Poor back muscle endurance is related to - Irsst

Pain catastrophizing and back muscle endurance 1

Title: Poor back muscle endurance is related to pain catastrophizing in patients with chronic low back pain. Authors and affiliations:

Christian Larivière, Ph.D. 1,4, Martin Bilodeau, PT, Ph.D.2, Robert Forget, PT, Ph.D. 3,4, Roger Vadeboncoeur, MD 4, Hakim Mecheri, Eng., M.Sc. 4

1 Occupational Health and Safety Research Institute Robert-Sauvé, Montreal, Quebec, Canada, H3A 3C2

2 Élisabeth Bruyère Research Institute and School of Rehabilitation Sciences, University of Ottawa, 451 Smyth, Ottawa (Ontario), Canada, K1H 8M5

3 School of Rehabilitation, Université de Montreal, C.P. 6128, Succursale Centre-Ville, Montreal, Quebec, Canada, H3C 3J7

4 CRIR - Montreal Rehabilitation Institute, Montreal, Canada Address for correspondence and reprints:

Christian Larivière, Ph.D. Institut de recherche Robert-Sauvé en santé et en sécurité du travail (IRSST) 505, boul. De Maisonneuve Ouest Montréal (Québec) H3A 3C2 Phone: (514) 288-1551 #217 Fax: (514) 288-6097 E-mail: [email protected]

Acknowledgements:

The present research project was funded by the Occupational Health and Safety Research Institute Robert-Sauvé (IRSST) of Quebec (Canada) and the Quebec Rehabilitation Research Network (REPAR). Special thanks to Flavia Dell’Oso and Marielle Trottier for data collection and to David McFadden and Jean-François Pilon for engineering assistance.

Disclosure statement: This work has not been funded by any of the following organizations: National Institutes of Health (NIH), Wellcome Trust; Howard Hughes Medical Institute (HHMI).


ABSTRACT Study Design. An experimental and comparative study of chronic low back pain (CLBP) patients and healthy controls. Objective. To use a motivation-independent electromyographic(EMG)-based test of back muscle capacity to determine whether back muscle deconditioning is present in CLBP patients and whether it is related to pain-related psychological variables. Summary of Background Data. The verification of the deconditioning syndrome in CLBP patients might be biased by the use of performance-based measures to assess physical fitness, especially in patients having fear of injury. Also, the use of lumbar-specific measures of physical fitness, such as back muscle strength and endurance, might be more sensitive to physical deconditioning than more general assessments such as aerobic capacity. Methods. A time-limited submaximal fatigue test was performed by 27 non-specific CLBP subjects (14 men) who had not had any surgery, and 31 healthy controls (17 men) while surface EMG signals were collected from back muscles. Motivation-independent EMG indices, that are sensitive to muscle fatigue or to activation patterns, were then computed and entered as input into previously developed regression equations to predict endurance (PTend) and Strength (PStrength). Between-group comparisons were completed with patients divided in subgroups based on a median split of pain intensity, fear of movement, or pain catastrophizing scores. Results. Differences between healthy and CLBP subgroups were mainly observed when patients were divided using pain catastrophizing scores (PCS). High-PCS patients showed significantly lower PTend than low-PCS patients. Various EMG indices showed comparable results to PTend. However, some of them also pointed out that the PCS-low patients were more fatigue-resistant and showed different activation patterns comparatively to healthy subjects. Conclusions. These results suggest that physical deconditioning that is specific to back muscle capacity was present in a subgroup of patients while the opposite was observed in another subgroup, pain catastrophizing being related to this outcome. These findings support previous theoretical models of pain/disability. Key Words: (1) strength, (2) fatigue, (3) physical deconditioning, (4) surface electromyography, (5) objective measures, (6) fear of movement, (7) lumbar impairment Key Points:

1. Predicted back muscle endurance was lower in CLBP patients showing high catastrophizing thoughts than in patients showing low catastrophizing

2. Various EMG variables further supported this finding 3. Some EMG variables also pointed out that the PCS-low patients had higher back muscle

endurance and different activation patterns than healthy subjects. 4. These findings have physiological meanings and can support previous theoretical models

of pain/disability for CLBP patients.


Mini Abstract/Précis (max 50 words ; presently 53 words)

A time-limited submaximal fatigue test was performed by back pain subjects and healthy controls while back muscles EMG signals were collected. Physical deconditioning that is specific to back muscle capacity was present in a subgroup of patients while the opposite was observed in another subgroup, pain catastrophizing being related to these findings.

INTRODUCTION

Physical deconditioning, the end-result of the so-called deconditioning syndrome, has been suspected to result from chronic low back pain (CLBP) as well as being a cause of perpetuating pain 1. This phenomenon would result from physical inactivity and is expected to be substantiated through the different domains of physical fitness including: aerobic capacity (VO2max), different muscle characteristics (strength, endurance, power, fiber-type composition and flexibility), motor control and body composition. However, to date, studies that attempted to verify these hypotheses are somewhat conflicting 2-8. This heterogeneity of findings can be partly explained by three factors: (1) the characteristic of CLBP subjects recruited in the study, (2) the choice of the domain of physical fitness assessed and (3) the degree of objectivity of selected outcome measures.

Individuals with CLBP are heterogeneous. Consequently, subgrouping CLBP subjects according to variables that are expected to influence the presence of physical deconditioning would help to highlight the deconditioning syndrome. According to the fear-avoidance model 9, pain-related psychological characteristics such as pain catastrophizing, fear of movement/(re)injury (or kinesiophobia) or fear-avoidance beliefs would represent credible factors. In line with this reasoning, low levels of physical activity have been associated with high levels of fear-avoidance beliefs and pain catastrophizing 10. Although the role of these potential factors has not been supported to explain the lower aerobic capacity of CLBP subjects 4;11, this might be different for other outcomes of physical fitness.

In this specific literature, while aerobic capacity has been favored as the domain of physical fitness that should be more related to physical activity, not all types of physical activity enhance cardiovascular fitness. In fact, many work or leisure activities only challenge the cardio-vascular system to a minor extent. Consequently, aerobic capacity might not be sensitive enough to detect differences in populations for whom physical activity is primarily performed at work and during leisure activities. We propose that measures such as back muscle strength and endurance, that evaluate more specifically the lumbar area, might be more relevant if CLBP subjects actually try to limit more specifically the use of their back.

Common measures of back muscle strength or relative endurance rely on the maximal performance of the patient. Unfortunately, the fact that some CLBP subjects are afraid to perform maximally during such assessments 12-14 represents a serious treat to their validity. Even electromyography (EMG)-based measures of back muscle fatigue are likely biased when the


submaximal load is determined as a proportion of the patient’s maximal strength 6;15;16. Testing the physical-deconditioning hypothesis under the fear-avoidance model requires to take into consideration that subjects who would present with lower physical activity and physical deconditioning, namely those that have fear of movement/(re)injury and catastrophizing thoughts, may also under-perform during assessments of physical fitness, which would affect the validity of the corresponding outcome measures. To bypass the detrimental effect of these pain-related psychological characteristics on the outcome measures, the superimposed electrical stimulation technique has been applied during the strength assessment of quadriceps 14. This allowed to elegantly substantiated the presence of physical deconditioning that is not necessarily specific to the lumbar area 14. Still, physical deconditioning could possibly be more apparent in back muscles but unfortunately, this technique is not realistically applicable to this muscle group.

More recently, our group has developed an EMG-based assessment of back muscle capacity with the use of surface EMG of eight back muscles, namely the functional endurance test (FET). The test is time-limited (5 and 10 min in women and men, respectively) and the submaximal load is absolute (90-Nm L5/S1 extension moment) rather than relative to strength (further details in the methods section), thus not requiring measuring a given patient/individual’s strength. So far, our work on healthy adults shows that the construct validity of the task (moderate intermittent contractions) is substantiated by showing that both strength and relative endurance have a complementary impact on the performance of the test, showing that this test measures back muscle capacity as a whole 17. This task also allowed to show the expected lower back muscle endurance of men relative to women, when the 90-Nm absolute load was translated to a relative load (40% of strength), which added to its construct validity 18. EMG indices of back muscles neuromuscular activation patterns 19 and fatigue 20 have been proposed and selected (criterion validity study 19;20) based on their significant correlations with strength (r reaching up to -0.72 to -0.81) and absolute endurance (r = -0.33 to 0.69) measures. Also, the day-to-day reliability of these EMG indices ranged from moderate to excellent (intraclass correlation coefficients ranging from 0.40 to 0.83) 19;20. Finally, gender specific predictive equations of back muscle strength (men: R2 = 0.72; Error = 9 %; women: R2 = 0.25, Error = 13 %) and absolute endurance (men: R2 = 0.76, Error = 9 %; women: R2 = 0.70, Error = 17 %) have been derived using body composition (body mass index, fat free mass or percent fat) and EMG indices as input variables 21. These predictions represent truly objective estimations of back muscle capacity.

The purpose of the present study was to determine whether pain-related psychological characteristics such as pain catastrophizing and kinesiophobia are related to physical deconditioning in CLBP subjects, relative to controls, as assessed with back muscle strength and endurance. It was hypothesized that only CLBP subjects showing more fear of movement and/or more pain catastrophizing would have lower back muscle strength and decreased endurance compared with control subjects.


METHODS

Subjects

Thirty-two individuals with CLBP (15 women) and 32 healthy controls (17 women) aged between 18 to 55 and having a body mass index (BMI) less than 31.5 kg/m2 (women) or 33 kg/m2 (men), were recruited. However, two subjects (1 healthy women, 1 men with CLBP) were excluded due to technical problems during the data collection and four women with CLBP were excluded because unable to reach the 90-Nm plateau during the FET. The remaining subjects are described in (Table 1). Inclusion criteria for the CLBP subjects were: lumbar or lumbosacral pain with or without proximal radicular pain (limited distally to the knees); and presence of chronic pain defined as a daily or almost daily pain for at least three months. Exclusion criteria for the healthy subjects were: back pain in the preceding year or exceeding one week; surgery of the pelvis or spinal column; pregnancy. Exclusion criteria for the CLBP subjects were: surgery of the pelvis or spinal column; scoliosis; systemic or degenerative disease; one positive response to the Physical Activity Readiness Questionnaire (Thomas et al. 1992). Ten out of the 27 CLBP subjects were not at work at the time of testing (7 because of the back pain problem). The ethics committee of the Centre for Interdisciplinary Research in Rehabilitation of Greater Montreal (CRIR) approved the study and consent form.

Questionnaires

Three pain-related psychological variables were assessed in CLBP subjects: (1) pain intensity with a 10-cm visual analog scale (VAS), (2) fear of movement or injury with the Tampa scale of kinesiophobia (TSK) 22 and (3) pain catastrophizing with the Pain catastrophizing scale (PCS) 23.

Tasks

The MVCs and FET were carried out with the use of a static triaxial trunk dynamometer (Figure 1B). Firstly, two to four submaximal trunk extension contractions were performed to warm-up and to familiarize with the apparatus and the visual feedback. Then, three MVCs that consisted of 7-s progressive ramp extension efforts were performed with verbal encouragement and with the use of a visual feedback (Figure 1C). The peak MVC represented the Strength criterion.

The FET consists of repeating an 8-s duration cycle comprising a 90-Nm (L5/S1 torque) static extension exertion (1.5 s ramp to reach 90 Nm + 5 s plateau at 90 Nm + 1.5 s rest; Figure 1D), again with the use of the visual feedback system (Figure 1C). Each 8-s cycle was repeated until exhaustion in healthy subjects from whom obtaining the mechanical criterion of absolute endurance (Tend = 8 s × number of cycles) was possible with the use of verbal encouragement. In CLBP subjects, the FET was time-limited to 5 min. (38 cycles) in females and 10 min. (75 cycles) in males.


Table 1. Demographic, anthropometric and clinical characteristics of the participants

Variable * Healthy controls CLBP patients P values *

Men (n = 17)

Women (n = 14)

Men (n = 14)

Women (n = 13) Pain

Status Gender

Age (yrs) 39 (10) 39 (10) 43 (10) 35 (9) 0.714 0.089

Height (m) 1.74 (0.06) 1.65 (0.09) 1.74 (0.06) 1.63 (0.09) 0.698 0.000

Mass (kg) 78 (8) 63 (7) 78 (13) 70 (14) 0.254 0.000

BMI (kg/m2) 26 (3) 23 (3) 26 (4) 26 (5) 0.128 0.247

% fat 18 (5) 24 (6) 19 (6) 27 (6) 0.325 0.000

FFM (kg) 63 (6) 48 (6) 63 (10) 50 (7) 0.471 0.000

VAS pain (cm) / / 3.6 (2.1) 3.1 (2.8) / 0.561

RDQ (%) / / 31 (19) 23 (24) / 0.577

TSK (score/68) / / 48 (8) 37 (10) / 0.004

PCS (score/52) / / 26 (10) 15 (13) / 0.017

* Probability values corresponding to the two-way ANOVAs used to assess the differences between healthy and CLBP subjects (PAIN-STATUS factor) and between sexes. T-tests were performed for the clinical variables (VAS, RDQ, TSK, PCS). BMI: Body Mass Index; % fat: percentage of body fat using skinfold measures 40; FFM: fat free mass; VAS: Visual Analog Scale of pain intensity ; RDQ: Rolland-Morris Disability Questionnaire; TSK: Tampa Scale of Kinesiophobia; PCS: Pain Catastrophizing Scale.

Figure 1. Experimental setup (details in the methods – measurement techniques). (A) Position of the surface electrodes at the different vertebral levels, over the multifidus at the L5 level, iliocostalis lumborum at L3 and longissimus at L1 and T10. The difficulty in capturing the multifidus muscle with surface electrodes 41 is acknowledged and therefore the validity of the electromyographic signal was assigned to the landmarked location rather than to the multifidus muscle itself; (B) The triaxial dynamometer; (C) Visual feedback displayed to the subject. For the purpose of this picture, the subject, represented by a small square, is positioned outside of the target; (D) Description of one work-rest cycle (8 s) of the functional endurance test (only 2 cycles are represented).


Surface electromyography and data processing

The EMG signals from four pairs of back muscles were collected (bandpass filter: 20-450 Hz; preamplification gain: 1000; sampling rate: 1024 Hz) with active surface electrodes (Delsys Inc., MA), as detailed elsewhere 24 and depicted in Figure 1A.

Briefly, from the EMG signals corresponding to the first 38 cycles (5-min. test) in women and to the first 75 cycles (10-min. test) in men, three types of EMG parameters were computed for each electrode site for each 5-s plateau: (1) the instantaneous median frequency (IMNF), (2) the signal intensity (power) contained in eight frequency bands (WINT3 to WINT10 covering the 27-364 Hz bandwidth) as calculated with the use of a wavelet filter bank, and (3) root mean square (RMS) amplitude values. Hereafter, the more generic term WINTn will be used to represent WINT3 to WINT10, n representing the wavelet number that ranged from 3 to 10. Then, EMG indices sensitive to muscle fatigue and to variable load sharing between back muscles synergists were calculated. These EMG indices are briefly described below but the reader is referred to previous papers 19;20 for a more thorough description.

EMG indices of fatigue are simply the slope values of linear regressions applied to IMNF and WINTn time-series (NIMNFslp and NWINTnslp), all being sensitive to the progressive compression of the EMG frequency spectrum to lower frequencies, which is mainly attributed to the decrease in motor unit action potential conduction velocity reflecting the accumulation of metabolic products on the surface area of the recruited muscle fibers 25. Lower NIMNFslp values indicate more fatigue. NWINTnslp is computed for each wavelet (n = 3, 4, 8, 9, 10), higher NWINT3slp and NWINT4slp values (representing lower frequencies of the power spectrum) and lower NWINT8slp, NWINT9slp and NWINT10slp values (representing higher frequencies of the power spectrum) indicating more fatigue. Three types of EMG load-sharing indices were computed from the RMS time-series, as detailed elsewhere 18;19. EMGVAR basically reflects the modulation in EMG level across samples and is assumed to be sensitive to load sharing between all synergist muscles. EMGALT quantifies alternating EMG activity between two muscles with the consideration of the proportion of total time (PALT) and amplitude of alternating activation. PALT was retained as the third index. Although all EMG indices were initially computed for the different muscles (electrode sites) individually, data from certain muscles/sites were combined to reduce the number of variables and increase their reliability. More specifically, NIMNFslp comprises four indices (X-L5, X-L1, X4, M4), the letters “X” and “M” representing the mean and maximum across the electrode sites considered (n = 4 in the present case: L5 and L1 bilaterally). EMGVAR and EMGALT and PALT have indices calculated from six electrode sites (L5, L3 et L1 bilaterally). For example, PALT – L5L1 is the percentage of alternating activity between electrode sites L5 and L1.

Finally, gender-specific multiple regression equations 21 were applied, using the above-described EMG indices as input, to predict Tend (PTend) and Strength (PStrength) for all healthy and CLBP subjects. In the remaining of the paper, Strength, Tend, PStrength, PTend will be called “mechanical variables” while purely EMG-based variables will be called “EMG variables”


Statistical analyses

For each sex separately, the CLBP subjects were first subdivided in two subgroups according to the score of the three questionnaires, using the median as thresholds values (for men and women respectively, VAS : 2.9 and 1.9 cm, TSK : 48 and 37/68, PCS : 27 and 14/52). The use of different threshold values between sexes was mandatory because the FET involves a 90-Nm load that represents a higher relative load in women (relative to their strength), which in turn speed up neuromuscular fatigue relative to men. Consequently, subgroups not matched for sex would have biased between-group comparisons.

Some of the variables were not normally distributed and thus were transformed before any statistical analyses (ANOVAs, ANCOVAs, Pearson correlations) were carried out. All analyses were done with NCSS statistical software (version 6.0 for Windows), using a significance level (alpha) of 0.05. Preliminary correlation analyses were carried out to investigate whether potential confounding variables (age, percent fat) would be associated with the mechanical and EMG variables. Only EMGVAR and age were significantly correlated (r = -0.31 to -0.42; P = 0.031 to 0.003 across the different muscles), which was accounted for using ANCOVAs, as described next. For each of the mechanical variable (Strength, PStrength, PTend), two-way ANOVAs (3 PAIN-STATUS × 2 SEX) were carried out to assess differences between the three above-defined groups (PAIN-STATUS factor). The SEX effect (but not the PAIN-STATUS×SEX interaction) was of interest here to substantiate the construct validity of the FET regression equations used to calculate PStrength and PTend. Effectively, the FET is expected leading to higher PStrength and PTend scores in men comparatively to women. For each EMG variable (NIMNFslp, NWINTnslp, EMGALT and PALT) except EMGVAR,, one-way ANOVAs were first carried out to test the PAIN-STATUS factor. For EMGVAR, ANCOVAs were used to adjust for age (used as covariate), which was significant (P<0.05) in all cases. Finally, again for each EMG variable, all the PAIN-STATUS comparisons were repeated (using ANCOVAs) but adjusting for Strength, and then for fat free mass as covariates, as further explained in the next section.

RESULTS

Healthy subjects were equal to both subgroups of CLBP patients (regardless of the subgrouping variable VAS, TSK or PCS), as regards to age, mass, height, percent fat and BMI.

Cross-validation of PTend and PStrength estimations

Direct comparisons between measured and predicted mechanical variables in healthy subjects (n = 31), for whom maximal performances can be trusted, revealed no significant difference (paired T-test) between Strength and PStrength (P = 0.146) but Tend was significantly higher (P = 0.026) than PTend. On the other hand, the corresponding correlations were not significant for strength (r = 0.06; P = 0.740) but significant for endurance (r = 0.71; P = 0.000).


Mechanical variables

Both Strength and PStrength showed that men were stronger (P<0.05) than women and that the different CLBP subgroups were as strong as healthy controls and this, regardless of how CLBP patients were divided (Table 2).

Table 2. Statistical results (ANOVA) corresponding to CLBP status (healthy vs CLBP subgroups) and sex difference for performance outcomes directly measured (Strength et Tend) or estimated through regression models (PStrength et PTend)

Variable * Healthy (1) CLBP (2)

Subgroup 1

CLBP (3)

Subgroup 2

P values (ANOVA) Post-hoc Tests

Men

(n = 17)

Women

(n = 14)

Men

(n = 6-7)

Women

(n = 6-7)

Men

(n = 6-7)

Women

(n = 6-7)

STATUS

(S)

SEX

(SX) S × SX

†

Subgrouping according to VAS scores

Strength 301 (60) 198 (64) 256 (87) 187 (70) 245 (67) 184 (36) 0.309 0.000 0.663 ♂>♀

PStrength 251 (127) 205 (14) 367 (40) 220 (28) 300 (106) 196 (16) 0.129 0.000 0.280 ♂>♀

Tend 16.1 (10.5) 7.6 (4.8) / / / / / 0.003 / ♂>♀

PTend 12.3 (8.1) 6.8 (4.3) 15.5 (5.2) 12.9 (11.6) 6.5 (6.4) 6.1 (4.9) 0.059 0.011 0.991 ♂>♀

Subgrouping according to TSK scores

Strength 301 (60) 198 (64) 231 (74) 224 (37) 271 (76) 152 (36) 0.180 0.000 0.009 ♂>♀

PStrength 251 (127) 205 (14) 336 (115) 206 (23) 325 (45) 207 (27) 0.398 0.001 0.326 ♂>♀

Tend 16.1 (10.5) 7.6 (4.8) / / / / / 0.003 / ♂>♀

PTend 12.3 (8.1) 6.8 (4.3) 11.9 (6.8) 10.6 (11.5) 12.8 (6.6) 8.1 (6.9) 0.903 0.013 0.920 ♂>♀

Subgrouping according to PCS scores

Strength 301 (60) 198 (64) 265 (78) 202 (59) 236 (74) 171 (43) 0.157 0.000 0.638 ♂>♀

PStrength 251 (127) 205 (14) 339 (101) 228 (19) 320 (74) 188 (6) 0.174 0.001 0.262 ♂>♀

Tend 16.1 (10.5) 7.6 (4.8) / / / / / 0.003 / ♂>♀

PTend ‡ 12.3 (8.1) 6.8 (4.3) 15.7 (6.8) 15.2 (10.1) 8.4 (3.0) 4.1 (3.0) 0.002 0.012 0.261 ♂>♀; 2>3

Statistically significant differences (P<0.05) are identified with bold characters while trends (0.05<P<0.10) are underlined. † Results from pairwise comparisons (Tukey-Kramer) corresponding to significant STATUS and/or SEX main effects. Comparisons corresponding to the STATUS effect are identified with numbers (ex. : 1 < 3 means that group 1 (healthy subjects) < group 3 (CLBP subjects – subgroup 2). Sex effects are identified with the appropriate symbols (♂ : Men; ♀ : Women). ‡ The corresponding results are illustrated in Figure 2.

In healthy subjects, for whom Tend was measured, men showed higher values than women, as expected. The same results were observed for PTend. For PTend, a STATUS effect was observed only when patients were divided using PCS scores (Table 2), where PCS-high patients showed significantly lower (P < 0.05) PTend than PCS-low patients (Figure 2). These results remained even when adjusting for pain intensity (VAS) in an ANCOVA. To determine whether these PTend results can be attributed to potentially unequal relative load imposed by the FET (e.g. 90 Nm represents a higher relative load in weaker individuals), ANCOVAs were used successively to adjust for Strength, and then for fat-free mass (FFM). Considering that Strength might be


affected by pain-related psychological variables, FFM was used as a proxy of Strength, the correlation between these variables ranging between 0.57 and 0.67 26;27. Again, the same conclusions were reached from the corresponding statistical results (P values). Interestingly, low catastrophizers (PCS-low) also showed a trend, although not statistically significant, towards higher endurance than healthy controls (Figure 2).

Figure 2. Predicted time to exhaustion (PTend) between sexes and between healthy controls and CLBP subgroups (PCS-low and PCS-high) divided on the basis of the median split of the scores from the pain catastrophizing scale (PCS). The horizontal line at the top identifies a significant difference according to pairwise comparisons.

Electromyographic variables

To summarize the main findings, only the results corresponding to EMG indices that were more consistently sensitive to the main effects are reported in Table 3 and in Figures 3 and 4.


Table 3. ANOVA and ANCOVA (adjusting for Strength) statistical results corresponding to EMG variables showing a statistically significant PAIN-STATUS (healthy vs CLBP subgroups) effect.

Subgroup Variable* Muscles ANOVA approach ANCOVA approach

variable STATUS Post-hoc Tests †

STATUS Strength Post-hoc Tests †

PCS NIMNFslp X4 < 0.001 1,2 > 3 < 0.001 0.008 2 > 1 > 3

(%/min) M4 < 0.001 1,2 > 3 < 0.001 < 0.001 1,2 > 3

PCS NWINT3slp X4 0.012 2 < 3 0.017 0.090 2 < 3

(27-49 Hz) M4 0.004 2 < 3 0.006 0.031 2 < 1,3

VAS NWINT9slp X4 0.047 N.S. 0.079 0.049 /

(242-301 Hz) M4 0.014 1 > 3 0.033 0.002 1 > 3

PCS NWINT9slp X4 0.031 1 > 3 0.082 0.085 /

(242-301 Hz) M4 0.002 1,2 > 3 0.011 0.004 1,2 > 3

VAS EMGVAR ‡ X6 0.014 1 > 3 < 0.001 0.301 1 > 3

TSK EMGVAR ‡ X6 0.009 1 > 2 0.007 0.364 1 > 2

PCS EMGVAR ‡ X6 < 0.001 2 < 1,3 < 0.001 0.540 2 < 1,3

PCS PALT L5-L1 0.026 1 < 2 0.022 0.306 1 < 2

L3-L1 0.018 1,3 < 2 0.018 0.456 1,3 < 2

Statistically significant differences (P<0.05) are identified with bold characters while trends (0.05<P<0.10) are underlined. N.S.: non significant.* Please see methods for EMG variable names (acronyms). † Results from pairwise comparisons (Tukey-Kramer) corresponding to a significant PAIN-STATUS main effect. Comparisons corresponding to the STATUS effect are identified with numbers [ex. : 1 < 3 means that group 1 (healthy subjects) < group 3 (CLBP subjects – subgroup 2)]. ‡ EMGVAR-M6 statistical results were not shown because they were very comparable to EMGVAR-X6.

EMG indices of muscle fatigue (NIMNFslp, NWINTnslp) showed a significant STATUS effect mostly using PCS subgrouping, VAS subgrouping leading to significant differences mainly for NWINT9slp (Table 3, Figure 3). Adjusting for pain intensity (ANCOVA with VAS as a covariate) did not change the results using PCS subgrouping. Post-hoc analyses revealed, in most cases, significantly higher fatigue in PCS-high patients than in healthy controls and/or PCS-low patients (Table 3). Again, to determine whether this effect can be attributed to potentially unequal relative load imposed by the FET, ANCOVAs were used successively to adjust for Strength, and then for FFM. As further depicted in Table 3, adjusting for Strength or for FFM led basically to the same results except for two variables (NIMNFslp-X4 and NWINT3slp-M4), where adjusting for Strength led to PCS-low patients showing significantly lower fatigue than healthy subjects. Results corresponding to ANCOVAs adjusting for FFM were not shown in Table 3 because almost identical to ANOVA results.


Figure 3. Averaged values corresponding to different EMG spectral indices of muscle fatigue corresponding to the mean (NIMNFslp-X4) and maximal (NWINT3slp-M4) values across L5 and L1 electrode sites (n = 4), for the different pain status (healthy and CLBP patients). The CLBP patients were divided into two subgroups (PCS-low and PCS-high) based on a median split of scores on the pain catastrophizing scale (PCS). The horizontal lines at the top of each plot identify significant differences according to pairwise comparisons following the ANCOVA adjusting for Strength. The corresponding statistical results are provided in Table 3.

Figure 4. Averaged values corresponding to different EMG amplitude indices of variable load sharing (EMGVAR-X6, PALT-L3L1) applied to different combinations of electrode sites, for the different pain status (healthy and CLBP patients). The CLBP patients were divided into two subgroups (PCS-low and PCS-high) based on a median split of scores on the pain catastrophizing scale (PCS). The horizontal lines at the top of each plot identify significant differences according to pairwise comparisons. The corresponding statistical results are provided in Table 3.


Among the EMG load-sharing indices (EMGVAR, EMGALT and PALT), EMGVAR and PALT disclosed statistically significant PAIN-STATUS effects (Table 3, Figure 4). This effect was almost always significant for EMGVAR variables, regardless of how CLBP were divided (using VAS, TSK or PCS), though the P values were lower for the PCS subgrouping. PCS-low patients showed lower EMG amplitude variations (EMGVAR) than healthy and PCS-high patients (Figure 4). For PALT, the PAIN-STATUS main effect was significant only with the PCS subgrouping, the PCS-low subgroup showing more alternating activity than healthy subjects and in one case (PALT -L3-L1), more than the PCS-high subgroup (Figure 4). Again, in both cases (EMGVAR and PALT), adjusting for VAS, for Strength or for FFM (used as covariate in an ANCOVA) did not change these statistical findings.

DISCUSSION

The main finding of the present study was that PCS-high patients showed lower back muscle endurance (PTend) than PCS-low patients, as determined with a methodology allowing to assess back muscle endurance without the detrimental influence of motivational factors. Unsurprisingly, the various EMG variables that are used to derive PTend further supported this finding. However, some of these EMG variables also pointed out that the PCS-low patients had lower back muscle fatigue (NIMNFslp, when adjusting for Strength), lower EMG amplitude variations (EMGVAR) and more alternating activity (PALT) than healthy subjects.

Cross-validation of predicted mechanical variables (PStrength, PTend)

It is relevant to ask ourselves whether PStrength and PTend estimates, derived from regression equations that were previously developed with another sample of subjects, are valid in other subjects. The results were somewhat ambiguous because comparisons between measured and predicted scores, using correlations and T-tests (bias), diverged. However, in both cases (strength and endurance), between-group comparisons led to the same conclusions, relative to PAIN-STATUS and SEX main effects, whatever the measured or predicted variables used (Table 2). This suggests that PStrength and PTend estimates cannot be used with confidence on an individual basis but that trustable conclusions can be drawn from between-group comparisons, which was the case here. In fact, even though there was a systematic bias between Tend and PTend, it is likely that this bias was applied on all subjects, allowing between-group comparisons.

In line with this reasoning PStrength and PTend estimates were also significantly higher in men than women, which further substantiates the construct validity of the FET. Effectively, men are inherently stronger than women and the PTend results were expected considering that the FET is more difficult for women because the 90-Nm load represents a higher proportion of their strength.


Mechanical variables

Back muscle absolute endurance (PTend) was significantly lower in PCS-high patients than PCS-low patients and adjusting for pain intensity (VAS used as a covariate in an ANCOVA) did not change this finding, thus excluding pain as a potential explanation. Furthermore, adjusting for Strength or FFM (ANCOVAs) also led to the same conclusions, suggesting that back muscle relative endurance (instead of absolute endurance) was involved here. Effectively, considering that the FET measures absolute endurance (because the load is absolute: 90 Nm), instead of relative endurance (where the load would be a fraction of Strength; e.g. 40% of strength), statistically adjusting for Strength allows to make such inferences. The fact that only endurance (PTend), not strength, reached statistical significance between PCS subgroups is in line with results showing that endurance is more discriminant than strength to differentiate between healthy subjects and individuals with CLBP 7;28;29 and with the fact that only poor endurance have shown some prognostic value with regards to predicting a first episode of back pain 30;31. This finding also supports the fear-avoidance model stating that patients showing more fear and pain catastrophizing decrease their physical activity and as a result, show more physical deconditioning 9. Interestingly, low catastrophizers (PCS-low) also showed a trend, although not statistically significant, towards higher back muscle endurance than healthy controls (Figure 2). EMG variables were apparently more sensitive to this effect, as outlined below.

Electromyographic variables

Different EMG fatigue indices (NIMNFslp and NWINTnslp indices; Figure 3) consistently showed comparable results to PTend (PCS-high patients more fatigable than PCS-low patients), which is not surprising considering that PTend is derived from various EMG variables. This concurs with previous findings showing that patients sub-classified as « avoiders » (using the Pain Behavior Checklist) were more fatigable than « confronters », according to an EMG-based assessment comparable to the one used here 32. These EMG findings mainly suggest that more metabolic products were accumulating at the surface of the recruited muscle fibers in PCS-high patients. This might likely be attributed to a reduced muscle oxidative metabolism following physical deconditioning 33. The larger proportion of type II fatigable muscle fibers previously observed in patients with CLBP, which correlates with pain duration 34, can further explain this reduced oxidative metabolism. Interestingly, when patients were divided using PCS scores only (but not VAS or TSK), PCS-low patients showed EMG signs of better back muscle endurance compared with healthy controls, which would substantiate the trend observed with the PTend mechanical variable. These findings support, in addition to the fear-avoidance model, a complementary model called the avoidance-endurance model of chronic pain 35. This model suggests a further subgroup of CLBP subjects (« endurance coping » patients, in addition to « fear avoidance coping » patients) who try to ignore pain sensations by attention diversion and who tend to finish all activities in spite of severe pain (pain-related behavioral endurance). Hasenbring et al. have


hypothesized [35 ;p. 2763] that this could lead to more physical activity than normal and to increased muscle strength. More physical activity was recently observed in some CLBP patients 2, but increased endurance, instead of increased strength, has been observed here. All in all, EMG fatigue indices would support both the fear-avoidance and avoidance-endurance models.

PCS-low patients showed lower EMG amplitude variations (EMGVAR) and more alternating activity (PALT) than PCS-high patients and healthy subjects. As discussed next, this would partly support the avoidance-endurance model of chronic pain 35, keeping in mind that there was no difference between healthy and PCS-high patients. The EMG indices of neuromuscular activation patterns (EMGVAR and PALT) are related to back muscle strength as well as back muscle endurance during the FET 21. However, considering that adjusting for Strength or FFM (using ANCOVAs) did not change the results, strength can be rejected as a possible explanation for the between-group differences observed for these EMG indices. This obviously concurs with the non-significant PStrength results. Regression coefficients corresponding to PALT in the PTend regression equations (for men and women) are positive, indicating that more between-muscles alternating activity (e.g. PALT results in Figure 4: PCS-low > PCS-high and healthy subjects) is associated with more endurance (e.g. PTend results in Figure 2: PCS-low > PCS-high), which concurs with previous findings 36. The rational is that sharing the 90-Nm load between back muscles throughout the FET allows to delay muscle fatigue. In other words, this might be the mechanism explaining why PCS-low patients were more fatigue resistant than PCS-high patients. More physical activity in PCS-low patients, as could be predicted by the avoidance-endurance model, might help to develop this beneficial neuromuscular activation pattern. In contrast, the regression coefficient corresponding to EMGVAR in the PTend regression equation (for men only) is negative, indicating that less EMG variability (e.g. EMGVAR results in Figure 4: PCS-low < PCS-high and healthy subjects) is associated with more endurance (e.g. PTend results in Figure 2: PCS-low > PCS-high). As discussed at length elsewhere 21, EMGVAR is proportional to the relative load [(90-Nm/Strength) × 100] sustained and the progression of muscle fatigue. Again, considering that Strength was equivalent between groups, only fatigue could explain EMGVAR findings. PCS-low patients were less fatigued, partly because producing more alternating activity, and consequently produced less EMG variability.

Pain catastrophizing has been confirmed as an important affect-related construct of pain 37, but is a rather complex construct comprizing three distinct dimensions (rumination, magnification and helplessness)23. The present findings could be explained by one or more of these dimensions, and this would represent a research question that should be addressed before tailoring treatment according to the “catastrophizing” profile of the patients. Exercise programs, which do not deliberately target cognitive factors, can reduce pain catastrophizing 38 in addition to address physical deconditioning. Moreover, the reduction in pain catastrophizing accompanying an exercise program alone, or an exercise program in combination with a cognitive-behavioral treatment, mediates clinical outcomes38. Consequently, adding a cognitive-behavioral element specifically targeting one or all dimensions of catastrophizing may further impact clinical


outcomes as well as back muscle impairments such as back muscle weakness and fatigability. As reviewed recently, the cognitive and affective processes captured within the construct of catastrophizing have been associated with a variety of neurophysiological responses (neuromuscular, cardiovascular, immune, neuroendrocrine), as well as with the activity in the pain neuromatrix within the brain 39. Clearly, further understanding of these relationships represents an essential step to substantiate the biopsychosocial model of pain experience and its consequences.

One important limitation of this study is that cause and effect relationships cannot be inferred from the present findings, considering the transversal design of the study. EMG findings are also limited to superficial back muscles. Finally, the present findings are limited to non-specific CLBP patients that had no prior surgery and that were able to perform the FET. We recall that four women with CLBP were unable to perform the FET, which can be partly explained by their disability and because women are generally weaker than men, which makes this test more challenging in women. Consequently, the results might be different in more disabled patients (who may or may not had surgery). On the other hand, the rationale supporting the objectivity of this test is that it is submaximal and should not be influenced by motivation if it is performed as required, that is to say if (1) attaining and sustaining each 90-Nm plateau and (2) reaching the 5- (women) or 10-min (men) duration. As a result, excluding subjects that did not fulfilled these requirements allowed to increase the objectivity of the results.

REFEREN CES

1. Mayer TG, Gatchel RJ. Functional restoration for Spinal disorders: the sports medicine approach. Philadelphia: Lea & Febiger, 1988.

2. Bousema EJ, Verbunt JA, Seelen HA, Vlaeyen JW, and Knottnerus JA. Disuse and physical deconditioning in the first year after the onset of back pain. Pain 2007;130:279-86.

3. Smeets RJ, Wittink H, Hidding A, and Knottnerus JA. Do patients with chronic low back pain have a lower level of aerobic fitness than healthy controls?: are pain, disability, fear of injury, working status, or level of leisure time activity associated with the difference in aerobic fitness level? Spine 2006;31:90-7.

4. Smeets RJ, van Geel KD, and Verbunt JA. Is the fear avoidance model associated with the reduced level of aerobic fitness in patients with chronic low back pain? Arch. Phys. Med. Rehabil. 2009;90:109-17.

5. Verbunt JA, Seelen HA, Vlaeyen JW et al. Disuse and deconditioning in chronic low back pain: concepts and hypotheses on contributing mechanisms. Eur. J. Pain 2003;7:9-21.

6. Crossman K, Mahon M, Watson PJ, Oldham JA, and Cooper RG. Chronic low back pain-associated paraspinal muscle dysfunction is not the result of a constitutionally determined "adverse" fiber-type composition. Spine 2004;29:628-34.


7. Hultman G, Nordin M, Saraste H, and Ohlsen H. Body composition, endurance, strength, cross-sectional area, and density of MM erector spinae in men with and without low back pain. J. Spinal Disord. 1993;6:114-23.

8. Smeets RJ, Wade D, Hidding A, Van Leeuwen PJ, Vlaeyen JW, and Knottnerus JA. The association of physical deconditioning and chronic low back pain: a hypothesis-oriented systematic review. Disabil. Rehabil. 2006;28:673-93.

9. Vlaeyen JWS and Linton SJ. Fear-avoidance and its consequences in chronic musculoskeletal pain: a state of the art. Pain 2000;85:317-32.

10. Elfving B, Andersson T, and Grooten WJ. Low levels of physical activity in back pain patients are associated with high levels of fear-avoidance beliefs and pain catastrophizing. Physiother. Res. Int. 2007;12:14-24.

11. Verbunt JA, Seelen HA, Vlaeyen JW, van der Heijden GJ, and Knottnerus JA. Fear of injury and physical deconditioning in patients with chronic low back pain. Arch. Phys. Med. Rehabil. 2003;84:1227-32.

12. Hirsch G, Beach G, Cooke C, Menard M, and Locke S. Relationship between performance on lumbar dynamometry and Waddell score in a population with low-back pain. Spine 1991;16:1039-43.

13. Ropponen A, Gibbons LE, Videman T, and Battie MC. Isometric back extension endurance testing: reasons for test termination. J Orthop. Sports Phys. Ther. 2005;35:437-42.

14. Verbunt JA, Seelen HA, Vlaeyen JW et al. Pain-related factors contributing to muscle inhibition in patients with chronic low back pain: an experimental investigation based on superimposed electrical stimulation. Clin. J.Pain 2005;21:232-40.

15. Kramer M, Ebert V, Kinzl L, Dehner C, Elbel M, and Hartwig E. Surface electromyography of the paravertebral muscles in patients with chronic low back pain. Arch. Phys. Med. Rehabil. 2005;86:31-6.

16. Lariviere C, Arsenault AB, Gravel D, Gagnon D, and Loisel P. Surface electromyography assessment of back muscle intrinsic properties. J. Electromyogr. Kinesiol. 2003;13:305-18.

17. Lariviere, C., Gravel, D., Gardiner, P., Arsenault, A. B., Gagnon, D., and Loisel, P. A submaximal test to assess back muscles capacity : Evaluation of construct validity. J. Electromyogr. Kinesiol. 2009; 19:e422-e429.

18. Lariviere C, Gravel D, Gagnon D, Gardiner P, Bertrand AA, and Gaudreault N. Gender influence on fatigability of back muscles during intermittent isometric contractions: A study of neuromuscular activation patterns. Clin. Biomech. 2006;21:893-904.

19. Lariviere C, Gravel D, Gagnon D, and Arsenault AB. The assessment of back muscle capacity using intermittent static contractions. Part II: Validity and reliability of biomechanical correlates of muscle fatigue. J. Electromyogr. Kinesiol. 2008;18:1020-31.


20. Lariviere C, Gagnon D, Gravel D, and Bertrand AA. The assessment of back muscle capacity using intermittent static contractions. Part I - Validity and reliability of electromyographic indices of fatigue. J. Electromyogr. Kinesiol. 2008;18:1006-19.

21. Lariviere C, Gravel D, Gagnon D, and Arsenault AB. Toward the development of predictive equations of back muscle capacity based on frequency- and temporal-domain electromyographic indices computed from intermittent static contractions. Spine J. 2009;9:87-95.

22. French DJ, Roach P, and Mayes S. Peur du mouvement chez les accidentés du travail: l'Échelle de Kinésiophobie de Tampa (EKT). Revue canadienne des sciences du comportement 2002;34:28-33.

23. Sullivan, M. J. L., Bishop, S. R., and Pivik, J. The Pain Catastrophizing Scale: Development and validation. Psychol. Assessment 7, 524-532. 1995.

24. Lariviere C, Arsenault AB, Gravel D, Gagnon D, and Loisel P. Median frequency of the electromyographic signal: effect of time-window location on brief step contractions. J. Electromyogr. Kinesiol. 2001;11:65-71.

25. Cifrek M, Medved V, Tonkovic S, and Ostojic S. Surface EMG based muscle fatigue evaluation in biomechanics. Clin. Biomech. 2009;24:327-40.

26. Lariviere C, Gravel D, Gagnon D, Arsenault AB, Loisel P, and Lepage Y. Back strength cannot be predicted accurately from anthropometric measures in subjects with and without chronic low back pain. Clin. Biomech. 2003;18:473-9.

27. Mannion AF, Adams MA, Cooper RG, and Dolan P. Prediction of maximal back muscle strength from indices of body mass and fat-free body mass. Rheumatology 1999;38:652-5.

28. Holmstrom E, Moritz U, and Andersson M. Trunk muscle strength and back muscle endurance in construction workers with and without low back disorders. Scand. J. Rehabil. Med. 1992;24:3-10.

29. Nicolaisen T and Jorgensen K. Trunk strength, back muscle endurance and low-back trouble. Scand. J. Rehabil. Med. 1985;17:121-7.

30. Biering-Sorensen F. Physical measurements as risk indicators for low-back trouble over a one-year period. Spine 1984;9:106-19.

31. Luoto S, Heliovaara M, Hurri H, and Alaranta H. Static back endurance and the risk of low-back pain. Clin. Biomech. 1995;10:323-4.

32. Biedermann HJ, Shanks GL, Forrest WJ, and Inglis J. Power spectrum analyses of electromyographic activity. Discriminators in the differential assessment of patients with chronic low-back pain. Spine 1991;16:1179-84.

33. Tartaglia MC, Chen JT, Caramanos Z, Taivassalo T, Arnold DL, and Argov Z. Muscle phosphorus magnetic resonance spectroscopy oxidative indices correlate with physical activity. Muscle Nerve 2000;23:175-81.


34. Mannion AF, Kaser L, Weber E, Rhyner A, Dvorak J, and Muntener M. Influence of age and duration of symptoms on fibre type distribution and size of the back muscles in chronic low back pain patients. Eur. Spine J. 2000;9:273-81.

35. Hasenbring MI, Plaas H, Fischbein B, and Willburger R. The relationship between activity and pain in patients 6 months after lumbar disc surgery: do pain-related coping modes act as moderator variables? Eur. J Pain 2006;10:701-9.

36. van Dieen JH, Vrielink HHE, and Toussaint HM. An investigation into the relevance of the pattern of temporal activation with respect to erector spinae muscle endurance. Eur. J. Appl. Physiol. 1993;66:70-5.

37. Mounce C, Keogh E, and Eccleston C. A Principal Components Analysis of Negative Affect-Related Constructs Relevant to Pain: Evidence for a Three Component Structure. J Pain (in press).

38. Smeets RJ, Vlaeyen JW, Kester AD, and Knottnerus JA. Reduction of pain catastrophizing mediates the outcome of both physical and cognitive-behavioral treatment in chronic low back pain. J Pain 2006;7:261-71.

39. Campbell CM and Edwards RR. Mind-body interactions in pain: the neurophysiology of anxious and catastrophic pain-related thoughts. Transl. Res. 2009;153:97-101.

40. Durnin JVGA and Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br. J. Nutrition 1974;32:77-97.

41. Stokes IA, Henry SM, and Single RM. Surface EMG electrodes do not accurately record from lumbar multifidus muscles. Clin. Biomech. 2003;18:9-13.

Poor back muscle endurance is related to - Irsst

Documents

Transcript of Poor back muscle endurance is related to - Irsst