Gluteus medius muscle activation patterns as a predictor of low back pain during standing

Available online at www.sciencedirect.com

Clinical Biomechanics 23 (2008) 545–553 www.elsevier.com/locate/clinbiomech

Gluteus medius muscle activation patterns as a predictor of low back pain during standing Erika Nelson-Wong, Diane E. Gregory, David A. Winter, Jack P. Callaghan * Department of Kinesiology, Faculty of Applied Health Sciences, University of Waterloo, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1 Received 31 August 2007; accepted 4 January 2008

Abstract Background. Low back pain is a primary source of disability and economic costs. Altered trunk muscle activation in people with low back pain, specifically agonist/antagonist co-activation, has been previously demonstrated. Prevailing theory considers this muscle activation pattern to be adaptive to low back pain. Muscle activation patterns prior to, and during, the development of low back pain in asymptomatic individuals, have not been well studied. Methods. Participants, without a history of low back pain, stood in a constrained area for 2 h. Continuous surface electromyography was collected from trunk and hip muscles. Participants rated their discomfort level on visual analog scale every 15 min. Cross-correlation analyses were used to determine co-activation patterns. Blind predictions were made to categorize participants into low back pain and non-low back pain groups, and comparisons made to visual analog scale scores. Findings. 65% of previously asymptomatic participants developed low back pain during the protocol. Co-activation of the bilateral gluteus medius muscles was found to be prevalent in the low back pain group (P = .002). 76% of the participants were correctly classified into low back pain and non-low back pain groups based on presence or absence of gluteus medius co-activation, with sensitivity = .87 and specificity = .50. Interpretation. Agonist–antagonist co-activation may not be entirely adaptive, and may in fact predispose some individuals to develop low back pain. Muscle activation patterns at the hip may be a useful addition for screening individuals to identify those at risk of developing low back pain during standing. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Low back pain; Occupational standing; Motor control; EMG

1. Introduction It is well known that low back pain (LBP) is a major contributor to escalating health care costs and disability in North America. It is estimated that 70–85% of all adults will experience a significant episode of LBP at some point in their lives (Giesecke et al., 2004). There are over 700,000 workers’ compensation claims for work-related LBP in the United States each year

*

Corresponding author. E-mail address: [email protected] (J.P. Callaghan).

0268-0033/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2008.01.002

(Waddell, 2004), and one third of worker’s compensation costs are related to LBP (Abenhaim et al., 2000). While occupational manual material handling tasks have been extensively identified as an important contributing factor in LBP and discomfort (Norman et al., 1998; Adams et al., 1999), static low magnitude exposures such as prolonged standing have also been associated with the development of LBP and discomfort (Macfarlane et al., 1997). Mechanisms for development of LBP are widely considered to be multi-factorial (Linton, 2000; Kumar, 2001; Waddell, 2004), therefore the effective prediction of who will develop LBP remains complex and problematic (Leboeuf-Yde et al., 1997).

546

E. Nelson-Wong et al. / Clinical Biomechanics 23 (2008) 545–553

1.1. Motor control/muscle activation patterns and low back pain Many researchers have reported differences in muscle activation patterns in people with LBP (Hungerford et al., 2003; van Dieen et al., 2003b; Ferguson et al., 2004; Pirouzi et al., 2006; Gregory et al., 2007). A primary finding that is consistent across studies is the presence of agonist–antagonist muscle co-activation in people who report LBP. One known consequence of increased muscle co-activation is a commensurate increase in spine loading (Granata and Marras, 2000). There has been an assumption that this altered muscle activation pattern in people with LBP is in response to, or an adaptation to their condition (van Dieen et al., 2003a, b; O’Sullivan, 2005). The bulk of the existing literature deals with cohort investigations (i.e., clinical LBP populations compared with healthy, asymptomatic populations). A limitation of this methodology is that it cannot be ascertained if altered muscle activation patterns were present in the clinical groups prior to the onset of their dysfunction, and therefore it remains unknown whether these patterns are adaptive, or in fact contributory to the problem of LBP. There are relatively few studies that have examined muscle activation patterns during prolonged exposures to static postures, such as those commonly seen in real-world workplace environments (Gregory et al., 2007). Typically, studies focused on examining LBP have not included the hip musculature in their analyses. However, hip function has been shown to be an important contributor to both trunk and spine function, and therefore likely plays a role in the development and response to LBP (Kankaanpaa et al., 1998; Leinonen et al., 2000; Nadler et al., 2001; Gombatto et al., 2006). Winter and colleagues (1996) identified synergistic muscle activity at the hip as a fundamental mechanism for medio-lateral postural control during standing, therefore the hip musculature should be considered during the study of any activity involving standing. 1.2. Purpose/aims The primary purpose of this work was to investigate differences in trunk and hip muscle activation patterns in asymptomatic individuals during prolonged standing while performing simulated occupational tasks. A secondary purpose was to determine whether identification of individuals who would develop LBP during prolonged standing could be achieved based upon their muscle activation patterns.

antagonist muscle co-activation in LBP developers. Furthermore, it was expected that these differences would be sufficient to enable separation of subjects into pain and non-pain groups prior to viewing their self-reported pain ratings. 2. Methods Data for this work was collected by two independent researchers, using identical protocols, and was analyzed retrospectively for this research study. 2.1. Participant description Twenty-three participants, 12 males and 11 females, from a university population volunteered for this study. Participant characteristics are reported in Table 1. Participants were required to have no history of LBP during the previous 12 months and were free of LBP at the time of initiation of the study. Written informed consent was obtained from all subjects in accordance with University Office of Research guidelines prior to their participation. 2.2. Experimental protocol Participants stood for 2 h in a constrained area (0.50 m  0.46 m) while they completed a series of four different simulated occupational tasks in 30 min blocks. These tasks were designed to mimic common occupational demands and represented jobs that require periods of prolonged standing. The tasks included assembly of retractable pens (assembly line worker), currency sorting (bank teller), grocery store checkout (cashier), and card dealing (casino dealer). The tasks were randomized to counter any potential order effects. Participants rated their level of neck, shoulder, upper back, and low back pain on a 100 mm visual analog scale (VAS) prior to the start of the standing protocol, every 15 min throughout the 2 h standing period, and at the end of the 2 h. For this study, only the LBP rating was considered as it was found to be the most consistently increased body area. The VAS is a self-rating of current level of perceived pain with end-point anchors of ‘no discomfort’ and ‘worst discomfort imaginable’ and is measured in millimeters. The VAS has been found to have good construct validity (Summers, 2001), as well as reliability (Revill et al., 1976).

1.3. Hypotheses A small percentage of the participants were expected to develop significant levels of LBP during the prolonged standing task. It was hypothesized that there would be differences in trunk and hip muscle activation patterns between those who developed pain and those who did not. Specifically, we expected to find increased agonist–

Table 1 Similarity of subject characteristics between groups (Mean, SD)

LBP

Height (m) Weight (kg) Age (yrs) BMI (kg/m2)

1.77, 74.6, 23.9, 23.6,

1.76 14.2 1.8 2.87

non-LBP

P (2-tailed t-test)

1.76, 75.2, 23.9, 24.3,

0.75 0.91 0.98 0.54

.08 10.3 2.3 1.64

E. Nelson-Wong et al. / Clinical Biomechanics 23 (2008) 545–553

An independent examiner (D.G.) collected the VAS scores, and withheld them until the data analysis was completed and a blinded classification of participants into LBP and non-LBP groups was accomplished. 2.3. Data collection and signal processing Continuous surface electromyography (EMG) was collected with disposable pre-gelled EMG Ag–AgCl electrodes (Blue Sensor, Medicotest, Inc., Olstykke, Denmark) that had an inter-electrode distance of 2 cm from the following bilateral muscles: lumbar erector spinae (LES), thoracic erector spinae (TES), rectus abdominus (RA), external oblique (EO) and gluteus medius (GM). Electrode placements for the trunk muscles were determined based on the work of McGill (1991). Placement of the GM electrodes was 15 cm inferior and 5 cm posterior of each iliac crest. Raw EMG was amplified (AMT-8, Bortec, Calgary, Canada; bandwidth = 10–1000 Hz, CMRR = 115 db at 60 Hz, input impedance = 10 GX) and collected with a sampling frequency of 2048 Hz using a 16-bit A/D card with a ±2.5 V range. EMG signals had systematic bias removed, and were full wave rectified prior to being dualpass filtered through a fourth order Butterworth filter with an effective cutoff frequency of 6 Hz (Winter, 2005). The resulting linear envelope signals were then normalized to maximal voluntary contractions (MVC). MVC’s were obtained for TES and LES through applying resistance in the Beiring–Sorensen position. MVC’s for RA and EO were obtained through a resisted modified sit-up with simultaneous twisting about the waist to ensure maximal contraction of the abdominal muscles. MVC’s for bilateral GM were obtained through resisted hip abduction in the side lying position. The processed EMG signals were down sampled to 32 Hz prior to further data analysis as a data reduction measure. 2.4. Data analysis Cross-correlation analyses were used to quantify the common signal between different EMG recording sites to provide muscle coordination information. Cross-correlation quantifies the similarity in shape and the phase delay between two time-varying waveforms, and has been used as a method of describing coordination with kinematic (Shum et al., 2005a, b) and muscle activation (Osu et al., 2002; Mogk and Keir, 2003; McDonnell et al., 2005) data previously. A custom program was written in Matlab R2006a version 7.2 (The Mathworks, Inc., Natick, MA, USA) to compute cross-correlation coefficients, Rxy, with the following equation Z 1 T xðtÞyðt þ sÞ dt ð1Þ Rxy ðsÞ ¼ T 0 Rxx ð0ÞRyy ð0Þ where Rxy(s) is the normalized cross-correlation of two signals, x(t) and y(t) at a phase shift s with a potential range

547

of values between 1 and +1, and T is the length of the record. A highly positive correlation is an indication that the two signals are acting together in phase, and a highly negative correlation indicates that one signal is at a maximum as the other is at a minimum (Winter and Patla, 1997). Therefore, a peak positive Rxy value near s = 0 indicates the two muscles are being activated and deactivated simultaneously. For this analysis, the Rxy function was examined at phase lags close to zero (±500 ms), with the interpretation that positive Rxy values with small phase lags (<500 ms) indicate muscles were being activated together, and negative Rxy values with small phase lags indicating one muscle was activated while the other was not. Pairs of EMG records of 15 min lengths were cross-correlated against each other and then the maximum and minimum Rxy values were recorded, and the largest absolute Rxy value extracted. Therefore the Rxy values represent an average correlation of the two signals over the 15 min window, and indicate a dominant pattern of co-activation (positive Rxy) or reciprocal firing (negative Rxy) for the muscle pair being considered. This analysis yielded eight Rxy values spanning the 2 h standing period for each pair of EMG signals under consideration. Table 2 details the pairs of EMG recordings that were used for cross-correlation analysis. Preliminary examination of the data revealed strikingly different patterns in Rxy values for left GM cross-correlated with right GM (Rxy–LGM–RGM) between participants. One participant with a strong positive Rxy–LGM–RGM value and one with a strong negative value were selected, and found to have high and low VAS scores, respectively. Based upon this finding, participants were predicted to fall into either a LBP or non-LBP category based upon the polarity of their Rxy–LGM–RGM values. Participants with positive Rxy values were predicted to fall into the LBP group, and those with negative Rxy values were predicted as non-LBP. For those participants who changed polarity of their Rxy value during the 2 h period, the average of the eight Rxy values was used for the group prediction. Table 2 Between group differences in cross-correlations between EMG recording pairs

LGM–RGM* LLES–RLES LTES–RTES LLES–LEO* RLES–REO LLES–LRA RLES–RLA LTES–LEO RTES–REO LTES–LRA RTES–RRA

Rxy LBP group (mean, SE)

Rxy non-LBP group (mean, SE)

P value (between groups ANOVA)

.187, .265, .380, .094, .134, .098, .116, .119, .110, .095, .104,

.068, .033 .112, .054 .241, .028 .160, .030 .038, .046 .039, .041 .022, .046 .089, .189 .161, .036 .072, .017 .063, .046

.002* NS NS .01* NS NS NS NS NS NS NS

* , significantly different. NS = non-significant.

.015 .019 .017 .014 .014 .015 .023 .015 .017 .012 .017

548

E. Nelson-Wong et al. / Clinical Biomechanics 23 (2008) 545–553

VAS ratings were then scored for the entire study population and the predicted pain category was compared with actual pain category from the previously blinded VAS. Participants were assigned to the pain group if they reported a VAS rating greater than 20 mm at any point during the study, and also had an overall average VAS rating greater than 10 mm. These threshold VAS values were chosen since 9 mm has been found to be the minimum clinically significant difference in VAS, representing a small treatment effect, with >20 mm differences representing a large treatment effect (Kelly, 1998). 2.5. Statistical treatment SPSS version 10.0 statistical software (SPSS, Inc., Chicago, IL, USA) was used for all statistical analysis. Twotailed independent t-tests were used to detect differences in the subject characteristics between the LBP and nonLBP groups (Table 1). Dependent variables of VAS were entered into a 3 way mixed model ANOVA with between-subjects factors of group (LBP or non-LBP), and gender, and a within-subjects factor of time. There was no effect of gender on VAS, and participants were balanced with regard to gender in their distribution across groups, therefore gender was not included in the remaining analyses. For Rxy values the between-subjects factors were group and task. There were no main effects of task, and no interactions, therefore the 4 levels of task were pooled for the remaining analyses. Significance was set at P 6 0.05 for all measures. Where the data did not meet sphericity assumptions, Huynh–Feldt epsilon corrections were used to determine the degrees of freedom.

3. Results 3.1. Prediction of low back pain development Seventeen of 23 participants were correctly predicted to have presence or absence of LBP during prolonged standing based upon the single factor of Rxy–LGM–RGM polarity. Figs. 1 and 2 show example EMG recordings for the left and right GM muscles with the corresponding cross-correlation values. Fig. 1 shows a typical pattern of synergistic GM muscle activity during standing from a non-LBP subject, with a corresponding negative Rxy value. Conversely, Fig. 2 shows a typical pattern of GM muscle co-activation during standing from a LBP subject, with a corresponding positive Rxy value. Using this algorithm for prediction, there were four false positives (i.e., pain was predicted when a participant did not actually have pain), and two false negatives (i.e. non-pain was predicted when a participant actually had pain). The prediction categorization based on Rxy– LGM–RGM polarity yielded a sensitivity (95% CI) of 0.87 (.58–.98) and a specificity of 0.50 (.17–.83). 3.2. Statistical analysis 3.2.1. VAS Fifteen of the 23 participants reported LBP. Participants clearly separated into two distinct groups with a LBP group reporting increasing pain over time and a non-LBP group remaining at low, stable levels of reported pain over time (Fig. 3). There were significant main effects of time (F4,86 = 27.621, P = 0.000), group (F1,21 = 40.419, P = 0.001), and a significant interaction between group and time (F4,86 = 10.572, P = 0.000) indicating that the LBP

16

LGM RGM

14

Rxy-LGM-RGM = -0.30

12 10 8 6 4 2 0 320

325

330

335

340

345

350

355

360

365

Time (s) Fig. 1. Linear enveloped EMG from left and right gluteus medius muscles in a non-LBP subject during standing shows a synergistic, reciprocal firing pattern. There is a corresponding negative Rxy–LGM–RGM value, indicating left GM is turning off as right GM is turning on.

E. Nelson-Wong et al. / Clinical Biomechanics 23 (2008) 545–553 16

549

LGM RGM

14

Rxy-LGM-RGM = 0.42

12 10 8 6 4 2 0 160

165

170

175

180

185

190

195

200

Time (s) Fig. 2. Linear enveloped EMG from left and right gluteus medius muscles in a subject with LBP during standing shows a co-activation pattern. There is a corresponding positive Rxy–LGM–RGM value, indicating left and right GM are turning on and off simultaneously.

40

35

P = .000

30

25

20

LBP non-LBP

15

10

5

0 Start

15

30

45

60

75

90

105

120

Time (minutes) Fig. 3. Self-reported pain on visual analog scale over time. Participants clearly separated into two groups. Pain increases particularly in the last 60 min of standing for the LBP group, while pain scores for the non-LBP group remain relatively constant over time.

group’s response over the 2 h was different from the nonLBP group. 3.3. Muscle co-activation patterns There were no significant differences between pain groups detected for co-activation patterns of the thoracic

erector spinae with rectus abdominus or external obliques, bilateral thoracic erector spinae, lumbar erector spinae with rectus abdominus, or bilateral lumbar erector spinae muscles (Table 2). Significant findings for co-activation of bilateral gluteus medius and lumbar erector spinae with external oblique muscles are reported below.

550

E. Nelson-Wong et al. / Clinical Biomechanics 23 (2008) 545–553

3.4. Bilateral gluteus medius

4. Discussion

When Rxy data was included from all 23 participants and classified into groups from the VAS scores, there was no main effect of time (F7,147 = .831, P = .563) or main effect of group (F1,21 = 2.299, P = .144). Since the predicted classification into LBP and non-LBP groups was relatively good, and there was a highly significant difference in VAS scores between these two groups, it was decided to run the same model of ANOVA, including data only from the 17 subjects who were correctly predicted. Within-subjects analysis detected no main effect of time. The between-subjects analysis found a main effect of group (F1,15 = 14.920, P = .002). Fig. 4 shows average Rxy–LGM–RGM values over time for the correctly predicted participants.

The primary aim of this work was to determine whether muscle activation patterns differed among asymptomatic individuals performing occupational tasks during standing. The secondary aim was to determine whether differences in muscle activation patterns could be used to differentiate individuals who would experience LBP during a lowdemand functional activity from those who would not. A large percentage of previously asymptomatic individuals (65%) in this study experienced significant levels of LBP during prolonged standing. Furthermore, participants clearly separated into two distinct groups, with different pain responses as time progressed. The predictive utility of examining a single factor of co-activation of the left and right GM muscles was relatively good, with 17/23 participants correctly predicted. There were four false positives and two false negatives, which yields a sensitivity (Sn) of 0.87 and a specificity (Sp) of 0.50 with a corresponding positive likelihood ratio (+LR) = 1.74 and a negative likelihood ratio (LR) = 0.26. The relatively strong Sn value (Fritz and Wainner, 2001), implies that examination of muscle co-activation patterns at the hip may show promise as a useful screening tool for assisting in the determination of who may be at risk for development of LBP during prolonged standing. Furthermore, the LR of 0.26 indicates a small, but potentially important shift in probability (Fritz and Wainner, 2001) that an individual has a decreased risk of developing LBP with standing when they do not exhibit co-activation of the hip musculature. The low-demand nature of prolonged standing does not require high level activation of the monitored trunk musculature, and it is possible that differences between pain

3.5. Lumbar erector spinae and external oblique There were no effects of time and no interactions when the lumbar erector spinae was cross-correlated with the external obliques (designated as Rxy–LLES–LEO for the left side and Rxy–RLES–REO for the right side). Similar to the cross-correlations between left and right GM, there were no statistically significant differences between groups when Rxy values were included from all participants (Rxy–LLES–LEO: F1,21 = 3.134, P = .091, Rxy–RLES– REO: F1,21 = 2.577, P = .123). However, when only the predicted participants were included, there was a significant main effect of group detected in Rxy–LLES–LEO (F1,15 = 8.606, P = .010) while Rxy–RLES–REO differences remained non-significant (F1,15 = 2.960, P = .106). Fig. 5 shows average Rxy–LLES–LEO values over time for correctly predicted participants. 0.3

Non-LBP LBP

0.2

0.1

0

-0.1

-0.2

-0.3 15

30

45

60

75

90

105

120

Time (min) Fig. 4. Rxy–LGM–RGM values over time for correctly predicted participants only. LBP group demonstrates consistent co-activation firing pattern from the early stages of standing.

E. Nelson-Wong et al. / Clinical Biomechanics 23 (2008) 545–553 0.3

551 Non-LBP LBP

0.2

0.1

0

-0.1

-0.2

-0.3

-0.4 15

30

45

60

75

90

105

120

Time (min) Fig. 5. Rxy–LLES–LEO values over time for correctly predicted participants only. LBP group demonstrates co-activation firing pattern from the early stages of standing.

groups in these muscles would become apparent with a higher-demand activity. The fact that the lumbar and thoracic erector spinae co-activation with rectus abdominus musculature showed no differences between groups was interesting in that, for these subjects at least, their pain during this low-demand task appears to be driven more by the hip musculature than the spine musculature. Decreased endurance and delayed firing in the hip extensors in people with chronic LBP has been found previously (Kankaanpaa et al., 1998; Leinonen et al., 2000), however the role of the hip abductors in LBP has not been extensively studied. Nadler and colleagues (2001) found asymmetrical hip extensor strength was related to future development of LBP in college athletes, however they found no such relationship with hip abductor asymmetry. Their study was based upon muscle strength measurements, and did not examine motor control and muscle activation patterns. The role of hip abductor function deserves more consideration as it has been shown here to be positively associated with the development of LBP. It has been shown through multiple studies that agonist/ antagonist muscle co-activation exists in people with LBP (Paquet et al., 1994; Hungerford et al., 2003; van Dieen et al., 2003a; Dankaerts et al., 2006), however it has not previously been viewed as a causal versus adaptive mechanism in the development of LBP. The participants in this study were pain-free upon entry, and yet those in the LBP developer group demonstrated co-activation of the left and right GM and left lumbar erector spinae/external oblique muscles from the beginning compared with nonLBP developers. This finding raises the question of whether there is a subset of individuals who are predisposed to development of LBP due to a pre-existing motor control

pattern. The underlying reasons for this co-activation pattern should also be further addressed in future work as it may be due to poor motor control or inefficient active stabilization further up the kinetic chain. Individuals in the non-LBP group showed greater variability in their muscle activation patterns than those in the LBP group. There was a subset of individuals (4) who were incorrectly predicted to be in the LBP group based on the presence of co-activation of left GM and right GM. It is probable that there are other factors that need to be considered in addition to co-activation at the hip to decrease this false positive rate and improve the predictive value of this technique. There are several apparent limitations of this work. The recording of surface EMG presents an obvious limitation in that the deeper musculature, such as quadratus lumborum and psoas, cannot be directly monitored. Due to their influence on lumbopelvic mechanics, these muscles may in fact play an important role in the development of LBP during standing. The terminology on the VAS used ‘discomfort’ instead of ‘pain’, and therefore may have led participants to either over- or underestimate their symptoms. However many clinical studies treat the two terms synonymously (Chapman and Dunbar, 1998; Tait and Chibnall, 2002; Schmader et al., 2007), while the ergonomics literature uses ‘discomfort’ to describe such disparate concepts as musculoskeletal pain as well as comfort ratings for seating (Wilder et al., 1994; Parakkat et al., 2007). Discomfort assessment tools have been validated against pain assessment tools with highly significant correlations (Crane et al., 2005). The prior experience level of the participants with prolonged standing activities was not documented. Participants with more experience with prolonged standing

552

E. Nelson-Wong et al. / Clinical Biomechanics 23 (2008) 545–553

tasks may have had different thresholds for reporting discomfort on the VAS. A clinical assessment was not performed on the participants, and although they did not have any overt symptoms of LBP during the previous year, it is possible that they would have exhibited clinical findings that would predispose them to developing LBP during standing. It is becoming well accepted that people with LBP syndromes do not constitute a single homogenous group, and when they are treated in this way there is a risk that significant findings may become washed out (Brennan et al., 2006; Dankaerts et al., 2006). It is possible that subgroups existed within this sample but were not detected. While it would be an oversimplification to assume the activation patterns of a single muscle group are the sole predictor of who will go on to develop LBP, the achieved prediction strength highlights that there is a basis for future work and a need to incorporate the hip musculature when examining LBP groups. 5. Conclusion Participants who developed LBP during prolonged standing demonstrated co-activation of the left and right GM muscles versus synergistic, reciprocal activation of these muscles in those who did not develop LBP. These participants were a non-clinical population, yet LBP was functionally induced through a low demand, common activity of standing. We were able to predict, with moderate success, those individuals who would develop LBP based upon the single factor of co-activation of hip musculature and correctly identified 74% of subjects into their respective pain or non-pain group. Since this muscle activation pattern was present prior to the onset of subjective pain complaint, it is possible that the presence of muscle co-activation should not be considered to be solely an adaptive response, and in fact may be causal, for development of LBP in some individuals. Acknowledgments The authors wish to acknowledge the Natural Sciences and Engineering Research Council Canada, AUTO21-Network of Centres of Excellence, Canadian Institute for the Relief of Pain and Disability, and Canadian Institutes for Health Research for financial support, as well as Sandy Sperling for her assistance with data collection. Dr. Jack Callaghan is also supported by a Canada Research Chair in Spine Biomechanics and Injury Prevention. Erika Nelson-Wong is supported in part by a scholarship through the Foundation for Physical Therapy, American Physical Therapy Association. References Abenhaim, L., Rossignol, M., Valat, J.-P., Nordin, M., Avouac, B., Blotman, F., Charlot, J., Dreiser, R.L., Legrand, E., Rozenberg, S., Vautravers, P., 2000. The role of activity in the therapeutic manage-

ment of back pain: report of the International Paris Task Force in Back Pain. Spine 25, 1S–33S. Adams, M.A., Mannion, A.F., Dolan, P., 1999. Personal risk factors for first-time low back pain. Spine 24, 2497–2505. Brennan, G.P., Fritz, J.M., Hunter, S.J., Thackeray, A., Delitto, A., Erhard, R.E., 2006. Identifying subgroups of patients with acute/ subacute ‘‘nonspecific” low back pain: results of a randomized clinical trial. Spine 31, 623–631. Chapman, C.R., Dunbar, P.J., 1998. Measurement in pain therapy: is pain relief really the endpoint? Current Opinion in Anaesthesiology 11, 533–537. Crane, B.A., Holm, M.B., Hobson, D., Cooper, R.A., Reed, M.P., Stadelmeier, S., 2005. Test-retest reliability, internal item consistency, and concurrent validity of the wheelchair seating discomfort assessment tool. Assistive Technology 17, 98–107. Dankaerts, W., O’sullivan, P., Burnett, A., Straker, L., 2006. Altered patterns of superficial trunk muscle activation during sitting in nonspecific chronic low back pain patients: importance of subclassification. Spine 31, 2017–2023. Ferguson, S.A., Marras, W.S., Burr, D.L., Davis, K.G., Gupta, P., 2004. Differences in motor recruitment and resulting kinematics between low back pain patients and asymptomatic participants during lifting exertions. Clinical Biomechanics 19, 992–999. Fritz, J.M., Wainner, R.S., 2001. Examining diagnostic tests: an evidencebased perspective. Physical Therapy 81, 1546–1564. Giesecke, T., Gracely, R.H., Grant, M.A.B., Nachemson, A., Petzke, F., Williams, D.A., Clauw, D.J., 2004. Evidence of augmented central pain processing in idiopathic chronic low back pain. Arthritis and Rheumatism 50, 613–623. Gombatto, S.P., Collins, D.R., Sahrmann, S.A., Engsberg, J.R., Van Dillen, L.R., 2006. Gender differences in pattern of hip and lumbopelvic rotation in people with low back pain. Clinical Biomechanics 21, 263–271. Granata, K.P., Marras, W.S., 2000. Cost-benefit of muscle cocontraction in protecting against spinal instability. Spine 25, 1398–1404. Gregory, D.E., Brown, S.H.M., Callaghan, J.P., 2007. Trunk muscle responses to suddenly applied loads: do individuals who develop discomfort during prolonged standing respond differently? Journal of Electromyography and Kinesiology. Hungerford, B., Gilleard, W., Hodges, P., 2003. Evidence of altered lumbopelvic muscle recruitment in the presence of sacroiliac joint pain. Spine 28, 1593–1600. Kankaanpaa, M., Taimela, S., Laaksonen, D., Hanninen, O., Airaksinen, O., 1998. Back and hip extensor fatigability in chronic low back pain patients and controls. Archives of Physical Medicine and Rehabilitation 79, 412–417. Kelly, A.-M., 1998. Does the clinically significant difference in visual analog scale pain scores vary with gender, age, or cause of pain? Academic Emergency Medicine 5, 1086–1090. Kumar, S., 2001. Theories of musculoskeletal injury causation. Ergonomics 44, 17–47. Leboeuf-Yde, C., Lauritsen, J.M., Lauritzen, T., 1997. Why has the search for causes of low back pain largely been nonconclusive? Spine 22, 877– 881. Leinonen, V., Kankaanpaa, M., Airaksinen, O., Hanninen, O., 2000. Back and hip extensor activities during flexion/extension: effects of low back pain and rehabilitation. Archives of Physical Medicine and Rehabilitation 81, 32–37. Linton, S.J., 2000. A review of psychological risk factors in back and neck pain. Spine 25, 1148–1156. Macfarlane, G.J., Thomas, E., Papageorgiou, A.C., Croft, P.R., Jayson, M.I.V., Silman, A.J., 1997. Employment and physical work activities as predictors of future low back pain. Spine 22, 1143–1149. Mcdonnell, M.N., Ridding, M.C., Flavel, S.C., Miles, T.S., 2005. Effect of human grip strategy on force control in precision tasks. Experimental Brain Research 161, 368–373. Mcgill, S.M., 1991. Electromyographic activity of the abdominal and low back musculature during the generation of isometric and dynamic

E. Nelson-Wong et al. / Clinical Biomechanics 23 (2008) 545–553 axial trunk torque: implications for lumbar mechanics. Journal of Orthopaedic Research 9, 91–103. Mogk, J.P.M., Keir, P.J., 2003. Crosstalk in surface electromyography of the proximal forearm during gripping tasks. Journal of Electromyography and Kinesiology 13, 63–71. Nadler, S.F., Malanga, G.A., Feinberg, J.H., Prybicien, M., Stitik, T.P., Deprince, M., 2001. Relationship between hip muscle imbalance and occurrence of low back pain in collegiate athletes: a prospective study. American Journal of Physical Medicine and Rehabilitation 80, 572–577. Norman, R., Wells, R., Neumann, P., Frank, J., Shannon, H., Kerr, M., 1998. A comparison of peak vs cumulative physical work exposure risk factors for the reporting of low back pain in the automotive industry. Clinical Biomechanics 13, 561–573. O’sullivan, P., 2005. Diagnosis and classification of chronic low back pain disorders: maladaptive movement and motor control impairments as underlying mechanism. Manual Therapy 10, 242–255. Osu, R., Franklin, D.W., Kato, H., Gomi, H., Domen, K., Yoshioka, T., Kawato, M., 2002. Short- and long-term changes in joint cocontraction associated with motor learning as revealed from surface EMG. Journal of Neurophysiology 88, 991–1004. Paquet, N., Malouin, F., Richards, C.L., 1994. Hip-spine movement interaction and muscle activation patterns during sagittal trunk movements in low back pain patients. Spine 19, 596–603. Parakkat, J., Yang, G., Chany, A.-M., Burr, D., Marras, W.S., 2007. The influence of lift frequency, lift duration and work experience on discomfort reporting. Ergonomics 50, 396–409. Pirouzi, S., Hides, J., Richardson, C., Darnell, R., Toppenberg, R., 2006. Low back pain patients demonstrate increased hip extensor muscle activity during standardized submaximal rotation efforts. Spine 31, E999–E1005. Revill, S.I., Robinson, J.O., Rosen, M., Hogg, M.I., 1976. The reliability of a linear analogue for evaluating pain. Anaesthesia 31, 1191–1198. Schmader, K.E., Sloane, R., Pieper, C., Coplan, P.M., Nikas, A., Saddier, P., Chan, I.S., Choo, P., Levin, M.J., Johnson, G., Williams, H.M.,

553

2007. The impact of acute herpes zoster pain and discomfort on functional status and quality of life in older adults. Clinical Journal of Pain 23, 490–496. Shum, G.L.K., Crosbie, J., Lee, R.Y.W., 2005a. Effect of low back pain on the kinematics and joint coordination of the lumbar spine and hip during sit-to-stand and stand-to-sit. Spine 30, 1998–2004. Shum, G.L.K., Crosbie, J., Lee, R.Y.W., 2005b. Symptomatic and asymptomatic movement coordination of the lumbar spine and hip during an everyday activity. Spine 30, E697–E702. Summers, S., 2001. Evidence-based practice part 2: reliability and validity of selected acute pain instruments. Journal of PeriAnesthesia Nursing 16, 35–40. Tait, R.C., Chibnall, J.T., 2002. Pain in older subacute care patients: associations with clinical status and treatment. Pain Medicine 3, 231– 239. Van Dieen, J.H., Cholewicki, J., Radebold, A., 2003a. Trunk muscle recruitment patterns in patients with low back pain enhance the stability of the lumbar spine. Spine 28, 834–841. Van Dieen, J.H., Selen, L.P.J., Cholewicki, J., 2003b. Trunk muscle activation in low-back pain patients, an analysis of the literature. Journal of Electromyography and Kinesiology 13, 333–351. Waddell, G., 2004. The Back Pain Revolution, second ed. Churchill Livingstone, Edinburgh, New York. Wilder, D., Magnusson, M.L., Fenwick, J., Pope, M.H., 1994. The effect of posture and seat suspension design on discomfort and back muscle fatigue during simulated truck driving. Applied Ergonomics 25, 66–76. Winter, D.A., 2005. Biomechanics and Motor Control of Human Movement, third ed. John Wiley & Sons, Inc., Hoboken. Winter, D.A., Patla, A.E., 1997. Signal Processing and Linear Systems for the Movement Sciences. Waterloo Biomechanics, Waterloo. Winter, D.A., Prince, F., Frank, J.S., Powell, C., Zabjek, K.F., 1996. Unified theory regarding A/P and M/L balance in quiet stance. Journal of Neurophysiology 75, 2334–2343.