Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 Oct 1.
Published in final edited form as: J Electromyogr Kinesiol. 2012 Apr 18;22(5):724–731. doi: 10.1016/j.jelekin.2012.03.004

Association between history and physical examination factors and change in lumbar multifidus muscle thickness after spinal manipulation in patients with low back pain

Shane L Koppenhaver 1, Julie M Fritz 2, Jeffrey J Hebert 3, Greg N Kawchuk 4, Eric C Parent 5, Norman W Gill 6, John D Childs 7, Deydre S Teyhen 8
PMCID: PMC3407297  NIHMSID: NIHMS367720  PMID: 22516351

Abstract

Understanding the clinical characteristics of patients with low back pain (LBP) who display improved lumbar multifidus (LM) muscle function after spinal manipulative therapy (SMT) may provide insight into a potentially synergistic interaction between SMT and exercise. Therefore, the purpose of this study was to identify the baseline historical and physical examination factors associated with increased contracted LM muscle thickness one week after SMT. Eighty-one participants with LBP underwent a baseline physical examination and ultrasound imaging assessment of the LM muscle during submaximal contraction before and one week after SMT. The relationship between baseline examination variables and 1-week change in contracted LM thickness was assessed using correlation analysis and hierarchical multiple linear regression. Four variables best predicted the magnitude of increases in contracted LM muscle thickness after SMT. When combined, these variables suggest that patients with LBP, (1) that are fairly acute, (2) have at least a moderately good prognosis without focal and irritable symptoms, and (3) exhibit signs of spinal instability, may be the best candidates for a combined SMT and LSE treatment approach.

Keywords: musculoskeletal manipulations, low back pain, muscle contraction, ultrasound, lumbar multifidus

1. Introduction

Low back pain (LBP) is one of the most costly and prevalent medical conditions in the world (Dagenais et al., 2008; Walker, 2000). Despite many recent advances in imaging and surgical technology, identifying a specific pathoanatomical cause is not possible in the majority of LBP patients (Deyo & Weinstein, 2001; Deyo et al., 2009). This lack of a consistent relation between pathoanatomy and symptoms has instigated more recent attempts to classify LBP patients according to the intervention with which gives them the greatest benefit. (Fritz et al., 2007a; Hebert et al., 2008). For example, clinical prediction rules have been developed to identify subgroups of patients likely to respond to spinal manipulation therapy (SMT) (Childs et al., 2004; Flynn et al., 2002) and lumbar stabilization exercise (LSE) (Hicks et al., 2005). While these clinical prediction rules appear to identify unique subgroups of patients who preferentially respond to SMT or LSE, they remain at various stages of validation (Haskins et al., 2012). Moreover, other evidence suggests that combining SMT and LSE results in superior clinical outcomes than either intervention alone (2004).

A growing body of evidence has reported an association between functional deficits of the lumbar multifidus (LM) muscle and LBP (Dickx et al., 2010; Hungerford et al., 2003; MacDonald et al., 2006; Wallwork et al., 2008). In previous work, we have found a relationship between the clinical factors identifying LSE responders and deficits in the LM muscle as measured by ultrasound imaging (Hebert et al., 2010). Most recently we have found changes in contracted LM thickness were associated with improved LBP-related disability one week after SMT (Koppenhaver et al., 2011). Together these findings lend support to the hypothesis that SMT may provide a facilitatory stimulus (“jump start”) to the LM muscle, which may help initiate clinical recovery from LBP (Gill et al., 2007; Konitzer et al., 2011). Of interest, the changes that were observed in contracted LM thickness after SMT were not seen in all patients with LBP; rather the direction and extent of change was highly variable and dependent upon the individual (Koppenhaver et al., 2011).

With these observations, understanding the range of clinical characteristics that describe patients with LBP who display improved LM function after SMT may provide further insight into a potentially synergistic interaction between SMT and LSE. Therefore, the purpose of this study was to identify the baseline historical and physical examination factors associated with increased LM muscle thickness during submaximal contraction after SMT.

2. Methods

2.1. Design overview

Data for this paper was collected originally from a prospective cohort study that examined the relation between improved disability and changes in resting and contracted abdominal and LM muscle thickness following SMT in LBP patients. The full details of the study have been published elsewhere (Koppenhaver et al., 2011) and showed that increased contracted LM thickness predicted improved LBP-related disability one week after SMT. While we also observed similar changes in other muscles (transverse abdominis and internal oblique), they were transient and unrelated to improvements in LBP related disability.

2.2. Subjects

Subjects were recruited from two geographic locations: (1) by responding to flyers posted around the University of Utah campus, and (2) from the physical therapy department of Brooke Army Medical Center in San Antonio, Texas. The participant selection criteria are listed in Table 1.

Table 1.

Study selection criteria

Inclusion Criteria Exclusion Criteria
Back pain located between the 12th rib and buttocks, that in the opinion of the screening examiner, was originating from the lumbar region Neurogenic pain defined by either a positive ipsilateral or contralateral straight leg raise (reproduction of symptoms at ≤ 45°) or reflex, sensation, or strength deficits in a pattern consistent with nerve root compression
Between the age of 18 and 60 years Prior surgery to the lumbosacral spine
Meet either ≥ 4 or ≤ 2 out of 5 of the spinal manipulation clinical prediction rule criteria (Flynn et al., 2002):

  1. symptoms fewer than 16 days

  2. no symptoms distal to the knee

  3. Fear Avoidance Behavior Questionnaire work subscale score < 19 points

  4. ≥ 1 lumbar segment graded as hypomobile

  5. ≥ 35° hip internal rotation range of motion on at least 1 side

 Medical ‘red flags’ of a potentially serious condition including cauda equina syndrome, major or rapidly progressing neurological deficit, fracture, cancer, infection, or systemic disease
Ability to lie prone and supine for a minimum of 20 Minutes Prior spinal manipulation to the lumbosacral spine or trunk muscle stabilization exercises performed in the previous 4 weeks
Modified Oswestry Disability score at least 20% Osteoporosis
Open in a new tab

2.3. Procedures

Subjects attended three sessions within one week. Session 1 included self-report questionnaires, baseline history and physical examination, SMT treatment, and pre- and post-SMT measures of LM muscle thickness on ultrasound images. Session 2 occurred 3–4 days after session 1 and included an additional SMT treatment. Session 3 occurred 1 week after session 1 and included repeat subjective questionnaires and measurements of the LM muscle thickness on ultrasound images.

2.3.1 Baseline examination

Demographic information including age, sex, past medical history, smoking, height, and weight were collected by self-report. Additional information regarding history of participants’ LBP was collected via interview and included duration and anatomical distribution of current symptoms, frequency of prior episodes, aggravating and relieving factors, and prior treatments.

Self-report questionnaires included: (1) an 11-point Numeric Pain Rating Scale (NPRS), to rate subjective pain intensity (Childs et al., 2005), (2) modified Oswestry Disability Questionnaire (ODI), a LBP-specific subjective measure of disability (Fritz & Irrgang, 2001), (3) Fear-Avoidance Beliefs Questionnaire (FABQ), a subjective questionnaire designed to measure a users beliefs about the relationship between physical activity, work, and their LBP (Waddell et al., 1993) and (4) treatment expectations. Similar to the procedures of Kalauokalani et al. (Kalauokalani et al., 2001). we asked participants to respond on a 5-point Likert type scale about their beliefs that different treatments will improve their LBP as expectations of a treatment have been shown to affect outcomes for patients with LBP.

A standardized physical examination was performed on each participant including all tests and measures associated with the Treatment Based Classification System (Fritz et al., 2007a; Hebert et al., 2008) by an examiner blinded to measures of LM muscle thickness. Lumbosacral range of motion of flexion, extension, and side bending was measured using a standard inclinometer (Fritz & Piva, 2003) as was straight leg raise (Hicks et al., 2005) and hip internal range of motion (Flynn et al., 2002). Aberrant movements (e.g. painful arc, instability catch, difficult return from flexion, reversal of lumbopelvic rhythm) were noted during flexion (Hicks et al., 2003). A repeated motion exam was performed by having standing participants bend ten times as far as possible into flexion and extension followed by sustained prone extension. Changes in symptoms were documented in terms of intensity (more or less pain) and location (centralize or peripheralize) (Fritz et al., 2007c). Segmental mobility was assessed by having the examiner apply manual pressure on each lumbar spinous process in a posterior to anterior (PA) direction. Intervertebral motion was judged to be normal, hypomobile, or hypermobile and pain as present or absent at each segment (L1 to L5) (Hicks et al., 2003). The prone instability test (PIT) was then performed and considered positive if patients reported less pain with PA pressure when they held their legs off the ground as opposed to resting with their feet touching the floor (Hicks et al., 2003). The active straight leg raise test (ASLR) was performed and considered positive when supine participants reported less difficulty in raising their leg when the examiner manually stabilized their pelvis (Mens et al., 2002).

2.3.2 Ultrasound examination

LM thickness was assessed at baseline and after one week with B-mode ultrasound imaging using a protocol with established reliability and validity (Hebert et al., 2009; Koppenhaver et al., 2009b, 2009a) by an experienced examiner blinded to the findings of the history and physical examination. Images on the LM muscle were obtained using a Sonosite Titan or Sonosite MicroMaxx (Sonosite Inc, Bothell, WA), with a 60- mm, 5-MHz curvilinear. The subject was placed prone with the abdomen supported as needed to ensure no more than 10o of lumbar lordosis. Ultrasonic images were taken of the LM during submaximal contraction which was achieved through a contralateral arm raise and a small hand weight previously shown to elicit approximately 30% of the maximal voluntary isometric contraction of the LM muscle (Fig. 1) (Kiesel et al., 2007). Images were acquired three times each on the participant’s more symptomatic side and averaged to reduce measurement error (Koppenhaver et al., 2009c).

Fig. 1.

Fig. 1

Open in a new tab

Imaging of the lumbar multifidus (LM) muscle was performed during a contralateral arm raise using one of 3 possible hand weight selected based on the participants body weight.

Images were imported into a computer and measured offline using Image J software (Wayne Rasband, National Institutes of Health, USA). Thickness was measured as the distance between the posterior-most portion of the L4/L5 facet joint and the fascial plane between the muscle and subcutaneous tissue (Fig. 2). By using Image J’s automatic measurement function, the examiner was additionally blinded to thickness values during measurement. Percent change over 1 week in contracted LM thickness was calculated as [(thickness1-week − thicknessbaseline) / thicknessbaseline ]* 100%.

Fig. 2.

Fig. 2

Open in a new tab

Parasagital ultrasound image of the lumbar multifidus (LM) muscle. Measurements were taken between the posterior-most portion of the L4/L5 facet joint and the fascial plane between the muscle and subcutaneous tissue.

2.3.3 Spinal manipulation treatment

Participants received lumbosacral SMT during sessions 1 and 2 by a licensed physical therapist or chiropractor who were blinded to both the findings of the baseline examination and the ultrasound image measurements. The technique involved maximally side-bending and rotating participants, and then providing an anterior to posterior high-velocity low-amplitude thrust to the participants’ anterior superior iliac spine (Fig. 3) (Childs et al., 2004; Flynn et al., 2002). Each participant received SMT to both right and left sides, one or two times each. The examiner then recommended participants stay as active as possible and avoid specific trunk strengthening.

Fig. 3.

Fig. 3

Open in a new tab

Spinal manipulation technique. The participant was maximally side-bent away and then rotated toward the examiner. A high-velocity low-amplitude thrust was then given to the participant’s anterior superior iliac spine (ASIS) in a posterior and inferior direction.

2.4 Statistical Analysis

Potential predictor variables were chosen based upon their theoretical relationship to muscular responses following SMT. Initially, the bivariate relationships between these predictors and change in contracted LM thickness following SMT were assessed using correlation coefficients (point biserial correlations for dichotomous variables, spearman rho correlations for ordinal variables, and pearson product moment correlations for continuous variables). Continuous variables with skewed distributions were converted to ordinal variables based on rank. Alpha was liberally set at 0.15 for each comparison to minimize the potential for type 2 error.

The multivariate relationships between the baseline variables and change in contracted LM thickness were then evaluated using stepwise hierarchical multiple linear regression. Variables representing demographic and historical information that were significantly (p < 0.15) related to change in LM thickness were entered into the model in the first step. Statistically significant physical examination variables were entered into the second step. During each step, variables were entered and then removed using a backwards stepwise fashion with a significance value of less than 0.05 for model entry and greater than 0.10 for removal.

3. Results

Eighty-one individuals with LBP were enrolled. Three of the 81 participants did not complete the study, one for receiving an epidural steroid injection, one due to family illness, and one that was unable to be contacted. Additionally, investigators were unable to identify the LM muscle boundaries of 5 individuals leaving complete data on 73 participants. Overall participants experienced a mean (SD) improvement in ODI score of 26.3 (32.5)% and a mean (SD) increase of 1.1 (7.6)% in percent LM thickness. An increased LM thickness was observed in 39 (53.4%) of the participants and a decreased LM thickness was seen in 34 (46.6%) of the participants after one week. Individual changes in LM thickness for each participant are displayed in Fig. 4. As previously reported, an increase in contracted LM thickness one week pot-SMT was associated significantly with an improved ODI score. These changes explained only 7% of the variance in ODI changes over 1 week after accounting for sex, age and body mass index (Koppenhaver et al., 2011).

Fig. 4.

Fig. 4

Open in a new tab

One-week change in contracted lumbar multifidus (LM) thickness for each study participant. Bars represent change in LM thickness as a percentage of baseline.

Descriptive statistics of all potential predictor variables, as well as the bivariate correlations of each with change in LM thickness, are listed in Tables 2 and 3. The distribution of the variable symptom duration (number of days) appeared non-normal and was converted into both dichotomous and ordinal scales. Six weeks (45 days) was used as the threshold to dichotomize symptom duration (acute ≤ 6 weeks, chronic > 6 weeks) based on previous literature (Fritz et al., 2007b, 2008; van Tulder et al., 2006; Wand et al., 2004). Four demographic and history variables (Table 2) and seven physical examination variables (Table 3) were found to be correlated to one-week change in LM thickness (rpbis = −0.41 to 0.18, p < 0.15). Symptom duration when dichotomized (rpbis = −0.41, p < 0.01) and the presence of pain with PA pressure on L5 (rpbis = −0.37, p < 0.01), demonstrated the strongest correlations with change in LM thickness. Specifically, larger increases in LM thickness changes were observed in acute participants and those that had no pain with PA pressure at L5.

Table 2.

Demographic and Historical Information Variables (N = 73)

Descriptive
statistics		 Correlation
Estimate		 p-value
Dichotomous Variables Frequency Point Biserial
Sex 56.2% female 0.08 0.49
Symptom duration (acute vs. chronic) 72.6% chronic −0.41 <0.01*
Smoking 8.2% yes −0.10 0.39
Current or history of anxiety disorder and/or depression 27.4% yes −0.10 0.41
Previous PT or chiropractic treatment 31.5% yes −0.19 0.11*
Pain below buttocks 28.8% yes −0.06 0.63
Pain below knee 21.9% yes −0.06 0.61
Aggravated by bending 60.3% yes 0.00 0.99
Aggravated by sitting 82.2% yes 0.04 0.73
Aggravated by walking 21.9% yes −0.02 0.86
Ordinal Variables Median/Mode Spearman Rho
Duration of symptoms (days) 154 (31, 709)† −0.20 0.10*
Oswestry Disability Score (%) 30 (22, 37)† −0.20 0.10*
Amount of missed work due to LBP 63.0% None†† −0.06 0.60
Believe SMT will help LBP 57.1%
Somewhat
agree†† 		0.08 0.50
Number of prior episodes of LBP 60.3% ≥ 10†† −0.11 0.35
Continuous Variables Mean (SD) Pearson’s
Age (years) 33.7 (12.3) −0.10 0.40
BMI (kg/m2) 25.3 (5.2) −0.15 0.22
FABQ- PA (0–42) 15.3 (4.3) 0.15 0.20
FABQ- work (0–24) 16.2 (9.4) −0.11 0.37
Numeric pain rating scale for back (0–10)** 5.0 (1.7) 0.03 0.78
Numeric pain rating scale for leg (0–10)** 1.4 (2.2) −0.17 0.15
Open in a new tab

PT: physical therapy, LBP: low back pain, SMT: spinal manipulative therapy, BMI: body mass index, kg/m: kilograms/meters, FABQ-PA: fear avoidance behavior questionnaire physical activity subscale

*

statistically significant at the 0.15 level

**

Reports the average of the worst, best and current scores for pain over the last 24 hours.

†

Median (interquartile range).

††

Most common response (Mode)

Table 3.

Physical Examination Variables

Physical Examination Variables Central
Tendency &
Variance		 Correlation
Estimate		 p-value
Dichotomous Variables Percent Point
Biserial
Pain with PA L4 73.6% yes −0.18 0.12*
Pain with PA L5 72.2% yes −0.37 <0.01*
Aberrant Movements 35.6% yes 0.16 0.19
Prone instability test 55.4% positive −0.17 0.18
Active straight leg raise test 50.7% positive 0.18 0.12*
Ordinal Variables Mode Spearman Rho
Mobility with PA L4† 50.7% hypo −0.02 0.85
Mobility with PA L5† 54.2% hypo −0.04 0.73
Pain Intensity with repeated flexion ROM†† 55.5% increased −0.23 0.05*
Pain Location with repeated flexion ROM††† 88.6% no effect −0.22 0.07*
Pain Intensity with repeated extension ROM†† 65.7% increased −0.10 0.40
Pain Location with repeated extension ROM††† 85.7% no effect −0.32 0.01*
Continuous Variables Mean (SD) Pearson’s
Number SMT criteria met (out of 5) 2.8 (1.3) 0.15 0.20
Number LSE criteria met (out of 4) 1.8 (1.1) 0.09 0.46
Flexion ROM (deg) 90.7 (27.5) −0.08 0.52
Extension ROM 29.5 (9.7) −0.07 0.56
Asymmetry in SB ROM (deg) 4.8 (4.0) −0.06 0.63
Average SLR ROM 75.6 (16.3) −0.23 0.05*
Average Hip IR ROM 35.3 (8.3) 0.00 0.99
Open in a new tab

PA: posterior-to-anterior, ROM: range of motion, deg: degrees, SMT: spinal manipulative therapy, LSE: lumbar stabilization exercise, SB: side bending, SLR: straight leg raise, IR: internal rotation

*

statistically significant at the 0.15 level

†

Coded as 1=hypomobile, 2=normal, 3=hypermobile

††

Coded as 1=decreased pain, 2=no effect, 3=increased pain

†††

Coded as 1=centralize, 2=no effect, 3=peripheralize

Results of the hierarchical linear regression analysis are presented in Table 4. Four of the 14 variables uniquely contributed to the multivariate model (p < 0.10). The strongest contributors to the model were LBP acuity and lack of pain response to posterior-to-anterior pressure on L5. Having a positive ASLR test, and symptoms that centralize, rather than peripheralize with repeated lumbar extension were also associated with larger increases in LM thickness. The four variables together resulted in an adjusted R2 of 0.27, representing 27% of the variance of change in LM thickness after adjusting for the number of variables in the model.

TABLE 4.

Hierarchical linear regression analysis predicting 1-week change in lumbar multifidus (LM) muscle thickness after spinal manipulation.

Variables Retained in Final Model Standardized β
coefficient		 Significance of
β coefficient		 Adjusted
R2 		Significance
of model fit
Acute low back pain 0.29 0.02
No pain with PA pressure to L5 0.29 0.01
Positive ASLR test 0.20 0.09
Centralize rather than peripheralize 0.20 0.07
with repeated extension
   Full Model 0.27 <0.001
Open in a new tab

PA: posterior-to-anterior, ASLR: active straight leg raise

4. Discussion

Previous work has identified the clinical characteristics of patients with LBP who are likely to experience success with SMT (Childs et al., 2004; Flynn et al., 2002) and LSE (Hicks et al., 2005). Other evidence, however, suggests that the best clinical outcome may occur when using the combination of SMT and LSE to treat patients with LBP (2004). The purpose of this study was to identify the baseline historical and physical examination factors associated with increased contracted LM muscle thickness after SMT in an effort to better understand a potentially synergistic interaction between SMT and LSE. The current analysis identified four variables comprising the most parsimonious set of predictors for LM thickness changes after SMT. Having acute rather than chronic LBP, no pain with PA pressure to L5, a positive ASLR test, and symptoms that centralize rather than peripheralize with repeated lumbar extension were associated with larger increases in contracted LM thickness. Knowing these variables may help us determine what type of patient with LBP might be expected to benefit from a combination of SMT and LSE as well as lend support to the hypothesis that SMT may “jump start” the LM muscle in these individuals (Gill et al., 2007; Konitzer et al., 2011).

Having acute rather than chronic LBP was the variable most strongly associated with increased LM thickness one week after SMT, both by itself and as part of the multivariate model. Symptom acuity has previously been found to be strongly prognostic for recovery of LBP both in general (Hancock et al., 2009) and specifically in patients that have been treated with SMT (Childs et al., 2004; Flynn et al., 2002; Fritz et al., 2005). The reason that patients with acute LBP have larger thickness changes than those with chronic LBP may be that chronic LBP often includes psychosocial components (Keeley et al., 2008), structural changes (atrophy and fatty infiltrate) (Yoshihara et al., 2001; Zhao et al., 2000) and/or central pain mechanisms (including neural adaptations and pain sensitization) that are likely less affected by SMT (Sharma et al., 2011; Young et al., 2003). In comparison, acute LBP largely concerns local and/or peripheral mechanisms, including neuromotor inhibition (Dickx et al., 2010; Kiesel et al., 2008), that can be altered by SMT (Bialosky et al., 2009).

The centralization phenomenon, like symptom acuity, has been shown to be strongly prognostic for recovery of LBP (Werneke et al., 2009) and occurs most commonly with repeated lumbar extension (Aina et al., 2004). The fact that we observed larger changes in LM thickness in participants who centralized with repeated lumbar extension may be due to an increased ability to contract the LM muscle due to their higher rate of clinical recovery. Because we did not include a control group, we cannot know whether these two prognostic factors (symptom acuity and centralization) predicted a larger LM response to SMT or merely quicker recovery regardless of treatment.

The ASLR test was initially advocated for use with pelvic girdle pain after pregnancy (Mens et al., 2002, 1999). More recently it’s been used as a test of general lumbar spine instability in patients with LBP and has been demonstrated to vary with both lumbar spine stability (Liebenson et al., 2009) and transverse abdominis muscle activity (Teyhen et al., 2009). The fact that participants who had a positive ASLR test demonstrated larger LM thickness changes after SMT may possibly be explained by the presence of lumbopevic motorcontrol deficits at baseline that were to some extent reversed by SMT.

Finally, the reason for an association between no pain with PA pressure at L5 with larger changes in LM thickness in the multivariate model is more difficult to ascertain. This finding is somewhat counterintuitive as pain with PA pressure at L5 would seem to indicate the presence of local L5 area irritation and/or dysfunction that would be indicative of treatment with regional SMT. Alternatively, it is possible that pain with PA pressure on L5 signified more severe irritation and/or pathology in the region. Evidence suggests that patients with nerve root/sciatic pain have a worse prognosis than patients with non-specific LBP (Bronfort et al., 2004), especially with SMT (Axen et al., 2005; Malmqvist et al., 2008). And while our screening process likely excluded individuals with more severe nerve root pathology, it is possible that pain with PA pressure on L5 identified participants with a higher degree of local pathology, like nerve root inflammation, that demonstrated less physiologic change to SMT.

Only one of the identified variables (symptom acuity) approximated a variable previously found to be predictive of clinical success with SMT (Childs et al., 2004; Flynn et al., 2002) or LSE (Hicks et al., 2005). This may be because the subgroup of patients that experiences the largest change in LM thickness after SMT is somewhat distinct from the subgroups of patients expected to experience clinical success after SMT or LSE alone. Alternatively, this could be due to the common overlap between treatment-based classifications, especially the SMT and LSE categories (Stanton et al., 2011). In the current study we purposefully avoided dichotomizing most of our variables, as we believe it would be premature to develop another clinical prediction rule or distinct subgroup of patients. Instead our focus was to describe the general type of LBP patient that might exhibit a meaningful increase in contracted LM thickness after SMT, and therefore, be an especially good candidate for SMT followed by LSE. Our results suggest that patients with LBP, (1) that are fairly acute, (2) have at least a moderately good prognosis without focal and irritable symptoms, and (3) exhibit signs of spinal instability, may be the best candidates for such a combined treatment approach.

Limitations of this study have been discussed elsewhere (Koppenhaver et al., 2011) including concerns for using ultrasound rather than electromyography to measure muscle function and using flyers to recruit participants rather than recruiting those seeking healthcare. Perhaps more important to the current study was and the fact that patients were only observed for one week and were not treated with LSE. Therefore any conjecture about the interaction between SMT and LSE is purely speculation that needs to be investigated in future research. Additionally the lack of the inclusion of a control group has been criticized in clinical prediction studies as it makes it impossible to ascertain whether the identified factors predicted a larger LM response to SMT or merely a larger LM response in general (Haskins et al., 2012).

Because we did not include a control group, we cannot know whether these two prognostic factors (symptom acuity and centralization) predicted a larger LM response to SMT or merely quicker recovery regardless of treatment.

Future studies should explore the interaction between SMT and LSE by measuring changes in LM muscle function during such interventions across longer time frames. It should additionally be noted that there might have been important predictors of LM thickness change that were not included in this study. Although the historical and physical examination used in this study was comprehensive and approximated a routine LBP examination, the analysis performed was retrospective in nature and did not include some potentially important predictors such as pain level during the contralateral arm raise procedure or manual assessment of LM contraction. Lastly, many statistical comparisons were performed in the current study using a liberal alpha level. As a result, it is possible that some of our findings may be spurious and attributable to Type I error. Our approach, however, is consistent with the exploratory nature of the study, as we were interested in identifying potentially important relationships that could be considered in future research regarding muscular changes associated with SMT.

5. Conclusions

Four baseline historical and physical examination variables best predicted the magnitude of increases in contracted LM thickness one week after SMT. When combined, these variables suggest that patients with LBP, (1) that are fairly acute, (2) have at least a moderately good prognosis without focal and irritable symptoms, and (3) exhibit signs of spinal instability, may be the best candidates for a combined SMT and LSE treatment approach. Such findings provide insight into the potentially synergistic interaction between SMT and LSE and may provide guidance for future studies regarding such a combined treatment approach.

Acknowledgement

The data for this study was collected using funding from the National Institutes of Health, National Center for Complementary and Alternative Medicine (R21 AT004221).

The authors would like to thank Dr. Leslie Hair and Mr. Faris Al-Odaibi for their assistance with data collection.

This study was approved by the Institutional Review Boards of the University of Utah and Brooke Army Medical Center.

The view(s) expressed herein are those of the author(s) and do not reflect the official policy or position of Brooke Army Medical Center, the U.S. Army Medical Department, the U.S. Army Office of the Surgeon General, the Department of the Army, Department of the Air Force, Department of Defense, or the U.S. Government.

Biographies

graphic file with name nihms367720b1.gif

Shane Koppenhaver is an Assistant Professor and the Research Director of the U.S. Army/Baylor University Doctoral Program in Physical Therapy. He received his Masters of Physical Therapy degree from the U.S. Army/Baylor University Graduate Program in 1998, and a PhD in Exercise Physiology from the University of Utah in 2009. He is a board-certified Orthopaedic Clinical Specialist and a Fellow in the American Academy of Orthopaedic and Manual Physical Therapists. His primary research interests concern mechanistic and clinical outcomes associated with manual therapy and dry needing, especially as they apply to clinical reasoning and management of patients with neuromuskuloskeletal conditions.

graphic file with name nihms367720b2.gif

Dr. John D. Childs is the Director of Musculoskeletal Research with the 81st Medical Group at Keesler Air Force Base and an Associate Professor in the U.S. Army-Baylor University Doctoral Program in Physical Therapy. He is a board-certified Orthopaedic Clinical Specialist and a Fellow in the American Academy of Orthopaedic and Manual Physical Therapists. Dr. Childs completed his PhD in Rehabilitation Science from the University of Pittsburgh. Dr. Childs has made numerous scientific presentations at national and international meetings, and his research agenda is focused on injury prevention, the identification of subgroups of patients with musculoskeletal disorders, the development of clinical prediction rules to better inform diagnostic and treatment decision-making, and comparative effectiveness research. He has received numerous research grants from federal and professional funding agencies and has published over 100 peer-reviewed manuscripts in leading scientific journals. Dr. Childs currently serves as an Associate Editor for both the Physical Therapy and Journal of Orthopaedic and Sports Physical Therapy journals. He has received numerous awards including the American Physical Therapy Associations’ Eugene Michels New Investigator Award, Jack Walker Research Award, Chattanooga Research Award, Rose Excellence in Research Award, and the Air Force’s Outstanding Achievement in Clinical Research Award and Physical Therapist of the Year (twice).

graphic file with name nihms367720b3.gif

Deydre S. Teyhen completed her PhD in Biomechanics from the University of Texas in 2004. She received a Master and Doctoral Degree in Physical Therapy from the U.S. Army-Baylor University in 1995 and 2008 and a Bachelor of Arts in Sports Science from Ohio Wesleyan University in 1993. Dr. Teyhen’s professional experience and training have been primarily in orthopaedic physical therapy. She is a board-certified Orthopaedic Clinical Specialist from the American Board of Physical Therapy Specialties, a certified Health Fitness Instructor from the American College of Sports Medicine, and a certified Health Promotion Director through the Cooper Institute. Dr. Teyhen’s primary research interests involve the rehabilitation of low back pain, the utilization of imaging to enhance rehabilitation, and prevention of musculoskeletal injuries. She has authored over 40 peer-reviewed publications in physical therapy, rehabilitation, orthopedic, and sports medicine journals. Dr. Teyhen’s research programs have been supported by Baylor University, Department of the Army, Department of Defense, and the National Institutes of Health.

graphic file with name nihms367720b4.gif

Norman Gill is the Director and Associate Professor of the Army-Baylor University Doctoral Fellowship in Orthopaedic Manual Physical Therapy. He holds an undergraduate degree from the U.S. Military Academy at West Point, and a Master of Physical Therapy and Doctor of Science from Baylor University in Texas. He is board certified in Orthopaedic Physical Therapy and a Fellow of the American Academy of Orthopaedic Manual Physical Therapists. He completed his postdoctoral research fellowship at the Walter Reed Army Medical Center Spine Biomechanics Lab, and his research interests are in spine imaging, manual/manipulative therapy, and osteoarthritis.

graphic file with name nihms367720b5.gif

Greg Kawchuk is an Associate Professor and Canada Research Chair in Spinal Function in the Department of Physical Therapy at the University of Alberta. Greg obtained his chiropractic training from the Canadian Memorial Chiropractic College in 1990 and PhD in bioengineering from the University of Calgary in 2000. Dr. Kawchuk’s research is aimed at defining the mechanisms that initiate and sustain spinal disorders with the long-term goal of creating meaningful strategies to prevent, diagnose and treat back pain. From his work, he has developed several novel technologies to measure spinal function that are now being used in clinical trials.

graphic file with name nihms367720b6.gif

Eric Parent is Assistant-Professor in the Department of Physical Therapy at the University of Alberta with a joint appointment as clinical-scientist at the Glenrose Rehabilitation Hospital within Alberta Health Services. Eric obtained a bachelor of Physiotherapy at in 1996, then completed a master in experimental medicine in 2000, both at Université Laval in Québec City. He completed a PhD Rehabilitation Science at University of Alberta on developing quantitative measures of disc contour abnormality on MRI images of the lumbar spine. His post-doctoral and recent work focuses on identifying predictors of good outcomes in response to repeated movement therapeutic exercises for the spine in patients with low back pain and to Schroth exercises for scoliosis. Dr Parent uses MRI and US imaging to study the mechanisms of action of therapeutic exercises for spinal disorders.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Aina A, May S, Clare H. The centralization phenomenon of spinal symptoms--a systematic review. Man Ther. 2004;9(3):134–143. doi: 10.1016/j.math.2004.03.004. [DOI] [PubMed] [Google Scholar]
  2. Axen I, Jones JJ, Rosenbaum A, Lovgren PW, Halasz L, Larsen K, et al. The Nordic Back Pain Subpopulation Program: validation and improvement of a predictive model for treatment outcome in patients with low back pain receiving chiropractic treatment. J Manipulative Physiol Ther. 2005;28(6):381–385. doi: 10.1016/j.jmpt.2005.06.008. [DOI] [PubMed] [Google Scholar]
  3. Bialosky JE, Bishop MD, Price DD, Robinson ME, George SZ. The mechanisms of manual therapy in the treatment of musculoskeletal pain: a comprehensive model. Man Ther. 2009;14(5):531–538. doi: 10.1016/j.math.2008.09.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bronfort G, Evans RL, Maiers M, Anderson AV. Spinal manipulation, epidural injections, and self-care for sciatica: a pilot study for a randomized clinical trial. J Manipulative Physiol Ther. 2004;27(8):503–508. doi: 10.1016/j.jmpt.2004.08.002. [DOI] [PubMed] [Google Scholar]
  5. Childs JD, Fritz JM, Flynn TW, Irrgang JJ, Johnson KK, Majkowski GR, et al. A clinical prediction rule to identify patients with low back pain most likely to benefit from spinal manipulation: a validation study. Ann Intern Med. 2004;141(12):920–928. doi: 10.7326/0003-4819-141-12-200412210-00008. [DOI] [PubMed] [Google Scholar]
  6. Childs JD, Piva SR, Fritz JM. Responsiveness of the numeric pain rating scale in patients with low back pain. Spine. 2005;30:1331–1335. doi: 10.1097/01.brs.0000164099.92112.29. [DOI] [PubMed] [Google Scholar]
  7. Dagenais S, Caro J, Haldeman S. A systematic review of low back pain cost of illness studies in the United States and internationally. Spine J. 2008;8(1):8–20. doi: 10.1016/j.spinee.2007.10.005. [DOI] [PubMed] [Google Scholar]
  8. Deyo RA, Weinstein JN. Low back pain. N Eng J Med. 2001;344(5):363–369. doi: 10.1056/NEJM200102013440508. [DOI] [PubMed] [Google Scholar]
  9. Deyo RA, Mirza SK, Turner JA, Martin BI. Overtreating chronic back pain: time to back off? J Am Board Fam Med. 2009;22(1):62–68. doi: 10.3122/jabfm.2009.01.080102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dickx N, Cagnie B, Parlevliet T, Lavens A, Danneels L. The effect of unilateral muscle pain on recruitment of the lumbar multifidus during automatic contraction. An experimental pain study. Man Ther. 2010;15(4):364–369. doi: 10.1016/j.math.2010.02.002. [DOI] [PubMed] [Google Scholar]
  11. Flynn T, Fritz J, Whitman J, Wainner R, Magel J, Rendeiro D, et al. A clinical prediction rule for classifying patients with low back pain who demonstrate short-term improvement with spinal manipulation. Spine. 2002;27(24):2835–2843. doi: 10.1097/00007632-200212150-00021. [DOI] [PubMed] [Google Scholar]
  12. Fritz JM, Childs JD, Flynn TW. Pragmatic application of a clinical prediction rule in primary care to identify patients with low back pain with a good prognosis following a brief spinal manipulation intervention. BMC Fam Pract. 2005;6(1):29. doi: 10.1186/1471-2296-6-29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fritz JM, Cleland JA, Childs JD. Subgrouping patients with low back pain: evolution of a classification approach to physical therapy. J Orthop Sports Phys Ther. 2007a;37(6):290–302. doi: 10.2519/jospt.2007.2498. [DOI] [PubMed] [Google Scholar]
  14. Fritz JM, Cleland JA, Brennan GP. Does adherence to the guideline recommendation for active treatments improve the quality of care for patients with acute low back pain delivered by physical therapists? Med Care. 2007b;45(10):973–980. doi: 10.1097/MLR.0b013e318070c6cd. [DOI] [PubMed] [Google Scholar]
  15. Fritz JM, Cleland JA, Speckman M, Brennan GP, Hunter SJ. Physical therapy for acute low back pain: associations with subsequent healthcare costs. Spine. 2008;33(16):1800–1805. doi: 10.1097/BRS.0b013e31817bd853. [DOI] [PubMed] [Google Scholar]
  16. Fritz JM, Lindsay W, Matheson JW, Brennan GP, Hunter SJ, Moffit SD, et al. Is there a subgroup of patients with low back pain likely to benefit from mechanical traction? Results of a randomized clinical trial and subgrouping analysis. Spine. 2007c;32(26):E793–E800. doi: 10.1097/BRS.0b013e31815d001a. [DOI] [PubMed] [Google Scholar]
  17. Fritz JM, Piva SR. Physical impairment index: reliability, validity, and responsiveness in patients with acute low back pain. Spine (Phila Pa 1976) 2003;28(11):1189–1194. doi: 10.1097/01.BRS.0000067270.50897.DB. [DOI] [PubMed] [Google Scholar]
  18. Fritz JM, Irrgang JJ. A Comparison of a Modified Oswestry Disability Questionnaire and the Quebec Back Pain Disability Scale. Phys Ther. 2001;81(2):776–788. doi: 10.1093/ptj/81.2.776. [DOI] [PubMed] [Google Scholar]
  19. Gill NW, Teyhen DS, Lee IE. Improved contraction of the transversus abdominis immediately following spinal manipulation: a case study using real-time ultrasound imaging. Man Ther. 2007;12(3):280–285. doi: 10.1016/j.math.2006.06.014. [DOI] [PubMed] [Google Scholar]
  20. Hancock MJ, Maher CG, Latimer J, Herbert RD, McAuley JH. Can rate of recovery be predicted in patients with acute low back pain? Development of a clinical prediction rule. Eur J Pain. 2009;13(1):51–55. doi: 10.1016/j.ejpain.2008.03.007. [DOI] [PubMed] [Google Scholar]
  21. Haskins R, Rivett DA, Osmotherly PG. Clinical prediction rules in the physiotherapy management of low back pain: a systematic review. Man Ther. 2012;17(1):9–21. doi: 10.1016/j.math.2011.05.001. [DOI] [PubMed] [Google Scholar]
  22. Hebert J, Koppenhaver S, Fritz J, Parent E. Clinical prediction for success of interventions for managing low back pain. Clin Sports Med. 2008;27(3):463–479. ix–x. doi: 10.1016/j.csm.2008.03.002. [DOI] [PubMed] [Google Scholar]
  23. Hebert JJ, Koppenhaver SL, Magel JS, Fritz JM. The relationship of transversus abdominis and lumbar multifidus activation and prognostic factors for clinical success with a stabilization exercise program: a cross-sectional study. Arch Phys Med Rehabil. 2010;91(1):78–85. doi: 10.1016/j.apmr.2009.08.146. [DOI] [PubMed] [Google Scholar]
  24. Hebert JJ, Koppenhaver SL, Parent EC, Fritz JM. A systematic review of the reliability of rehabilitative ultrasound imaging for the quantitative assessment of the abdominal and lumbar trunk muscles. Spine (Phila Pa 1976) 2009;34(23):E848–E856. doi: 10.1097/BRS.0b013e3181ae625c. [DOI] [PubMed] [Google Scholar]
  25. Hicks GE, Fritz JM, Delitto A, McGill SM. Preliminary development of a clinical prediction rule for determining which patients with low back pain will respond to a stabilization exercise program. Arch Phys Med Rehabil. 2005;86:1753–1762. doi: 10.1016/j.apmr.2005.03.033. [DOI] [PubMed] [Google Scholar]
  26. Hicks GE, Fritz JM, Delitto A, Mishock J. The reliability of clinical examination measures used for patients with suspected lumbar segmental instability. Arch Phys Med Rehabil. 2003;84:1858–1864. doi: 10.1016/s0003-9993(03)00365-4. [DOI] [PubMed] [Google Scholar]
  27. Hungerford B, Gilleard W, Hodges P. Evidence of altered lumbopelvic muscle recruitment in the presence of sacroiliac joint pain. Spine. 2003;28(14):1593–1600. [PubMed] [Google Scholar]
  28. Kalauokalani D, Cherkin DC, Sherman KJ, Koepsell TD, Deyo RA. Lessons from a trial of acupuncture and massage for low back pain: patient expectations and treatment effects. Spine. 2001;26(13):1418–1424. doi: 10.1097/00007632-200107010-00005. [DOI] [PubMed] [Google Scholar]
  29. Keeley P, Creed F, Tomenson B, Todd C, Borglin G, Dickens C. Psychosocial predictors of health-related quality of life and health service utilisation in people with chronic low back pain. Pain. 2008;135(1–2):142–150. doi: 10.1016/j.pain.2007.05.015. [DOI] [PubMed] [Google Scholar]
  30. Kiesel KB, Uhl T, Underwood FB, Nitz AJ. Rehabilitative ultrasound measurement of select trunk muscle activation during induced pain. Man Ther. 2008;13(2):132–138. doi: 10.1016/j.math.2006.10.003. [DOI] [PubMed] [Google Scholar]
  31. Kiesel KB, Uhl TL, Underwood FB, Rodd DW, Nitz AJ. Measurement of lumbar multifidus muscle contraction with rehabilitative ultrasound imaging. Man Ther. 2007;12(2):161–166. doi: 10.1016/j.math.2006.06.011. [DOI] [PubMed] [Google Scholar]
  32. Konitzer LN, Gill NW, Koppenhaver SL. Investigation of Abdominal Muscle Thickness Changes After Spinal Manipulation in Patients Who Meet a Clinical Prediction Rule for Lumbar Stabilization. J Orthop Sports Phys Ther. 2011 doi: 10.2519/jospt.2011.3685. [DOI] [PubMed] [Google Scholar]
  33. Koppenhaver SL, Fritz JM, Hebert JJ, Kawchuk GN, Childs JD, Parent EC, et al. Association between changes in abdominal and lumbar multifidus muscle thickness and clinical improvement after spinal manipulation. J Orthop Sports Phys Ther. 2011;41(6):389–399. doi: 10.2519/jospt.2011.3632. [DOI] [PubMed] [Google Scholar]
  34. Koppenhaver SL, Hebert JJ, Fritz JM, Parent EC, Teyhen DS, Magel JS. Reliability of rehabilitative ultrasound imaging of the transversus abdominis and lumbar multifidus muscles. Arch Phys Med Rehabil. 2009a;90(1):87–94. doi: 10.1016/j.apmr.2008.06.022. [DOI] [PubMed] [Google Scholar]
  35. Koppenhaver SL, Hebert JJ, Parent EC, Fritz JM. Rehabilitative ultrasound imaging is a valid measure of trunk muscle size and activation during most isometric sub-maximal contractions: a systematic review. Aust J Physiother. 2009b;55(3):153–169. doi: 10.1016/s0004-9514(09)70076-5. [DOI] [PubMed] [Google Scholar]
  36. Koppenhaver SL, Parent EC, Teyhen DS, Hebert JJ, Fritz JM. The effect of averaging multiple trials on measurement error during ultrasound imaging of transversus abdominis and lumbar multifidus muscles in individuals with low back pain. J Orthop Sports Phys Ther. 2009c;39(8):604–611. doi: 10.2519/jospt.2009.3088. [DOI] [PubMed] [Google Scholar]
  37. Liebenson C, Karpowicz AM, Brown SH, Howarth SJ, McGill SM. The active straight leg raise test and lumbar spine stability. PM R. 2009;1(6):530–535. doi: 10.1016/j.pmrj.2009.03.007. [DOI] [PubMed] [Google Scholar]
  38. MacDonald DA, Moseley GL, Hodges PW. The lumbar multifidus: does the evidence support clinical beliefs? Man Ther. 2006;11(4):254–263. doi: 10.1016/j.math.2006.02.004. [DOI] [PubMed] [Google Scholar]
  39. Malmqvist S, Leboeuf-Yde C, Ahola T, Andersson O, Ekstrom K, Pekkarinen H, et al. The Nordic back pain subpopulation program: predicting outcome among chiropractic patients in Finland. Chiropr Osteopat. 2008;16:13. doi: 10.1186/1746-1340-16-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mens JM, Vleeming A, Snijders CJ, Koes BW, Stam HJ. Validity of the active straight leg raise test for measuring disease severity in patients with posterior pelvic pain after pregnancy. Spine. 2002;27(2):196–200. doi: 10.1097/00007632-200201150-00015. [DOI] [PubMed] [Google Scholar]
  41. Mens JM, Vleeming A, Snijders CJ, Stam HJ, Ginai AZ. The active straight leg raising test and mobility of the pelvic joints. Eur Spine J. 1999;8(6):468–473. doi: 10.1007/s005860050206. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Sharma NK, McCarson K, Van Dillen L, Lentz A, Khan T, Cirstea CM. Primary somatosensory cortex in chronic low back pain - a H-MRS study. J Pain Res. 2011;4:143–150. doi: 10.2147/JPR.S19297. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Stanton TR, Fritz JM, Hancock MJ, Latimer J, Maher CG, Wand BM, et al. Evaluation of a treatment-based classification algorithm for low back pain: a cross-sectional study. Phys Ther. 2011;91(4):496–509. doi: 10.2522/ptj.20100272. [DOI] [PubMed] [Google Scholar]
  44. Teyhen DS, Williamson JN, Carlson NH, Suttles ST, O'Laughlin SJ, Whittaker JL, et al. Ultrasound characteristics of the deep abdominal muscles during the active straight leg raise test. Arch Phys Med Rehabil. 2009;90(5):761–767. doi: 10.1016/j.apmr.2008.11.011. [DOI] [PubMed] [Google Scholar]
  45. United Kingdom back pain exercise and manipulation (UK BEAM) randomised trial: effectiveness of physical treatments for back pain in primary care. BMJ. 2004;329(7479):1377. doi: 10.1136/bmj.38282.669225.AE. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. van Tulder M, Becker A, Bekkering T, Breen A, del Real MT, Hutchinson A, et al. Chapter 3. European guidelines for the management of acute nonspecific low back pain in primary care. Eur Spine J. 2006;15(Suppl 2):S169–S191. doi: 10.1007/s00586-006-1071-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Waddell G, Newton M, Henderson I, Somerville D, Main CJ. A Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and disability. Pain. 1993;52:157–168. doi: 10.1016/0304-3959(93)90127-B. [DOI] [PubMed] [Google Scholar]
  48. Walker BF. The prevalence of low back pain: a systematic review of the literature from 1966 to 1998. J Spinal Disord. 2000;13(3):205–217. doi: 10.1097/00002517-200006000-00003. [DOI] [PubMed] [Google Scholar]
  49. Wallwork TL, Stanton WR, Freke M, Hides JA. The effect of chronic low back pain on size and contraction of the lumbar multifidus muscle. Man Ther. 2008 doi: 10.1016/j.math.2008.09.006. [DOI] [PubMed] [Google Scholar]
  50. Wand BM, Bird C, McAuley JH, Dore CJ, MacDowell M, De Souza LH. Early intervention for the management of acute low back pain: a single-blind randomized controlled trial of biopsychosocial education, manual therapy, and exercise. Spine. 2004;29(21):2350–2356. doi: 10.1097/01.brs.0000143619.34308.b4. [DOI] [PubMed] [Google Scholar]
  51. Werneke MW, Hart DL, George SZ, Stratford PW, Matheson JW, Reyes A. Clinical outcomes for patients classified by fear-avoidance beliefs and centralization phenomenon. Arch Phys Med Rehabil. 2009;90(5):768–777. doi: 10.1016/j.apmr.2008.11.008. [DOI] [PubMed] [Google Scholar]
  52. Yoshihara K, Shirai Y, Nakayama Y, Uesaka S. Histochemical changes in the multifidus muscle in patients with lumbar intervertebral disc herniation. Spine. 2001;26(6):622–626. doi: 10.1097/00007632-200103150-00012. [DOI] [PubMed] [Google Scholar]
  53. Young S, Aprill C, Laslett M. Correlation of clinical examination characteristics with three sources of chronic low back pain. Spine J. 2003;3(6):460–465. doi: 10.1016/s1529-9430(03)00151-7. [DOI] [PubMed] [Google Scholar]
  54. Zhao WP, Kawaguchi Y, Matsui H, Kanamori M, Kimura T. Histochemistry and morphology of the multifidus muscle in lumbar disc herniation: comparative study between diseased and normal sides. Spine. 2000;25(17):2191–2199. doi: 10.1097/00007632-200009010-00009. [DOI] [PubMed] [Google Scholar]

RESOURCES