adolescent Patients With Borderline Dysplasia of the Hip: Minimum 2-Year Follow-Up. Arthroscopy. 2017;33(8):1530-1536.

34. Domb BG, Stake CE, Lindner D, El-Bitar Y, Jackson TJ. Arthroscopic capsular plication and labral preservation in borderline hip dysplasia: two-year clinical outcomes of a surgical approach to a chal-lenging problem. Am J Sports Med. 2013;41(11):2591-2598.

35. Cvetanovich GL, Levy DM, Weber AE, et al. Do Patients With Borderline Dysplasia Have Inferior Outcomes After Hip Arthroscopic Surgery for Femo-roacetabular Impingement Compared With Patients With Normal Acetabular Coverage? Am J Sports Med. 2017;45(9):2116-2124.

36. Chaharbakhshi EO, Perets I, Ashberg L, Mu B, Lenkeit C, Domb BG. Do Ligamentum Teres Tears Portend Inferior Outcomes in Patients With Borderline Dysplasia Undergoing Hip Arthroscopic Surgery? A Match-Controlled Study With a Minimum 2-Year Follow-up. Am J Sports Med. 2017;45(11):2507-2516.

37. Domb BG, Chaharbakhshi EO, Perets I, Yuen LC, Walsh JP, Ashberg L. Hip Arthroscopic Sur-gery With Labral Preservation and Capsular Plication in Patients With Borderline Hip Dysplasia: Minimum 5-Year Patient-Reported Outcomes. Am J Sports Med. 2018;46(2):305-313.

38. Byrd JW, Jones KS. Arthroscopic management of femoroacetabular impingement: minimum 2-year follow-up. Arthroscopy. 2011;27(10):1379-1388.

39. Nho SJ, Magennis EM, Singh CK, Kelly BT. Out-comes after the arthroscopic treatment of femoroac-etabular impingement in a mixed group of high-level athletes. Am J Sports Med. 2011;39 Suppl:14S-19S.

40. Waterman BR, Ukwuani G, Clapp I, Malloy P, Neal WH, Nho SJ. Return to Golf After Arthroscop-ic Management of Femoroacetabular Impingement Syndrome. Arthroscopy. 2018;34(12):3187-3193 e3181.

41. Parvizi J, Bican O, Bender B, et al. Arthroscopy for labral tears in patients with developmental dys-plasia of the hip: a cautionary note. J Arthroplasty. 2009;24(6 Suppl):110-113.

42. Kirsch JM, Khan M, Bedi A. Does Hip Arthroscopy Have a Role in the Treatment of Developmental Hip Dysplasia? J Arthroplasty. 2017;32(9S):S28-S31.

43. Adler KL, Giordano BD. The Utility of Hip Ar-throscopy in the Setting of Acetabular Dysplasia: A Systematic Review. Arthroscopy. 2019;35(1):237-248.

44. Grammatopoulos G, Davies OLI, El-Bakoury A, Gill HS, Pollard TCB, Andrade AJ. A Traffic Light Grading System of Hip Dysplasia to Predict the Success of Arthroscopic Hip Surgery. Am J Sports Med. 2017;45(12):2891-2900.

45. Larson CM, Ross JR, Stone RM, et al. Arthroscop-ic Management of Dysplastic Hip Deformities: Pre-dictors of Success and Failures With Comparison to an Arthroscopic FAI Cohort. Am J Sports Med. 2016;44(2):447-453.

46. Minkara AA, Westermann RW, Rosneck J, Lynch TS. Systematic Review and Meta-analysis of Out-comes After Hip Arthroscopy in Femoroacetabular Im-pingement. Am J Sports Med. 2018:363546517749475.


ABSTRACTBackground: Metacarpal and phalanx fractures

are common among professional athletes. There is a paucity of data to guide team physicians on expected return to play after hand fractures. The purpose of this study was to examine the epidemi-ology and return to play times after hand fractures in NCAA athletes. We hypothesized that surgical management of fractures may expedite return to play times.

Methods: The NCAA Injury Surveillance Pro-gram database was queried for metacarpal and phalanx fractures during the 2009-2014 seasons in all sports. Injury rates per 100,000 athlete-exposures (AEs) were calculated. Student’s t-test, Wilcoxon Rank sum tests, Chi-Squared tests, and Fisher Exact Test were used. Statistical significance was set to p<0.05.

Results: Sports with the highest rates of phalanx and metacarpal fractures included Men’s Football, Men’s Ice Hockey, Men’s Wrestling, and Women’s Field Hockey. Multiple sports had participants with no hand fractures over the study period. Male stu-dent-athletes with metacarpal fractures treated op-eratively returned to play at a mean of 31.8±29.4 days versus 13.8±23.6 days for those treated non-operatively. 92% of male student-athletes were able to return to sport in the same season without operative management versus 67% with operative management. Female student-athletes had a co-hort too small for statistical analysis. Return to play times for male student-athletes with phalanx fractures were not significantly different between operative and non-operative groups (16.1±21.5 days versus 7.1±13.3 days).

Conclusions: Hand fractures are relatively com-mon among NCAA student-athletes participating in

contact sports. Student-athletes with metacarpal fractures returned to play at an average of 2-4 weeks after injury; those with phalanx fractures returned at an average of 1-2 weeks. The return to play times illustrated within this study can be used to counsel athletes, athletic trainers, and coaches.

Level of evidence: IVKeywords: hand, fracture, athlete, NCAA, epi-

demiology

INTRODUCTIONFractures of the phalanges and metacarpals of the

hand are common among athletes,1-9 particularly among male athletes who participate in contact sports.1,3,5,6,8,9 Recent data estimate that 54% of sports-related hand injuries involve the phalanges, and 34% involve the meta-carpals.1 Sports-related hand fractures may be managed non-operatively with splints or protective casts in the vast majority of patients (estimated 80–90% of cases).1,5-9 Non-operative management was originally thought to allow athletes to return to sport with little time-loss and without the risk of complications of wound healing or development of tendon or joint adhesion that may follow operative fixation of metacarpal and phalanx fractures. 5,6,9 Recent studies suggest that operative fixation of metacarpal fractures in athletes has increased over time, from an estimated 5–10% of cases in 20039 to upwards of 20% in 2014.7 Stern10 and Fufa et al.5 have also made note of the recent increase in procedures available for metacarpal fracture fixation.

Current literature surrounding operative fixation of metacarpal and phalanx fractures is dichotomous. There are multiple studies5,6,8,11 suggesting that open reduction and internal fixation (ORIF) of metacarpal fractures may provide increased fracture stability relative to a splint or cast, thereby allowing for earlier range of motion at the surrounding joints, decreased joint stiffness, and ulti-mately, a reduction in time-loss from sport participation. Other studies have suggested the opposite, demonstrat-ing increased time-loss following operative management of metacarpal and phalanx fractures.12

There has been minimal investigation into the epidemiology of hand fractures and time-loss following fractures of the phalanges and metacarpals in National Collegiate Athletic Association (NCAA) student-athletes.

DESCRIPTIVE EPIDEMIOLOGY AND RETURN TO SPORT AFTER HAND FRACTURES IN NCAA ATHLETES

Christopher N. Carender, MD1, Joseph A. Buckwalter V, MD, PhD1, Natalie A. Glass, PhD1, Robert W. Westermann, MD1

1University of Iowa Hospitals and Clinics, Department of Orthopedics and RehabilitationCorresponding Author: Christopher N Carender; Phone: 319-356-3699. Email: [email protected]: The authors report no potential conflicts of interestrelated to this study.Sources of Funding: No sources of funding declared.

C.N. Carender, J.A. Buckwalter V, N.A. Glass, R.W. Westermann


There is a paucity of data to guide NCAA team physi-cians on expected time-loss and return to play after hand fractures treated operatively or non-operatively. The purpose of this study was to examine the epidemiol-ogy and return to play time frames after hand fractures in NCAA student-athletes. We hypothesized that rates of hand fractures would be higher amongst student-athletes participating in contact sports. Additionally, we hypothesized that surgical management of phalanx and metacarpal fractures may shorten time-loss and expedite return to play.

METHODSThe study was approved by the Institutional Review

Board at our institution, as well as by the NCAA. The NCAA Injury Surveillance Program database was que-ried for metacarpal, phalanx, and volar plate/avulsion fractures during the 2009–2010 to 2013–2014 academic years in all available contact and non-contact sports. Game and practice scenarios were queried. The NCAA Injury Surveillance Program (ISP) is a prospectively collected injury surveillance database managed by the Datalys Center for Sports Injury and Prevention. The methodology of the NCAA-ISP has been previously de-scribed.13-15 Data collected before the 2009 season were available in the NCAA-ISP but were not utilized in this study due to methodological differences in collection methods that had the potential to bias analysis of return to sport rates and times.13-15

Data Collection The NCAA-ISP collects data from over 150 national

colleges and universities that field collegiate athletic teams at the Division I–III levels. Data are collected and reported to the NCAA-ISP by certified Athletic Train-ers (ATs) at participating institutions that are present during school-sanctioned games and practices.13-15 Such data is collected at the time of the injury.19 Data that is reported by the ATs includes exposure variables for all athletes participating in each respective event and injury variables for athletes that sustained an injury during each respective event. Additionally, ATs are able to return to records for the purposes of updating them as needed, i.e., if an athlete were to return to full participation af-ter undergoing treatment for an injury.19 The validation process for data accepted into the NCAA-ISP has been described in detail in previous studies. 13-15

The database was queried for exposure and injury variables pertaining to all available metacarpal, phalanx, and volar plate fractures in all available sports: Men’s Baseball (BA-M), Men’s and Women’s Basketball (BB-M, BB-W), Men’s and Women’s Cross Country (CC-M, CC-W), Men’s Football (FB-M), Women’s Field Hockey (FH-W), Women’s Gymnastics (GY-W), Men’s and Women’s

Ice Hockey (IH-M, IH-W), Men’s and Women’s Lacrosse (LX-M, LX-W), Women’s Softball (SB-W), Men’s and Women’s Soccer (SO-M, SO-W), Men’s and Women’s Swimming (SW-M, SW-W), Men’s and Women’s Tennis (TE-M, TE-W), Men’s and Women’s Indoor Track (IT-M, IT-W), Men’s and Women’s Outdoor Track (OT-M, OT-W), Women’s Volleyball (VB-W), and Men’s Wres-tling (WR-M). Exposure variables queried included the academic year (season) and sport in which the athlete was participating. Injury variables collected included the academic year (season) and sport in which the injury occurred, number of days lost from participation due to the injury, if participant returned to play within the same season, type of injury (discussed below), and if surgery resulted from injury.

DefinitionsHand fractures were defined as metacarpal fractures

of the fingers and thumb, phalanx fractures of the fingers and thumb, and volar plate/avulsion fractures of fingers. Return to sport was defined as a continuous variable, ranging from 0 days (no time lost from participation) to being out of participation for the remainder of the season. Return to sport within the same season as injury was defined as a binary yes or no variable. An athlete exposure was defined as one student-athlete’s participa-tion in a game or practice event in which he or she was exposed to potential injury.

Statistical Analysis Injury rates per 100,000 athlete-exposures (AEs) and

return to play times for athletes treated with operative and non-operative management of the above fractures were calculated. Student’s t-test and Wilcoxon Rank sum tests were used for continuous variable and Chi-Squared tests and Fisher Exact Test were used for categorical variables to determine significance, set to p<0.05.

RESULTSEpidemiology and Rates of Hand Fractures

From the 2009–2010 to 2013–2014 seasons, 288 hand fractures (225 men, 63 women) were reported in NCAA athletes in 3,739,004 AEs [Table 1]. Sports with the highest rate of hand fractures were Men’s Ice

Table 1. Epidemiology of Hand Fractures in NCAA Athletes

Metacarpal Phalanx Volar Plate/

Avulsion

All FracturesFinger Thumb Finger Thumb

Male 83 27 83 25 7 225

Female 16 3 32 10 2 63

Total 99 30 115 35 9 288


Descriptive Epidemiology and Return to Sport after Hand Fractures in NCAA Athletes

Hockey, Men’s Football, and Men’s Wrestling [Table 2]. Participants in Men’s and Women’s Cross Country, Women’s Swimming, Men’s and Women’s Tennis, Men’s Indoor Track, and Men’s and Women’s Outdoor Track had no reported hand fractures during the 2009–2010 to 2013–2014 seasons. Sports with the highest rates of phalanx fractures were Men’s Ice Hockey, Women’s Field Hockey, and Men’s Baseball [Table 2]. Sports with the highest rate of metacarpal fractures were Men’s Wrestling, Men’s Football, and Men’s Basketball [Table 2]. Sports with the highest rate of volar plate/avulsion fractures were Women’s Ice Hockey, Men’s Wrestling, and Men’s Soccer [Table 2].

Time-Loss and Rates of Return to SportOf the 288 hand fractures reported, 225 (78.1%) oc-

curred in male student-athletes; 110 of these fractures were metacarpal fractures, 108 were phalanx fractures, and 7 were volar plate/avulsion fractures. Overall, 181 of these fractures had data pertaining to time-loss avail-able. Rates of operative treatment and time-loss in male student-athletes are demonstrated in Table 3. Female student athletes had 63 (21.9%) reported hand fractures; 19 of these fractures were metacarpal fractures, 42 were phalanx fractures, and 2 were volar plate/avulsion frac-tures. In total, 45 of these fractures had data pertaining to time-loss available. Rates of operative treatment and time-loss for female student-athletes are demonstrated in Table 4.

Examining metacarpal fractures of the thumb in

male student athletes, 19 fractures had time-loss data available; 12 of 19 (63.2%) were treated non-operatively, and 7 of 19 (36.8%) were treated operatively. Fractures treated non-operatively returned at a mean of 22.6±41.9 days (median, 4.0 days; range, 0–148.0 days); those treated operatively returned at a mean of 27.3±30.7 days (median, 23.0 days; range, 0–87 days) (p=0.55). Data of return to sport within the same season was available for 24 fractures of the thumb. Male student-athletes were able to return at similar rates, with 13 of 15 (86.7%) and 7 of 9 (77.8%) metacarpal thumb fractures treated non-operatively and operatively were able to return in the same season, respectively (p=0.62).

Rates of return to sport in the same season for metacarpal and phalanx fractures in male and female student-athletes are demonstrated in Table 5.

Male student-athletes sustained 7 volar plate/avul-sion fractures, while female student-athletes sustained 2 of these fractures. All 9 volar plate/avulsion fractures were treated non-operatively. Only 5 fractures in male student-athletes had time-loss data available. The mean time-loss for male student-athletes was 0.2 days (range, 0–1 days), and the mean for female student athletes was 2.0 days (range, 1–3 days). All (9 of 9) male and female student-athletes that sustained volar plate/avulsion inju-ries returned to sport in the same season.

DISCUSSIONHand fractures are relatively common among NCAA

student-athletes, with an incidence of 4.90 injuries per

Table 2. Rates of Hand Fractures by Sport

Sport Metacarpal Phalanx Volar Plate/Avulsion

Rate* 95% CI Rate* 95% CI Rate* 95% CIBA-M 1.69 (1.68, 1.69) 6.18 (6.17, 6.19) 0.56 (0.56, 0.57)BB-M 4.17 (4.16, 4.18) 1.85 (1.85, 1.86) 0.00 --BB-W 1.54 (1.53, 1.55) 4.62 (4.61, 4.63) 0.00 --FB-M 5.12 (5.11, 5.12) 3.23 (3.22, 3.23) 0.45 (0.44, 0.45)GY-W 2.21 (2.19, 2.22) 0.00 -- 0.00 --FH-W 0.00 -- 10.73 (10.70, 10.76) 0.00 --IH-M 3.89 (3.88, 3.89) 12.01 (12.00, 12.02) 0.00 --IH-W 1.77 (1.76, 1.78) 0.88 (0.88, 0.89) 1.77 (1.76, 1.78)LX-M 3.74 (3.73, 3.75) 1.25 (1.24, 1.25) 0.00 --LX-W 0.00 -- 1.89 (1.89, 1.90) 0.00 --SB-W 1.86 (1.85, 1.87) 4.34 (4.33, 4.35) 0.00 --SO-M 1.88 (1.87, 1.89) 1.88 (1.87, 1.89) 0.63 (0.62, 0.63)SO-W 2.78 (2.77, 2.79) 1.85 (1.85, 1.86) 0.00 --SW-M 0.00 -- 1.03 (1.03, 1.04) 0.00 --IT-W 0.00 -- 0.99 (0.98, 0.99) 0.00 --VB-W 0.64 (0.63, 0.64) 2.55 (2.54, 2.56) 0.00 --WR-M 6.35 (6.33, 6.37) 1.27 (1.26, 1.28) 1.27 (1.26, 1.28)

*Rate of respective fracture per 100,000 AE’s. Sports with no fractures are omitted from the table. CI - confidence interval; M - mens’s; W - women’s; BA - baseball; BB - basketball; FB - football; GY - gymnastics; FH - field hockey; IH - ice hockey; LX - lacrosse; SB - softball; SO - soccer; SW - swimming; IT - indoor track; VB - volleyball; WR - wrestling



100,000 AEs (range, 0.00–21.19 injuries/100,000 AEs) across all sports combined. However, there is a large degree of sport-to-sport and gender variation. Student-athletes participating in contact sports, such as ice hockey, football, and wrestling, as well as sports that may involve some degree of incidental contact with other players or apparatuses, such as basketball and field hockey, had a higher incidence of hand fractures relative to student-athletes participating in non-contact sports. Additionally, some sports (Men’s and Women’s Cross Country, Women’s Swimming, Men’s and Women’s Tennis, Men’s Indoor Track, and Men’s and Women’s Outdoor Track) had participants with no hand fractures over the five seasons of study. Male student-athletes participating in basketball and ice hockey experienced higher rates of metacarpal and phalangeal fractures rela-tive to their female counterparts; however, rates of hand fractures in the remaining sports were similar between genders. These findings are in contrast to prior stud-ies that have demonstrated an increased rate of hand fractures in male athletes relative to female athletes.7,11

In male student-athletes, we found a statistically sig-nificant difference in return to play times in metacarpal fractures treated non-operatively relative to those treated operatively (13.8±23.6 days versus 31.8±29.4 days, respec-tively). No significant difference was noted in return to play times in phalangeal fractures treated operatively and non-operatively. Currently, there is no consensus regarding treatment modality and early return to play times in metacarpal and phalangeal fractures.5,6,8,11,12 For stable metacarpal and phalangeal fractures, non-operative management in well-padded cast or splint with return to play as tolerated has been demonstrated to be an ef-

fective and relatively ubiquitous treatment option.1,5,16,17 However, operative fixation of metacarpal and phalangeal fractures in athletes has been increasing in popularity in recent years.5,9,16,18 Morgan et al.16 and Capo et al.18 sug-gested that ORIF of unstable metacarpal fractures may allow rigid fixation leading to earlier return to play rela-tive to non-operative management for a similarly unstable fracture. Additional studies have further supported this suggestion.5,19 However, there has been a recent trend towards consideration of ORIF for any metacarpal or phalangeal fracture in contact sports athletes, stability notwithstanding.6,7 Geissler and McCraney19 described ORIF of metacarpal fracture using a cage plate, followed by return to play in a protective splint 1–2 weeks after surgery. A series of 8 patients has been described using this technique, with adequate fracture healing achieved in all patients.6 Kodama et al.7 described a series of 20 athletes with metacarpal and phalangeal fractures who underwent ORIF with plate and/or screw fixation and who were allowed to return to full-sport participation 2 weeks following surgery. All 20 patients went on to radiographic union within 3 months. A recent review by Fufa et al.5 discussed aiming for athlete return to play at 2–3 weeks following ORIF of a metacarpal shaft fracture.

The metacarpal of the thumb has greater range of mo-tion relative to metacarpals of the other digits. As such, larger deformity in metacarpal fractures may be toler-ated without a loss of function.5,6 However, intra-articular fractures of the thumb metacarpal, including Bennett and Rolando fractures, often require extended periods of immobilization and delays in return to sport relative to fractures of the base of non-thumb metacarpals. In the present study, there was no difference in time or rate

Table 3. (a-b) – Rates of Operative Treatment and Time-loss in Male Student-Athletesa. Metacarpal fractures of the fingers

Time-loss (days)n Mean SD Median Range

Non-operative 67 13.8 23.6 3.0 0.0-148.0Operative 12 31.8 29.4 25.0 0.0-87.0

p=0.02, statistically significant. SD – standard deviation.

Table 4. (a-b) – Rates of Operative Treatment and Time-loss in Female Student-Athletesa. Metacarpal fractures



No statistical analysis perfomed secondary to small sample size. SD – standard deviation.

b. Phalanx fractures



p=0.17, no significant difference between groups. SD – standard deviation.

b. Phalanx fractures



No statistical analysis perfomed secondary to small sample size. SD – standard deviation.


Descriptive Epidemiology and Return to Sport after Hand Fractures in NCAA Athletes

of return to sport within the same season between op-eratively and non-operatively treated thumb metacarpal fractures. Metacarpal fractures of the thumb did have a higher rate of operative fixation relative to non-thumb metacarpal fractures (37% versus 15%); however, this comparison is significantly limited by a small cohort of thumb metacarpal fractures. Because of this, no statisti-cal analysis between the groups was performed, and we report only descriptive statistics. Additionally, without data on individual fractures, the contribution of each of these phenomena on the effect of return to play rates and times in thumb metacarpal fractures is unknown.

Treatment of hand fractures is largely contingent on injury and patient-specific factors, including fracture pat-tern, fracture stability, athlete handedness, and athlete sport. No information regarding fracture pattern and fracture stability is available in the NCAA-ISP database, and is therefore not included in this study. As such, the effect of fracture type and stability on return to play time in this study is unknown, but it is thought to be significant. Additional studies examining the influence of fracture characteristics on return to play are needed before specific treatment recommendations can be made.

Rates of return to play within the same season also warrant consideration when making treatment deci-sions regarding metacarpal and phalangeal fractures in student-athletes. The present study found that 93% of male student-athletes with metacarpal fractures treated non-operatively were able to return to play in the same season versus 67% of those treated with operative man-agement; this difference was statistically significant. A difference in rates of return to play within the same season was not seen in phalangeal fractures in male student-athletes. As noted previously, selection bias in the form of more severe injuries being selected for surgi-cal management is likely to contribute to this disparity. However, a recent study by Bannasch et al.20 showed a complication rate of 15% following ORIF of metacarpal and phalangeal fractures. These complications, in tan-dem with the aforementioned selection bias, may con-tribute to lower rates of return to play within the same season seen in metacarpal fractures treated with ORIF relative to fractures treated non-operatively. Common complications with ORIF of metacarpal fractures include

metacarpophalangeal joint stiffness, implant dissociation, non-union, and need for additional surgery.3,9,20

There are several limitations to this study. First, this was a retrospective study utilizing data from multiple institutions, thereby making data vulnerable to inac-curacies in data collection. Sample sizes were relatively limited, especially in regards to metacarpal fractures of the thumb. This study did not include any data regarding an index of fracture severity or stability as noted previ-ously. Further, there may be a treatment bias present in the current study design, as the worst fractures may have been treated operatively. Injury severity, therefore, may be a confounder contributing to increased return to play rates after operative management. The impact on handedness and the methods of stabilization in the non-operative groups could not be assessed. Recommen-dations regarding return to play following treatment of metacarpal and phalangeal fractures are physician-depen-dent and likely to be highly variable, and such variation may have a direct influence on the study results.

CONCLUSIONThis study demonstrates that hand fractures are rela-

tively common among NCAA student-athletes participat-ing in contact sports. Student-athletes with metacarpal fractures returned to play at an average of 2-4 weeks after injury; those with phalanx fractures returned at an average of 1-2 weeks. Additional studies examining the influence of fracture characteristics on return to play are needed before specific treatment recommendations can be made.

REFERENCES1. Aitken S, Court-brown CM. The epidemiol-

ogy of sports-related fractures of the hand. Injury. 2008;39(12):1377-1383.

2. Chung KC, Spilson SV. The frequency and epide-miology of hand and forearm fractures in the United States. J Hand Surg Am. 2001;26(5):908-915.

3. Cotterell IH, Richard MJ. Metacarpal and pha-langeal fractures in athletes. Clin Sports Med. 2015;34(1):69-98.

Table 5. (a-b) – Rates of Return to Sport Within the Same Season for Male and Female Student-Athletesa. Male student-athletes

Rate of Return to Sport Within Same Season

Location Metacarpal PhalanxNon-operative 92.2% (71/77)* 97.9% (92/94)^Operative 66.7% (14/21)* 90.9% (10/11)^

*Metacarpal fractures p=0.01, statistically significant.^Phalanx fractures p=0.29, not statistically significant.

b. Female student-athletes

Rate of Return to Sport Within Same SeasonLocation Metacarpal Phalanx

Non-operative 92.3% (12/13) 96.6% (28/29)

Operative 60.0% (3/5) 50% (2/4)

No statistical analysis perfomed for metacarpal or phalanx fractures secondary to small sample size.



4. De jonge JJ, Kingma J, Van der lei B, Klasen HJ. Fractures of the metacarpals: A retrospective analysis of incidence and aetiology and a review of the English-language literature. Injury. 1994;25(6):365-369.

5. Fufa DT, Goldfarb CA. Fractures of the thumb and finger metacarpals in athletes. Hand Clin. 2012;28(3):379-388.

6. Geissler WB. Operative fixation of metacarpal and phalangeal fractures in athletes. Hand Clin. 2009;25(3):409-421.

7. Kodama N, Takemura Y, Ueba H, Imai S, Mat-susue Y. Operative treatment of metacarpal and phalangeal fractures in athletes: early return to play. J Orthop Sci. 2014;19(5):729-736.

8. Rettig AC, Ryan R, Shelbourne KD, Mccarroll JR, Johnson F, Ahlfeld SK. Metacarpal fractures in the athlete. Am J Sports Med. 1989;17(4):567-572.

9. Singletary S, Freeland AE, Jarrett CA. Metacarpal fractures in athletes: treatment, rehabilitation, and safe early return to play. J Hand Ther. 2003;16(2):171-179.

10. Stern PJ. Management of fractures of the hand over the last 25 years. J Hand Surg Am. 2000;25(5):817-823.

11. Kovacic J, Bergfeld J. Return to play issues in upper extremity injuries. Clin J Sport Med. 2005;15(6):448-452.

12. Morse KW, Hearns KA, Carlson MG. Return to Play After Forearm and Hand Injuries in the Na-tional Basketball Association. Orthop J Sports Med. 2017;5(2):2325967117690002.

13. Kerr ZY, Dompier TP, Snook EM, et al. National collegiate athletic association injury surveillance sys-tem: review of methods for 2004-2005 through 2013-2014 data collection. J Athl Train. 2014;49(4):552-560.

14. Westermann RW, Kerr ZY, Wehr P, Amendola A. Increasing Lower Extremity Injury Rates Across the 2009-2010 to 2014-2015 Seasons of National Col-legiate Athletic Association Football: An Unintended Consequence of the “Targeting” Rule Used to Prevent Concussions? Am J Sports Med. 2016;44(12):3230-3236.

15. Zuckerman SL, Kerr ZY, Yengo-kahn A, Wasser-man E, Covassin T, Solomon GS. Epidemiology of Sports-Related Concussion in NCAA Athletes From 2009-2010 to 2013-2014: Incidence, Recurrence, and Mechanisms. Am J Sports Med. 2015;43(11):2654-2662.

16. Morgan WJ, Slowman LS. Acute hand and wrist injuries in athletes: evaluation and management. J Am Acad Orthop Surg. 2001;9(6):389-400.

17. Toronto R, Donovan PJ, Macintyre J. An alterna-tive method of treatment for metacarpal fractures in athletes. Clin J Sport Med. 1996;6(1):4-8.

18. Capo JT, Hastings H. Metacarpal and phalangeal fractures in athletes. Clin Sports Med. 1998;17(3):491-511.

19. Geissler WB, McCraney WO. Operative manage-ment of metacarpal fractures. In: Ring DC, Cohen MS, editors. Fractures of the Hand and Wrist. New York: Informa Healthcare USA, Inc.; 2007. p. 75–89.

20. Bannasch H, Heermann AK, Iblher N, Momeni A, Schulte-mönting J, Stark GB. Ten years stable internal fixation of metacarpal and phalangeal hand fractures-risk factor and outcome analysis show no increase of complications in the treatment of open compared with closed fractures. J Trauma. 2010;68(3):624-628.


ABSTRACTBackground: To compare functional and radio-

graphic outcomes of radius fractures distal to the watershed line treated with variable-angle volar rim locking compression plates (VA-LCP) with traditional fixed-angle volar rim locking compres-sion plates (FA-LCP).

Methods: A retrospective review of patients who underwent open reduction and internal fixa-tion (ORIF) using either VA-LCP (19 wrists) or traditional fixation with FA-LCP (28 wrists). The average follow-up period was 14.5 months (range 11-16 months) for the VA-LCP group and 15.8 months (range 12-18 months) for the FA-LCP group. Clinical outcomes were evaluated using the Modified Mayo wrist score (MMWS), Disabilities of the Arm, Shoulder, and Hand (DASH) score, wrist range of motion (ROM) and grip strength relative to the uninjured side, and signs of flexor tendon irritation. Radiographic evaluation included radial height, radial inclination, volar tilt, and volar tear drop angle. All outcomes were assessed at 3, 6, and 12 months postoperatively.

Results: MMWS and DASH scores improved with time postoperatively in both groups. Relative ROM was improved in VA-LCP compared to the FA-LCP at 12 months. VA-LCP was associated with a decreased incidence of flexor tendon irritation compared to FA-LCP. VA-LCP also better held the volar tilt reduction compared the FA-LCP.

Conclusion: VA-LCP shows improved clinical and radiographic outcomes throughout the follow up period when compared to traditional fixation. VA-LCP may be an effective alternative to tradi-tional fixation methods to treat radius fractures distal to the watershed line.

Keywords: distal radius fracture, volar rim plate, wrist surgery, treatment outcomes

INTRODUCTIONThe application of a volar buttress plate in open

reduction and internal fixation (ORIF) for distal radius fractures provides both construct stability and recovery of wrist function.1 A subset of distal radius fractures may propagate distal to the watershed line and involve the volar rim. These fractures have challenged current indications for fixed-angle plates.2 The most distal end of the radius is relatively flat with a volarly sloped lip that forms the lunate facet. The geometry of the lunate facet causes it to be inadequately supported by conventional fixation techniques, which may result in a loss of reduc-tion. Sufficient stabilization by buttressing this fragment requires placement of conventional volar locked plates distal to the watershed region. This far distal fixation strategy leads to plate prominence that may cause flexor tendon irritation. Furthermore, positioning these fixed angle devices far distally may also lead to inadvertent wrist joint penetration by distally directed screws.3

Various methods and techniques have been applied to stabilize these fractures such as tension band wires, distal buttress plates, and external fixators.4-6 However, there is no consensus on the optimal fixation strategy.7,8 A variable-angle volar rim locking compression plate system (VA-LCP; Depuy-Synthes, West Chester, PA) was designed to be placed distal to the watershed line with a low profile contour to prevent flexor tendon irritation. The VA-LCP system also has 15o off axis variable angle screws that may assist in avoiding penetrating the wrist joint. Furthermore, VA-LCP has distal radial and ulnar “teardrop” holes that may be used to augment fixation of the radial styloid, lunate facet, and distal radial-ulnar joint (Figure 1). However, there is a paucity of literature examining the outcomes using VA-LCP fixation for distal radius fractures involving the volar rim. The purpose of this study is to compare functional and radiographic out-comes of VA-LCP to traditional fixation strategies with fixed angle volar locking compression plates (FA-LCP).

METHODSWe conducted a retrospective review of a consecutive

series of patients with an intra-articular distal radius

VARIABLE-ANGLE LOCKING COMPRESSION PLATE FIXATION OF DISTAL RADIUS VOLAR RIM FRACTURES

Mengcun Chen, MD1; Daniel J. Gittings, MD2; Shuhua Yang, MD1; Guohui Liu, MD, PhD1; Tian Xia, MD, PhD1

1Department of Orthopaedics Surgery, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China2Department of Orthopaedics, Hospital of University of Pennsylvania, Philadelphia, Pennsylvania, United States of AmericaCorresponding Author: Tian Xia; Email: [email protected]: The authors report no potential conflicts of interestrelated to this study.Sources of Funding: No sources of funding declared.

M. Chen, D.J. Gittings, S. Yang, G. Liu, T. Xia


fracture that extended distal to the volar rim who un-derwent ORIF with a VA-LCP or FA-LCP system at a single institution between January and October 2015. Institutional review board approval was obtained for this study. All patients during this period had preoperative radiographs and computed tomography (CT) imaging of the wrist joint. Patients were included if their fracture pattern was an Orthopaedic Trauma Association (OTA) Type-B3 or Type-C fracture, which were distal radius fractures that extended distal to the watershed region. Patients were excluded from the study if they were less than eighteen years old, had an open fracture, or had all fracture fragments smaller than three millimeters (mm) in size. Forty-seven patients were identified who met these criteria including 19 patients (19 wrists) who underwent ORIF with VA-LCP and 28 patients (28 wrists) with FA-LCP.

All patients underwent an extended volar approach of the distal radius to enable adequate visualization of the cortical rim of the distal radius.9 Distal fracture fragments were reconstructed using the volar cortical bone of the middle column of the distal radius as a point of reference. Kirschner wires (K-wires) were utilized for provisional fixation to hold the reduction. In the FA-LCP group, a conventional volar locking plate was positioned as far distally to stabilize the VMF. Auxiliary K-wires were added to augment fixation if stability was inadequate. In the VA-LCP group, a 2.4 mm VA-LCP low profile plating system was positioned straddling the watershed line, and held preliminarily in place with a K-wire in the distal end of the plate. The distal variable angle screws were placed as a “row of nails” to raft and support the articular surface. Additional variable angle locking screws were

placed in the radial and ulnar distal “teardrop” holes to augment fixation (Figure 2). After fixation in both groups, the reduction and hardware position were checked with fluoroscopy. Stability was assessed with passive wrist range of motion (ROM) and stress examination by load-ing the lunate facet. Patients in the FA-LCP group were immobilized with a plaster splint for six weeks before initiating rehabilitation. Patients in VA-LCP group wore a wrist brace for one week and then started gradual re-habilitation thereafter.

All the patients were routinely evaluated during the first month after surgery. Patients were also assessed at three, six, and twelve months postoperatively with functional outcome measures and radiographic imaging. Patient reported functional outcomes were assessed using the modified Mayo wrist score (MMWS) and the disabili-ties of the arm, shoulder and hand (DASH) score.10,11 Wrist ROM and grip strength were recorded as a percentage of the contralateral extremity’s function.12 Postoperative complications including numbness, pain, and discomfort during ROM or grip strength evaluation were also noted. Tendon irritation was defined as swelling or palpable crepitus during thumb and finger flexion and exten-sion. Anterior-posterior and lateral wrist radiographs were used to assess interval healing and maintenance of reduction. Postoperative CTs were obtained when neces-sary based upon surgeon preference. Radiographs were

Figure 1. Variable angle locking compression plate for fragment-specific fracture fixation of fractures distal to the watershed line of the radius. (a) The pre-contoured variable angle volar rim plate (VA-LCP); (b) A 2.4 mm cortex screw and two 2.4 mm variable angle screws; (c) Low profile of the plate with holes allow up to 15° off-axis screw angulation in all directions.

Figure 2. Radiograph obtained pre-operatively and two days post-operatively in a 49-year-old woman who had undergone ORIF using a VA-LCP. (a) Pre-operative images demonstrate an intra-articular distal volar marginal rim fracture with extension into the dorsal surface; (b) Postoperative images after ORIF with VA-LCP.


VA-LCP Fixation of Distal Radius Volar Rim Fractures

A

reviewed by independent radiologists to measure radial inclination and height, volar tilt angle, and volar teardrop angle using previously described methods.13,14 Statistical analysis was performed using SPSS version 18.1. Student t-test and chi-square test were used for data analysis. Statistical significance was set at p=0.05.

RESULTSIn the VA-LCP group, there were 19 patients (4 male,

15 female) with an average age of 52.9 years (range 42-65 years). The patients’ dominant wrists were involved in 47.4% (9/19) of the injuries. Average time between injury and surgery was 2.8 days (range 2-5 days). The average postoperative follow-up period was 14.5 months (range 11–16 months). In the FA-LCP group, there were 28 patients (6 male, 22 female) with an average age of 53.5 years (range 39-66 years). 42.9% (12/28) of the injuries involved the patient’s dominant wrist. Average time between injury and surgery was 2.9 days (range 2-5 days). The average postoperative follow-up period was 15.8 months (range 12–18 months) (Table 1).

The results of MMWS and DASH scores throughout follow up are summarized in Table 2. The mean MMWS and DASH improved in both groups throughout follow up. All patients obtained a satisfactory outcome. The VA-LCP group had significantly improved MMWS and DASH scores throughout follow up compared to the FA-LCP group. At three months follow up, the VA-LCP MMWS and DASH were 76.3±5.7 and 24.4±2.1 respectively. The FA-LCP group had an MMWS of 67.4±5.3 and DASH of 31.7±2.9. The differences between these two groups at early follow up were significant (p = 0.02 for MMWS and p = 0.04 for DASH). Furthermore, the VA-LCP group con-tinued to have significantly improved MMWS and DASH outcomes a latest follow up evaluation at 12 months with MMWS 93.8±4.8 and DASH 9.2±1.0 compared to FA-LCP with MMWS 83.5±3.8 and DASH 12.8±1.6 (p < 0.01 for

MMWS and p = 0.02 for DASH). ROM showed recovery of 94.8% of flexion-extension

and 93.8% of supination-pronation movements as com-pared to the contralateral side in the VA-LCP group at 12 months follow up (Figure 3). In contrast, the FA-LCP group recovered 82.8% of flexion-extension and of 84.5% supination-pronation movements compared to the con-tralateral side at 12 months follow up. This difference in ROM was significantly different (p<0.01). Grip strength as relative to the contralateral side at 3 months follow up was 78.8% for the VA-LCP group and 75.7% for the FA-LCP group (p =0.03). At 12 months follow up percent grip strength was not different between groups (91.3% VA-LCP, 92.3% FA-LCP) (Figure 4). There was no soft tissue necrosis, infection, nerve palsy, vascular occlusion, or compartment syndrome observed in all patients. Two patients (2/19, 10.5%) in the VA-LCP group displayed flexor tendon irritation, but one of the two patients had their symptoms resolve within five days after their opera-tion. Another patient complained of mild irritation while forcefully gripping, but their symptoms were not severe enough for the patient to elect to undergo removal of hardware. In the FA-LCP group, six patients (6/28, 21.4%) suffered from intermittent or persistent flexor tendon ir-ritation. Two of these patients elected to undergo removal of hardware secondary to their symptoms. The incidence of flexor tendon irritation between the two groups was statistically significant (p=0.02).

Radiographic results showed a loss of reduction over the follow up period with decreases in radial inclination, radial height, and volar tilt in both groups between initial postoperative radiographs and 12 month follow up. There was a significant difference in volar tilt lost between the VA-LCP and FA-LCP groups at 3(p=0.01), 6(p=0.03) and 12 month(p=0.02) follow up visits. The initial volar tilt measured postoperatively was 10.05 degrees and 9.86 degrees in the VA-LCP and FA-LCP groups respectively.

Table 1. Comparison of Patient Demographics Between Both Groups

Number of Patients Gender(Male/Female) Age(year) Dominant hand injured

Time between injury and surgery(day)

VA-LCP 19 4M / 15F 52.9 (42-65) 9 2.8 (2-5)

FA-LCP 28 6M / 22F 53.5 (39-66) 12 2.9 (2-5)

p value - 0.77 0.74 0.73 0.89

Table 2. MMWS and DASH Functional Outcome Scores Over the 12 Month Follow Up Visit3 Months 6 Months 12 Months

MMWS DASH MMWS DASH MMWS DASH

VA-LCP 76.3(70-85) 24.4(20-28) 93.3(80-100) 10.5(2-14) 93.8(85-100) 9.2(2-12)

FA-LCP 67.4(60-75) 31.7(25-34) 80.6(70-90) 14.1(8-20) 83.5(75-90) 12.8(6-18)

p value 0.02 0.04 <0.01 0.03 <0.01 0.02



A B C

The latest follow up volar tilt was 8.37 degrees and 7.11 degrees in the VA-LCP and FA-LCP groups respectively (p=0.02). There was no difference in radial height, radial inclination, and volar tear drop angles detected between groups at all time points (Figure 5).

DISCUSSIONIntra-articular distal radius fractures that involve the

volar rim are challenging to manage. In these cases, it is paramount to restore the lunate facet as it is indispensable to maintain length, alignment, and stability of the wrist joint. Failure to stabilize the lunate facet may lead to carpal subluxation or dislocation.7,15 A volar marginal frag-ment (VMF) may be either too small or too distal to the watershed line to be adequately supported with traditional fixation strategies.2 Although conventional volar locked plates may provide a stable reduction,16,17 the geometry of the lunate facet poses exceptional challenges. In order to

support the entire volar surface effectively using conven-tional locked plates, the plate must be placed distal to the watershed line, which may lead to tendon irritation and intra-articular screw penetration.18-20 Furthermore, loss of fixation of the lunate facet may still occur. Harness et al. reported loss of fixation of a volar lunate facet fragment with carpal dislocation in a series of seven patients with an average of 24-months of follow-up.15 The VA-LCP low profile plating system was designed to be placed distal to the watershed line with an anatomically pre-contoured geometry that allows supplemental fixation into the lu-nate facet. Thus, the purpose of this study is to compare outcomes of VA-LCP to traditional fixation strategies with FA-LCP.

This study has several limitations. First, we recognize the limitations of our retrospective review. We had a lim-ited number of patients with a relatively short follow-up period that may not capture long term outcomes including

Figure 3. Clinical photographs demonstrating ROM of the 49-year-old female patient (Figure 2) at thirteen months postoperatively. Relative wrist flexion-extension (a) and supination-pronation (b) motion were near 100% compared with the contralateral side. (c) Incision demonstrat-ing the successful treatment of her right-sided injury.

Figure 4. Relative range of motion and grip strength throughout the 12 month follow up period. (a) Relative flexion-extension ROM (b) Rela-tive supination-pronation ROM (c) Relative grip strength. *p<0.05 when comparing between the two groups.



development of post traumatic osteoarthritis and flexor tendon rupture. Second, our differences in DASH scores throughout the follow up period, although statistically significant, were relatively small (<10 points). Previous studies have determined the threshold for minimal clini-cally important difference in DASH scores to be greater than 10 points.21 Furthermore, although there was a significant decrease of volar tilt in the FA-LCP group than the VA-LCP group, the volar tilt in both groups were within range of acceptable alignment. However, other functional assessments, including flexor tendon irritation and need for secondary surgery to remove hardware, were significantly greater in FA-LCP. This higher incidence of re-operation in the FA-LCP group compared to the VA-LCP group is clinically significant. Orbay et al. emphasized that salvage procedures may lead to further impaired functional outcomes in this patient population.2 Thus, the results of this study in aggregate demonstrate a meaningful clinical difference between VA-LCP and FA-LCP.

This study shows differences in functional outcomes between VA-LCP and FA-LCP. MMWS and DASH were improved at all time points in the VA-LCP compared to

FA-LCP group. As discussed in the limitations section, although these results were statistically significant, the clinical significance of this difference is less apparent. When assessing the relative wrist ROM, the VA-LCP had better recovery compared to FA-LCP. This difference in relative wrist ROM may contribute to the difference in MMWS.22 However, the difference in relative wrist ROM may be confounded by different postoperative immobiliza-tion protocols. The FA-LCP group was immobilized for a longer period of time (six weeks for FA-LCP versus one week for VA-LCP) because there was concern the fixation construct was less stable than that used for VA-LCP. This increased period of immobilization and delayed rehabilita-tion may have introduced a lag time bias. Longer follow up is needed to whether FA-LCP patients will recover more ROM to catch up to the VA-LCP group.

Relative grip strength was also improved in VA-LCP compared to FA-LCP at early follow up at three months. Again, this difference may have been related to the lon-ger immobilization protocol and delayed rehabilitation in FA-LCP. Interestingly, the FA-LCP group recovered grip strength more rapidly after three months and we were unable to detect a difference in grip strength between

Figure 5. Postoperative radiographic measurements of radial height, radial inclination, volar tilt, and volar tear drop angle throughout the 12 month follow up period; *p<0.05 when comparing between the two groups.



both groups at 12 month follow up. However, there was an apparent difference in flexor tendon irritation between both groups. The FA-LCP group had a high incidence of flexor tendon irritation (21.4% FA-LCP versus 10.5% VA-LCP) and a re-operation rate for removal of hard ware (7.1% FA-LCP versus 0% VA-LCP). In the FA-LCP group, the conventional plate had to be positioned distal to the watershed line in order to adequately stabilize the lunate facet, which may have contributed to flexor tendon irrita-tion. Furthermore, the supplemental K-wires used for fixa-tion may have also injured the tendons. A cadaveric study by Chia et al, found that volar radial styloid, transverse radial, and dorso-ulnar K-wires may penetrate both ten-dons and nerves about the wrist.23 Although VA-LCP also straddled the watershed line, the low profile anatomically contoured design with highly polished cambered surface may have contributed to a lower incidence of flexor tendon irritation. Other studies describing various techniques us-ing fragment-specific fixation, such as volar hook plates, found satisfactory outcomes in their series.2,8,16 However, O’Shaughnessy et al. reported 20% (5/25) of patients with an average follow up of nine months required removal of hardware secondary to flexor tendon irritation.8 This reported incidence is higher than what we observed with VA-LCP (0%, 0/19, average 15.8 months follow up). These results suggest that the design of the VA-LCP system may be successful at reducing tendon irritation at midterm follow up.

In addition to differences in clinical outcomes, we observed differences in radiographic outcomes. Radial height, radial inclination, and volar tear drop angle were not found to be different between groups at all time points. There was a larger loss of volar tilt in FA-LCP group com-pared to VA-LCP. The differences in stability of fixation holding the reduction can be explained by prior studies. In the FA-LCP group, although K-wire augmentation has been reported to be effective to enhance stability of the reduction in volar rim fractures, they are biomechanically inferior to a volar buttress for these fragments.6 Thus additional immobilization periods after conventional fixa-tion may be necessary to augment stability and prevent a greater loss of reduction. In contrast, the VA-LCP system firmly buttressed the volar rim with its additional options for fixation into the radial styloid, lunate facet, and distal radial-ulnar joint with variable angle screws in the radial and ulnar “tear drops” holes. However, the application of the VA-LCP device is not without limitations. Highly comminuted intra-articular fractures and extremely distal fractures that preclude capture with screws may not be well suited for VA-LCP. In these cases, external fixation may be the only salvage operation possible. Although there was difference of 1.3 degrees of volar tilt between groups, the clinical significance of this difference is likely minimal.

In conclusion, we report favorable clinical and radio-graphic results using the VA-LCP system compared to the FA-LCP. These results may be in part attributed to the VA-LCP system design with its low profile, ana-tomic contour, and multiple options for fixation that may decrease the incidence of joint penetration and improve lunate facet stability. Further research assessing the biomechanical properties of this system may further elucidate the mechanical properties of this plate and af-fect is has on the overlying flexor tendons. Furthermore, a long-term prospective study is needed to assess long term clinical and radiographic implications when using this device compared to conventional plates. Surgeons may consider the VA-LCP system as an alternative to conventional plates when treating radius fractures distal to the watershed region.

REFERENCES1. Schneppendahl J, Windolf J, Kaufmann RA

(2012) Distal radius fractures: current concepts. J Hand Surg Am 37 (8):1718-1725. doi:10.1016/j.jhsa.2012.06.001

2. Orbay JL, Rubio F, Vernon LL (2016) Pre-vent Collapse and Salvage Failures of the Volar Rim of the Distal Radius. J Wrist Surg 5 (1):17-21. doi:10.1055/s-0035-1570745

3. Kachooei AR, Tarabochia M, Jupiter JB (2016) Distal Radius Volar Rim Fracture Fixation Using DePuy-Synthes Volar Rim Plate. J Wrist Surg 5 (1):2-8. doi:10.1055/s-0035-1570740

4. Khanchandani P, Badia A (2013) Functional outcome of arthroscopic assisted fixation of distal radius fractures. Indian J Orthop 47 (3):288-294. doi:10.4103/0019-5413.109872

5. Chin KR, Jupiter JB (1999) Wire-loop fixation of volar displaced osteochondral fractures of the distal radius. J Hand Surg Am 24 (3):525-533. doi:10.1053/jhsu.1999.0525

6. Mellstrand Navarro C, Ahrengart L, Tornqvist H, Ponzer S (2016) Volar Locking Plate or External Fixation With Optional Addition of K-Wires for Dor-sally Displaced Distal Radius Fractures: A Random-ized Controlled Study. J Orthop Trauma 30 (4):217-224. doi:10.1097/bot.0000000000000519

7. Brogan DM, Richard MJ, Ruch D, Kakar S (2015) Management of Severely Comminuted Distal Radius Fractures. J Hand Surg Am 40 (9):1905-1914. doi:10.1016/j.jhsa.2015.03.014

8. O’Shaughnessy MA, Shin AY, Kakar S (2016) Stabilization of Volar Ulnar Rim Fractures of the Distal Radius: Current Techniques and Re-view of the Literature. J Wrist Surg 5 (2):113-119. doi:10.1055/s-0036-1579549



9. Orbay JL, Badia A, Indriago IR, Infante A, Khouri RK, Gonzalez E, Fernandez DL (2001) The extended flexor carpi radialis approach: a new perspective for the distal radius fracture. Tech Hand Up Extrem Surg 5 (4):204-211

10. Cooney WP, Linscheid RL, Dobyns JH (1994) Triangular fibrocartilage tears. J Hand Surg Am 19 (1):143-154. doi:10.1016/0363-5023(94)90238-0

11. Hudak PL, Amadio PC, Bombardier C (1996) Development of an upper extremity outcome measure: the DASH (disabilities of the arm, shoulder and hand) [corrected]. The Upper Extremity Collaborative Group (UECG). Am J Ind Med 29 (6):602-608. doi:10.1002/(sici)1097-0274(199606)29:6<602::aid-ajim4>3.0.co;2-l

12. Tanaka H, Hatta T, Sasajima K, Itoi E, Aizawa T (2016) Comparative study of treatment for distal radius fractures with two different palmar lock-ing plates. J Hand Surg Eur Vol 41 (5):536-542. doi:10.1177/1753193415625830

13. Medoff RJ (2005) Essential radiographic evaluation for distal radius fractures. Hand Clin 21 (3):279-288. doi:10.1016/j.hcl.2005.02.008

14. Fujitani R, Omokawa S, Iida A, Santo S, Tanaka Y (2012) Reliability and clinical importance of tear-drop angle measurement in intra-articular distal radius fracture. J Hand Surg Am 37 (3):454-459. doi:10.1016/j.jhsa.2011.10.056

15. Harness NG, Jupiter JB, Orbay JL, Raskin KB, Fernandez DL (2004) Loss of fixation of the volar lunate facet fragment in fractures of the distal part of the radius. J Bone Joint Surg Am 86-a (9):1900-1908

16. Bakker AJ, Shin AY (2014) Fragment-specific volar hook plate for volar marginal rim fractures. Tech Hand Up Extrem Surg 18 (1):56-60. doi:10.1097/bth.0000000000000038

17. Orbay JL, Fernandez DL (2002) Volar fixation for dorsally displaced fractures of the distal radius: a preliminary report. J Hand Surg Am 27 (2):205-215

18. Cross AW, Schmidt CC (2008) Flexor tendon inju-ries following locked volar plating of distal radius frac-tures. J Hand Surg Am 33 (2):164-167. doi:10.1016/j.jhsa.2007.11.011

19. Limthongthang R, Bachoura A, Jacoby SM, Os-terman AL (2014) Distal radius volar locking plate de-sign and associated vulnerability of the flexor pollicis longus. J Hand Surg Am 39 (5):852-860. doi:10.1016/j.jhsa.2014.01.038

20. Majima M, Horii E, Matsuki H, Hirata H, Genda E (2008) Load transmission through the wrist in the extended position. J Hand Surg Am 33 (2):182-188. doi:10.1016/j.jhsa.2007.10.018

21. Sorensen AA, Howard D, Tan WH, Ketchersid J, Calfee RP (2013) Minimal clinically important differ-ences of 3 patient-rated outcomes instruments. J Hand Surg Am 38 (4):641-649. doi:10.1016/j.jhsa.2012.12.032

22. Slutsky DJ (2013) Outcomes assessment in wrist sur-gery. J Wrist Surg 2 (1):1-4. doi:10.1055/s-0033-1333892

23. Chia B, Catalano LW, 3rd, Glickel SZ, Barron OA, Meier K (2009) Percutaneous pinning of distal radius fractures: an anatomic study demonstrating the proximity of K-wires to structures at risk. J Hand Surg Am 34 (6):1014-1020. doi:10.1016/j.jhsa.2009.04.004


ABSTRACTBackground: Adolescent idiopathic scoliosis

(AIS) is defined as a lateral curvature of the spine of unknown etiology with a Cobb angle of greater than 10 degrees with vertebral rotation. Bracing, specifically with a rigid thoracolumbosacral ortho-sis (TLSO), decreases the risk of curve progres-sion to over 50 degrees, the threshold for surgical intervention. Some authors have suggested that 30-50% in-brace correction of the Cobb angles is required to prevent significant curve progression. The purpose of the study is to evaluate the cur-rent bracing protocol at the University of Iowa as a quality control exercise for the treatment team.

Methods: AIS patients (n = 61; 8 male, 53 female) who received a Rosenberger TLSO at the University of Iowa Department of Orthopaedics and Rehabilitation from 2016-2017 were included in the study. Inclusion criteria include presence of pre-brace and in-brace x-rays within 3 months of initiating brace treatment. Patients with other diagnoses were excluded. Radiographic indicators of brace effectiveness, such as the Cobb angle, were measured.

Results: The in-brace x-rays of 46 (76%) pa-tients showed less than 30% correction. Minimal changes from the pre- to in-brace x-ray were ob-served in other radiographic measures.

Conclusions: Results indicate that if the 30-50% correction recommended by the literature is valid, then modifications to the process of measuring, fabricating or modifying our current TLSO’s for AIS are warranted.

Level of evidence: IIIKeywords: brace, cobb angle, adolescent idio-

pathic scoliosis

INTRODUCTIONAdolescent idiopathic scoliosis (AIS) is defined as a

lateral curvature of the spine of unknown etiology with a Cobb angle of greater than 10 degrees with vertebral rotation.1 The prevalence of AIS is approximately 3% in children under the age of 16, with 0.3-0.5% having wors-ening curves requiring treatment.2 Bracing, specifically with a rigid thoracolumbosacral orthosis (TLSO), has been shown to decrease the risk of curve progression to over 50 degrees, one commonly used threshold for surgical consideration.1 Preventing surgery through ef-fective bracing is an important goal not only for patients, but also for the United States healthcare system. In 2012, spinal fusion for AIS resulted in the highest aggregate (approximately $341 million) costs for procedures in children between the ages of 10 and 14.3

Given a patient’s baseline characteristics (e.g. magni-tude of the curvature and skeletal maturity) the two most important factors thought to be associated with bracing success are the amount of correction obtained in the brace, and the number of hours of wear time per day.1,4,5 The literature suggests braces must correct the curve by at least 30-50% in order to prevent significant curve progression.6,7,8 In addition to correction of the coronal Cobb angle, components of the deformity (e.g. kyphosis, lordosis, coronal compensation) are increasingly being considered. For example, hypokyphosis independent of pelvic parameters is common in AIS and is associated with decreased pulmonary function.9,10 Sagittal imbal-ance, particularly seen in spinal fusion patients, leads to increased cervical and lumbar complications, includ-ing low back pain.10 Surgical correction in the coronal and sagittal planes have beneficial effects on center of mass and center of pressure, ultimately improving gait.11 Apical vertebral rotation is an important prognostic fac-tor for curve progression in both thoracic and lumbar curves.12,13

The purpose of the study is to evaluate the immediate (in-brace) outcomes of a recent sample of patients be-ing treated for AIS. Specifically, we analyze the amount of correction achieved in the brace at the beginning of treatment and compare it to recommendations from the literature.

EVALUATION OF PREDICTORS AND OUTCOMES OF BRACING WITH EMPHASIS ON THE IMMEDIATE EFFECTS OF IN-BRACE CORRECTION

IN ADOLESCENT IDIOPATHIC SCOLIOSIS

Tzu Chuan Yen, MD1 and Stuart L. Weinstein, MD2,3

1University of Missouri - Columbia, Department of Orthopaedic Surgery, Columbia MO USA2University of Iowa Department of Orthopaedics and Rehabilitation, Iowa City IA USA3University of Iowa Department of Pediatrics, Iowa City IA USACorresponding author: Stuart L. Weinstein, MD; Phone (319) 356-1872; [email protected]: The authors report no potential conflicts of interest related to this study.Sources of Funding: No sources of funding declared.

T. C. Yen, S. L. Weinstein


METHODSParticipants

AIS patients who received a Rosenberger TLSO brace at the University of Iowa Department of Orthopaedics and Rehabilitation from 2016-2017 and had both pre-treat-ment radiographs and a radiograph in the brace within 3 months were included. Exclusion criteria include patients receiving a TLSO for a diagnosis other than AIS.

Data CollectionDemographics were collected from the medical

record. The pre-treatment hand, posterioranterior and lateral full-spine, and the in-brace radiographs were downloaded for measurement. The Cobb angle(s), tho-racic kyphosis, lumbar lordosis, apical vertebral rotation, sagittal balance, and coronal balance were measured by the first author and spot-checked by the third author for accuracy. The Risser grade and Sanders skeletal maturity score (SSMS) were measured to assess skeletal maturity.14,15

Data AnalysisSummary statistics included the mean with standard

deviation, median, range, and 95% confidence intervals. The obtained measurements were compared to the ranges of normal (Table 1). Data analysis was performed using Microsoft Excel (2016).

RESULTSData and images from a total of 61 patients (8 male,

53 female) with a mean age of 12.81±1.37 years were extracted from the EPIC™ medical record and included in the study (Table 2).

Pre-brace vs. in-brace radiographic measurementsMost of the patients were skeletally immature, with

59% at less than Risser 2 and 59% had an SSMS bone age less than 4. The mean Cobb angle of the largest curve prior to treatment was 27.98±7.32 degrees compared to 22.46±7.96 degrees in-brace (Table 3). The average curve correction was 5.5 degrees. The in-brace correction in 48% of patients was less than 15%, 28 % was between 15-30% correction, 16% with 30-50% correction, and 8% had over 50% correction (Figure 1). The coronal balance was restored in the brace in 5% of patients, whereas it continued to be abnormal 5% and 5% of patients became decompensated in the brace (Table 4). The average change in coronal balance was 3.2 x 10-2 cm. 82% had normal sagittal balance (< 5 cm deviation) initially and in-brace, with only 10% of patients improving from abnor-mal to normal and 5% worsening. The average change in sagittal balance was 0.35 cm. The thoracic kyphosis was within normal range (20 to 45 degrees) in 61% of patients both at baseline and in the brace, 16% became hypokyphotic, and kyphosis was increased into the nor-mal range in 13%. The average change in kyphosis was 1.8 degrees. 61% of patients were hyperlordotic at base-line and in-brace, 21% improved from hyperlordotic to normal (less than 45 degrees), while lordosis increased in 8% who became hyperlordotic. The average change in lordosis was 3.1 degrees. Apical vertebral rotation was consistently normal (5 degrees or less) both pre-brace and in-brace in 36% of patients; 3% developed abnormal rotation and rotation was decreased in 13% (improved from abnormal to normal). The average change in apical vertebral rotation was 2.6 degrees.

Table 1. Definitions of Normal for Radiographic Measurements

Definitions

Thoracic kyphosis 20 to 45 degrees

Lumbar lordosis < 45 degrees

Apical vertebral rotation ≤ 5 degrees

Sagittal balance < 5 cm deviation

Coronal balance ≤ 2 cm deviation

Table 2. Demographics of AIS patients Mean with SD Median Range

Age 12.8 ± 1.4 12.7 12.7 - 17.1

Height (cm) 158.0 ± 9.2 158 131.3 – 180.1

Height per-centile 64.3 ± 26.8 71.0 1.3 – 99.6

Weight (kg) 46.7 ± 12.0 44.0 27.2 – 97.3

Weight percentile 52.0 ± 27.1 52.5 2.1 – 99.6

Table 3. Mean with Standard Deviation (range) and Average Change of Pre-Brace and In-Brace

Radiographic Measurements

Pre-brace In-brace Average Change

Cobb angle (degrees) 28.0 ± 7.3 (10-43) 22.5 ± 8.0 (5-41) 5.5 ± 6.0

Thoracic kyphosis (degrees) 29.0 ± 10.1 (7-55) 27.2 ± 9.2 (4-46) 1.8 ± 7.2

Lumbar lordosis (degrees) 54.0 ± 10.1 (28-80) 50.9 ± 9.3 (27-72) 3.1 ± 8.4

Apical vertebral rotation (degrees) 10.5 ± 6.9 (0-30) 8.4 ± 6.8 (0-30) 2.1 ± 4.8

Sagittal balance (cm) 2.6 ± 1.8 (0.1-6.3) 2.2 ± 1.6 (0.4-1.6) 0.35 ± 2.1

Coronal balance (cm) 1.4 ± 0.9 (0.1-3.9) 1.3 ± 1.2 (0.1-4.6) 3.2 x 10-2

± 1.1

Risser 59% less than 2 (0-4) - -

Sanders 59% less than 4 (2-6) - -


Evaluation of Predictors & Outcomes of Bracing in Adolescent Idiopathic Scoliosis

A

DISCUSSIONPrevious studies have shown that Cobb angle cor-

rection with a TLSO significantly influences curve progression. Katz et al. (2001) showed that a 25% or greater in-brace Cobb angle reduction was associated with a 73% success rate (curve progression of <6 de-grees) in a study of Boston braces.16 Similarly, Emans et al. (1983) found that the greater the in-brace Cobb angle correction, the better the outcome, suggesting a goal of 50% correction to halt curve progression.6 In a retrospective study, Landauer et al. (2003) showed that over 40% curve correction was significantly related to successful outcomes in patients using a Chêneau-type brace.8 Goodbody et al. (2016) found that 45% or greater in-brace correction was associated with bracing success.7 The braces in our study fell short of these standards, as only 24% resulted in more than 30% correction. It must be acknowledged, however, that the achievable in-brace correction is likely limited by the flexibility of the curve, amount of vertebral rotation, skeletal maturity and other factors. In literature, both the SSMS bone age and Risser stage is significantly correlated with curve progression.17 Additionally, the contribution of in-brace correction to the final result relative to other factors such as maturity, curve pattern and wear time has not been established. Little change was observed in the other radiographic parameters associated with AIS. New Com-puter Aided Design and Computer Aided Manufacturing (CAD/CAM) techniques have allowed a 3D approach to designing braces that could improve other parameters, such as kyphosis, lordosis, and coronal balance, along with Cobb angles.18 With the CAD/CAM system, the optimal brace shape and placement of pressure pads for immediate in-brace correction can be simulated.18 These findings indicate that substantial changes should be instituted in our bracing program. Modifications to the measurement system, the fabrication process, and or the fitting adjustments of our braces should be instituted.

The limitations of this study include a small sample size and the fact that changes may have been made to the orthosis after review of the in-brace x-ray. Modifi-cations were not immediately followed by an additional in-brace x-ray, therefore, the measurements here might not accurately reflect the corrective forces of the brace after that point.

Future research could track wear time and curve progression all the way to skeletal maturity in addition to obtaining radiographic measurements. Because of this study, patients with AIS at the University of Iowa are scheduled to see the prosthetist before all x-rays and clinic appointments to ensure optimal brace fit. Posteroanterior and lateral x-rays are required at brace delivery and repeated at six weeks per standard of care. Further studies performed after these adjustments will assess whether or not effective in-brace Cobb angle correction is now achieved in AIS patients.

REFERENCES1. Weinstein SL, Dolan LA, Wright JG, Dobbs MB.

Effects of bracing in adolescents with idiopathic scoliosis. N Engl J Med. 2013 Oct 17;369(16):1512-21. PubMed PMID: 24047455; NIHMSID: NI-HMS542731; PubMed Central PMCID: PMC3913566.

2. Nachemson A. Report of the prevalence and natural history committee of the Scoliosis Research Society. the Annual Meeting of the Scoliosis Research Society; September 22, 1982; Denver, CO. 1982.

3. Rank Order of Clinical Classification Software (CCS) Procedures by Aggregate Costs, National, Children Only: Age Group 10-14. Healthcare Cost and Utiliza-tion Project (HCUP). 2012. Agency for Healthcare Research and Quality, Rockville, MD.

Figure 1. Distribution of in-brace percent correction (Initial Cobb angle – in-brace angle)/Initial Cobb angle*100.

Table 4. Number and Percentage of Patients with Radiographic Changes from Pre-Brace to In-Brace

Maintained Normal Range

Became Abnormal

in the Brace

Became Normal in

Brace

Remained Abnormal

Thoracic kyphosis 37 (61%) 10 (16%) 8 (13%) 6 (10%)

Lumbar lordosis 6 (10%) 5 (8%) 13 (21%) 37 (61%)

Apical vertebral rotation

22 (36%) 2 (3%) 8 (13%) 29 (48%)

Sagittal balance 50 (82%) 3 (5%) 6 (10%) 2 (3%)

Coronal balance 46 (75%) 6 (10%) 3 (5%) 6 (10%)

T. C. Yen, S. L. Weinstein


A B C

4. Karol LA, Virostek D, Felton K, Jo C, Butler L. The Effect of the Risser Stage on Bracing Outcome in Adolescent Idiopathic Scoliosis. J Bone Joint Surg Am. 2016 Aug 3;98(15):1253-9. PubMed PMID: 27489315.

5. Katz DE, Herring JA, Browne RH, Kelly DM, Birch JG. Brace wear control of curve progression in adolescent idiopathic scoliosis. J Bone Joint Surg Am. 2010 Jun;92(6):1343-52. PubMed PMID: 20516309.

6. Emans JB, Kaelin A, Bancel P, Hall JE, Miller ME. The Boston bracing system for idiopathic sco-liosis Follow-up results in 295 patients. Spine (Phila Pa 1976). 1986 Oct;11(8):792-801. PubMed PMID: 3810295.

7. Goodbody CM, Asztalos IB, Sankar WN, Flynn JM. It’s not just the big kids: both high and low BMI impact bracing success for adolescent idiopathic scoliosis. J Child Orthop. 2016 Oct;10(5):395-404. PubMed PMID: 27501808; PubMed Central PMCID: PMC5033782.

8. Landauer F, Wimmer C, Behensky H. Estimating the final outcome of brace treatment for idiopathic thoracic scoliosis at 6-month follow-up. Pediatr Re-habil. 2003 Jul-Dec;6(3-4):201-7. PubMed PMID: 14713586.

9. Clément JL, Geoffray A, Yagoubi F, Chau E, Solla F, et al. Relationship between thoracic hypokypho-sis, lumbar lordosis and sagittal pelvic parameters in adolescent idiopathic scoliosis. Eur Spine J. 2013 Nov;22(11):2414-20. PubMed PMID: 23771577; PubMed Central PMCID: PMC3886492.

10. Lonner BS, Lazar-Antman MA, Sponseller PD, Shah SA, Newton PO, et al. Multivariate analysis of factors associated with kyphosis maintenance in adolescent idiopathic scoliosis. Spine (Phila Pa 1976). 2012 Jul 1;37(15):1297-302. PubMed PMID: 22228329.

11. Paul JC, Patel A, Bianco K, Godwin E, Naziri Q, et al. Gait stability improvement after fusion surgery for adolescent idiopathic scoliosis is influenced by corrective measures in coronal and sagittal planes. Gait Posture. 2014 Sep;40(4):510-5. PubMed PMID: 25023225.

12. Weinstein SL, Ponseti IV. Curve progression in idiopathic scoliosis. J Bone Joint Surg Am. 1983 Apr;65(4):447-55. PubMed PMID: 6833318.

13. Spoonamore MJ, Dolan LA, Weinstein SL. Use of the Rosenberger brace in the treatment of progres-sive adolescent idiopathic scoliosis. Spine (Phila Pa 1976). 2004 Jul 1;29(13):1458-64. PubMed PMID: 15223939.

14. Risser JC. The classic: The iliac apophysis: an in-valuable sign in the management of scoliosis 1958. Clin Orthop Relat Res. 2010 Mar;468(3):643-53. PubMed PMID: 19763720; PubMed Central PMCID: PMC2816762.

15. Sanders JO, Khoury JG, Kishan S, Browne RH, Mooney JF 3rd, et al. Predicting scoliosis progres-sion from skeletal maturity: a simplified classification during adolescence. J Bone Joint Surg Am. 2008 Mar;90(3):540-53. PubMed PMID: 18310704.

16. Katz DE, Durrani AA. Factors that influence out-come in bracing large curves in patients with adoles-cent idiopathic scoliosis. Spine (Phila Pa 1976). 2001 Nov 1;26(21):2354-61. PubMed PMID: 11679821.

17. Sitoula P, Verma K, Holmes L Jr, Gabos PG, Sanders JO, et al. Prediction of Curve Progression in Idiopathic Scoliosis: Validation of the Sanders Skel-etal Maturity Staging System. Spine (Phila Pa 1976). 2015 Jul 1;40(13):1006-13. PubMed PMID: 26356067.

18. Desbiens-Blais F, Clin J, Parent S, Labelle H, Aubin CE. New brace design combining CAD/CAM and biomechanical simulation for the treatment of adolescent idiopathic scoliosis. Clin Biomech (Bristol, Avon). 2012 Dec;27(10):999-1005. PubMed PMID: 22989479.


ABSTRACTBackground: The impact of surgery on the

quality of life of adolescents with idiopathic sco-liosis (AIS) remains to be clarified as most of the studies are retrospective and few include quality of life questionnaires completed in the pre- and postoperative periods.

Methods: Operated patients with AIS who com-pleted preoperative and postoperative SRS-22 questionnaires were selected for evaluation. The demographic data were collected and quality of life improvement was assessed by comparing deviation to the means with standard deviation at both mo-ments. Using the Minimal Important Clinical Dif-ference (MICD) concept, individual improvement was also assessed.

Results: 28 patients (27 females) with an aver-age age of 14.4 years (min: 12 – max: 18) and an average Cobb angle of 61.46º (min: 35º– max: 98º) were retrospectively reviewed. The correction rate was 68% with a final Cobb angle of 18.97º (min: 7.37º – max: 37.6º). The Global SRS22 score and all sub domains showed a significant variation (p<0,01). Self-image, followed by mental health, were the subdomains where the mean dif-ferences was more relevant, with the highest effect dimension given by d-Cohen analyses (Self-image d-Cohen 2.51; Mental health –d-Cohen 0.86). Both mean score differences as well as global SRS22 score reached the MICD but, while 96.4% of the patients did so for self-image ,the same only hap-pened in 50% of the patients in Mental health and global score. No clinical relevant change occurred in pain or activity domains.

Conclusion: Taking into consideration the AIS natural history and the fact that the most relevant change after surgery occurs in patients’ self-esteem, the advantages of a surgical treatment should be thoroughly evaluated not only based on curve severity but also by looking at which quality of life subdomains are mostly affected and which are expected to improve in order to meet proper patient expectations.

Level of evidence: IIIKeywords: idiopathic scoliosis, surgery, quality

of life, SRS-22

INTRODUCTIONAdolescent idiopathic scoliosis (AIS) is a tridimensional

deformity of the spine which usually manifests itself at the beginning of puberty through an esthetical alteration of the trunk. This disease has a 2-3% prevalence in the general population and predominantly affects the female gender, with 10% of the patients requiring treatment and only 0.1% surgical correction.1-3 Weinstein et al., who stud-ied a population of patients with scoliosis untreated for 50 years and compared it with an equivalent population with no scoliosis, revealed a similar survival rate and dyspnea on exertion in 22% of the population with scoliosis (15% in the control population, odds ratio=9.25 scoliosis > 80º and thoracic apex). Regarding pain, 61% of the scoliosis patients presented lumbar pain (35% in the control popula-tion), even though the pain was considered mild to moder-ate in most of the cases. Self-image and the perception of functional limitation were the most affected areas, despite the almost overlapping of the psychosocial indexes of the two populations.4 Conversely, other retrospective studies suggest that untreated patients refer reduced activity level, pain and psychosocial problems, including lack of self-esteem and depression.5-7

In fact, even though scoliosis is easily recognizable, the greatest difficulty lies in realizing to what extent deformity affects the quality of life of adolescents. Fre-idel K et al. mentioned that adolescents are less enthusi-astic with their lives and present greater tendency for depression and physical incapacity when compared to peers.8 Payne et al. stated that, regardless of treatment phase, scoliosis in itself was considered a risk factor for psychiatric disorder,9 which may explain the differences

IMPACT OF SURGERY ON THE QUALITY OF LIFE OF ADOLESCENT IDIOPATHIC SCOLIOSIS

Pedro Fernandes1; Joaquim Soares Do Brito1; Isabel Flores2; Jacinto Monteiro1

1Orthopedics Department of the North Lisbon Hospital Center, EPE; Santa Maria Hospital, Lisbon, Portugal2Instituto Universitario de Lisboa, Lisbon, PortugalCorresponding author: Joaquim Soares do Brito; Email: [email protected]; Phone: +351 926755558Disclosures: The authors report no potential conflicts of interest related to this study.Sources of Funding: No sources of funding declared.

P. Fernandes, J. Soares Do Brito, I. Flores, J. Monteiro


in young people’s adherence to conservative treatments with braces, or even in the interpretation of the clinical results of the surgery.11-13 In recent years an effort has been made to include evaluation scales filled in by patients themselves in clinical research interpretation, particu-larly those from the Scoliosis Research Society (SRS),14 The validity of the SRS scale with 22 questions (SRS-22) was well demonstrated by Arsher et al. in the context of idiopathic scoliosis, and it has been validated for several idioms and is amply used.15-17 Yet there are still difficulties in the interpretation of the studies which applied these scales due to a relative lack of data on normality and the fact that statistically significant differences might not be clinically relevant. Therefore, the concept of Minimal Im-portant Clinical Difference (MICD) emerged, which helps to compare populations with and without scoliosis, as well as to determine the clinical outcome of the treatment.18-20

AIS surgery aims to stop deformity progression by preventing the emergence of degenerative disorders of the spine, as well as the compromise of the cardiore-spiratory function present in progressive deformities of the thoracic spine. Even though recent studies show that 23-54% of the patients report spine pain, this disease is still considered asymptomatic and rarely is pain the main surgical indication.21-25 And if surgical correction is usually effective in reducing deformity, an objective improvement in the quality of life of the adolescent is not always easily foreseen, as there is no direct correlation with the radiographic result. Hence, the importance of clarification for both the adolescent and their parents on the purpose of surgical indication and, above all, on what to expect from scoliosis surgery, is paramount, as most of the patients are highly functional.

This study aimed to assess the impact of AIS surgery on adolescent quality of life by applying the SRS-22 questionnaire before and after the surgery, as well as by trying to identify the subdomains where this impact is more relevant according to well established MICDs.

METHODSFrom all AIS patients who underwent surgery in our

department, the patients who completed the SRS-22 ques-tionnaire on quality of life before and after the surgery at the date of the last assessment at orthopedics outpatient clinic were the ones selected for our study. All deformities were measured in pre- and postoperative long standing radiographs at the last follow-up appointment (Figure 1).

The SRS-22 questionnaire applied to the patients presents 20 questions distributed among four subdomains: Function, Pain, Self-image and Mental health, and two additional questions regarding satisfaction with received treatment. Each subdomain consists of five questions scored from one (worst) to five (best). The questionnaires

were answered by the adolescents, alone or in the presence of their parents, out of the consultation office. Data were recorded in a Microsoft Excel spreadsheet and statistical analysis was performed using the IBM SPSS Statistics 23 software.

T-Student tests were used in matched samples to analyze the results, as the goal was to measure the dif-ference of the mean of patients’ answers before and after the surgery by using mean values and standard deviation at both moments. D-Cohen was used as measure of effect size, since it expresses the differences between the means in standard deviations. Differences whose D-Cohen scores were above 0.2 were considered relevant, in compliance with the interpretative table for this indicator. A new variable (Zvariation=global SRS score or the score of each final subdomain - global SRS score or each preoperative subdomain) was created for this assessment, so as to deter-mine the proportion of patients who improved, remained the same or deteriorated, as the information was consid-ered complementary to the interpretation of the differ-ences of the means. Finally, the differences obtained with surgery and the Minimal Important Clinical Difference (MICD), i.e. the difference with impact on the patient’s quality of life18.19 in each subdomain, were compared. The MICD values used, resulting from the method based on the relevant mean effect, which considered the standard deviation and the instrument correlation coefficient for the four subdomains (0.6 for Pain, 0.8 for Function, 0.5 for Self-image and 0.4 for Mental health), were in line with the recommendations of Bago et al.20 Regarding the global SRS-22 score, the difference based on improvement perception tests was applied, as this was higher than the value resulting from the analysis based on the mean ef-fect. Satisfaction results were also analyzed, despite the

Figure 1. Sixteen year old at preoperative (a) with marked deformity of hte trunk presenting a SRS-22 global score of 3.52 (Function: 4; Pain: 3.8; Self-image: 1.8; Mental health: 2.8); 14-year-old at postop-erative (b) with corrected deformity presenting a SRS-22 global score of 4.54 (Function: 4.2; Pain: 3.8; Self-image: 5; Mental health: 4.6).


Impact of Surgery on the Quality of Life of Adolescent Idiopathic Scoliosis

A

absence of a Minimal Important Clinical Difference.

RESULTSThis study analyzed 28 adolescents (27 females) with

a mean age of 14.4 years (min: 12 years – max: 18 years) suffering from a mean deformity of 61.46º (Cobb angle: min: 35 – max: 98º) whose AIS surgery was performed by the same surgeon and who completed their SRS-22 questionnaires before and after the surgery. The patients presented relatively long spine instrumentations with an average of 11 fixated vertebrae per instrumentation (min: 5 – max: 15). The deformity correction rate obtained with surgery, i.e. the final Cobb angle – initial Cobb angle multiplied by 100, was 68.2% (min: 40 – max: 90.7%), and a mean deformity of the final Cobb angle of 18.97º (min: 7.37 – max: 37.6º) was obtained. Thoracoplasty with the partial resection of three to five ribs at the deformity convexity level was resorted to in 11 cases so as to improve the esthetical result of the surgery. Spine instrumentations were preferably performed with screws to fix vertebrae pedicles and two rods to attach the screws, with a mean implant density (number of screws used per included vertebra multiplied by two, as each vertebra has two pedicles, multipied by 100) of 68% (min: 46% – max: 100%). In 13 patients only the major curve was instrumented. In no case was there any need to return to the operating

room to resolve complications. All the patients answered the SRS-22 questionnaire both before and after surgery with a mean follow-up of 36 months (min: 24 – max: 72).

When comparing the questionnaires completed pre-operatively and those completed postoperatively at the time of the last assessment, a statistically significant improvement (P≤0.01) was observed both in the global score and in the subdomains. Nevertheless, only the global SRS score, Self-image and Satisfaction, and, to a lesser extent, Mental health, exhibited a meaningful dimen-sion, becoming more relevant (Table 1). Therefore, the difference observed with the surgery in the subdomains Pain (D-Cohen=0.68) and Function (D-Cohen=0.75) were less relevant, with a small mean variance of 0.39 and 0.32, respectively, since high preoperative scores were already present. Self-image and Mental health were the subdomains where the adolescents presented the lowest preoperative scores, which translates the impact that de-formity has on the psychosocial sphere (Table 1). On the other hand, Self-image was the subdomain where the most significant variation introduced by surgery was observed, with a mean difference of 1.44 and an effect size of 2.51.

The evolution of the mean of each score after the surgery is presented in Figure 2 (a), where the number of patients which improved, maintained or reduced their

Table 1. Clinical Result of Adolescent Idiopathic Scoliosis Surgical Treatment Comparing SRS-22 Questionaire Scores in Pre and Post Operative Periods with a Mean Follow Up of 36 Months

SRS22 Domains Pre-operative Post-operative S (p) Mean Difference D-Cohen

SRS Global score 3.79 (0.43) 4.46 (0.32) 0.000 0.67 1.76

Function 4.04(0.45) 4.36 (0.39) 0.002 0.32 0.75

Pain 4.01 (0.61) 4.4 (0.54) 0.01 0.39 0.68

Self-image 2.96 (0.7) 4.4 (0.41) 0.000 1.44 2.51

Mental Health 3.72(0.68) 4.25 (0.54) 0.001 0.53 0.86

Satisfaction 4.15 (0.73) 4.87 (0.49) 0.000 0.72 1.15

Figure 2. a) evolution of the means for each score after the surgery; b) number of patients which improved, maintained or reduced the score in the various domains.

0

0,5

1

1,5

2

2,5

3

3,5

4

4,5

5

SRS Global Function Pain Self-image Mental Health Satisfaction

Pre operative Post Operative

1819

2122

26

28

4

7

2 1

61

5 5 2

0

5

10

15

20

25

Pain Satisfaction Mental Health Function SRS Global Self image

Improved The same Worse0

0,5

1

1,5

2

2,5

3

3,5

4

4,5

5

SRS Global Function Pain Self-image Mental Health Satisfaction

Pre operative Post Operative

1819

2122

26

28

4

7

2 1

61

5 5 2

0

5

10

15

20

25

Pain Satisfaction Mental Health Function SRS Global Self image

Improved The same Worse



scores in the various subdomains (b) can be observed. The same figure also shows that all the patients were satisfied with the surgery, and just in two cases (7.1%) did the SRS-22 score diminish after the surgical intervention. Only Pain and Function correlated with one another (r=0.389, p=0.041), and no relevant direct correlation was observed between Self-image and Mental health.

When calculating the variation of each score pre- and post-surgery (Zvariation) and comparing this with the Minimal Important Clinical Difference, we verified that even though 98% of the patients improved their global SRS-22 score, only 50% reached a MICD of 0.6 (Figure 3). Moreover, only four patients (14.2%) exceeded the MICD threshold in the Function domain (Figure 4a) and nine patients (32%) reached the MICD in the Pain domain (Figure 4b). Improvement was observed in 100% of the patients in Self-image, where 27 adolescents (96.4%) reached the MICD. The same did not, however, occur in Mental health, where greater variability was registered in the evolution of the score, as only 14 patients (50%) reached the MICD (Figures 5a and 5b). Furthermore, a decrease in the global SRS-22 score was observed in two patients, although the MICD was not exceeded (Figure 2b). Notably, a relevant number of patients reported lower scores in the subdomains Pain, Function and Mental health after the surgery. Finally, the satisfaction analysis of the questionnaires completed by the adolescents lead us to state that all the cases presented a higher level of satisfaction postoperatively, when compared to the pre-operative period (Figure 6).

DISCUSSIONThe results emerging from this study can be consid-

ered particularly relevant to both the patients and their

families and the surgeons who treat scoliosis. This study clearly states that with a deformity correction rate of 68%, which overlaps those published in literature, surgery sta-tistically significantly improved all the domains assessed by the SRS-22 quality of life questionnaire. Nevertheless, only the results of Self-image and Mental health deviate from the mean, with a relevant effect size which also conditions the variation of the global score of the question-naire. Even though thoracic deformity is quite visible, its repercussions in the quality of life of adolescents is not linear, which may hinder the appreciation of the impact of surgery on the psychometric domains of clinical outcome scales26 and, therefore, explain these results.

Rushton et al. assessed the impact of scoliosis on the quality of life of adolescents by comparing SRS-22 scores of adolescents with scoliosis with scores of the same questionnaire completed by adolescents without scoliosis. 81% of the published series showed that the adolescents with scoliosis registered a reduction in the Pain score, even though only in 5% of the series was this difference clinically relevant. On the other hand, Self-image de-creased in 91% of the series, of which 73% were clinically relevant reductions. Moreover, neither Mental health nor Function presented clinically important differences.27 Similarly, in our population, Self-image was the only clinically decreased subdomain during the preoperative period, whereas the remaining subdomains reflected the scores reported by the normal population published by Verma et al.28 Hence, the need arises to clearly understand the value which may eventually be added with scoliosis surgery. Even though 40% of the adolescents do not seem to be affected by the diagnosis and the eventual treatment of the disease, Danielsson et al. referred that 25% to 43% of the adolescents may present symptoms of depression and isolation, thus translating the impact of the deformity on their psychic sphere.29 In line with this association, Andersen et al. showed that, irrespective of the treat-ment performed, there is significant improvement in the psychologic outcome when the procedure is performed before the age of 16 years. Therefore, surgery should not be long postponed, especially in cases of children with low psychometric evaluations in the Self-image and Mental health subdomains.30

As shown, adolescent idiopathic scoliosis surgery aims to prevent deformity progression, thus preserving respiratory function, delaying the appearance of degen-erative alterations, and, at the same time, improving the appearance of the trunk. In essence, the surgery which treats adolescent idiopathic scoliosis seeks to substan-tially improve the quality of life of young people and is usually recommended for deformities whose Cobb angle is above 40/50o. It is, therefore, the responsibility of the surgeon to convincingly demonstrate that this goal may

Figure 3. Value of individual variations of the SRS-22 score, resulting from the subtraction of the initial score from the final score, with horizontal line representing MICD value.



be achieved with the surgery. For this reason, clinical outcome studies should include not only radiographic as-pects, but also features related to the quality of life of the adolescents, namely the activity, pain, mental health and cosmetic levels. Moreover, patient expectations should be thoroughly included in the study, for, if not very clear, they might not be achieved and might affect the clinical outcome of the procedure.

After comparing the means of each score of the SRS-22 questionnaire, both before and after the surgery, the Minimal Important Clinical Difference (MICD)18,19 concept was applied to the differences recorded for each patient in each subdomain. A gain in Self-image becomes, then, evident, as 98% of the patients exceeded MICD. The same

did not occur for Mental health, where only 50% of the patients presented a clinically relevant positive variation. For Pain and Function, our data showed that, based on a sample of 28 patients without any complications, only four (14%) referred an increased activity level and nine (32%) mentioned a clinically relevant improvement in Pain. Regarding the global SRS-22 score, 50% of the patients achieved a clinically relevant improvement. However, as abovementioned, this is mostly due to an improvement in Self-image, which, on its own, conditioned the great level of satisfaction. Bago et al. reached the same conclusion when they reported an improvement above the MICD in the SRS-22 score in only 52% of the adolescents after a 2-year follow-up.20 On the other hand, Upasani et al. reg-

Figure 4. Value of individual variations of Function (a) and Pain (b) scores resulting from the subtraction of the initial score from the final score, represented with horizontal line representing MICD value.

Figure 5. Value of individual variations of Self-image (a) and Mental health (b) scores resulting from the subtraction of the initial score from the final score, with horizontal line representing MICD value.



istered a deterioration of the scores in the period between two and five years after surgery, namely in Pain, even though the majority of the patients kept a high level of satisfaction.31

The component Pain will probably require reconsidera-tion both of the Minimal Important Clinical Differences for this subdomain and of the content of the questions which provide its score in the SRS-22 questionnaire. For example, in our sample, a variation above 0.42 in this subdomain (preoperative standard deviation multiplied by a D_Cohen of 0.7 – which is considered a relevant effect) would disclose a more expressive improvement in pain. Rushton et al. even questioned if, in the context of idiopathic scoliosis, Pain was not being underrated, since 82% of the published series reported a statistically significant decrease of pain with the surgery. According to Landman et al., the prevalence of pain in idiopathic sco-liosis was of 77.9% in a universe of 1,433 adolescents, and it is more correlated with high thoracic curves, obesity and perception of deformity.32 The patients with a lower score in the Pain domain were indeed the ones who, according to Danielsson AJ et al., benefitted more from the surgery, as most pain complaints were resolved.33

Given the relative benignity of the natural history of adolescent idiopathic scoliosis and the insufficient knowledge of the impact of deformity progression on the quality of life of the patients, we also share Daniels-son AJ et al.’s concern, who state that the benefit of the surgery reported by the different authors has been, to a certain extent, overvalued.33 When analyzing the goals of surgery in mild to moderate deformities, we verify that, ever so often, we will be prophylactically intervening in the probable progression of scoliosis with no obvious short- to medium-term impact. Moreover, we also empha-

size that two patients of our series presented decreased global SRS-22 scores, which represents a reduction in the quality of life of the adolescents, even in the absence of complications, thus showing that sometimes we may fall short of what was expected.

Even though the prevalence of pain in the population with scoliosis and its negative impact on the self-image, with possible repercussions in metal health, is well known, in some cases we may have to face preoperative high scores in all SRS-22 domains. Providing that adding a clinically relevant gain may not be very realistic, these are borderline cases where the benefits of surgical treatment should be carefully assessed and the natural history of the disease, the subdomains affected by the clinical condition and the patient’s expectations should be considered.

REFERENCES1. Rogala, E.J., Drummond, D.S., and Gurr, J.

Scoliosis: incidence and natural history. A prospective epidemiological study. J Bone Joint Surg Am 1978; 60: 173–176.

2. Brooks, H.L., Azen, S.P., Gerberg, E., Brooks, R., Chan, L. Scoliosis: a prospective epidemiological study. J Bone Joint Surg Am 1975; 57: 968–972.

3. Dickson, R. Scoliosis in the community. British Medi-cal Journal (Clinical Research Ed.) 1983; 286, 615–618.

4. Weinstein SL, Dolan L, Spratt KF, Peterson KK, Spoonamore, MJ, Ponseti IV. Health and Function of Patients With Untreated Idiopathic Scoliosis. JAMA 2003; 289: 559–567

5. Ascani E, Bartolozzi P, Logroscino CA, et al. Natural history of untreated idiopathic scoliosis after skeletal maturity. Spine 1986;11:784–9.

6. Andersen MO, Thomsen K, Kyvik KO. Perceived health status in self-reported adolescent idiopathic scoliosis: a survey based on a population of twins. Spine 2010; 35:1571–74 .

7. Haefeli M , Elfering A, Kilian R, et al. Nonop-erative treatment for adolescent idiopathic scoliosis: a 10- to 60-year follow-up with special reference to health-related quality of life. Spine 2006;31:355 – 66.

8. Freidel K, Petermann F, Reichel D, et al. Quality of life in women with idiopathic scoliosis. Spine 2002; 27:E87 – 91.

9. Payne WK, Ogilvie JW, Resnick MD, Kane RL, Transfeldt EE, Blum RW. Does scoliosis have a psychological impact and does gender make a differ-ence? Spine 1997; 22: 1380–1384.

10. Clayson, D., and Levine, D.B. Adolescent scoliosis patients. Personality patterns and effects of corrective surgery. Clin Orthop Relat Res 1976; 116: 99–102.

Figure 6. Level of satisfaction with the treatment, both before and after the surgery.



11. Clayson D, Mahon B, Levine DB. Preoperative personality characteristics as predictors of postopera-tive physical and psychological patterns in scoliosis. Spine 1981;6: 9–12.

12. MacLean WE, Green NE, Pierre CB, Ray DC. Stress and coping with scoliosis: Psychological effects on adolescents and their families. J Pediatr Orthop B 1989; 0: 257–261.

13. Koch KD, Buchanan R, Birch JG, Morton AA, Gatchel RJ, Browne RH. Adolescents undergoing surgery for idiopathic scoliosis: How physical and psy-chological characteristics relate to patient satisfaction with the cosmetic result. Spine 2001; 26: 2119–2124.

14. http//www.srs.org/professionals/SRS_outcomes/srs22.pdf

15. Asher M, Min Lai S, Burton D, Manna B. The reliability and concurrent validity of the scoliosis re-search society-22 patient questionnaire for idiopathic scoliosis. Spine 2003; 28: 63–69.

16. Asher M, Lai SM, Burton D, Manna B. Scoliosis Research Society-22 patient questionnaire: Respon-siveness to change associated with surgical treatment. Spine 2003; 28: 70–73.

17. Asher M, Min Lai S, Burton D, Manna B. Dis-crimination validity of the Scoliosis Research Soci-ety-22 patient questionnaire: Relationship to idiopathic scoliosis curve pattern and curve size. Spine 2003; 28: 74–77.

18. Carreon LY, Sanders JO, Diab M, Sucato DJ, Sturm PF, Glassman SD. The minimum clinically important difference in scoliosis research society-22 appearance, activity, and pain domains after surgical correction of adolescent idiopathic scoliosis. Spine 2010; 35: 2079–2083.

19. Verma, Kushagra et al. Demographic Factors Affect Scoliosis Research Society-22 Performance in Healthy Adolescents: A Comparative Baseline for Adolescents with Idiopathic Scoliosis. Spine 2010; 35: 2134–2139.

20. Bagó, J., Pérez-Grueso, F.J.S., Les, E., Hernán-dez, P., Pellisé, F. Minimal important differences of the SRS-22 Patient Questionnaire following surgical treatment of idiopathic scoliosis. Eur Spine J 2009; 18: 1898–1904.

21. Mayo NE, Goldberg MS, Poitras B, et al. The Ste-Justine adolescent idiopathic scoliosis cohort study. Part III: Back pain. Spine 1994;19:1573-81.

22. SatoT, Hirano T, Ito T, et al. Back pain in adoles-cents with idiopathic scoliosis: epidemiological study for 43,630 pupils in Niigata City, Japan. Eur Spine J 2011;20:274-9.

23. Smorgick Y, Mirovsky Y, Baker KC, et al. Pre-dictors of back pain in adolescent idiopathic scoliosis surgical candidates. J Pediatr Orthop 2013; 33: 289-92.

24. Ramirez N, Johnston CE, Browne RH. The preva-lence of back pain in children who have idiopathic scoliosis. J Bone Joint Surg Am 1997; 79(3):364-8.

25. Weinstein SL, Zavala DC, Ponseti IV. Idiopathic scoliosis: long-term follow-up and prognosis in un-treated patients. J Bone Joint Surg Am 1981:63:702-12.

26. D’Andrea LP, Betz RR, Lenke LG, Clements DH, Lowe TG. Merola Haher T, Harms J, Huss GK, Blanke K, et al. Do radiographic parameters cor-relate with clinical outcomes in adolescent idiopathic scoliosis? 2000 Spine; 25: 1795–1802.

27. Rushton PRP, Grevitt MP. Comparison of untreat-ed adolescent idiopathic scoliosis with normal controls: A review and statistical analysis of the literature. Spine 2013; 38: 778–785.

28. Verma, Kushagra et al. “Demographic Factors Affect Scoliosis Research Society-22 Performance in Healthy Adolescents: A Comparative Baseline for Adolescents with Idiopathic Scoliosis.” Spine 2010; 35: 2134–2139.

29. Danielsson AJ, Wiklund I, Pehrsson K, Nachem-son A. Health-related quality of life in patients with adolescent idiopathic scoliosis: a matched follow-up at least 20 years after treatment with brace or surgery. Eur Spine J 2001; 10: 278–288.

30. Andersen MO, Andersen GR, Thomsen K, Chris-tensen SB. (2002). Early weaning might reduce the psychological strain of Boston bracing: a study of 136 patients with adolescent idiopathic scoliosis at 3.5 years after termination of brace treatment. Journal of Pediatric Orthopaedics. 2002; 11: 96–99.

31. Upasani VV, Caltoum C, Petcharaporn M., Bastrom, T. P., Pawelek, J. B., Betz, R. R., ... & Newton, P. O. Adolescent idiopathic scoliosis patients report increased pain at five years compared with two years after surgical treatment. Spine 2008; 33(10): 1107-1112.

32. Landman, Z., Oswald, T., Sanders, J., Diab, M., & Spinal Deformity Study Group. (2011). Preva-lence and predictors of pain in surgical treatment of adolescent idiopathic scoliosis. Spine 36(10): 825-829.

33. Danielsson, Aina J. What Impact Does Spinal De-formity Correction for Adolescent Idiopathic Scoliosis Make on Quality of Life?. Spine 2007; 32: S101-8.


ABSTRACTPigmented Villonodular Synovitis (PVNS) of the

shoulder is rare. We present a patient with locally recurrent tonsillar squamous cell carcinoma (SCC) and newly identified soft tissue lesion of the right shoulder exhibiting similar focal FDG-PET CT uptake. Biopsy demonstrated features consistent with pigmented villonodular synovitis (PVNS). We discuss imaging features and describe a ‘ two needle,’ coaxial biopsy technique.

Keywords: PVNS, shoulder, metastasis, mimic

INTRODUCTIONPigmented villonodular synovitis (PVNS) is a chronic

mono-articular neoplastic disease that typically involves a single joint.1 It usually affects the knee or hip with char-acteristic overgrowth of the synovial lining. While not malignant, local joint tissue proliferation can be difficult to manage surgically and can cause chronic joint swelling leading to arthritis.1

Although rare, PVNS is known to be FDG-avid on PET/CT, with limited reports describing shoulder in-volvement.2 We present a patient with locally recurrent tonsillar SCC and newly identified soft tissue lesion of the right shoulder exhibiting similar focal FDG-PET CT uptake. MRI features were atypical of soft tissue metastasis, therefore CT-guided biopsy was performed. Biopsy confirmed characteristic histological features of PVNS. We used a two needle, coaxial biopsy technique not widely described in current literature; a technique we believe to be ideal for surgically inaccessible lesions or for deep lesions within the shoulder.

Case ReportA sixty-year-old female with a background of tonsil-

lar squamous cell carcinoma (SCC), treated with radical chemo-radiotherapy, presented with local neck recurrence after three years. She was re-staged in anticipation of a

salvage radical neck dissection, segmental mandible resec-tion and flap reconstruction. Restaging CT was carried out as any metastasis outside of the neck would exclude her from the proposed radical surgery. There was no other relevant past medical history of note.

FDG-PET/CT demonstrated intense uptake within the locally recurrent tonsillar SCC and also intense uptake adjacent to the right anterior glenoid with a SUV max of 6. Both soft tissue lesions showed similar imaging appear-ances on FDG-PET CT (Fig. 1a-d). However, the patient was asymptomatic in her right shoulder. Following multidisciplinary discussion, SCC metastasis or soft tissue sarcoma was considered as the main differential diagnosis.

Magnetic resonance imaging (MRI) (Fig. 2a-f) showed a 2x3cm oval shaped lesion anterior to the right gle-noid, deep to the subscapularis muscle. The lesion was isointense to muscle on T1 and slightly hyperintense on fat suppressed imaging. Unfortunately a gradient echo sequence was not performed, as PVNS was not on the list of differential diagnoses. Ultrasound demonstrated a solid mass with mild increased signal on Doppler. As imaging features were inconclusive, CT-guided biopsy was performed to obtain a definitive sample for histology and exclude metastatic disease prior to radical surgery.

Under CT guidance, an 8-gauge T-lok bone marrow biopsy needle (Argon medical devices, USA)3 was first introduced into the right shoulder, via the deltopectoral groove, and advanced deep to the subscapularis muscle with the tip proximal to the lesion, anterior to the glenoid. A 14-gauge supercore soft tissue needle (Argon medical devices, USA)3 was then placed through the outer 8-gauge needle (Fig. 3a-c). Multiple core biopsies were taken of the lesion and sent for histology (Fig. 4a-b). The procedure was performed under conscious sedation, using ropivicaine for local anaesthetic, without complication. Histology revealed typical features of PVNS (Fig. 5a-d). She was managed symptomatically for her PVNS. She underwent radical neck dissection for recurrent tonsillar SCC.

DISCUSSIONPVNS can be focal or diffuse and can present as an intra

or extra-articular form. It usually presents in the third to fifth decades of life, with varied clinical symptoms includ-ing pain, soft-tissue mass, swelling and joint dysfunction.3 Radiographs and Computed tomography (CT) are often

PIGMENTED VILLONODULAR SYNOVITIS MASQUERADING AS A METASTASIS: IMAGING FEATURES AND COAXIAL NEEDLE

BIOPSY TECHNIQUE

Ramanan Rajakulasingam FRCR1, Jennifer Murphy MB BCh1, Idris Badreddine MB ChB1, Steven James FRCR1, Rajesh Botchu FRCR1

1Royal Orthopaedic Hospital NHS Foundation Trust, Birmingham, UKCorresponding Author: Rajesh Botchu FRCR; Phone: 00441216854000; Email: [email protected]: The authors report no potential conflicts of interestrelated to this study.Sources of Funding: No sources of funding declared.

R. Rajakulasingam, J. Murphy, I. Badreddine, S.L. James, R. Botchu


non-specific. If there is intra-articular involvement, radio-graphs may show erosions of the affected joint.3

MRI is certainly more specific than radiographs or CT. MRI may demonstrate several features suggestive of PVNS including non-specific intermediate low signal in-tensity on T1 and T2, and a blooming artifact on diffusion gradient echo sequences caused by hemosiderin deposits.3 However, in our case no specific imaging features to indi-cate aetiology were present, so biopsy was performed as it would dramatically alter patient management.

There is limited literature discussing the imaging features of PVNS on nuclear medicine modalities. PVNS is known to be FDG-avid, a SUV max of 11.3 has previ-ously been reported.4 Moreover there are a number of case

Figure 2. MRI T2 axial (A), STIR axial (B,D), T1 coronal (C) se-quences shows an iso intense 2x3cm well defined lesion in relation to the anterior part of the glenoid but not directly involving the scapula. T1 axial fat saturated (E) and coronal (F) post contrast images show a modest enhancement pattern. The overall MRI appearances were non specific.

Figure 3. (A) shows the 8 gauge bone marrow biopsy and 14 gauge soft tissue biopsy needle on the left and right respectively. (B) shows how the soft tissue needle is placed through the bone biopsy needle. (C) shows a close up view of the trochar needle tip.

Figure 4. (A and B) show CT guided biopsy of the anterior glenoid le-sion with T-lock bone biopsy needle and supercore soft tissue needle.

Figure 5. (A) Low power magnification of biopsy showing features of tenosyno-vial giant cell tumour (pigmented villonodular synovitis) composed of sheet like arrangement of mononuclear cells with scattered mul-tinucleated giant cells (HE stain x100 original magnification). (B) Higher power view of the tumour showing the oval mononuclear cells with round nuclei lacking cytological atypia or atypical mitoses and admixed multinucleated giant cells (HE stain x400 original magnification). (C) Focal areas in the tumour showing foamy his-tiocytic infiltrate (HE stain x200 original magnification). (D) Focal haemosiderin pigment deposition is also noted in the tumour (HE stain x200 original magnification).

Figure 1. (A and C) shows a focal area of increased uptake in the left tonsil in keeping with recurrent tonsillar squamous cell carcinoma. (SUV max of six). (B and D) shows an avid area of uptake anterior to the right glenoid. Both lesions showed similar SUV max values of 6.


PVNS Masquerading as a Metastasis: Imaging Features and Coaxial Needle Biopsy Technique

A

reports of PVNS as an incidental findings on FDG-PET/CT in the setting of known malignancy.5 As in our case, this posed a diagnostic challenge to differentiate between metastasis and other potential disease entities.

Studies have also shown no statistical difference in SUV max between PVNS subtypes including intra-articular PVNS, diffuse PVNS, or giant cell tumour of the tendon sheath.5 PVNS has also been shown to respond to cancer treatment with subsequent decrease in metabolic activity.6 These findings can complicate patient work up, more so in conditions such as melanoma or sarcoma where metastatic musculoskeletal disease manifesta-tions are common. Depending on the form (localized vs. diffuse), radiation may be considered in conjunction with surgical excision. In diffuse PVNS, synovectomy is usu-ally necessary.

The described biopsy technique using a two needle coaxial technique is an alternative method to obtain diagnostic tissue. Percutaneous biopsy using a single co-axial needle system method is commonly used currently. The coaxial needle technique is an alternative and more efficient method, which we have used successfully as in this case without any complication. It facilitates multiple soft tissue cores to be obtained rapidly with minimal soft tissue trauma as the outer needle stays in place between samples being taken.

Learning Points• There are only a handful of case reports describ-

ing PVNS in the shoulder, implying it is a very rare location.

• A metabolically active lesion near but not directly involving the scapula should raise the suspicion for a coexistent disease process. Despite the known history of cancer, a metastatic lesion should not be conclusively diagnosed from PET imaging alone if imaging features are not characteristic.

• We present a two needle coaxial biopsy method us-ing CT which can facilitate rapid tissue sampling.

CONCLUSIONThere are a number of case reports of PVNS as an inci-

dental finding on FDG-PET/CT in the setting of known malignancy. With the ever increasing use of PET imaging within oncology, it is likely that potentially benign but metabolically active lesions will be identified more fre-quently. This can complicate treatment and delay patient work up. If typical MRI imaging features are present, then PVNS can be confidently diagnosed. However, if atypical or in an unusual location, we describe a two needle CT guided biopsy technique to confirm the diagnosis.

REFERENCES1. Goldman AB, DiCarlo EF. Pigmented villonodular

synovitis. Diagnosis and differential diagnosis. Radiol Clin North Am. 1988;26:1327–1347

2. Serra TQ, Morais J, Gonçalves Z, Agostinho F, Melo G, Henriques M . An unusual case of diffuse pigmented villonodular synovitis of the shoulder: A multidisciplinary approach with arthroscopic syno-vectomy and adjuvant radiotherapy. European Journal of Rheumatology. 2017;4(2):142-144.

3. Murphey MD, Rhee JH, Lewis RB, Fanburg-Smith JC, Flemming DJ, Walker EA. Pigmented villonodular synovitis: radiologic-pathologic correla-tion. Radiographics. 2008;28(5):1493-1518.

4. Kitapci MT, Coleman RE. Incidental detection of pigmented villonodular synovitis on FDG PET. Clin Nucl Med. 2003;28(8):668-669.

5. Paul JC, Unnanuntana A, Goldsmith SJ, Land JM. Extra-articular knee lesion with high fluorode-oxyglucose-uptake on positron emission tomography. Bull Hosp Jt Dis. 2013;71(2):170-4.

6. Broski SM, Murdoch NM, Skinner JA, Wenger DE. Pigmented villonodular synovitis: Potential pit-fall on oncologic 18F-FDG PET/CT. Clin Nucl Med. 2016;41:e24–31


ABSTRACTBackground: The primary aim of this study was

to determine the prevalence of asymptomatic pes planus and cavovarus foot deformities using the tripod index (TI).

Methods: A retrospective study was conducted on 122 adult subjects over the age of 18 from January 2010 to December 2016 with symptom-atic pes planus (n=78) or cavovarus (n=44) foot deformities. We subdivided both groups into sub-jects who presented with unilateral symptomatic deformities (pes planus unilateral symptomatic; cavovarus unilateral symptomatic) and bilateral symptomatic foot deformities (pes planus bilateral symptomatic feet and cavovarus bilateral symp-tomatic feet). The severity of TI was compared between sides.

Results: The prevalence of asymptomatic pes planus and cavovarus foot deformities was 52% and 67.6%, respectively. Subjects with unilateral symptomatic foot deformities had significantly more severe TI values for the symptomatic cavovarus foot -98.96% (-288.89 to 0%) compared to asymp-tomatic cavovarus -67.41% (-270.59 to 14.71%) (p=0.015). Subjects with unilateral symptomatic pes planus deformity also had more severe TI on the symptomatic foot 57.49 (-9.38 to 141.67%) compared to the asymptomatic foot 30.43 (-51.52 to 119.23%) (p<0.01). Subjects with bilateral symptomatic foot deformities had no significant difference in severity of Tripod Index between feet.

Conclusion: Although half of subjects with unilateral symptomatic deformities had a foot de-formity on the contralateral side, the severity of deformity between symptomatic and asymptomatic feet was significantly different for both pes pla-nus and cavovarus feet. Further studies should prospectively follow postoperative radiographs to determine whether a correction in foot alignment directly improves symptoms.

Level of evidence: IIIKeywords: tripod index, cavovarus, pes planus

INTRODUCTIONBilateral pes planus and cavovarus deformities are

commonly seen in foot deformity patients, occurring in 23.7% and 48.9% of foot deformity cases, respectively.1

Muscle imbalance is usually the cause of these deformi-ties, typically arising from either idiopathic, neurologic, inflammatory, vascular, traumatic, or congenital causes.2-7

The overall prevalence of pes planus deformity has been reported at 2-5%8,9 and is characterized by a combination of posterior tibialis tendon deficiency, collapse of the medial longitudinal arch, foot abduction at the talonavicular joint, and hindfoot valgus due to subtalar joint eversion.10-18 In contrast, cavovarus occurs in 10% of the population and is characterized by hindfoot varus, subtalar joint inversion, midfoot cavus, plantar flexion of the first metatarsal, and forefoot adduction.4,8,10,11,19 While subjects with severe pes planus and cavovarus deformities can be completely asymptomatic, those with symptomatic deformities can present with impingement, pain, instability, callosities, dif-ficulty with shoe wear, and subsequent arthritis.4,15,16,20-23 Conservative treatment for these conditions includes stretching, proprioceptive training, strength training, nonsteroidal anti-inflammatory medications, ankle brac-ing, and orthotics. However, disease progression can lead to worsening symptoms and necessitate operative correction.24,25

The Tripod Index (TI) is an aggregate measure of the position of the talar head in relation to the “foot tripod” on a standing AP radiograph.10 The foot tripod is the center of the heel (designated by a fitted hemispherical marker on AP radiographs), the medial border of the medial sesa-moid, and the lateral border of the 5th metatarsal head. TI uses easily identifiable landmarks, to integrate multiple

EVALUATION OF ASYMPTOMATIC CONTRALATERAL FOOT DEFORMITIES USING THE TRIPOD INDEX

Courtney Carlson, MD1, Craig Akoh, MD2, Chamnanni Rungprai, MD3, Phinit Phisitkul, MD4

1Department of Orthopedic Surgery, Mayo Clinic, Rochester, MN USA2Department of Orthopedics and Rehabilitation, University of Wisconsin School of Medicine, Madison, WI USA3Phramongkutklao Hospital and College of Medicine, Thung Phaya Thai, Ratchathewi, Bangkok, Thailand4 University of Iowa Hospitals and Clinics, Department of Orthopedics and Rehabilitation. Iowa City, IA USACorresponding author: Craig C. Akoh, MD; Phone: (608) 263-0888; Email: [email protected]: Phinit Phisitkul is a paid consultant for Arthrex and Restor 3D, receives royalties from Arthrex; and has stock/stock options in First Ray and Mortise Medical.Sources of Funding: No sources of funding declared.

C. Carlson, C Akoh, C. Rungprai, P. Phisitkul


planes of the foot with excellent intraobserver (0.99) and interobserver (0.98) reliability.10,11 The more medial the center of the talar head is to the foot tripod, the more positive the TI, as seen in pes planus deformities. Nega-tive TI values correspond to the center of the talar head located more lateral to the foot tripod, as seen in cavovarus deformities. Cutoffs for diagnosing symptomatic pes planus and cavovarus deformities were determined to be 26% and -39%, respectively.10 Using these values, the sen-sitivity and specificity of TI in pes planus and cavovarus was 100% and 93%, and 96% and 95%, respectively.11 There has been a paucity of data investigating radiographic predictors of developing symptoms on the asymptomatic foot. Determining the relationship between radiographic severity and asymptomatic foot deformities may allow physicians to diagnose and appropriately counsel patients before contralateral symptoms develop.

The purpose of this study was to use the previously validated tripod index radiographic measurement10,11 to determine the prevalence of pes planus and cavovarus on the asymptomatic contralateral foot in subjects indicated for foot reconstruction surgery. This study also aimed to compare the radiographic severity of deformity between the symptomatic and asymptomatic sides. We hypoth-

esized the following: 1) subjects indicated for pes planus or cavovarus foot reconstruction have a deformity on the asymptomatic contralateral side; 2) the TI measurement would be more severe in symptomatic feet compared with contralateral, asymptomatic feet.

METHODSCohort

This study was approved by the hospital’s institutional review board (IRB) using a standard protocol. Subjects were identified in the senior author’s practice through a clinic billing database using the ICD-9 code for pes planus (754.6 and 726.7) and cavovarus (736.7 and 754.7). Our bill-ing search yielded 245 consecutive subjects from January 2010 to December 2016 with symptomatic pes planus (n=147) or cavovarus (n=98) foot deformities indicated for reconstructive surgery. Exclusion criteria included subjects under the age of 18, diagnosis of Charcot Marie Tooth (CMT), other neuromuscular conditions, previous hindfoot or midfoot surgeries, or lack of hemispheri-cal marker around the heel during AP bilateral weight bearing foot films (n=123). A total of 122 subjects were included in our cohort (Figure 1). We subdivided both groups into subjects who reported unilateral symptoms

Figure 1. Study cohort.


Evaluation of Asymptomatic Contralateral Foot Deformities Using the Tripod Index

A

and those who reported bilateral symptoms at initial pre-sentation. Symptomatic pes planus deformity was defined as clinical flattening of the medial longitudinal arch in subjects that reported pain (i.e. posterior tibial tendonitis or subfibular impingement) at the initial clinical presen-tation. Symptomatic cavovarus deformity was defined as clinical hindfoot varus and high medial longitudinal arch in subjects that reported pain in the lateral column or peroneal pathology. In the bilateral symptomatic popula-tions, the operative foot was defined as the symptomatic foot. All clinical diagnoses were made by one of three senior foot and ankle surgeons who also performed all

reconstruction foot surgeries. Subjects were followed a minimum of one year after their index surgery or until their contralateral reconstructive surgery.

Radiographic MeasurementsPreoperative bilateral AP weight-bearing radiographs

with a hemispherical marker around the heel, lateral, and hindfoot alignment films were obtained in a standardized manner as described previously.10,11 The tripod index of symptomatic and asymptomatic feet was measured in all subjects using the iSite Enterprise 3.5 image analysis soft-ware (Philips Medical Systems Nederland B.V., Best, The Netherlands). The tripod index was measured by a single fellowship-trained foot and ankle surgeon (CR) using the center of the heel, center of the talar head, the medial first metatarsal sesamoid bone, and the lateral border of the fifth metatarsal. The measurement technique is further illustrated in Figure 2. Tripod index values greater than 26% or less than -39% were used as the cut-off values for radiographic diagnosis of pes planus and cavovarus feet, respectively (Figure 3).10,11

Statistical AnalysisCavovarus and pes planus groups were evaluated

separately. Descriptive analyses of subject characteristics and radiographic measurements, including mean±SD or median (min-max) for continuous variables and frequency for categorical variables, were completed separately for subjects with and without bilateral symptoms. Nonpara-metric tests were completed due to the distribution of

Figure 2: Measurement technique for the Tripod Index. The Tripod Index is measured from the center of the heel using the hemispheric marker (A). (B) is the meial edge of the medial sesamoid. (C) is the lateral edge of the fifth metatarsal. (D) is the center of the talar head. E is the lateral tripod angle created by the intersection of AB and AC. F is the medial tripod angle created by the intersection of AB and AD. The Tripod Index (%) = (angle F/angle E) x 100. If the AD line is medial to the AB line, the index is positive; if the AD line is lateral to the AB line, the index is negative.

Figure 3: Tripod Index measured on a patient with pes planus (a) and cavovarus foot (b).



A B C

Table 1. Results Comparing Bilateral vs Unilateral Symptomatic Pes Planus ParticipantsVariable Bilateral Symptoms (n=27) Unilateral Symptoms (N=50) p-value

Age (mean±SD, yrs) 47.8±18.746.0 (18.0-75.0)

49.6±16.852.5 (18.0-83.0)

0.5897

BMI (mean±SD, kg/m2) 33.8±7.034.6 (19.6-46.3)

32.8±7.131.7 (22.2-53.4)

0.5384

Sex (n, % women) 18 (66.7) 27 (54.0) 0.2818

HTN (n, % yes) 8 (29.6) 15 (30.0) 1.0000

Depression (n, % yes) 4 (14.8) 5 (10.0) 0.7119

Fibromyalgia (n, % yes) 2 (7.4) 0 0.1200

Depression or fibromyalgia (n, % yes) 6 (22.2) 5 (10.0) 0.1790

Diabetes (n, % yes) 0 6 (12.0) 0.0853

Neuropathy (n, % yes) 2 (7.4) 2 (4.0) 0.6089

Inflammatory arthropathy (n, % yes) 3 (11.1) 3 (6.0) 0.6591

Smoker (n, % yes) 2 (7.4) 4 (8.0) 1.0000

Manual Labor (n, % yes) 2 (7.4) 3 (6.0) 1.0000

Neuro condition (n, % yes) 0 (0.0) 0 (0.0) -

Workers Comp (n, % yes) 0 5 (10.0) 0.1527

Diagnosis by TI, Index Side 23 (85.2) 38 (76.0) 0.3432

Diagnosis by TI, Contralateral Side 21 (77.8) 26 (52.0) 0.0269

Table 2. Comparing Index to Contralateral Side (paired tests) in Unilateral Patients and Bilateral Patients with Pes Planus Deformity

Unilateral patients (n=50)

Variable Side Min Max Med Mean SD p-value

TI Index -9.38 141.67 61.65 57.49 37.67 <0.0001

Contralateral -51.52 119.23 31.38 30.43 35.95

N, (% Yes) p-value

Flatfoot Diagnosis by TI

Index 38 (76.0) 0.0075

Contralateral 26 (52.0)

Bilateral patients (n=28)


TI Index -6.25 116.67 58.62 62.82 34.76 0.1140

Contralateral -55.56 108.57 57.14 56.21 34.92

N, (% Yes) p-value


Index 23 (85.2%) 0.6875

Contralateral 21 (77.8%)

Wilcoxon signed rank sum test (nonparametric paired t-test) for TI, F angle, E angle



continuous variables. To test the hypothesis that the TI measurement would be more severe in symptomatic feet compared with contralateral, asymptomatic feet, we utilized the Wilcoxon Signed Rank test to compare TI values between paired feet. These analyses were repeated in subjects with bilateral symptoms. Data analyses were performed by a statistician using SAS software, version 9.3 (SAS Institute, Inc. Cary, North Carolina). A p-value of <0.05 was considered significant.

RESULTSPes planus group

Overall demographic characteristics for pes planus are outlined in Table 1. The mean age for pes planus de-formity was 49.0±17.4 yo (median = 52.0 yo; min-max =18 to 83 yo). Subjects presented with bilateral symptomatic deformities in 28 out of 78 (35.9%) of cases. When compar-ing participants with unilateral or bilateral symptoms, there were no significant differences in age, gender, body mass index (BMI), smoking status, hypertension (HTN), depression, fibromyalgia, diabetes, or neuropathy. The ra-diographic values for subjects diagnosed with pes planus are shown in Table 2. Using the TI cutoff values, 76% of symptomatic unilateral patients met the radiographic di-agnosis for pes planus. Additionally, 52% (n=26) of subjects with unilateral symptoms were diagnosed with pes planus on the contralateral asymptomatic side. Subjects with unilateral symptoms had significantly greater TI values in the symptomatic foot [(57.49 ± 37.67% (61.65; -9.38 to 141.67%)] compared to the contralateral asymptomatic foot [(30.43 ± 35.95% (31.38; -51.52 to 119.23%)] (p < 0.01). In subjects with bilateral symptomatic pes planus, there was no significant difference in the severity of TI between the operative foot (66.27 ± 38.68% (61.13; -6.25 to 159.38%)) and symptomatic contralateral foot (57.98 ± 35.53% (63.87; -55.56 to 108.57%)) (p=0.07). Of note, six out of 50 subjects (12%) with unilateral symptoms went on to develop symp-toms in the asymptomatic contralateral foot at a mean time of two years after their initial presentation. Fifteen out of 28 subjects (53.6%) with bilateral symptomatic feet had subsequent contralateral reconstructive surgery at a mean time of 1.25 years after the index surgery.

Cavovarus groupForty-four subjects were clinically diagnosed with

symptomatic cavovarus foot deformity (Table 3). The mean age of subjects diagnosed with cavovarus defor-mity was 48.3±13.8 years (48.5; 18.0 to 85.0). Of these subjects, seven out of 44 subjects (15.9%) had bilateral symptomatic feet at presentation. Except for BMI, there were no significant differences in the demographic char-acteristics between subjects with unilateral and bilateral symptomatic feet. The radiographic values for unilat-

eral symptomatic and bilateral symptomatic subjects are shown in Table 4. Using the TI, 78.4% of unilateral symp-tomatic cavovarus deformity met radiographic criteria. Sixty-seven percent of subjects with unilateral symptoms were diagnosed with cavovarus foot deformity on the contralateral asymptomatic side. Subjects with unilateral symptoms had significantly more severe TI values on the symptomatic foot [-98.96 ± 76.32% (-82.05; -288.89 to 0%)] than the asymptomatic contralateral foot [-67.41 ± 58.34% (-57.58; -270.59 to 14.71%)] (p=0.015). Subgroup analysis in subjects with bilateral symptomatic feet demonstrated no significant difference in TI values between the operative and the symptomatic contralateral foot [84.91±73.12% (-73.53; -188.24 to 0%)] and [-81.16±78.12% (-68.57%; -194.12 to 40.63%)] (p=0.813), respectively. One out of 37 subjects with unilateral symptoms (2.7%) and three out of seven subjects (42.9%) with bilateral symptoms went on to have subsequent contralateral reconstructive surgery at a mean time of 1.1 years after their index surgery.

DISCUSSIONThe TI has been previously shown to be a reliable

method of measuring foot alignment and diagnosing both pes planus and cavovarus foot in symptomatic sub-jects.10-11 Although previous studies have shown that the TI has a higher sensitivity for flatfoot deformity (100%) compared to cavovarus deformity (96%), quantifying both deformities using the TI have a high positive and negative predictive value.11 We found that 76% and 78.4% of patients with unilateral symptomatic deformity met radiographic diagnosis for pes planus and cavovarus deformities, respectively. A significant number of sub-jects with pes planus and cavovarus deformities had TI values that failed to meet the diagnostic TI parameters for their respective deformity. Given that the tripod index accounts for the position of the talar neck in relation to the foot tripod, patients with clinical diagnosis may have lacked talar uncoverge and overcoverage. Additionally, compensatory forefoot deformity could potentially reduce the tripod index from reaching the defined diagnostic cutoff values. Additionally, our study found that over half of subjects with unilateral symptomatic pes planus or cavovarus foot deformities had a similar, though less severe, deformity on the asymptomatic contralateral foot. Thus, it is important for clinicians to both clinically and radiographically diagnose bilateral foot deformities. Sur-geons may potentially use this consolidated information to guide treatment of each individual foot, especially in bilateral deformity cases.

Our pes planus cohort showed that symptomatic feet presented with more severe radiographic deformity when compared to the asymptomatic side. Our results for deformity severity are in contrast to Dyal’s study,



Table 3. Results Comparing Bilateral vs Unilateral Symptomatic Cavovarus ParticipantsVariable Bilateral Symptoms (n=7) Unilateral Symptoms (N=37) p-value

Age (mean±SD, yrs) 39.1±15.042 (18-56)

50.0±13.149.0 (22-85)

0.1311

BMI (mean±SD, kg/m2) 27.7±7.025.4 (20.3-37.4))

34.2±6.634.6 (21.9-54.1)

0.0400

Sex (n, % women) 3 (42.9) 16 (43.2) 1.0000

HTN (n, % yes) 1 (14.3) 16 (43.2) 0.2204

Depression (n, % yes) 1 (14.3) 6 (16.2) 1.0000

Fibromyalgia (n, % yes) 0 1 (2.7) 1.0000

Depression or fibromyalgia (n, % yes) 1 (14.3) 7 (18.9) 1.0000

Diabetes (n, % yes) 1 (14.3) 5 (13.5) 1.0000

Neuropathy (n, % yes) 0 0 -

Inflammatory arthropathy (n, % yes) 1 (14.3) 3 (8.1) 0.5135

Smoker (n, % yes) 1 (14.3) 4 (10.8) 1.0000

Manual Labor (n, % yes) 2 (28.6) 6 (16.2) 0.5934

Neuro condition (n, % yes) 0 0 -

Workers Comp (n, % yes) 1 (14.3) 9 (24.3) 1.0000

Diagnosis by TI, Index Side 5 (71.4) 29 (78.4) 0.6490

Diagnosis by TI, Contralateral Side 5 (71.4) 25 (67.6) 1.0000

Table 4. Comparing Index to Contralateral Side (paired tests) in Unilateral Patients and Bilateral Patients with Cavovarus Deformity

Unilateral patients (n=37)


TI Index -288.89 0 -82.05 -98.96 76.32 0.0150

Contralateral -270.59 14.71 -57.58 -67.41 58.34

N, (% Yes) p-value


Index 29 (78.4%) 0.3877


Bilateral patients (n=7)


TI Index -188.24 0 -73.53 -84.91 73.12 0.8125

Contralateral -194.12 40.63 -68.57 -81.16 78.12

N, (% Yes) p-value


Index 5 (71.4%) -


Wilcoxon signed rank sum test (nonparametric paired t-test) for TI, F angle, E angle



which found similar deformity severity in the asymp-tomatic and symptomatic pes planus population.13 Unlike our study, Dyal utilized several individual measurements to assess for pes planus deformity. The talometatarsal, calcaneometatarsal, and cuneiform to fifth metatarsal height measure the foot in one direction on a single plane and level of the foot. The angles alone are not able to inte-grate information across dimensions; therefore, multiple angles must be used to try and accurately determine the severity of foot deformity—requiring multiple radiograph views to be taken. The TI uses one radiograph view and can integrate information across multiple planes and levels of the foot. Previous studies have shown that the tripod index is a highly sensitive and specific test for pes planus deformity.10,11 As a result, asymptomatic pes planus deformity occurred in 50% of our cohort. The TI’s ability to reliably diagnose pes planus deformities will allow physicians to consider preventative measures before the progression of symptoms.

About two-thirds of subjects were radiographically diagnosed with cavovarus deformity with TI on the as-ymptomatic contralateral side. Cavovarus deformity com-monly presents with bilateral deformity due to underlying neuromuscular conditions. In the general population, a diagnosis of CMT is present in 78% of bilateral cavovarus deformity cases.26 The progression of forefoot cavus and hindfoot inversion can lead to locking of the subtalar and Chopart joints. Once this occurs, the intrinsic muscula-ture and plantar plate become contracted and bilateral rigid deformities develop. Additionally, patients often experience lateral forefoot overloading and increased tension on the lateral structures, predisposing them to instability and injury.23,27 Our study excluded 28% of our initial cohort with a diagnosis of Charcot-Marie-Tooth or other neurologic conditions. Although we attempted to control for neurologic conditions, individuals in our study with cavovarus deformity may have asymptom-atic neuropathy, therefore making it difficult to detect symptoms in the setting of severe deformity. The bilat-eral rigid deformity of cavovarus feet likely explains why our cohort of asymptomatic feet had high rates of radiographic diagnosis using the TI. Our study showed that nearly 50% of subjects with bilateral symptomatic cavovarus deformity underwent bilateral surgery. This is similar to Azmaipairashvili’s study in which CMT patients with cavovarus deformity underwent bilateral surgery 64% of the time.28 The progression of bilateral idiopathic cavovarus is poorly understood when compared to neurologic cases.19,29 Much like pes planus patients, adequate treatment of both feet in bilateral cases with orthotics, bracing, and potentially surgery may lead to improved function and quality of life.

This study introduces the question whether the as-ymptomatic foot in bilateral deformities should warrant

close clinical follow-up. Our pes planus cohort showed that 12% of subjects with bilateral radiographic deformity developed symptoms in the asymptomatic contralateral foot at two years follow up. Saltzman et al reported on 47 subjects that underwent nonoperative treatment for stage I and II pes planus deformities with an ankle foot orthotic and home physical therapy.30 At a minimum follow up of one year, 83% of the cohort had significant improvement in strength testing and 89% of the cohort avoided surgical intervention. Flemister et al performed a randomized con-trol trial of 36 subjects with stage II pes planus deformities being treated with prefabricated orthosis.31 One group received home stretching therapy while the second group receive a home strengthening protocol for six weeks. The strengthening group minimal improvements in Short Musculoskeletal Functional Assessment (SMFA) with mobility and function scores without significant improve-ments in pain scores when compared to the stretching group. The authors concluded that home physical therapy has minimal effect on outcomes during orthotic wear for pes planus deformity. Lin et al performed a long-term retrospective study on 32 subjects with stage II pes planus deformity being treated with double upright ankle foot orthosis.32 At 8.6 years follow up, they found that 84.3% of their cohort avoided surgery.

Regarding cavovarus foot deformity, our cohort showed that 2.7% of subjects developed symptoms on the previously asymptomatic side at 2 years follow up. LoPic-colo et al published his series of 93 subjects with cavus foot deformity that were treated with custom orthotics.33

At two years follow up, the administered questionnaire showed a significant decrease in pain scores from 7.22 to 2.41. Additionally, 92% of subjects reported improvements in ankle stability with orthotic wear. Diagnosing asymp-tomatic deformities may allow physicians to appropriately council and educate patients regarding the contralateral foot deformity. This may assist in earlier recognition and implementation of non-operative treatment should patients go on to develop symptoms.

Our study is not without limitations. Firstly, the nature of our retrospective study did not allow us to definitively determine whether severe deformity directly caused symptoms and may have missed subjects with bilateral symptoms. Although we followed subjects up to six years, we may not have been able to capture all the asymptom-atic foot deformities that would eventually become symp-tomatic. We also did not prospectively follow subjects clinically over time and were unable to directly determine whether surgical intervention improved radiographic foot alignment and clinical symptoms. Although our study reported various comorbidities within our cohort, we did not directly assess the influence of other factors that may predispose subjects to develop symptoms such as BMI, gender, diabetes, generalized laxity, foot size, or



worker’s compensation. We utilized static radiographic measurements to determine significant pes planus and cavovarus deformities. Our static measurements do not account for the dynamic loading or rotational abnormali-ties of symptomatic subjects as potential causes of symp-toms. Although our study only partially accounted for dynamic deformity using weight-bearing radiographs, a previous study by Coughlin showed that dynamic Harris mat prints significantly correlate with radiographic and clinical flatfoot deformity.34 A number of subjects had to be excluded from this study because an adequate bilateral AP view radiograph was not taken with the hemisphere heel marker before the subject underwent surgery on their symptomatic foot.

Lastly, although our study was adequately powered for unilateral symptomatic subjects, our study may have been underpowered to detect significant differences in the bilateral symptomatic cohort. A priori power analysis based on a previous prevalence study of unilateral and bilateral pes planus and cavovarus deformities1 and at least 90% detection of these deformities11 was performed. We estimated that a minimum of 17 pairs with unilateral pes planus and 32 pairs of unilateral cavovarus deformities would provide 80% power to detect a significant difference between paired symptomatic and asymptomatic feet at an alpha level of 0.05.

CONCLUSIONTripod Index has previously been shown to be a reli-

able tool for analyzing foot deformities. The results of our study suggest that bilateral foot deformities are common, and the severity of deformity between asymptomatic and symptomatic feet is significantly different in pes planus and cavovarus foot deformities. Future studies should correlate radiographic correction with symptom improvement. Additionally, future studies should assess whether radiographic severity of deformity warrants closer clinical observation.

REFERENCES1. Nguyen US, Dufour AB, Positano RG, et al. The

occurrence of ipsilateral or contralateral foot disorders and hand dominance: the Framingham foot study. Journal of the American Podiatric Medical Association 2013;103(1):16-23.

2. Brewerton DA, Sandifer PH, Sweetnam DR. “IDIOPATHIC” PES CAVUS: AN INVESTIGATION INTO ITS AETIOLOGY. British medical journal 1963;2(5358):659-61.

3. Haddad SL, Deland JT. Pes Planus. In: Coughlin MJ, Saltzman MD, Anderson RB, eds. Mann’s Surgery of the Foot and Ankle. Philadelphia, PA Elsevier Inc; 2014:1292-1360.

4. Holmes JR, Hansen ST, Jr. Foot and ankle manifes-tations of Charcot-Marie-Tooth disease. Foot & ankle 1993;14(8):476-86. 20. Alexander IJ, Johnson KA. Assessment and management of pes cavus in Charcot-Marie-tooth disease. Clinical orthopaedics and related research 1989(246):273-81.

5. Kettelkamp DB, Alexander HH. Spontaneous rup-ture of the posterior tibial tendon. The Journal of bone and joint surgery American volume 1969;51(4):759-64.

6. Krause FG, Guyton GP. Pes Cavus. In: Coughlin MJ, Saltzman MD, Anderson RB, eds. Mann’s Surgery of the Foot and Ankle. 2. Philadelphia, PA: Elsevier Inc; 2014:1361-1382.

7. Mosier SM, Lucas DR, Pomeroy G, Manoli A, 2nd. Pathology of the posterior tibial tendon in posterior tibial tendon insufficiency. Foot & ankle international 1998;19(8):520-4. doi: 10.1177/107110079801900803.

8. Sachithanandam V, Joseph B. The influence of footwear on the prevalence of flat foot. A survey of 1846 skeletally mature persons. The Journal of bone and joint surgery British volume 1995;77(2):254-7.

9. Spratley EM, Matheis EA, Hayes CW, Adelaar RS, Wayne JS. A population of patient-specific adult acquired flatfoot deformity models before and after surgery. Annals of biomedical engineering 2014;42(9):1913-22. doi: 10.1007/s10439-014-1048-y.

10. Arunakul M, Amendola A, Gao Y, Goetz JE, Femino JE, Phisitkul P. Tripod index: a new radiographic parameter assessing foot alignment. Foot & ankle international 2013;34(10):1411-20. doi: 10.1177/1071100713488761.

11. Arunakul M, Amendola A, Gao Y, Goetz JE, Femino JE, Phisitkul P. Tripod Index: diagnostic accuracy in symptomatic flatfoot and cavovarus foot: part 2. The Iowa orthopaedic journal 2013;33:47-53.

12. Chi TD, Toolan BC, Sangeorzan BJ, Hansen ST, Jr. The lateral column lengthening and medial column stabilization procedures. Clinical orthopaedics and related research 1999(365):81-90.

13. Dyal CM, Feder J, Deland JT, Thompson FM. Pes planus in patients with posterior tibial tendon insufficiency: asymptomatic versus symptomatic foot. Foot & ankle international 1997;18(2):85-8.

14. Funk DA, Cass JR, Johnson KA. Acquired adult flat foot secondary to posterior tibial-tendon pathol-ogy. The Journal of bone and joint surgery American volume 1986;68(1):95-102.

15. Gould N. Evaluation of hyperpronation and pes planus in adults. Clinical orthopaedics and related research 1983(181):37-45.

16. Kitaoka HB, Luo ZP, An KN. Subtalar arthrodesis versus flexor digitorum longus tendon transfer for severe flatfoot deformity: an in vitro biomechanical analysis. Foot & ankle international 1997;18(11):710-5.



17. Myerson MS. Adult acquired flatfoot deformity: treatment of dysfunction of the posterior tibial tendon. Instructional course lectures 1997;46:393-405.

18. Pinney SJ, Hansen ST, Jr., Sangeorzan BJ. The effect on ankle dorsiflexion of gastrocnemius reces-sion. Foot & ankle international 2002;23(1):26-9.

19. Manoli A, 2nd, Graham B. The subtle cavus foot, “the underpronator”. Foot Ankle Int 2005;26(3):256-63.

21. Malicky ES, Crary JL, Houghton MJ, Agel J, Hansen ST, Jr., Sangeorzan BJ. Talocalcaneal and subfibular impingement in symptomatic flatfoot in adults. The Journal of bone and joint surgery American volume 2002;84-a(11):2005-9.

22. Mizel MS, Hecht PJ, Marymont JV, Temple HT. Evaluation and treatment of chronic ankle pain. Instructional course lectures 2004;53:311-21.

23. Saltzman CL, Salamon ML, Blanchard GM, et al. Epidemiology of ankle arthritis: report of a consecu-tive series of 639 patients from a tertiary orthopaedic center. The Iowa orthopaedic journal 2005;25:44-6.

24. Vulcano E, Deland JT, Ellis SJ. Approach and treat-ment of the adult acquired flatfoot deformity. Current reviews in musculoskeletal medicine 2013;6(4):294-303. doi: 10.1007/s12178-013-9173-z.

25. Burns J, Landorf KB, Ryan MM, Crosbie J, Ouvrier RA. Interventions for the prevention and treatment of pes cavus. Cochrane Database Syst Rev 2007(4):Cd006154. doi: 10.1002/14651858.CD006154.pub2.

26. Nagai MK, Chan G, Guille JT, Kumar SJ, Scavina M, Mackenzie WG. Prevalence of Charcot-Marie-Tooth disease in patients who have bilateral cavovarus feet. Journal of pediatric orthopedics 2006;26(4):438-43. doi: 10.1097/01.bpo.0000226278.16449.c4.

27. Berkowitz MJ, Kim DH. Fibular position in relation to lateral ankle instability. Foot & ankle international 2004;25(5):318-21.

28. Azmaipairashvili Z, Riddle EC, Scavina M, Kumar SJ. Correction of cavovarus foot deformity in Charcot-Marie-Tooth disease. Journal of pediatric orthopedics 2005;25(3):360-5.

29. Deben SE, Pomeroy GC. Subtle cavus foot: diag-nosis and management. The Journal of the American Academy of Orthopaedic Surgeons 2014;22(8):512-20. doi: 10.5435/jaaos-22-08-512.

30. Alvarez RG, Marini A, Schmitt C, Saltzman CL. Stage I and II posterior tibial tendon dysfunc-tion treated by a structured nonoperative manage-ment protocol: an orthosis and exercise program. Foot & ankle international 2006;27(1):2-8. doi: 10.1177/107110070602700102.

31. Houck J, Neville C, Tome J, Flemister A. Random-ized Controlled Trial Comparing Orthosis Augmented by Either Stretching or Stretching and Strengthening for Stage II Tibialis Posterior Tendon Dysfunction. Foot & ankle international 2015;36(9):1006-16. doi: 10.1177/1071100715579906.

32. Lin JL, Balbas J, Richardson EG. Results of non-surgical treatment of stage II posterior tibial tendon dysfunction: a 7- to 10-year followup. Foot & ankle international 2008;29(8):781-6. doi: 10.3113/fai.2008.0781.

33. LoPiccolo M, Chilvers M, Graham B, Manoli A, 2nd. Effectiveness of the cavus foot orthosis. Journal of surgical orthopaedic advances 2010;19(3):166-9.

34. Coughlin MJ, Kaz A. Correlation of Harris mats, physical exam, pictures, and radiographic measure-ments in adult flatfoot deformity. Foot & ankle inter-national 2009;30(7):604-12. doi: 10.3113/fai.2009.0604.

Volume 39 85

ABSTRACTBackground: In the setting of outpatient ortho-

paedic surgery, this pilot study utilized automated mobile messaging to assess (1) the feasibility of and interaction rates with a software delivered cognitive behavior therapy (CBT) intervention for postoperative opioid utilization, (2) the reliability of patient reported opioid utilization through our platform, (3) daily patient reported pain and opi-oid utilization within the first two postoperative weeks, and (4) the effect of software delivered CBT intervention on patient reported opioid utilization.

Methods: Musculoskeletal tumor patients sched-uled for outpatient surgery were randomized into two study groups. Control patients received stan-dard postoperative communication limited to a two-week postoperative follow-up visit. The intervention group received automated daily text-messages regarding pain, opioid utilization, and a daily CBT intervention. Interventional group patients also completed a patient satisfaction questionnaire at their two-week follow-up. Completion rates of all software delivered questions were determined in the interventional group. Median values of opioid utilization and interquartile range (IQR) were de-termined to compare utilization between groups. Spearman correlation coefficients were used to determine reliability of patient reported opioid utilization in the interventional group.

Results: Fourteen patients completed the pilot study (seven controls, seven intervention). Patients in the intervention arm completed 90% of pain and opioid questions. Intervention group patients utilized less of their daily prescribed opioid medi-

cation (20%, IQR:10%-27%) compared to controls (50%, IQR:4%-68%). Correlation between in-office pill counts and patient reported opioid medication utilization via our software messaging system was high (r=0.90, p=0.037).

Conclusion: Automated mobile phone mes-saging in outpatient tumor surgery yielded high interaction rates. Patient reported opioid utiliza-tion obtained through our platform demonstrated a high correlation with in-office pill counts. CBT delivered via automated mobile phone messaging demonstrated decreased opioid utilization in this pilot investigation.

Level of evidence: IIKeywords: cognitive behavior therapy (cbt), pain

management, opioids, mhealth, musculoskeletal tumors

INTRODUCTIONOpioid prescriptions rose sharply in the 1990s in a

response to calls for better patient pain control.1-3 This historical increased use of opioids led to the epidemic we are experiencing today, with over 60% of drug overdose deaths involving the use of opioids in recent years.4 Opi-oids continue to be a consistently utilized analgesic for relief of post-operative pain, and opioid misuse poses a great challenge for orthopaedic surgeons as they provide the third most opioid prescriptions among physicians.1, 5

Opioid medications are particularly important in pa-tients with musculoskeletal tumors as these patients not only suffer from immediate post-operative pain, but also chronic pain that is multifactorial in nature.5 Patients with cancer, especially those with bone involvement, suf-fer from pain due to the sensitization and stimulation of nerve fibers that innervate bone.6 Chronic opioid use can lead to hyperalgesia and increased narcotic demand in the postoperative setting, warranting increased diligence to minimize the risks of opioid overdose.7

Understanding of pain and pain medication utilization has been difficult following orthopaedic surgical interven-tions, in part due to a lack of standardized daily interac-tions or communication between patients and their surgi-cal team to assess and advise on postoperative pain and narcotic use. Previous work has investigated automated mobile phone text messaging as a means to communicate

AUTOMATED MOBILE PHONE MESSAGING UTILIZING A COGNITIVE BEHAVIORAL INTERVENTION: A PILOT INVESTIGATION

Edward O. Rojas, BS1; Chris A. Anthony, MD1; Jill Kain, RN1; Natalie Glass, PhD1; Apurva S. Shah, MD, MBA2; Tammy Smith, MS1; Benjamin J. Miller, MD, MS1

1Department of Orthopaedics and Rehabilitation, University of Iowa Hospitals and Clinics, Iowa City, IA, USA.2 Division of Orthopaedics, Children’s Hospital of Philadelphia, Philadelphia, PA, USA. Corresponding author: Edward O. Rojas, Phone: 562-405-2654Fax: 319-353-6754, Email: [email protected]: The authors report no potential conflicts of interestrelated to this study.Sources of Funding: This study was made possible through grant funding from the University of Iowa Sarcoma Multidisciplinary Oncology Group.

E. O. Rojas, C. A. Anthony, J. Kain, N. Glass, A. S. Shah, T. Smith, B. J. Miller


with patients on a daily basis, with favorable results.8-10 This technology can deliver predefined questions, remind-ers, and instructions to patients directly to their mobile device, and is an effective way to collect patient data at a minimal cost.9, 11-13

Cognitive behavior therapy (CBT) is a common psychological intervention that has been shown to help patients cope with chronic pain.14 Prior work suggests that CBT improves post-surgical pain and that mobile phone delivery of CBT interventions may improve pain catastrophizing.15-17

In the setting of patients with musculoskeletal tumors undergoing outpatient orthopaedic surgery, this pilot study aimed to use automated mobile phone messaging to assess (1) the feasibility of and interaction rates with a software delivered CBT intervention for postoperative opioid utilization, (2) the reliability of patient reported opi-oid medication utilization through our designed software platform, (3) daily patient reported pain and opioid medi-cation utilization in this patient population in the first two postoperative weeks, and (4) the effect of a software delivered cognitive behavior therapy (CBT) intervention on patient reported opioid medication utilization.

METHODSThis pilot investigation was approved by our insti-

tutional review board and deemed HIPAA compliant. Patients presenting to musculoskeletal tumor clinic were approached for possible inclusion in the study if they were undergoing an outpatient musculoskeletal tumor procedure. Patients provided informed consent. Patients choosing to participate were then randomized to one of two study arms utilizing a random number generator. The control arm received standard of care postoperative communication consisting of discharge from the outpa-tient surgery center and subsequent two-week postopera-tive follow-up appointment. No software messaging was sent to the control group. The intervention arm received daily text messages inquiring about pain and opioid use for that day, as well as a CBT intervention consisting of daily messages giving general postoperative guidance and encouragement. The CBT intervention messages also encouraged decreasing opioid medication utilization over the first two postoperative weeks (Table 1). The intervention group also attended the standard two-week postoperative follow-up visit. Patients in both groups were instructed to bring their opioid pill bottles with them to clinic at their two-week postoperative visit. Additionally, all patients in both groups were instructed according to

Table 1. Automated Mobile Message Content and Chronology

Postoperative Day Morning Cognitive Behavior Intervention Message

1 Its normally to be in pain after surgery. You are just beginning a healing process! Consider elevating your operative extremity to help with any pain and swelling you might be having.

2 Surgery sometimes causes temporary symptoms including numbness, tingling, cold, burning, or pain in and around the area of your procedure. These symptoms should decrease day by day!

3 Remember to pace your activities right after surgery. We want you to be active but remember to not overdo it right after surgery! Slowly progress your activity so that you are not having any unnecessary pain.

4 Remember post-surgical symptoms are temporary, think positive thoughts today and remember that your body is recovering and healing!

5 Know that you are supported after surgery. You are not alone! Get together with family or friends today if you have time. We are always here for you if you need us!

6 You are now 1 week out from surgery. Your healing process continues! Consider reducing your pain medication use if you have not already done this.

8 Your body continues to heal itself. Bone, muscle, and skin all make significant healing gains over the first postoperative week. Keep up the good work!

10 You are now 1.5 weeks out from your surgical procedure, congratulations you’ve made it through the toughest part! Take time today to reflect on how far you’ve come since surgery.

12 You are now almost 2 weeks into the recovery process! We look forward to seeing you in clinic soon and discussing the progress you have made!

Evening Pain Rating Question

0-14 Please rate your pain on a scale of 0-10, with 0 being no pain and 10 being the most pain you have ever felt.

Evening Opioid Pill Utilization Question

0-14 How many tablets of prescription pain medication have you taken in the past 24 hours?

Volume 39 87

Automated Text-Messaging Using a CBT Intervention

A

our standard postoperative pain control protocol, which includes advising patients to elevate their operative ex-tremity, apply ice as needed, and call our clinic with any concerning signs including intractable pain, excessive swelling or consistent fevers.

The messaging protocol completion rate was calculated for the intervention group. The average pain and opioid use reported by the patients in the intervention group was evaluated by day. The percentage of the prescribed, maxi-mum allotted daily opioid medication patients reported taking was assessed via median values and interquartile range (IQR) across all participants in the intervention group utilizing patient reported pill counts through our software and compared to median value manual pill counts obtained from the control group at two-week follow-up. We elected to utilize median values and IQR to describe our groups given the pilot nature of this study and the small sample size. Pill counts in the intervention group reported via mobile phone messaging were compared to their manual pill count at two-week follow-up and evaluated utilizing Spearman correlation coefficients. Correlation was interpreted based on previously reported definitions of excellent (>0.70), excellent-good (0.61-0.70), good (0.31-0.60), and poor (0.2-0.3).18 Statistical analyses were performed using Microsoft Excel (Microsoft Corp., Redmond, WA). Alpha level for significance was set a priori at 0.05.

RESULTSA total of 17 patients were enrolled with nine random-

ized to the control group and eight patients randomized to the intervention group. The control group consisted

of all mass excisions, the intervention group consisted of six mass excisions and one curettage and bone grafting (Table 2). Average age was 45±13 years and nine of the 17 patients were female. Two of the patients in the control group did not return with their pill bottles and thus could not be further evaluated regarding opioid utilization. One patient from the messaging group decided not to partici-pate in the study after signing up. After exclusion of these patients, there were 14 total patients with seven patients in both the control and interventional groups.

There was a 90% completion rate of all pain and opioid questions through two weeks among patients in the inter-vention arm. The two-week total of daily patient reported pill counts in the intervention group demonstrated good correlation with the pill counts from the returned bottles at the two-week postoperative visit (r=0.90, p=0.037) and is presented in Table 3. In the intervention arm, 86% of patients felt the automated software messages made them feel “somewhat” or “much more” connected to their care team, 71% preferred postoperative communication via messaging versus phone (voice) or email, and 57% of patients would ask for the messaging system again if they were to undergo another procedure.

Daily median pain and opioid utilization (Figure 1) decreased in the intervention group over the course of the first two postoperative weeks. On post-operative day (POD) one the median reported pain in the interventional group was five (range 0-8) and 3.5 tablets of pain medica-tion (range 0-9). On POD 10 the median reported pain was zero (range 0-5) with zero tablets of opioid medica-tion utilized (range 0-4). The intervention group utilized 20% (IQR:10%-28%) of their prescribed narcotic over the first two postoperative weeks versus 50% (IQR:4%-68%) utilization in the control group. In the intervention group, five of seven patients were no longer utilizing opioids on POD 11 and six of seven were no longer utilizing opioids on POD 13.

Table 2. Procedures Undergone by Participating Patients

Intervention Group

Count Procedures

4 Knee mass excision

1 Arm mass excision

1 Osteochondroma excision

1 Curettage and bone grafting

Control Group

1 Chest mass excision

1 Leg sarcoma excision

2 Knee mass excision

1 Elbow lipoma excision

2 Thigh mass excision

1 Foot mass excision

1 Arm soft tissue excision

Table 3. Interventional Group Patients Clinic Visit vs. Reported Pill Count

Participant Clinic Visit Pill Count SMS Reported Pill Count

1 ** 2

2 24 26

3 ** 4

4 0 16

5 3 4

6 59 59

7 17 17

SMS, Short Messaging Service; **Patient did not return prescrip-tion opioid medication bottle at 2-week follow-up.



A B C

DISCUSSIONMobile phone and software communication platforms

are beneficial in the delivery of patient reported outcomes and treatment of many conditions including alcohol use disorder, HIV antiretroviral treatment, and diabetes.10,

11, 19, 20 Previous work found mobile phone communica-tion provided better data collection in comparison to traditional methods in patients answering post-surgical pain questions.21 Additionally, the patients reported the post-surgical mobile pain surveys were more convenient than traditional communication methods.21 Understanding of software interaction rates in the orthopaedic tumor population is important given evidence that increased communication, including via text messaging, in cancer patients helps this patient population cope better with their treatment and is ultimately associated with bet-ter treatment outcomes.13, 22 Our group has previously utilized automated mobile phone messaging to interact with different patient populations finding interaction rates that approach 90%.23, 24 The current investigation showed that patients undergoing outpatient procedures for musculoskeletal tumors demonstrated a high interac-tion rate (90%) with an automated mobile phone messag-ing platform. We also found there was a high correlation between patient reported opioid medication utilization via our software platform and manual pill counts in clinic in this pilot investigation (r=0.90). In the post study follow-up questionnaire, patients predominantly stated they preferred text messaging communication versus other modern forms of telephone (voice) communication

or email in the postoperative period. Additionally, four of seven patients stated they would want to utilize the mobile phone messaging platform again if they were to undergo another procedure. We found that patients were generally accepting of our postoperative communication platform and interacted with the software at a high rate. Our results also suggested a good correlation between opioid medication data collected via automated mobile phone messaging and manual pill counts in clinic, though, further formal investigations are required to validate this finding. Given our findings, subsequent investigations can utilize our software communication methods when attempting to obtain postoperative pain and opioid data and also when trialing possible CBT interventions across different patient populations.

Controlling patient’s pain, specifically patients with cancer, contributes to their ability to cope with their treatment and achieve a better quality of life.5, 25 Current understanding of the daily and short term utilization of opioid medication after outpatient orthopaedic tumor surgery is limited. This data, in addition to the natural history of patient reported pain after outpatient muscu-loskeletal procedures, would be beneficial during patient counseling and benchmarking for normative values. Patients in the intervention group of this study showed a gradual reduction in patient reported pain scores and opioid utilization over the two-week study period (Figure 1). Though this investigation is limited in scope due the nature of the pilot investigation, our results suggested a down trending course of pain and opioid medication utili-

Figure 1. Daily median pain and opioid demand in interventional group patient.

Volume 39 89


zation over the first two postoperative weeks with nearly all patients reporting they were no longer utilizing opioid medications by the end of the second postoperative week.

Acute and chronic pain is affected by psychological factors including pain catastrophizing, which ultimately results in interference with the patient’s activities of daily living and a lower quality of life. Mobile app and text messaging communication with patients using the principles of CBT is noted to be helpful and efficacious in the treatment of several psychological conditions and cog-nitive problems, including pain catastrophizing.16, 17 This pilot investigation demonstrated lower patient reported opioid utilization in the experimental group receiving automated software communication that was designed on the principles of CBT, compared to a control group that did not receive software CBT messaging. Although our pilot investigation was not powered to comment on the statistical significance of our results, we report that delivery of automated mobile phone messaging protocols implementing CBT may be a way to decrease opioid uti-lization in the acute postoperative period. Further work with a larger patient population is indicated to determine the true efficacy of mobile phone and software delivered CBT in the orthopaedic tumor patient population and other orthopaedic patient populations.

This study was not without limitations. First, the study was a single center and single surgeon feasibility study which may not necessarily make our results reproducible across other healthcare settings. Second, our study was designed to test the feasibility and patient acceptance of daily automated mobile phone messages and a CBT intervention and thus had a small sample size and was not sufficiently powered to report findings of possible statisti-cal significance. We also identify the patient population we studied was undergoing relatively low morbidity, outpatient surgery. This potentially skews our results and warrants further investigation utilizing similar methods in patients who are undergoing more extensive procedures, those with chronic pain, and patient popula-tions who have been previously identified to be at risk for postoperative opioid utilization.26-29 Finally, the two-week follow up period was short and longer studies would be needed to observe effects long-term on patient outcomes.

CONCLUSIONHigh mobile phone messaging interaction rates with

our software driven communication platform and CBT intervention suggest this is a feasible method and commu-nication intervention to test in other patient populations. Interaction rates are high in orthopaedic tumor patients using automated mobile phone messaging to assess post-operative daily pain score and opioid use. Obtaining post-operative opioid medication utilization data via automated

mobile phone messaging suggested a good correlation with in-office pill counts at two weeks. CBT delivered via our software may reduce opioid use in orthopaedic tumor patients, though further work with appropriately powered studies are needed to properly assess this observation.

REFERENCES1. Morris BJ, Mir HR. The Opioid Epidemic: Im-

pact on Orthopaedic Surgery. JAAOS - Journal of the American Academy of Orthopaedic Surgeons. 2015;23(5):267-71. doi: 10.5435/jaaos-d-14-00163. PubMed PMID: 00124635-201505000-00001.

2. Owen GT, Burton AW, Schade CM, Passik S. Urine drug testing: current recommendations and best practices. Pain Physician. 2012;15(3 Suppl):Es119-33. Epub 2012/07/20. PubMed PMID: 22786451.

3. Mandell BF. The fifth vital sign: A complex story of politics and patient care. Cleve Clin J Med. 2016;83(6):400-1. Epub 2016/06/10. doi: 10.3949/ccjm.83b.06016. PubMed PMID: 27281246.

4. Rudd RA, Seth P, David F, Scholl L. Increases in Drug and Opioid-Involved Overdose Deaths - United States, 2010-2015. MMWR Morb Mortal Wkly Rep. 2016;65(5051):1445-52. Epub 2016/12/30. doi: 10.15585/mmwr.mm655051e1. PubMed PMID: 28033313.

5. Anghelescu DL, Oakes LL, Hankins GM. Treat-ment of Pain in Children after Limb-Sparing Surgery: an Institution’s 26-year Experience. Pain Manag Nurs. 2011;12(2):82-94. doi: 10.1016/j.pmn.2010.02.002. PubMed PMID: 21620310; PubMed Central PMCID: PMCPMC3104211.

6. Mantyh P. Bone cancer pain: causes, consequenc-es, and therapeutic opportunities. Pain. 2013;154 Suppl 1:S54-62. Epub 2013/08/07. doi: 10.1016/j.pain.2013.07.044. PubMed PMID: 23916671.

7. Bannister K. Opioid-induced hyperalgesia: where are we now? Curr Opin Support Palliat Care. 2015;9(2):116-21. Epub 2015/04/15. doi: 10.1097/spc.0000000000000137. PubMed PMID: 25872113.

8. Anthony CA, Polgreen LA, Chounramany J, Fos-ter ED, Goerdt CJ, Miller ML, et al. Outpatient blood pressure monitoring using bi–directional text messaging. J Am Soc Hypertens. 2015;9(5):375-81. doi: https://doi.org/10.1016/j.jash.2015.01.008.

9. Anthony CA, Lawler EA, Ward CM, Lin IC, Shah AS. Use of an Automated Mobile Phone Messaging Robot in Postoperative Patient Monitoring. Telemed J E Health. 2018;24(1):61-6. Epub 2017/06/18. doi: 10.1089/tmj.2017.0055. PubMed PMID: 28622079.



10. Marcolino MS, Oliveira JAQ, D’Agostino M, Ribeiro AL, Alkmim MBM, Novillo-Ortiz D. The Impact of mHealth Interventions: Systematic Review of Systematic Reviews. JMIR Mhealth Uhealth. 2018;6(1):e23. Epub 2018/01/19. doi: 10.2196/mhealth.8873. PubMed PMID: 29343463; PubMed Central PMCID: PMCPMC5792697.

11. Berrouiguet S, Baca-Garcia E, Brandt S, Walter M, Courtet P. Fundamentals for Future Mobile-Health (mHealth): A Systematic Review of Mobile Phone and Web-Based Text Messaging in Mental Health. J Med Internet Res. 2016;18(6):e135. Epub 2016/06/12. doi: 10.2196/jmir.5066. PubMed PMID: 27287668; PubMed Central PMCID: PM-CPMC4920962.

12. Jones KR, Lekhak N, Kaewluang N. Using mobile phones and short message service to deliver self-man-agement interventions for chronic conditions: a meta-review. Worldviews Evid Based Nurs. 2014;11(2):81-8. Epub 2014/03/07. doi: 10.1111/wvn.12030. PubMed PMID: 24597522.

13. Rico TM, Dos Santos Machado K, Fernandes VP, Madruga SW, Noguez PT, Barcelos CRG, et al. Text Messaging (SMS) Helping Cancer Care in Patients Undergoing Chemotherapy Treatment: a Pilot Study. J Med Syst. 2017;41(11):181. Epub 2017/10/11. doi: 10.1007/s10916-017-0831-3. PubMed PMID: 28990135.

14. Benzon HT. Raj’s practical management of pain. Fourth edition.. ed. Benzon HT, Raj PP, editors: Phila-delphia : Mosby-Elsevier; 2008.

15. Nicholls JL, Azam MA, Burns LC, Englesakis M, Sutherland AM, Weinrib AZ, et al. Psychological treatments for the management of postsurgical pain: a systematic review of randomized controlled trials. Patient Relat Outcome Meas. 2018;9:49-64. Epub 2018/02/07. doi: 10.2147/prom.S121251. PubMed PMID: 29403322; PubMed Central PMCID: PM-CPMC5783145.

16. Shapiro JR, Bauer S. Use of short message service (SMS)-based interventions to enhance low intensity CBT. Oxford guide to low intensity CBT interventions. 2010:281-6.

17. Rathbone AL, Clarry L, Prescott J. Assessing the Efficacy of Mobile Health Apps Using the Basic Principles of Cognitive Behavioral Therapy: System-atic Review. J Med Internet Res. 2017;19(11):e399. Epub 2017/12/01. doi: 10.2196/jmir.8598. PubMed PMID: 29187342; PubMed Central PMCID: PM-CPMC5727354.

18. Shoukri MM. Statistical methods for health sciences. Second edition.. ed. Pause CA, editor: Boca Raton, Fla. : CRC Press; 1999.

19. Hall AK, Cole-Lewis H, Bernhardt JM. Mobile text messaging for health: a systematic review of reviews. Annu Rev Public Health. 2015;36:393-415. Epub 2015/03/19. doi: 10.1146/annurev-publ-health-031914-122855. PubMed PMID: 25785892; PubMed Central PMCID: PMCPMC4406229.

20. Anthony CA, Lawler EA, Glass NA, McDon-ald K, Shah AS. Delivery of Patient-Reported Outcome Instruments by Automated Mobile Phone Text Messaging. Hand (N Y). 2017;12(6):614-21. Epub 2017/11/02. doi: 10.1177/1558944716672201. PubMed PMID: 29091492; PubMed Central PMCID: PMCPMC5669321.

21. Stomberg MW, Platon B, Widén A, Wallner I, Karlsson O. Health Information: What Can Mobile Phone Assessments Add? Perspectives in Health In-formation Management / AHIMA, American Health Information Management Association. 2012;9(Fall):1d. PubMed PMID: PMC3510647.

22. Tian J, Jia LN, Cheng ZC. Relationships between patient knowledge and the severity of side effects, daily nutrient intake, psychological status, and per-formance status in lung cancer patients. Curr Oncol. 2015;22(4):e254-8. Epub 2015/08/25. doi: 10.3747/co.22.2366. PubMed PMID: 26300675; PubMed Central PMCID: PMCPMC4530822.

23. Anthony CA, Peterson AR. Utilization of a text-messaging robot to assess intraday variation in con-cussion symptom severity scores. Clin J Sport Med. 2015;25(2):149-52. Epub 2014/06/07. doi: 10.1097/jsm.0000000000000115. PubMed PMID: 24905538.

24. Anthony CA, Volkmar A, Shah AS, Willey M, Karam M, Marsh JL. Communication with Orthope-dic Trauma Patients via an Automated Mobile Phone Messaging Robot. Telemed J E Health. 2017. Epub 2017/12/21. doi: 10.1089/tmj.2017.0188. PubMed PMID: 29261036.

25. Caltagirone C, Spoletini I, Gianni W, Spalletta G. Inadequate pain relief and consequences in onco-logical elderly patients. Surg Oncol. 2010;19(3):178-83. Epub 2009/12/18. doi: 10.1016/j.suronc.2009.11.009. PubMed PMID: 20015635.

26. Bedard NA, Pugely AJ, Dowdle SB, Duchman KR, Glass NA, Callaghan JJ. Opioid Use Following Total Hip Arthroplasty: Trends and Risk Factors for Prolonged Use. J Arthroplasty. 2017;32(12):3675-9. Epub 2017/09/18. doi: 10.1016/j.arth.2017.08.010. PubMed PMID: 28917616.

Volume 39 91


27. Bedard NA, Pugely AJ, Westermann RW, Duch-man KR, Glass NA, Callaghan JJ. Opioid Use After Total Knee Arthroplasty: Trends and Risk Factors for Prolonged Use. J Arthroplasty. 2017;32(8):2390-4. Epub 2017/04/18. doi: 10.1016/j.arth.2017.03.014. PubMed PMID: 28413136.

28. Anthony CA, Westermann RW, Bedard N, Glass N, Bollier M, Hettrich CM, et al. Opioid Demand Before and After Anterior Cruciate Ligament Re-construction. Am J Sports Med. 2017;45(13):3098-103. Epub 2017/08/15. doi: 10.1177/0363546517719226. PubMed PMID: 28806097.

29. Westermann RW, Anthony CA, Bedard N, Glass N, Bollier M, Hettrich CM, et al. Opioid Consumption After Rotator Cuff Repair. Arthroscopy. 2017;33(8):1467-72. Epub 2017/06/03. doi: 10.1016/j.arthro.2017.03.016. PubMed PMID: 28571723.


ABSTRACTBackground: The Oswestry Disability Index

(ODI) is among the most widely used patient re-ported outcome measures for the assessment of spinal conditions. Traditionally, the ODI has been administered in outpatient clinics on a face-to-face basis, which can be expensive and time consum-ing. Furthermore, the percentage of patients lost to clinical follow-up is high, particularly after 2-5 years. Thus, telephonic administration of the ODI, if valid, could be a convenient way of capturing pa-tient outcomes and increasing follow-up rates. The objective of this study was to validate telephonic administration of the ODI compared to face-to-face administration.

Methods: A convenience sample of individuals with and without back pain in an academic medi-cal center were recruited for this study. Face-to-face administration of the ODI was completed and retested 24 hours later via phone. Test-retest reliability was determined by calculating the intra-class correlation coefficient.

Results: 22 individuals completed the ODI ques-tionnaire face-to-face, then via telephone 24 hours later. There was a mean 2% (± 3) intra-rater ODI score difference (range: 0% to 12%). The intra-class correlation coefficient overall was 0.98 (95% CI: 0.96, 0.99, p<0.001) with a range of 0.95 to 1.0, revealing near-perfect test-retest reliability.

Conclusions: Administration of the ODI ques-tionnaire over the phone has excellent test-retest reliability when compared to face-to-face adminis-tration. Telephone administration is a convenient and reliable option for obtaining follow-up out-comes data.

Clinical Relevance: Telephonic administration of the ODI is scientifically valid, and should be an accepted method of data collection for state-level and national-level outcomes projects.

Keywords: Oswestry Disability Index; ODI; spine; low back pain; disability; assessment; questionnaire; quality of life; outcomes; validation; telephone

INTRODUCTIONThe measurement of disability is a critical factor in

caring for patients with spinal conditions. Although multiple measurement tools exist, the Oswestry Dis-ability Index (ODI) is among the most widely used.1 Traditionally, the ODI was collected as a patient-admin-istered questionnaire during face-to-face clinic visits.1 However, administration in clinic requires time and clinical resources. Furthermore, patients may be unable or unwilling to return to clinic, and the percentage lost to clinical follow-up is high, particularly by 2-5 years.2,3 Thus, telephonic methods of administering ODI data forms would be helpful both in terms of convenience for the patient and provider, and also in increasing follow-up participation rates. Although telephone administration of the ODI has been reported,4 it has not been extensively validated.

Recently, increased emphasis has been placed on public reporting of patient-outcomes data.5 The goal of these initiatives is to facilitate practice-based learning, transparency, and shared decision making for patients. However, there exist both inducements for success and penalties for poor results, and physicians and hospitals risk financial loss if they fail to comply.5 Specifically in Minnesota, the Minnesota Community Measurement organization (MNCM) has begun collecting and publicly releasing patient-reported outcome measures (PROM) from health care centers across the state.6 Furthermore, they provide a data collection guide which outlines acceptable forms of administration for various PROM tools, including the ODI.7 In recent editions, telephonic administration was deemed not acceptable due to a lack of validation.7

Thus, further validation of the ODI administered via telephone is needed. Successful validation of this tool should be useful to promote its utility for payors and

OSWESTRY DISABILITY INDEX: IS TELEPHONE ADMINISTRATION VALID?

Christopher T. Martin, MD1, Alexandra K. Yaszemski, MBA2, Charles G. T. Ledonio, MD3, Tara C. Barrack, PA-C4, David W. Polly, Jr., MD1

1Department of Orthopaedic Surgery, University of Minnesota, Minneapolis, MN2Uniformed Services University of the Health Sciences, Bethesda, MD3Innovative Surgical Designs, Inc., Wayne, PA4Summit Orthopedics, Woodbury, MNCorresponding Author: Christopher T. Martin; Phone: (612) 273-1177; Email: [email protected] Disclosures: The authors report no potential conflicts of interestrelated to this study.Sources of Funding: No sources of funding declared.

C.T. Martin, A.K. Yaszemski, C.G T. Ledonio, T.C. Barrack, D.W. Polly, Jr.


governmental organizations, and should provide evi-dence for a more convenient method of administering the survey for clinicians.

METHODSPatient Selection

A convenience sample of 22 individuals with and without back pain in an academic medical center was recruited to participate in this study. Convenience sam-pling is a type of non-probability sampling that involves the sample being drawn from the population close at hand, and the only inclusion criteria is a willingness to participate. Patients with symptomatic back pain (10 of 22) were recruited from the senior author’s clinic during a scheduled visit for a spinal condition. The remaining study subjects (12 of 22) were asymptomatic support staff who worked in the clinic and agreed to participate. Mean age was 48.6 years (standard deviation ± 14.7 years) and most subjects were women (73%). Participants were asked to complete the ODI questionnaire (ODI v 2.1a) in standard fashion.1 Twenty-four hours later, the test was re-administered via telephone by a single individual. The ODI questionnaire asks patients about symptoms they are experiencing “now.” Thus, the 24-hour interval was selected in order to remain as close as possible to the observation period for assessing those current symptoms.

Administration, Scoring, and Data AnalysisBoth telephone and in-person administration were

performed in accordance with the MAPI Research Trust administration guidelines.8 Test-retest reliability and intra-class correlation coefficient (ICC) were calculated using SPSS Statistics Desktop, version 20.0.0 (IBM Corporation 2011, Armonk NY, USA).

RESULTSTable 1 reports the patient demographic results in this

study. The mean scaled score for both the written and telephone versions was 25% (range 0-70% for written and 0-76% for telephone). There was a mean 2% (± 3) intra-rater ODI score difference (range 0% to 12%) (Figure 1). The intra-class correlation coefficient overall was 0.98 (95% CI: 0.96, 0.99, p<0.001) with a range of 0.95 to 1.0, revealing near-perfect test-retest reliability.

DISCUSSIONMaintaining clinical follow-up rates beyond 6-12

months is particularly difficult, and even well conducted clinical trials with substantial administrative support frequently report less than 50% follow-up rates beyond 5 years.3 Thus, physicians need methods of collecting patient outcome data that do not require face-to-face clinical visits, and telephonic administration holds sub-stantial promise in that regard. However, given that many of the original PROM measures were designed to be conducted in-person, administration over the telephone must be validated before the method gains widespread acceptance.7

In this study, we found very high reliability for tel-ephonic administration of the ODI with an ICC of 0.98, suggesting near perfect test-retest reliability. In the origi-nal publication of the ODI by Fairbank et al., reliability was assessed with repeated written administrations 24 hours apart (n=22, r = 0.99).8 With an ICC of 0.98, our study suggests that telephonic administration is just as effective.

Two primary advantages exist for administration of the ODI via telephone. First, it should be easier to ob-tain clinical follow-up if telephonic administration is an accepted method of data collection. In a study evaluating the performance characteristics of the verbal QuickDash, follow-up could not be obtained for 17% of patients for the written survey, but that number dropped to 0% for the verbal method.9 Second, telephonic administration is more convenient for patients. In a study by Bokshan et al., which investigated the validity of spine based PROMs, the majority of patients (57%) preferred the telephone over in-person administration (29%, with 14% expressing no preference).4

A limitation to this study is that all of the patients had

Table 1: Patient Demographics

Sample Size n = 22

Female gender 16 (73%)

Scheduled clinic visit 10 (45%)

Age 48.6 ± 14.7 years

Figure 1. Oswestry Disability Index scores obtained via written and telephone administration.


ODI Validity Via Phone

A

the ODI administered face-to-face prior to the telephone attempt, so they were familiar with the written format. De novo telephone testing may not have the same reproduc-ibility. Other studies have used a wider interval between the written and telephone versions of the survey to try and minimize this recall bias.4 However, the ODI specifi-cally asks patients about symptoms they are experienc-ing “now.” Thus, we felt that a narrow window between the two survey administrations was most appropriate. Furthermore, telephone administration, while more convenient for the patient, may increase the burden on the physician’s office staff.

CONCLUSIONOverall, telephonic administration is a well-recognized

and validated method for multiple other PROMs across multiple medical disciplines,10,11 as well as for other PROMs in spine.4 Our data adds to this bulk of literature supporting this method of data collection. Telephonic ad-ministration of the ODI is scientifically valid, and should be an accepted method of data collection for state-level and national-level outcomes projects.

REFERENCES1. Fairbank JC, Pynsent PB. The Oswestry Disability

Index. Spine (Phila Pa 1976). 2000;25(22):2940-2952; discussion 2952.

2. Solberg TK, Sorlie A, Sjaavik K, Nygaard OP, Ingebrigtsen T. Would loss to follow-up bias the outcome evaluation of patients operated for degen-erative disorders of the lumbar spine? Acta Orthop. 2011;82(1):56-63.

3. Lurie JD, Tosteson TD, Tosteson A, et al. Long-term outcomes of lumbar spinal stenosis: eight-year results of the Spine Patient Outcomes Research Trial (SPORT). Spine (Phila Pa 1976). 2015;40(2):63-76.

4. Bokshan SL, Godzik J, Dalton J, Jaffe J, Lenke LG, Kelly MP. Reliability of the revised Scoliosis Research Society-22 and Oswestry Disability Index (ODI) questionnaires in adult spinal deformity when administered by telephone. Spine J. 2016;16(9):1042-1046.

5. Bekelis K, McGirt MJ, Parker SL, et al. The present and future of quality measures and pub-lic reporting in neurosurgery. Neurosurg Focus. 2015;39(6):E3.

6. MN Community Measurement. Our Story. 2017; http://mncm.org/about-us/our-story/#the-need-for-mncm. Accessed 30 Aug 2017.

7. MN Community Measurement. Spinal surgery: Functional status and quality of life outcome mea-sures 2015. 2015; http://mncm.org/wp-content/uploads/2013/04/Spinal-Surgery-Data-Collection-Guide-2015-FINAL-v1.pdf. Accessed 30 Aug 2017.

8. Fairbank JC, Couper J, Davies JB, O’Brien JP. The Oswestry low back pain disability questionnaire. Physiotherapy. 1980;66(8):271-273.

9. London DA, Stepan JG, Boyer MI, Calfee RP. Performance characteristics of the verbal Quick-DASH. J Hand Surg Am. 2014;39(1):100-107.

10. Farzanfar R, Hereen T, Fava J, Davis J, Vachon L, Friedman R. Psychometric properties of an auto-mated telephone-based PHQ-9. Telemed J E Health. 2014;20(2):115-121.

11. Geller EJ, Barbee ER, Wu JM, Loomis MJ, Visco AG. Validation of telephone administration of 2 condition-specific quality-of-life questionnaires. Am J Obstet Gynecol. 2007;197(6):632 e631-634.

IOWAORTHOPAEDIC JOURNAL

VOLUME 35

Christopher Martin, MDRobert Westermann, MD

Jesse Otero, MD, PhD

2015

Department of Orthopedics

Arthur Steindler, 1912-1949

Theodore Willis, 1917-1918

Joseph Milgram, 1926-1932

Ernest Freund, 1932-1936

Thomas Waring, 1932-1939

James Vernon Luck, 1936-1939

Ignacio Ponseti, 1946-2009

Eberly Thornton, 1946-1952

Robert Newman, 1948-1956

Michael Bonfiglio, 1950-1995

Carroll Larson, 1950-1978

Adrian Flatt, 1956-1979

Reginald Cooper, 1962-2016

Howard Hogshead, 1964-1965

Maurice Schnell, 1964-1965

Richard Johnston,

1967-1970, 1998-2016

Donald Kettelkamp, 1968-1971

Gerald Laros, 1968-1971

Richard Stauffer, 1970-1972

John Albright, 1972-present

Doug Mains, 1972-1973

Bruce Sprague, 1972-1979

Richard Brand, 1974-2002

Mike Mickelson, 1976-1981

Stuart Weinstein, 1976-present

Thomas Lehmann, 1978-1987

Joseph Buckwalter IV, 1979-present

William Pontarelli,

1980-1984; 1999-2004

Charles Clark, 1980-present

William Blair, 1980-1997

Barbara Campbell, 1982-1984

James Weinstein, 1983-1996

James Nepola, 1984-present

Fred Dietz, 1984-2015

Curtis Steyers, 1985-2006

Ernest Found, 1987-2017

Lawrence Marsh, 1987-present

David Tearse, 1989-2000

John Callaghan, 1990-2018

Charles Saltzman, 1991-2005

Brian Adams, 1993-2014

Mindy Trotter, 2018-present

Joshua Holt, 2018-present

Matthew Hogue, 2018-present

Kyle Duchman, 2018-present

Timothy Brown, 2017-present

Joseph Buckwalter V, 2017-present

Philip Chen, 2017-present

Jesse Otero, 2017-present

Brendan Patterson, 2017-present

Robert Westermann, 2017-present

Timothy Fowler, 2016-present

Andrew Pugely, 2016-present

Chester Pelsang, 2016-2019

Lindsey Caldwell, 2015-present

Cassim Igram, 2015-present

Michael Willey, 2014-present

Heather Kowalski, 2014-present

Eric Aschenbrenner, 2012-present

Apurva Shah, 2012-2015

Melissa Willenborg, 2012-2016

Mederic Hall, 2011-present

Carolyn Hettrich, 2011-2017

Ryan Ilgenfritz, 2011-2014

Matthew Karam, 2011-present

Matthew Bollier, 2010-present

Benjamin Miller, 2010-present

Heather Bingham, 2008-present

Phinit Phisitkul, 2008-2017

Nicolas Noiseux , 2007-present

Robert Yang, 2007-2010

Ericka Lawler, 2006-present

John Femino, 2005-present

Joseph Smucker, 2005-2014

Neil Segal, 2004-2014

Valerie Keffala, 2004 -present

Brian Wolf, 2003-present

Michael O’Rourke, 2003-2007

Sergio Mendoza, 2003-2015

Jose Morcuende, 2001-present

Annunziato Amendola, 2001-2015

Joseph Chen, 2000-2018

Todd McKinley, 1999-2012

R. Kumar Kadiyala, 1998-2004

Leon Grobler, 1996-1999

The University of Iowa Roy J. and Lucille A. Carver College of Medicine

VOLUM 3 ISSUE 2 - University of Iowa...Corresponding Author: Jessica E. Goetz, PhD; Phone: (319)...

Documents

Transcript of VOLUM 3 ISSUE 2 - University of Iowa...Corresponding Author: Jessica E. Goetz, PhD; Phone: (319)...