Age related change in strength, joint laxity, and walking patterns. are they related to knee oa

Age-Related Changes in Strength,Joint Laxity, and Walking Patterns: AreThey Related to Knee Osteoarthritis?Katherine S Rudolph, Laura C Schmitt, Michael D Lewek

Background and PurposeAging is associated with musculoskeletal changes and altered walking patterns. Thesechanges are common in people with knee osteoarthritis (OA) and may precipitate thedevelopment of OA. We examined age-related changes in musculoskeletal structuresand walking patterns to better understand the relationship between aging and kneeOA.

MethodsForty-four individuals without OA (15 younger, 15 middle-aged, 14 older adults) and15 individuals with medial knee OA participated. Knee laxity, quadriceps femorismuscle strength (force-generating capacity), and gait were assessed.

ResultsMedial laxity was greater in the OA group, but there were no differences between themiddle-aged and older control groups. Quadriceps femoris strength was less in theolder control group and in the OA group. During the stance phase of walking, the OAgroup demonstrated less knee flexion and greater knee adduction, but there were nodifferences in knee motion among the control groups. During walking, the oldercontrol group exhibited greater quadriceps femoris muscle activity and the OA groupused greater muscle co-contraction.

Discussion and ConclusionAlthough weaker, the older control group did not use truncated motion or higherco-contraction. The maintenance of movement patterns that were similar to thesubjects in the young control group may have helped to prevent development ofknee OA. Further investigation is warranted regarding age-related musculoskeletalchanges and their influence on the development of knee OA.

KS Rudolph, PT, PhD, is AssistantProfessor, Department of PhysicalTherapy and Program in Biome-chanics and Movement Science,University of Delaware, 301McKinly Lab, Newark, DE 19716(USA). Address all correspondenceto Dr Rudolph at: [email protected].

LC Schmitt, PT, PhD, is Post-Doctoral Fellow, Department ofPediatrics, University of Cincin-nati, College of Medicine; andPhysical Therapist, Sports Medi-cine Biodynamics Center, Cincin-nati Children’s Hospital MedicalCenter, Cincinnati, Ohio.

MD Lewek, PT, PhD, is AssistantProfessor, Center for HumanMovement Science, Division ofPhysical Therapy, University ofNorth Carolina at Chapel Hill,Chapel Hill, NC.

[Rudolph KS, Schmitt LC, LewekMD. Age-related changes instrength, joint laxity, and walkingpatterns: are they related to kneeosteoarthritis? Phys Ther. 2007:87:1422–1432.]

© 2007 American Physical TherapyAssociation

Research Report

Post a Rapid Response orfind The Bottom Line:www.ptjournal.org

1422 f Physical Therapy Volume 87 Number 11 November 2007

Symptomatic knee osteoarthritis(OA) is a worldwide problem1–4

that produces substantial dis-ability in middle-aged and olderadults and leads to a tremendouseconomic burden on society.5 Theprevalence of OA among older indi-viduals has led some authors6 to re-gard its development as a normalpart of aging. Loeser and Shakoor,6

however, suggested that age-relatedchanges in musculoskeletal tissue,such as muscle weakness and liga-ment laxity, do not directly causeOA, but may predispose individualsto develop the disease. It is possiblethat the manner in which people re-spond to these age-related changesin musculoskeletal tissues about theknee may be related to whether ornot OA develops in the knees ofolder adults.

The features that are similar betweenolder adults and people with kneeOA include quadriceps femoris mus-cle weakness and altered knee move-ment during walking. Sarcopenia iswell known in older adults and leadsto quadriceps femoris muscle weak-ness,7–11 which has been noted inpeople as early as 40 years of age.9

Because both the development ofknee OA12 and quadriceps femorismuscle strength changes9 are initi-ated during middle age, it is not sur-prising that quadriceps femoris mus-cle weakness has been implicated inthe development of knee OA.13–15

Quadriceps femoris muscle weak-ness also is associated with adapta-tions in walking patterns that are the-orized to put articular cartilage atrisk. For instance, subjects with kneeOA who have weaker quadricepsfemoris muscles exhibit less stance-phase knee motion during walking.16

At self-selected walking speeds, it isthe role of the quadriceps femorismuscles to control knee flexion dur-ing weight acceptance while thehamstring and gastrocnemius mus-cles are typically silent.17 However,

in the presence of quadriceps femo-ris weakness, which occurs with ag-ing,7–11 and in the presence of kneeOA, either the hamstring or the gas-trocnemius muscles may be requiredto assist with knee control.

Activation of muscles surroundingthe knee can occur selectively andwith precise timing that allows fornormal knee motion, or, alterna-tively, activation can occur moregenerally as a global co-contractionpattern that could limit joint motion.We, therefore, have defined themovement strategy that involves bothincreased muscle co-contraction andreduced knee flexion during walkingas a “stiffening strategy.” Excessivemuscle co-contraction can lead toexcessive joint contact forces,18 andreduced knee motion during weightacceptance can cause higher impactloads in the knee.19,20 Older adultsare known to walk with less kneeflexion,21 but whether they do soas a result of higher muscle co-contraction is unknown. However,not all older adults develop knee OA.If older adults who have not devel-oped knee OA walk with a knee stiff-ening strategy, then the combinationof reduced knee flexion and muscleco-contraction alone, in the pres-ence of quadriceps femoris muscleweakness, is unlikely to contributeto the development of knee OA.

Another possible precursor to kneeOA is excessive frontal-plane laxity,which is common in people withexisting knee OA.22–24 Specifically,greater frontal-plane knee laxity isobserved in both the involved anduninvolved knees of people with OAcompared with control subjects, sug-gesting that laxity may precede thedevelopment of knee OA.22 Addi-tionally, a significant correlation be-tween frontal-plane laxity and agehas been observed in individualswithout evidence of knee OA.22 Thisfinding is consistent with the find-ings of other studies,22,25,26 indicat-

ing that the material properties ofligaments of older adults can lead toexcessive joint laxity. Becausefrontal-plane laxity has been relatedto high muscle co-contraction in in-dividuals with knee OA,24 it is plau-sible that normal age-related in-creases in joint laxity also maycontribute to higher muscle co-contraction patterns and predisposeindividuals to develop knee OA.Whether older adults have greaterfrontal-plane knee laxity coupledwith higher muscle co-contraction isnot known.

In this study, we investigated theknee laxity, quadriceps femoris mus-cle strength (force-generating capac-ity), walking patterns, and muscle ac-tivation patterns in 3 age groups ofpeople without symptomatic or ra-diographic knee OA to examine fac-tors that are thought to contribute tothe development of knee OA. Theresults are discussed in relation tocharacteristics of a group of peoplewith knee OA. Because the olderadults in our study did not have jointdegeneration, we hypothesized thatolder adults who are healthy willhave weaker quadriceps femorismuscles and increased frontal-planeknee laxity but will not exhibitgreater muscle co-contraction pat-terns compared with young ormiddle-aged people.

MethodSubjectsFifty-nine people were recruitedfrom the community or were re-ferred by a local orthopedic surgeonto participate in the study. All sub-jects signed an informed consentstatement approved by the HumanSubjects Review Board of the Univer-sity of Delaware. Forty-four partici-pants who reported no history ofknee OA (confirmed by radiograph)or previous lower-extremity injurycomprised 3 control groups (15younger individuals [ages 18–25years], 15 middle-aged individuals

Age-Related Changes in Strength, Joint Laxity, and Walking Patterns

November 2007 Volume 87 Number 11 Physical Therapy f 1423

[ages 40–59 years], and 14 older in-dividuals [ages 60–80 years])(Tab. 1). The middle-aged individualswere matched by age and sex to 15people with symptomatic, medialknee OA (Tab. 1). The subjects withmedial knee OA were part of a largerstudy of people who were going toundergo a high tibial osteotomy.They had no history of knee liga-ment injury; however, those individ-uals with a history of meniscectomywere included. Data on some ofthe people with knee OA and dataon the middle-aged control grouphave been reported previously.24

Radiographic information, isometricquadriceps femoris strength, and ki-nematic, kinetic, and electromyo-graphic (EMG) data during walkingwere collected from the more-involved limb of the subjects withOA and a randomly chosen limb ofthe control subjects. The test limb ofthe control subjects was chosen ran-domly to avoid any possible influ-ence of limb dominance.

ProcedureRadiographs. The diagnosis of OAis based on the presence of kneepain in conjunction with age over 50years and either radiographic evi-dence of OA (eg, osteophytes) orother symptoms such as stiffness orcrepitus.27 Although none of ourcontrol subjects complained of kneepain or stiffness, standing posterior-

anterior (approximately 30° of kneeflexion) radiographs of the knees ofthe middle-aged and older controlgroups were obtained as an addedprecaution to rule out the presenceof knee OA. Radiographs were nottaken of the knees of the young sub-jects because they had no kneesymptoms and no history of kneeinjury and were unlikely to have un-diagnosed knee OA based on theabove definition.

Varus and valgus stress radiographswere taken of the tested lower ex-tremity in the middle-aged and oldercontrol groups as well as the OAgroup. Subjects were positioned su-pine on a radiograph table with theknee flexed to 20 degrees and thepatella facing anteriorly. The x-raytube was centered approximately100 cm above the knee joint. ATELOS* stress device was used to ap-ply a 150-N force in the varus orvalgus direction (Fig. 1). Medial andlateral joint spaces were measuredat the narrowest location in bothcompartments using calipers. X-raybeams were adjusted for magnifica-tion using a known distance from theTELOS device that was visible in ev-ery image. Medial and lateral jointlaxities were calculated as describedin Figure 1.28 Interrater reliability

was assessed by repeated testing ona subset of 8 subjects and showedhigh reliability for medial (intraclasscorrelation coefficient [ICC]�.96)and lateral laxity (ICC�.98).

* Austin & Associates, 1109 Sturbridge Rd,Fallston, MD 21047.

Figure 1.Setup for stress radiographs. The top im-ages show the limb alignment in theTELOS device (top left) and the resultingradiograph (top right), and the method ofcalculating medial laxity is shown in thelower images. Lateral laxity was calculatedsimilarly but with subtraction of the lateraljoint space in valgus from lateral jointspace in varus.

Table 1.Group Characteristics

YoungControl Group(n�15)

Middle-agedControl Group(n�15)

OlderControl Group(n�14)

OsteoarthritisGroup(n�15)

Age (y), X (range) 20.6 (18–25) 49.2 (40–57) 68.8 (60–80) 49.2 (39–57)

Sex (female/male) 8/7 7/8 10/4 7/8

Body mass index (kg/m2), X (SD) 24.3 (2.8)a,b 28.7 (5.5)a 24.7 (2.5)c 30.7 (4.8)b.c

Alignment (°), X (SD) Not tested 0.1 (1.58)d valgus 1.0 (2.09)d varus 6.33 (2.39)d varus

a P�.030.b P�.001.c P�.002.d P�.001.



Skeletal alignment of the tested limbwas measured with a standing longcassette radiograph for the middle-aged and older control groups andthe OA group. Subjects stood, with-out footwear, with the tibial tuber-cles facing forward and the x-raybeam centered at the knee from adistance of 2.4 m. Alignment wasmeasured as the angle formed by theintersection of the mechanical axesof the femur and tibia.29–31 A kneewas in varus alignment when the in-tersection of the lines was �0 de-grees in the varus direction and wasin valgus alignment when the inter-section of the lines was �0 degreesin the valgus direction.30

Quadriceps femoris muscles func-tion. Quadriceps femoris muscleforce output was measured with anisokinetic dynamometer (Kin-Com500H).† Each subject sat with theknee and hip flexed to 90 degrees,the knee joint axis aligned with thedynamometer axis, and the trunkfully supported. Thigh and hip strapssecured each subject in the seat,while an ankle strap secured theshank to the dynamometer. Subjectsperformed a maximal volitional iso-metric contraction (MVIC) on whicha supramaximal burst of electricalcurrent (Grass S48 stimulator‡) (100pulses/second, 600-microsecondpulse duration, 10-pulse tetanictrain, 130 V) was applied. The burstsuperimposition was used for themeasurement if the subjects wereproviding maximum activation ofthe quadriceps femoris muscles.32 Acentral activation ratio (CAR) is a ra-tio between the highest volitionalforce (measured as the peak forcebefore the electrical burst was ap-plied) and the force achieved duringthe electrically elicited burst. Maxi-mum volitional isometric contrac-

tion was measured as the highestvolitional force (N) during the con-traction and was normalized to bodymass index (BMI) (N/BMI). In testson 10 subjects who were healthy,repeated testing of the MVIC re-vealed an intraclass correlation coef-ficient (2,1) of .98.33

Gait and electromyographic(EMG) data. To determine kneemotion during walking, the motionsof the lower-extremity segmentswere collected by a 6-camera, pas-sive, 3-dimensional motion analysissystem (Vicon 512)§ at 120 Hz. Cam-eras were calibrated to detect mark-ers within a volume that was 1.5 �2.4 � 1.5 m. Calibration residualswere kept below 0.6 mm. The cam-eras detected retroreflective markers(2.5 cm in diameter) placed on thetested lower extremity. Markerswere placed bilaterally over thegreater trochanters, the lateral femo-ral condyles, and lateral malleoli foridentification of appropriate jointcenters. Thermoplastic shells with 4rigidly attached markers were usedto track segment motion. The shellswere secured on the posterior-lateralaspects of the thigh and shank. Pre-vious work in our laboratory (unpub-lished data collected April 2002) hasrevealed good reliability for kine-matic variables with ICCs rangingfrom .6343 to .9969. The ICCs forthe kinematic variables used in thepresent study ranged from .9721 to.9969. Errors in estimating bonemovement from skin mounted mark-ers for sagittal and frontal plane mo-tions are approximately 2 to 3 de-grees during the stance phase ofwalking.34,35 Vertical, medial-lateral,and anterior-posterior ground reac-tion forces were collected from a6-component force platform (Bertecforce platform, model 60905�) and

sampled at 1,920 Hz. Ground reac-tion force data were used to calcu-late moments about the knee and fordetermination of heel-strike andtoe-off.

Electromyographic data were col-lected with a 16-channel electromyo-graphy system (model MA-300-16#)sampled at 1,920 Hz. After skin prep-aration, surface electrodes with par-allel, circular detection surfaces(1.14 cm in diameter, 2.06 cm apart),a common mode rejection ratio (100dB at 65 Hz), and a signal detectionrange of less than 2 �V for thebuilt-in preamplifier were placedover the mid-muscle bellies of thelateral quadriceps femoris (LQ), me-dial quadriceps femoris (MQ), lateralhamstring (LH), medial hamstring(MH), lateral gastrocnemius (LG),and medial gastrocnemius (MG) mus-cles. Electromyographic data wererecorded for 2 seconds at rest andduring an MVIC for each musclegroup for normalization.

Motion, force, and EMG data werecollected simultaneously as subjectswalked at a self-selected speed alonga 9-m walkway for 10 trials. Walkingspeed was recorded from 2 photo-electric beams to ensure that speeddid not vary more than 5% from theirself-selected speed during the trials.Trials were only accepted if the sub-ject walked at a consistent speed andwalked across the force platformwithout adjusting their stride in anyway to contact the force platform.

Data management. Marker tra-jectories and ground reaction forceswere collected over the stance phaseof one limb (heel-strike to toe-offon the force platform) and were fil-tered with a second-order, phase-corrected Butterworth filter with acutoff frequency of 6 Hz for thevideo data and 40 Hz for the force-† Isokinetic International, 6426 Morning Glory

Dr, Harrison, TN 37341.‡ Grass Instrument Division, Astro-Med Inc,600 East Greenwich Ave, West Warwick, RI02893.

§ Oxford Metrics, 14 Minns Business Park,West Way, Oxford OX2 0JB, United Kingdom.� Bertec Corp, 6171 Huntley Rd, Ste J, Colum-bus, OH 43229

# Motion Lab Systems, 15045 Old HammondHwy, Baton Rouge, LA 70816.



plate data. Sagittal- and frontal-planeknee angles and external knee mo-ments were calculated with Eulerangles and inverse dynamics, respec-tively (Visual 3D**). Data were ana-lyzed using custom-written com-puter programs based on strictcriteria (eg, thresholds for initial con-tact, time of peak adduction mo-ment) to eliminate tester bias. Datawere analyzed during the loading in-terval, which we defined as from 100milliseconds prior to initial contact(to account for electromechanicaldelay)36 through the first peak kneeadduction moment. Data during theloading interval were time normal-ized to 100 data points and averagedacross each subject’s trials. Knee mo-ments were normalized to bodymass � height and are expressed asexternal moments. In addition to dis-crete variables, we calculated kneeflexion excursion (from initial con-tact to peak knee flexion) and kneeadduction excursion during loading.

All EMG data were band-pass filteredfrom 20 to 350 Hz. A linear envelopewas created with full-wave rectifica-tion and filtering with a 20-Hz low-pass filter (eighth-order, phase-corrected Butterworth filter). Thelinear envelope was normalized tothe maximum EMG signal obtainedduring a MVIC for each muscle.

Custom-designed software (Labview,version 8.0††), using the same kine-matic and kinetic events as statedabove, was used to analyze all EMGdata. Magnitude of muscle activityand co-contraction between oppos-ing muscle groups were analyzedover the loading interval after it wastime normalized to 100 points. Mag-nitude of muscle activity was ex-pressed as the average rectified valueacross the loading interval. Co-

contraction was operationally de-fined as the simultaneous activationof a pair of opposing muscles andwas calculated using an equation de-veloped in our laboratory37:

Average co-contraction value �

� �i � 1

nlowerEMGi

higherEMGi�lowerEMGi � higherEMGi��

n

where i is the sample number and nis the total number of samples in theinterval. Co-contraction values wereaveraged across the trials, and theaverage was used for analysis. Thismethod does not identify whichmuscle is more active; rather, it rep-resents a relative activation of 2 mus-cles while accounting for the magni-tudes of both muscles. Co-contraction was calculated betweenthe LQ and LH (LQH), MQ and MH(MQH), LQ and LG (LQG), and MQand MG (MQG) muscles.

Data AnalysisGroup means and standard devia-tions were calculated for all data.One-way analysis of variance(ANOVA) was used to detect groupdifferences in BMI, quadriceps fem-oris muscle strength and CAR, andradiograph variables. Because walk-ing speed can influence lower-extremity kinematic and kineticdata,38–40 analysis of covariance (AN-COVA), with walking velocity as acovariate, was used to detect groupdifferences in kinematic and kineticvariables. For strength, radiograph,and kinematic and kinetic data, sig-nificance was established whenP�.05. To detect group differencesin magnitudes of muscle activity andin co-contraction variables, 95% con-fidence intervals were used to evalu-ate differences in the mean values.

ResultsSubjects in the middle-aged controlgroup had greater BMI values thanthe subjects in the young control

group (P�.030), and subjects in theOA group had higher BMI valuesthan the subjects in the young andolder control groups (P�.002)(Tab. 1). The knees of the subjects inthe middle-aged and older controlgroups were in less varus than theknees of the subjects in the OAgroup (P�.001) (Tab. 1).

Knee LaxitySubjects in the OA group had signif-icantly greater medial laxity than thesubjects in the middle-aged and oldercontrol groups (P�.001) (Fig. 2).There were no differences in laterallaxity between the subjects in theOA group and the subjects in themiddle-aged and older controlgroups (P�.272) (Fig. 2).

Quadriceps FemorisMuscle StrengthThe subjects in the young controlgroup produced greater volitionalquadriceps femoris muscle forcethan the subjects in the older controlgroup (P�.001) or the subjects inthe OA group (P�.001) (Fig. 3). Sub-jects in the middle-aged controlgroup generated more force than thesubjects in the older control group(P�.002) or the subjects in the OAgroup (P�.003) (Fig. 3). There wereno differences between the sub-jects in the older control group andthe subjects in the OA group(P�1.0) or between the subjects inthe young and middle-aged controlgroups (P�.974) (Fig. 3). No dif-ferences in CAR were observedamong the control groups (young�0.93�.038 [X�SD], middle-aged�0.93�.027, older�0.94�.052; P�.84).

Gait CharacteristicsThe subjects in the middle-aged con-trol group walked faster than thesubjects in the OA group (P�.023)and there were no other statisticaldifferences among the groups(young control group�1.39�0.08[X�SD] m/s, middle-aged controlgroup�1.51�0.15 m/s, older con-

** C-Motion Inc, 15821-A Crabbs Branch Way,Rockville, MD 20855.†† National Instruments, 11500 N MopacExpwy, Austin, TX 78759-3504.



trol group�1.45�0.10 m/s, OAgroup�1.38�0.12 m/s).

Results of kinematic and kinetic vari-ables are shown in Table 2. Kneeflexion excursion was not differentamong the young, middle-aged, and

older control groups, but the OAgroup showed less knee flexion ex-cursion compared with all 3 controlgroups (P�.036). The peak kneeflexion moment was no differentamong the subjects in the young,middle-aged, and older control

groups, but was reduced in subjectsin the OA group compared with thesubjects in the young (P�.006) andolder (P�.039) control groups.There were no differences in frontal-plane knee motions or momentsamong the young, middle-aged, andolder control groups. The subjects inthe OA group exhibited greater ad-duction compared with the subjectsin the young, middle-aged, and oldercontrol groups at initial contact(P�.004) and at peak adduction dur-ing loading (P�.001). The OA groupshowed greater adduction excursioncompared with the young controlgroup (P�.049) and the older con-trol group (P�.055); however, thelatter was not statistically significantat the P�.05 level. The OA groupshowed greater peak knee adductionmoments compared with the young,middle-aged, and older controlgroups (P�.002).

Muscle ActivityThere was a large degree of variabil-ity in the muscle activation and co-contraction as is evident in the largerange of the 95% confidence inter-vals shown in Figures 4 and 5. Dur-ing loading response, there was atendency for the subjects in theolder control group to use higherlateral quadriceps femoris activitythan the subjects in the young andmiddle-aged control groups and atendency for higher medial gastroc-nemius muscle activity in the sub-jects in the OA group and the oldercontrol group than in the subjects inthe young and middle-aged controlgroups. However, the overlap of the95% confidence intervals indicatethat a larger sample size is needed todetermine with more certaintywhether the population means aredifferent. In terms of muscle co-contraction, there were no differ-ences among the control groups, al-though the OA group showed higherco-contraction than the subjects inthe young control group in the LQGand MQG muscle pairs (Fig. 5).

Figure 2.Medial and lateral joint laxity. MA�middle-aged control group, O�older control group,OA�group with osteoarthritis. * P�.001. Error bars represent standard deviation.

Figure 3.Quadriceps femoris muscle force production. Y�young control group, MA�middle-aged control group, O�older control group, OA�group with osteoarthritis,N/BMI�highest volitional force during contraction normalized to body mass index.* P�.000, † P�.000, ‡ P�.002, § P�.003. Error bars represent standard deviation.



Discussion and ConclusionsMost studies of age-related differ-ences in movement and muscle acti-vation patterns include samples ofyoung subjects in their 20s and olderadults over 60 years of age; yet, age-related changes in characteristicssuch as muscle strength or neuro-muscular responses can occur inmiddle age7–11 and may coincidewith the development of knee OA.As a result, we intended to investi-gate characteristics in individualswho are healthy that are purportedto be associated with the develop-ment of knee OA across a range ofages, including middle age. Thenovel nature of this approach andthe findings of this study providesome insights into how changes inmusculoskeletal function might es-

tablish an environment in which OAcould develop. The results set thestage for future research into howage-related musculoskeletal changesmight influence the development ofknee OA.

The results of this study indicate thathealthy aging was associated with aconsiderable loss of quadriceps fem-oris muscle strength in the olderadults, although we did not observeincreased frontal-plane laxity inthose subjects. Despite quadricepsfemoris muscle weakness, the olderadults participating in this study didnot adopt a knee stiffening strategy(ie, reduced knee motion and highmuscle co-contraction) that we spec-ulate may contribute to damage ofarticular cartilage. Despite the small

sample size, these findings suggestthat the older adults included in thisstudy demonstrate movement strate-gies similar to those of younger indi-viduals, which may have helpedto prevent the development ofknee OA as they aged; these find-ings, however, warrant furtherinvestigation.

As age-related muscle weakness de-velops, individuals must adapt theirmovements and muscle activity pat-terns to accommodate the dimin-ished force-generating capacity oftheir aging muscles to maintain a cer-tain level of function. As such, wepropose that adaptations allowingfor the continuation of normalizedjoint mechanics and muscle activa-tion patterns are less likely to predis-

Table 2.Mean Values (Adjusted for Walking Speed) and 95% Confidence Interval (in Parentheses) for Sagittal- and Frontal-PlaneKinematics and Kinetics

YoungControl Group(n�15)

Middle-agedControl Group(n�15)

OlderControl Group(n�14)

Osteoarthritis Group(n�15)

P

Kinematics (°)

Sagittal-plane knee angleat initial contact(negative is flexion)

�4.97 (�7.76, �2.18) �3.59 (�6.48, �0.70) �5.56 (�8.41, �2.70) �4.68 (�7.50, �1.85) .800

Knee flexion excursionduring loading

16.75 (14.42, 19.09) 16.97 (14.55, 19.39) 17.94 (15.55, 20.33) 11.98 (9.62, 14.35) �.036a

Frontal-plane knee angleat initial contact(positive is adduction)

�0.434 (�2.5, 1.64) �2.30 (�4.45, �0.15) �0.833 (�2.96, 1.29) 4.83 (2.73, 6.93) �.004a

Peak frontal-plane kneeangle during loading(positive is adduction)

2.62 (0.40, 4.84) 2.35 (0.05, 4.65) 2.18 (�0.10, 4.45) 9.93 (7.68, 12.18) �.001a

Knee adductionexcursion duringloading

3.06b (1.99, 4.12) 4.65 (3.54, 5.76) 3.00c (1.92, 4.10) 5.10b,c (4.02, 6.18) .049b

.055

Kinetics (N�m/kg�m)

Peak knee flexionmoment duringloading

�0.36b (�0.29, �0.44) �0.27 (�0.19, �0.35) �0.33c (�0.25, �0.41) �0.17b,c (�0.09, �0.25) .006b

.039c

Peak knee adductionmoment duringloading

0.28a (0.23, 0.32) 0.33a (0.28, 0.37) 0.26a (0.21, 0.31) 0.45a (0.41, 0.50) �.002a

a Osteoarthritis group different from all control groups.b Osteoarthritis group different than young control group.c Osteoarthritis group different than older control group.



pose the joint to articular cartilagedamage. A failure to adapt tostrength declines might contributeto the development of movementpatterns similar to individuals withquadriceps femoris muscle weak-ness due to knee joint patholo-gy.41–43 Because the older adults inthis study exhibited similar move-ment and muscle activity patterns tothose in the younger age groups, itappears that they have discovered asuccessful approach to maintainingnormal knee function despite theirquadriceps femoris muscle strengthdecline.

In particular, the older control sub-jects exhibited significantly weakerquadriceps femoris muscles com-pared to the younger cohorts, yetthey showed no differences in kneemotion during weight acceptancecompared with the young controlsubjects. Quadriceps femoris muscleweakness has previously been asso-ciated with reduced knee motionduring walking in the presence ofjoint pathology.33,44 Electromyo-graphic data suggest that the olderadults in this study have compen-sated for the quadriceps femorismuscle weakness by selectively in-creasing quadriceps activity duringloading. Although adequate muscleactivity is necessary to ensure jointstability, too much activation can re-sult in limited knee flexion and in-creased impact load on the knee.19

Whether increased activation wouldbe a positive or negative adaptationduring walking, therefore, would de-pend on the end result of the muscleactivity. The older adults in thisstudy were able to maintain normal-ized knee motion, comparable toyounger subjects, with increasedquadriceps femoris activity. The abil-ity to maintain normalized knee jointmechanics may have contributed tothe lack of knee OA in this olderadult cohort. Our conclusions arelimited by the cross-sectional designof this study. A longitudinal study

Figure 4.Mean electromyographic (EMG) muscle activation during loading and 95% confidenceinterval (indicated by bars). MVIC�maximal voluntary isometric contraction,LQ�lateral quadriceps femoris muscle, MQ�medial quadriceps femoris muscle,LH�lateral hamstring muscle, MH�medial hamstring muscle, LG�lateral gastrocne-mius muscle, MG�medial gastrocnemius muscle.

Figure 5.Mean muscle co-contraction index during loading and 95% confidence interval (indi-cated by bars). LQH�lateral quadriceps femoris-lateral hamstring, MQH�medial quad-riceps femoris-medial hamstring, LQG�lateral quadriceps femoris-lateral gastrocne-mius, and MQG�medial quadriceps femoris-medial gastrocnemius muscle pairs.



would be required to further investi-gate the effect of age-related muscu-loskeletal changes on movementstrategies in terms of the develop-ment of knee OA.

Despite prior evidence of reducedstiffness and ligament strength withadvancing age,25 we were unable todetect increases in frontal plane lax-ity with aging in the control subjects.The OA group, however, exhibitedincreased frontal-plane laxity. Al-though subjects were carefullyscreened for a history of ligamentinjury, we included individuals witha history of meniscal damage in theOA group. The subjects with OA hadno history of an incident ligamentousinjury, and studies45,46 suggest thatmeniscal injury in the absence of atraumatic event is a part of the de-generative process of knee OA. Inindividuals with knee OA, the pres-ence of increased frontal plane laxityis known to degrade the relationshipbetween strength and physicalfunction.47

Because the older adults who werehealthy did not have to cope withstrength loss in a lax joint, they mayhave had the ability to adopt move-ment strategies that remain normal-ized and may be “joint sparing.” Wecan speculate that the subjects withOA did not have such an option, be-cause they had to contend withstrength loss in a lax joint, makingjoint stabilization a primary determi-nant in their adopted control strat-egy. These findings suggest thatquadriceps femoris muscle weak-ness is associated with reducedstance-phase knee motion in thepresence of other factors, such asincreased knee laxity or pain, as wasevident in the OA group. Such aspeculation would suggest that age-related changes to musculoskeletaltissues alone are insufficient to leadto the development of knee OA, pro-vided the aging individual has the

means to maintain normal move-ment strategies.

The similarity in the knee motionamong the control groups might beunexpected because other research-ers21 have shown that older adultswalk with less knee motion duringloading when walking at the samespeed as younger subjects. It is pos-sible that our finding of similarsagittal-plane knee kinematicsamong the young, middle-aged, andolder control groups is due to a smallnumber of subjects in our sample orrelated to our method of subject re-cruitment. Some of the older adultsin our study were recruited from lo-cal fitness and senior centers andmay represent a more active olderadult compared with a typical olderadult, and this may have enabled theolder subjects in our study to bettercontrol more knee motion as thelimb accepted weight. Future studiesmay consider measuring daily activ-ity levels to account for potential in-fluences on walking speeds andmovement patterns.

It is interesting to note that differ-ences in muscle co-contraction val-ues were found only between thesubjects with OA and young controlsubjects. In this study, the subjectswith OA used greater lateral gas-trocnemius muscle activity duringloading and greater quadricepsfemoris-gastrocnemius muscle co-contraction on the medial and lateralsides compared with the young con-trol subjects, but no differenceswere observed between the subjectswith OA and the middle-aged controlsubjects. This is in contrast to otherwork in our lab in which subjectswith OA were found to use higherco-contraction between the quadri-ceps femoris-gastrocnemius muscleson the medial side only comparedwith age-matched control subjects.24

It is possible that people with patho-logic conditions in the knee maylimit knee motion through different

muscle co-contraction strategies.41,43

Additional research is required to de-lineate whether consistent differ-ences in muscle activation patternsexist in people with knee OA orwhether there are several strategiesthat people use to help control theknee in the face of a pathologic con-dition. Whether one strategy is moredetrimental than another remains tobe seen and should be furtherinvestigated.

There are several limitations to thestudy design that the readers shouldtake into consideration. First, testerswere not blinded to group assign-ment, which may have created biasin recording data. However, the useof the TELOS device to apply uni-form stress during the stress radio-graphs reduced the influence oftester bias for this measure. Testersattempted to provide equivalent ver-bal encouragement to all subjectsequally when testing quadricepsfemoris muscle strength. In addition,the discomfort of the superimposedburst was motivation for all subjectsto perform to their best ability toavoid repeat testing. During themovement analysis testing, similarinstructions were provided to allsubjects to walk at a comfortablespeed. Custom-written computer al-gorithms were used to determinedata points used in the analysis ofkinematics, kinetics, and EMG datato reduce tester bias. Second, thedistribution of male and female sub-jects in the groups was not the sameand we did not account for the levelof physical activity in the subjects ineach group, both of which couldhave influenced the results. Finally,the subjects all walked faster thanhas been reported elsewhere,48 andwalking speed—although used as acovariate—may have influenced theresults.

The finding that the older adults inour study used what can be consid-ered a favorable movement pattern



may suggest why they did not de-velop knee OA. We speculate thatthe manner in which middle-aged in-dividuals compensate for age-relatedneuromuscular changes might influ-ence the future integrity of theknee’s articular cartilage. An alterna-tive interpretation of our results isthat the process of knee OA maycause changes in movement andmuscle activation patterns. The ab-sence of reduced knee motion andhigher co-contraction in the middle-aged and older control subjects mayhave been a consequence of not hav-ing developed OA rather than thereason they did not develop OA. Res-olution of such a question would en-tail a large-scale longitudinal study totrack changes in movement and mus-cle activation patterns along with ar-thritic changes in large numbers ofsubjects.

Dr Rudolph and Dr Schmitt provided writingand project management. Dr Schmitt and DrLewek provided data collection. All authorsprovided data analysis. Dr Rudolph providedfund procurement, facilities/equipment, andinstitutional liaisons. Dr Lewek provided con-sultation (including review of manuscript be-fore submission). The authors acknowledgeWilliam Newcomb, MD, for providing sub-ject recruitment and Laurie Andrews, RTR,for assistance with radiographs.

This study was approved by the HumanSubjects Review Board of the Universityof Delaware.

This study was funded, inpart, by a grant (P20-RR16458) from the Na-tional Center for Research

Resources (KSR), a component of the Na-tional Institutes of Health (NIH); a grant(T32HD007490) from the NIH (LCS, MDL); aPODS II grant from the Foundation for Phys-ical Therapy (LCS, MDL); and a Doctoral Re-search Grant from the American College ofSports Medicine (MDL).

This work was presented as a platform pre-sentation at the Combined Sections Meetingof the American Physical Therapy Associa-tion; February 14, 2003; Tampa, Fla; as aplatform presentation at XVth Biannual In-ternational Society of Electrophysiology andKinesiology Conference; June 20, 2004; Bos-ton, Mass; and at the Functional Fitness and

the Science of Joint Preservation Course, Uni-versity of Louisville, Louisville, Ky, April 28–29, 2006.

The contents of this article are solely theresponsibility of the authors and do not nec-essarily represent the official views of theNCRR or NIH.

This article was submitted May 12, 2006, andwas accepted July 11, 2007.

DOI: 10.2522/ptj.20060137

References1 Carmona L, Ballina J, Gabriel R, Laffon A;

EPISER Study Group. The burden ofmusculoskeletal diseases in the generalpopulation of Spain: results from a na-tional survey. Ann Rheum Dis. 2001;60:1040–1045.

2 Felson DT, Zhang Y, Hannan MT, et al.Risk factors for incident radiographic kneeosteoarthritis in the elderly: the Framing-ham study. Arthritis Rheum. 1997;40:728–733.

3 Kannus P, Jarvinen M, Kontiala H, et al.Occurrence of symptomatic knee osteo-arthrosis in rural Finland: a prospectivefollow up study. Ann Rheum Dis.1987;46:804–808.

4 McAlindon TE, Wilson PW, Aliabadi P, etal. Level of physical activity and the risk ofradiographic and symptomatic knee osteo-arthritis in the elderly: the Framinghamstudy. Am J Med. 1999;106:151–157.

5 Peat G, McCarney R, Croft P. Knee painand osteoarthritis in older adults: a reviewof community burden and current use ofprimary health care. Ann Rheum Dis.2001;60:91–97.

6 Loeser RF, Shakoor N. Aging or osteoar-thritis: which is the problem? Rheum DisClin North Am. 2003;29:653–673.

7 Akima H, Kano Y, Enomoto Y, et al. Mus-cle function in 164 men and women aged20–84 yr. Med Sci Sports Exerc. 2001;33:220–226.

8 Frontera WR, Hughes VA, Lutz KJ, EvansWJ. A cross-sectional study of musclestrength and mass in 45- to 78-yr-old menand women. J Appl Physiol. 1991;71:644–650.

9 Lindle RS, Metter EJ, Lynch NA, et al. Ageand gender comparisons of musclestrength in 654 women and men aged20–93 yr. J Appl Physiol. 1997;83:1581–1587.

10 Roos MR, Rice CL, Connelly DM, Vander-voort AA. Quadriceps muscle strength,contractile properties, and motor unit fir-ing rates in young and old men. MuscleNerve. 1999;22:1094–1103.

11 Hurley MV, Rees J, Newham DJ. Quadri-ceps function, proprioceptive acuity andfunctional performance in healthy young,middle-aged and elderly subjects. Age Age-ing. 1998;27:55–62.

12 Felson DT, Zhang Y. An update on theepidemiology of knee and hip osteoarthri-tis with a view to prevention. ArthritisRheum. 1998;41:1343–1355.

13 Slemenda C, Brandt KD, Heilman DK, et al.Quadriceps weakness and osteoarthritis ofthe knee. Ann Intern Med. 1997;127:97–104.

14 Brandt KD, Heilman DK, Slemenda C, et al.Quadriceps strength in women with radio-graphically progressive osteoarthritis ofthe knee and those with stable radio-graphic changes. J Rheumatol. 1999;26:2431–2437.

15 Hurley MV. The role of muscle weaknessin the pathogenesis of osteoarthritis.Rheum Dis Clin North Am. 1999;25:283–298, vi.

16 Bennell KL, Hinman RS, Metcalf BR. Asso-ciation of sensorimotor function withknee joint kinematics during locomotionin knee osteoarthritis. Am J Phys Med Re-habil. 2004;83:455–463; quiz 464–456,491.

17 Perry J. Gait Analysis: Normal and Patho-logical Function. Thorofare, NJ: Slack;1992.

18 Hodge WA, Fijan RS, Carlson KL, et al.Contact pressures in the human hip jointmeasured in vivo. Proc Natl Acad SciU S A. 1986;83:2879–2883.

19 Lafortune MA, Hennig EM, Lake MJ. Dom-inant role of interface over knee angle forcushioning impact loading and regulatinginitial leg stiffness. J Biomech. 1996;29:1523–1529.

20 Cook TM, Farrell KP, Carey IA, et al. Ef-fects of restricted knee flexion and walk-ing speed on the vertical ground reactionforce during gait. J Orthop Sports PhysTher. 1997;25:236–244.

21 DeVita P, Hortobagyi T. Age causes a re-distribution of joint torques and powersduring gait. J Appl Physiol. 2000;88:1804–1811.

22 Sharma L, Lou C, Felson DT, et al. Laxity inhealthy and osteoarthritic knees. ArthritisRheum. 1999;42:861–870.

23 Wada M, Imura S, Baba H, Shimada S. Kneelaxity in patients with osteoarthritis andrheumatoid arthritis. Br J Rheumatol.1996;35:560–563.

24 Lewek MD, Rudolph KS, Snyder-MacklerL. Control of frontal plane knee laxity dur-ing gait in patients with medial compart-ment knee osteoarthritis. OsteoarthritisCartilage. 2004;12:745–751.

25 Noyes FR, Grood ES. The strength of theanterior cruciate ligament in humans andRhesus monkeys. J Bone Joint Surg Am.1976;58:1074–1082.

26 Woo SL, Hollis JM, Adams DJ, et al. Tensileproperties of the human femur-anteriorcruciate ligament-tibia complex: the ef-fects of specimen age and orientation.Am J Sports Med. 1991;19:217–225.

27 Altman R, Asch E, Bloch D, et al. Develop-ment of criteria for the classification andreporting of osteoarthritis: classification ofosteoarthritis of the knee. Diagnostic andTherapeutic Criteria Committee of theAmerican Rheumatism Association. Ar-thritis Rheum. 1986;29:1039–1049.



28 Moore TM, Meyers MH, Harvey JP Jr. Col-lateral ligament laxity of the knee: long-term comparison between plateau frac-tures and normal. J Bone Joint Surg Am.1976;58:594–598.

29 Chao EY, Neluheni EV, Hsu RW, Paley D.Biomechanics of malalignment. OrthopClin North Am. 1994;25:379–386.

30 Hsu RW, Himeno S, Coventry MB, ChaoEY. Normal axial alignment of the lowerextremity and load-bearing distribution atthe knee. Clin Orthop Relat Res. 1990;225:215–227.

31 Tetsworth K, Paley D. Malalignment anddegenerative arthropathy. Orthop ClinNorth Am. 1994;25:367–377.

32 Kent-Braun JA, Le Blanc R. Quantitation ofcentral activation failure during maximalvoluntary contractions in humans. MuscleNerve. 1996;19:861–869.

33 Snyder-Mackler L, Binder-Macleod SA, Wil-liams PR. Fatigability of human quadricepsfemoris muscle following anterior cruciateligament reconstruction. Med Sci SportsExerc. 1993;25:783–789.

34 Reinschmidt C. Three-Dimensional Tibio-calcaneal and Tibiofemoral KinematicsDuring Human Locomotion � MeasuredWith External and Bone Markers [doc-toral dissertation]. Calgary, Alberta, Can-ada: University of Calgary; 1996.

35 Benoit DL, Ramsey DK, Lamontagne M, etal. Effect of skin movement artifact onknee kinematics during gait and cuttingmotions measured in vivo. Gait Posture.2006;24:152–164.

36 Vos EJ, Harlaar J, van Ingen Schenau GJ.Electromechanical delay during knee ex-tensor contractions. Med Sci Sports Exerc.1991;23:1187–1193.

37 Rudolph KS, Axe MJ, Snyder-Mackler L.Dynamic stability after ACL injury: whocan hop? Knee Surg Sports Traumatol Ar-throsc. 2000;8:262–269.

38 Andriacchi TP, Ogle JA, Galante JO. Walk-ing speed as a basis for normal and abnor-mal gait measurements. J Biomech. 1977;10:261–268.

39 Crowinshield RD, Brand RA, Johnston RC.The effects of walking velocity and age onhip kinematics and kinetics. Clin OrthopRelat Res. 1978;132:140–144.

40 Kirtley C, Whittle MW, Jefferson RJ. Influ-ence of walking speed on gait parameters.J Biomed Eng. 1985;7:282–288.

41 Chmielewski TL, Rudolph KS, FitzgeraldGK, et al. Biomechanical evidence sup-porting a differential response to acuteACL injury. Clin Biomech (Bristol, Avon).2001;16:586–591.

42 Mundermann A, Dyrby CO, Andriacchi TP.Secondary gait changes in patients withmedial compartment knee osteoarthritis:increased load at the ankle, knee, and hipduring walking. Arthritis Rheum. 2005;52:2835–2844.

43 Rudolph KS, Axe MJ, Buchanan TS, et al.Dynamic stability in the anterior cruciateligament deficient knee. Knee Surg SportsTraumatol Arthrosc. 2001;9:62–71.

44 Lewek M, Rudolph K, Axe M, Snyder-Mackler L. The effect of insufficient quad-riceps strength on gait after anterior cru-ciate ligament reconstruction. ClinBiomech (Bristol, Avon). 2002;17:56–63.

45 Englund M. Meniscal tear: a feature of os-teoarthritis. Acta Orthop Scand Suppl.2004;75:1–45, backcover.

46 Englund M, Roos EM, Lohmander LS. Im-pact of type of meniscal tear on radio-graphic and symptomatic knee osteo-arthritis: a sixteen-year follow-up of menis-cectomy with matched controls. ArthritisRheum. 2003;48:2178–2187.

47 Sharma L, Hayes KW, Felson DT, et al.Does laxity alter the relationship betweenstrength and physical function in kneeosteoarthritis? Arthritis Rheum. 1999;42:25–32.

48 Lopopolo RB, Greco M, Sullivan D, et al.Effect of therapeutic exercise on gaitspeed in community-dwelling elderly peo-ple: a meta-analysis. Phys Ther. 2006;86:520–540.



Age related change in strength, joint laxity, and walking patterns. are they related to knee oa

Health & Medicine

Transcript of Age related change in strength, joint laxity, and walking patterns. are they related to knee oa