Spine

SPINE Volume 39, Number 1, pp 39-47 ©2013, Lippincott Williams &C Wilkins

CERVICAL SPINE

Differential Changes in Muscle Composition Exist in Traumatic and Nontraumatic Neck Pain James M. Elliott, PT, PhD,* Ashley R. Pedler, PT, PhD,t Gwendolen A. Jull, PT, PhD,* Luke Van Wyk, BSc (Hons) Physiotherapy,* Graham G. Galloway, PhD,§ and Shaun P. O'Leary, PT, PhD*1

Study Design. A population based cross-sectional study. Objective. To clarify relative constituents of viable muscle in 2-dimensional cross-sectional area (CSA) measures of ventral and dorsal cervical muscles in patients with chronic whiplash-associated disorders (WAD), idiopathic neck pain, and healthy controls. Summary of Background Data. Previous data using T l weighted magnetic resonance image demonstrated large amounts of neck muscle fat infiltration and increased neck muscle CSA in patients with chronic WAD but not in idiopathic neck pain or healthy controls. Methods. Magnetic resonance images were obtained for 14 cervical muscle regions in 135 females, including 79 with chronic whiplash, 23 with chronic idiopathic neck pain, and 34 healthy controls. Results. Without fat removed, relative CSA of 7 of 14 muscle regions in the participants with chronic WAD was larger, 3 of 14 smaller and 4 of 14 similar to healthy individuals. When T l weighted signal representing the lipid content of these muscles was removed, 8 of 14 relative muscle CSA in patients with whiplash were similar, 5 of 14 were smaller and only 1 of 14 was larger than those observed in healthy controls. Removal of fat from the relative CSA measurement did not alter findings between participants with idiopathic neck pain and healthy controls. From tbe 'Department of Physical Therapy and Human Movement Sciences, Northwestern University, Feinberg School of Medicine, Chicago, IL, and School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane, Queensland, Australia; tCentre of National Research on Disability and Rehabilitation Medicine, Royal Brisbane and Women's Hospital, Herston, Brisbane, Queensland, Australia; Í N H M R C Centre for Clinical Research Excellence in Spinal Pain, Injury and Health, The University of Queensland, Brisbane, Queensland, Australia; §Centre for Advanced Imaging, The University of Queensland, Brisbane, Queensland, Australia; and 1 Physiotherapy Department, Royal Brisbane and Women's Hospital, Herston, Brisbane, Queensland, Australia. Acknowledgment date: April 9, 2013. First revision date: June 11, 2013. Second revision date: August 26, 2013. Acceptance date: September 16, 2013. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. Relevant financial activities outside the submitted work: grants/grants pending, royalties, payment for development of educational presentations, and travel/ accommodations/meeting expenses. Address correspondence and reprint requests to lames M. Elliott, PT, PhD, Department of Physical Therapy and Human Movement Sciences, Northwestern University, Eeinberg School of Medicine, 645 N. Michigan Ave, Ste 1100, Chicago, IL 60611; E-mail: [email protected]ü DOI: 10.1097/BRS.000000OO0000O033 Spine

Conclusion. These findings clarify that previous reports of increased relative CSA in patients with chronic whiplash represent cervical muscle pseudohypertrophy. Relative muscle CSA measures reveal atrophy in several muscles in both patients with WAD and idiopathic neck pain, which supports inclusion of muscle conditioning in the total management of these patients. Keywords: muscle, whiplash, pain, magnetic resonance imaging, idiopathic, neck pain. Level of Evidence: 3 Spine 2014;39:39-47

P

ersistent pain-related disability in persons with chronic whiplash-associated disorders (WAD) constitutes a sizeable financial and economic cost for society, and conservative management of these patients remains a challenge.'-^ Nearly 50% of patients with whiplash will never fully recover^ and approximately 25% will develop chronic pain-related disability with complex clinical signs of poor functional recovery.'* Physical and psychological factors have been widely implicated in poor functional outcomes, including signs of central nervous system hyperexcitability (widespread hyperalgesia, sympathetic nervous system dysfunction) and post-traumatic stress.^^"** One physical factor that seems to contribute to or at least be associated with poor recovery after whiplash is magnetic resonance imaging (MRI) findings of cervical muscle degeneration (e.g., fatty inflltrates).'-'" These degenerative changes are present within weeks of the injury but only in those demonstrating poor functional recovery, suggesting the occurrence of a more severe injury.' Heightened levels of fatty infiltrate have been shown in both the dorsal" and ventral'- neck muscles with the deeper muscle layers, including the suboccipital muscles, being most affected." Heightened levels of cervical muscle fat are not present in those with milder symptoms from whiplash,' chronic nontraumatic neck pain," or in those with no history of neck pain,"-^^ suggesting that traumatic factors play a role. Although provisional at this stage, muscle degeneration may be an objective biological marker of the physical'" and psychological'"^'* factors commonly reported in patients with poor functional recovery."•'"'"'^ Coinciding with elevated levels of muscle fatty infiltrate are changes in neck muscle cross-sectional area (CSA). Contrary to expectations, the CSA of many of the cervical muscles in www.spinejournal.com

39

Spine

CERVICAL SPINE

patients with chronic WAD was significantly hypertrophied when compared with healthy controls.'^-^^ This hypertrophy could be attributed to physiological depositions of fat and connective tissue, similar to the characteristic pseudohypertrophy of the lower extremity muscles in patients with muscular dystrophy.'' Although interesting, findings of true hypertrophy of neck muscle are contrary to the usual and expected response of muscle atrophy after injury^""^* and at even further odds with recent reports of no observed changes in neck muscle size after whiplash.-'-'" Such uncertainties warrant precise examination of fatty infiltrates and reports of pseudohypertrophy of neck muscle in WAD. The purpose of this study was to further interpret our previous reports of increased muscle CSA in WAD'^-'* by reevaluating the data with specific measures of relative fat and relative muscle CSA. Comparisons of relative neck muscle CSA (rmCSA) with fat removed were performed between WAD, idiopathic neck pain, and healthy control populations. A secondary purpose of the article was to explore whether these new indices of CSA were related to specific patient population characteristics. This study represents a comprehensive comparison of neck muscle morphology in traumatic neck pain and control populations.

MATERIALS AND METHODS Study Design, Setting, and Participants A group-based cross-sectional study of 136 females who underwent magnetic resonance imaging of their cervical spine at dedicated research facilities in the United States and Australia. Participants included 79 individuals with chronic neck pain and disability after a motor vehicle collision, 23 individuals with chronic mechanical idiopathic neck pain (nontraumatic origin), and 34 healthy controls. Details of the participants are displayed in Table 1. Participants in the chronic WAD group all reported a history of chronic pain-related disability of 3-months to 3-years duration and were classified within the Quebec Task Force category of grade II (neck pain with limited cervical spine range of motion and point tenderness on the neck).'' Participants in the idiopathic group were included if they reported neck pain of at least 3-months duration that was not associated with a motor vehicle collision or other traumatic injury. The participants in the healthy control population were included if they reported no history of neck pain requiring treatment. Local and institutional ethics committees granted ethical approval for all studies, and written informed consent was obtained from all subjects prior to their inclusion.

MRI Protocol and Procedures For JMRI acquisition, participants in the United States were scanned using a conventional spin-echo pulse sequence (656 ms TR [repetition time] and 14 ms TE [echo time]) with a Horizon LX General Flectric 1.5 T scanner (Milwaukee, WI). In Australia, a SONATA 1.5 T magnet (448 ms TR and 14 ms TE) (Siemens, Erlangen, Germany) was used under an identical measurement methodology. Careful attention was taken 40

www.spinejournal.com

Differential Changes in Muscle Composition • Elliott et al

TABLE 1 .

, 1 filU Vt] 1] |]Tt|| Bra fel 1 CÏ'Él Ä%'| i l p l RTH Î ^ ^ ^ B B

iroious Onset Netk'Rai^n, and ::,«^Jj ;ontrol^Subiects.,;:.::!:.=H:«*«á*-.ítp==*5flHB Group

Age (yr) Mean (SD)

BMI (kg/m^)

Control

26.9(5.6)

23.0(4.4)

Idiopathic

29.2 (6.8)*

23.3 (4.8)

WAD

29.8(7.8)*

NDI (%)

Duration (mo)

21.9(7.3)

34.4 (20.5)

25.2 (5.7)*t 45.5(15.8)t

20.3 (9.6)t

'Significantly different from the control group at the level P < 0.001. tSignificantly different from the idiopathic group at the level P < 0.001. BMI indicates body mass index; NDi, neck disability index; WAD, whiplashassociated disorder.

to replicate the protocol used in Australia as exactly as possible, and there were no systematic differences found between images acquired between the 2 sites. Axial images of the cervical spine were obtained from the midpoint of the cerebellum through to the Tl segmental level to ensure comprehensive capture of the neck flexor and extensor muscles. A neutral position of the head/neck was visually determined by ensuring that a horizontal position of the forehead to the chin was parallel to the JVIRI table. A research assistant was present with all scans to ensure that the subject's head was in this position prior to and after placing the neck coil over the head and the neck. For image analysis, Tl-weighted magnetic resonance (JVIR) parameters were chosen to provide images of reasonable tissue contrast between fat and soft-aqueous skeletal muscle.'^ The images were securely stored on a laptop computer and analyzed post hoc with JMRIcro software (vi^ww.mricro. com).^- Analysis was accomplished by manually tracing defined regions of interest (ROI) within the fascial borders of the muscles. ROI were previously established from images obtained at the C2-C3 and C5-C6 intervertebral levels for the fiexor (longus capitis/longus coUi [LCa/LCo] collectively, sternocleidomastoid [SCM]), and extensor [cervical multifidus (JVIult), semispinalis cervicus (SCe) and capitis (SCa), and splenius capitis (SpC)] muscles (Figure 1). ROI were established for the suboccipital muscles (rectus capitis posterior minor [RCPmin] and major [RCPmaj]) at the C1-C2 intervertebral level.

Muscle Morphological Measurements The following sequence of measurements was undertaken for all images recorded for the flexor and extensor muscles at the C2-C3 and C5-C6 levels and the suboccipital extensor muscles at the C1-C2 level. For relative CSA (rCSA), the axial iVIR slices were positioned parallel to the C2-C3 intervertebral disc, which produced a slight measurement error for the CSA measures of muscles above and below. This was consistent for all muscles bilaterally at each vertebral segment and between each subject. As a result, rCSAs are reported. The rCSAs of the muscles were calculated by the number of pixels under each ROI January 2014

CERVICAL SPINE

Differential Changes in Muscle Composition • Elliott et al

at the C2 level, thus allowing comparisons between individuals. This measure, based on the contrasting pixel intensities of fat and muscle on the Tl-weighted imaging (bright and dark, respectively), permitted intramuscular fat to be quantified relative to a recognized area of intermuscular fat (Eigure 2A, B). The MEI measures have been shown to have high inter- and intrarater reliability."* Eor relative muscle GSA (rmCSA), to determine a value representative of the nonfat elements of the muscle and, therefore, potentially tnore representative of the contractile elements of the muscle, the rmCSA was calculated using the formula: rmCSA = rCSA X (1-MEI). Therefore, the fat portion of the muscle was subtracted from the rCSA measure.

DATA ANALYSIS

Figure 1. Outline of the bilateral sternocleidomastoid muscles, longus colli muscles, and extensor muscles (deep to superficial—multifidus, semispinalis cervicus, semispinalis capitis, and splenius capitis) on axial T1-weighted spin-echo images (TR/TE—448/14). TR indicates repetition time, TE, echo time.

in the X- (mm) and y-axes (mm) with MRIcro software. This process ensured consistent representation of rGSA measures for each muscle at the vertebral levels examined between the subjects and has been previously reported in detail.^^ Eor Muscle Eat Index (MEI), a measure of fat in muscle, which has been previously reported in detail,'''•'*'^^'^'' was created from the pixel intensity profile with MRIcro software. The measure quantified the pixel intensity profiles of individually traced muscle ROIs ratioed to that of the pixel intensity profile from a standardized region of intermuscular fat

Figure 2. A, Regions of interest of the left cervical multifidus (outlined in grey) on axial Tl-weighted spin-echo image (TR/TE—448/14) with histogram of pixel intensity profile and B, regions of interest (Grey circle) of intermuscular fat at the inferior level of C2 on axial T1 -weighted spin-echo image (TR/TE448/14) with histogram of pixel intensity profile. The same procedure was performed for each muscle at each vertebral level.TR indicates repetition time, TE, echo time. Spine

Eor rGSA and rmGSA, a linear mixed model was used to compare the WAD, idiopathic and control groups (factor = group) across all muscles (factor = muscles) and across suboccipital (for rectus capitis major and minor muscles only), and upper (G2-G3) and lower (G5-G6) cervical levels (factor = level). Data for left and right-side MEI, rGSA, and rmGSA were compared for each muscle. No statistically significant differences were observed (P > 0.05), and a mean of left and right data was used for further analysis. Based on previous findings,^-' body mass index (BMI) was included as a covariate and the partial and interactive effects of BMI on rGSA and rmGSA were explored. Initial model iterations included age as a covariate; however, no statistically significant partial or interactive effects of age on either rmGSA or rGSA were observed (P > 0.05), and age was subsequently removed as a factor in the final model iterations. Model validity was assessed for both rGSA and rmGSA through examination of the distribution of the residuals for normality and assessment of the association between the observed and predicted values using Pearson correlation coefficient and scatter plots of

4095

0

(A)

Pixel intensity

0

(B)

4095 Pixel intensity

www.spinejournal.com

41

Spine

CERVICAL SPINE

Differential Changes in Muscle Composition • Elliott et al

500 -SOO

< 300 • u •o 200 .a 100 O 100

200 300 Predicted rmCSA (mm2 )

400

500

100

200 300 Predicted rmCSA (nini2 )

400

500

500

'Î400 < 300 U •o 200 .2 100

o

Figure 3. Scatter plot of observed values versus predicted values based on linear mixed model analysis for rmCSA (top panel) and rCSA (bottom panel). rmCSA indicates relative muscle cross-sectional area.

predicted versus observed values for each model. Estimated marginal means of the final model (adjusted for BMI) were used to perform post hoc tests between groups at each cervical level for each muscle using a t statistic. For post hoc tests, a Bonferroni correction was applied to adjust for multiple comparisons. Significance levels were set at P value of less than 0.05 for all analyses. Data analyses were performed using PROC MIXED in SPSS version 20.0 (Chicago, IL). RESULTS Baseline demographic data for each group were presented in Table 1. The control participants were significantly younger than the WAD and idiopathic groups (P < 0.001). The WAD group had a significantly higher mean BMI than both the control and idiopathic groups (P < 0.001). Of the neck pain groups, the idiopathic group had a significantly longer duration of symptoms and significantly lower neck pain-related disability than the WAD group (P < 0.001 all comparisons). Using a linear mixed-model approach, BMI was found to have a significant interactive effect with level and muscle for both rCSA and rmCSA. Exploration of the differential effects of BMI on rCSA and rmCSA across level and muscle was not an aim of this study, and, therefore, this interactive effect was included as a random effect in the final model. The fixed effects included in the models for rmCSA and rCSA were, therefore, group, muscle and level, and BMI. The interactive effects 42

www.spinejournal.com

between group and muscle and level were also included. Eor both rCSA and rmCSA, all fixed effects included in the model were statistically significant (Table 2). Scatter plots of predicted versus observed values for each model showed good model validity (Eigure 3), and this was confirmed through excellent correlation between predicted and observed values (rCSA: r = 0.98, P < 0.001; rmCSA: r = 0.99, P < 0.001). The significance of the effect of the interaction between group and level and muscle for both rmCSA and rCSA indicates that between-group differences in muscle CSA were not consistent across muscles or within muscles across cervical levels. This finding indicates that general conclusions about changes in muscle CSA in patients with WAD and idiopathic neck pain in comparison with asymptomatic controls (i.e., general increase or reduction of CSA for all muscles or for all muscles at a cervical level) cannot be made at either the muscle or group level. Examination of between-group differences for each muscle at each level was, therefore, required to fully elucidate between-group differences in rmCSA and rCSA. Post hoc comparisons of the estimated marginal means for rCSA and rmCSA between groups for each muscle at each cervical level were, therefore, performed. Comparison of rCSA and rmCSA between groups at each muscle across each level using estimated marginal means from the linear mixed model analysis showed varying results as indicated in Tables 3 and 4 and Eigure 4A-G. Between-Group Findings for rCSA and rmCSA Eor WAD versus controls, compared with the control group, the rCSAs of the WAD group were significantly larger in 7 of 14 muscle regions (C5-C6; multifidus, C2-C3; and semispinalis capitis, splenius capitis, deep cervical flexors [DCF], and SCM), significantly smaller in 3 of 14 regions (C2-C3 and C5-C6 semispinalis cervicus, and C5-C6 semispinalis

TABLE 2. B u i t s

wr lests OT pxea tneetrror TOctors Included in the Linear Mixed Model An;ilvsi

Differential changes in muscle composition exist in traumatic and nontraumatic neck pain.

A population based cross-sectional study...
10MB Sizes 0 Downloads 0 Views