Can the morphology or neuromuscular control of lumbar multifidus (LM) and transversus abdominis (TrA) predict clinical outcomes/recurrence of back pain in patients with nonspecific low back pain (LBP) and is this predictive value treatment-specific?

Whether the values of morphology or neuromuscular control of LM and TrA in predicting clinical outcomes/recurrence of back pain in LBP patients will vary with the chronicity of LBP?

What is the prognostic/treatment modification value of LM and TrA in predicting clinical outcomes/recurrence of LBP at different age groups?

Whether the value of LM and TrA in predicting clinical outcomes/recurrence of LBP will vary between “inception” and “survival” cohort?

A thorough search of MEDLINE, EMBASE, PEDro, SPORT Discus, CINAHL, Cochrane Library without limitation of published date. Citation chaining will be performed using Scopus and Web of Science. Languages will be restricted to English, French, Portuguese, and Chinese.

Inclusion Criteria:

Systematic review/meta-analysis of cohort studies or follow-up studies; validation of clinical decision rule; follow up of untreated control groups in randomized controlled trials; prospective/retrospective cohort studies; observational studies; case series with > 10 subjects; non-duplicated information from multiple reports of the same patient cohorts.

Exclusion Criteria:

Single cross-sectional studies; multiple reports that included duplicated results of same patient case series (only include the study with the largest series).

Lumbar multifidus (LM) and transversus abdominis (TrA) have been advocated as the major spinal stabilizers for controlling spinal stability and preventing recurrence of low back pain (LBP). Preliminary research shows that the morphology or neuromuscular control of these stabilizers are associated with LBP symptoms and clinical outcomes. Specifically, the first-week change in contracted LM thickness as measured by ultrasound predicts 7% of the variability of the first week modified Oswestry disability index score of LBP patients following spinal manipulative therapy. Similarly, the onset latency of TrA explains 17% of the variance of pain scores in patients with chronic LBP.

Given that the morphology and neuromuscular control of LM and TrA may predict the clinical outcomes of LBP patients who may or may not receive back treatment, a prognostic systematic review on this topic will help clinicians monitor their treatment effectiveness and predict prognosis of their patients.

Inclusion Criteria:

Self-ambulatory adult patients of any gender and ethnicity (>18 years of age) with acute (< 6 weeks), subacute (6 to 12 weeks) or chronic (> 12 weeks) nonspecific LBP with no definite diagnosis of any pathological condition.

Exclusion Criteria:

individuals with pregnancy related back pain; LBP patients from sources outside the back (nonspinal LBP); fibromyalgia or myofascial pain syndrome in >20% of the overall sample size; spondylolisthesis; spondylolysis; stenosis; back surgery; cancer; neural disease (e.g. multiple sclerosis or Guillian Barre Syndrome); acute major trauma; spinal cord injury; congenital lumbar/sacral condition; spasticity resulting from spinal cord injury or neural diseases; cauda equina syndrome; infection; solely osteoporosis; vertebral compression fracture; rheumatoid arthritis; scoliosis; stroke; cerebral palsy; disc extrusion or sequestration with or without spinal cord compression.

Inclusion Criteria:

Any kinds of non-surgical treatment with explicit description of therapeutic dosage (e.g drugs, physical therapy modalities or manual therapy)

Exclusion Criteria:

Surgical treatment for back pain

Asymptomatic individuals, individuals without treatment or waiting for treatment, individuals without improvement following an intervention.

No restriction will be applied to the context of the conservative interventions. They can be exercise, chiropractic treatment, physiotherapy or pharmaceutical intervention.

Primary outcomes

At least one of the following LBP clinical outcome measures: Quebec Back Pain Disability scale; Roland Morris disability questionnaire (RMDQ); Oswestry Disability Index (ODI); Japanese orthopedic association score; McGill pain questionnaire; Patient-Specific Functional Scale; pain score (e.g. visual analog scale, numerical pain rating scale); subjective improvement of symptoms (global rating of change); return-to-work/school; days off work/school; recurrent rate; quality of life (e.g. short-form 36, short-form 12, EQ-5D, etc.); subjective complaints.

Secondary outcomes

None.

Two reviewers who are blinded to authors, journals and institutions of the articles will independently extract the data from each selected study using a data extraction form. If the studies have assessed multiple outcomes with different statistical methods, we will solely extract the information which is relevant to our research questions. After the two reviewers have extracted data of five included articles, the inter-reviewer reliability will be assessed by percentage agreement on the extracted information. If agreement is less than 90%, a thorough discussion and modification of instruction will be made and recorded, and the entire data extraction process will be redone. The two reviewers will attempt to resolve all the data extraction disagreements. If the discrepancy persists, a third reviewer will make the final decision.

Data to be extracted: participants’ characteristics (age, gender, type and duration of LBP), sample size, intervention (type, dose, duration and follow up frequency), prognostic factors, outcomes (functional outcome score, work status and recurrence), data analysis (regression, odds ratio, etc.) and results (relevant results at multiple assessment periods).

There is no standardized assessment tool for appraising the risk of bias in prognostic studies. Based on the recommendations on evaluation of prognostic studies, a risk of bias assessment tool, which has been used in another systematic review of spinal prognostic studies, will be modified and adopted for this review. This tool covers six potential bias areas including patient population, follow up of patients, prognostic factors, outcome measurement, confounding measurement and statistical analysis. A detailed explanation of each criterion in this tool will be given. The maximum score of this checklist is 22. The cut-off point for distinguishing a high quality study is 50% of maximum score. Since the cut-off point is arbitrary, a sensitivity analysis using 60% and 70% cut-off point will be conducted to test the robustness of this cut-off point on the conclusion of our primary objective.

Although no systematic review related to our research question has been published at the time of writing this draft, Assessment of Multiple Systematic Reviews (AMSTAR) will be used, if necessary, for evaluating the risk of bias in selected systematic reviews.

Two reviewers will independently conduct the risk of bias assessment on five sample studies with authors, institutions and journals being blinded. The inter-reviewer reliability for risk of bias assessment will be analyzed by percentage agreement and Kappa coefficients of the score of each question and final scores. Similar to the data extraction procedure, if the agreement is less than 90%, modification of assessment instructions will be made and recorded. The risk of bias assessment will start over. If the disagreement cannot be resolved, the third reviewer will make the final decision. The assessment result for an individual study will be tabulated to improve the reporting transparency.

The conclusions related to the primary objective will be drawn based on the “level of evidence” on the ability of the morphology or neuromuscular control of LM and TrA in predicting each clinical outcome (pain, functional ability, quality of life, recurrent rate, etc). If there is substantial heterogeneity on review of the data extraction table, a narrative synthesis will be conducted. If there is no substantial heterogeneity of data, a meta-analysis of pooled dichotomous and/or continuous data will be performed separately to answer the research questions.

Subgroup identification will be determined based on clinical homogeneity with reference to the secondary objectives. If subgroups have been identified and analyzed separately in a primary study, the results will be directly combined with those respective subgroups identified within our secondary objectives. If >80% of the sample population in a primary study fits in a particular subgroup, the study will be classified into that group for analysis. If a study does not provide sufficient information for subgroup classification, it will be excluded. The conclusions of secondary objectives will be drawn based on the level of evidence.

1. Treatment types will be grouped into stabilization exercise, general exercise, spinal mobilization, spinal manipulation and traction. The value of morphology or neuromuscular control of LM and TrA in predicting different treatment outcomes will be compared.

2. Chronicity of LBP will be grouped into acute, sub-acute and chronic. The corresponding predictive value of morphology or muscle activity of LM and TrA will be compared.

3. Patients will be grouped into three subgroups: 18 to 40, 41 to 60 and above 60. The corresponding prognostic value of LM and TrA in predicting clinical outcomes of LBP will be compared.

4. Patients will be divided into “inception” and “survival cohort and the corresponding prognostic value of LM and TrA in predicting back pain clinical outcomes will be evaluated.

The systematic review will be submitted to a peer-review journal for publication. The results will also be presented in local and international conferences.

Arnold Wong

3-48, Corbett Hall

University of Alberta

Edmonton, AB

Canada, T6G 2G4

yulok@ualberta.ca

Department of Physical Therapy, Faculty of Rehabilitation Medicine, University of Alberta

http://www.physicaltherapy.ualberta.ca/

Mr Arnold Wong, Department of Physical Therapy, Faculty of Rehabilitation Medicine, University of Alberta
Ms Martha Funabashi, Department of Physical Therapy, Faculty of Rehabilitation Medicine, University of Alberta
Dr Eric Parent, Department of Physical Therapy, Faculty of Rehabilitation Medicine, University of Alberta
Dr Gregory Kawchuk, Department of Physical Therapy, Faculty of Rehabilitation Medicine, University of Alberta

Ms Jeanette Buckingham, librarian, University of Alberta

The authors have no earlier version of the systematic review on this topic.

23 July 2012

31 December 2012

Arnold Wong is supported by the Alberta Innovates Health Solutions graduate studentship. Greg Kawchuk is supported by the Canadian Research Chair in Spinal Function.

None known

Not applicable

English

Canada

Subject indexing assigned by CRD

Abdominal Muscles; Humans; Low Back Pain; Recurrence

The review team has no relevant ongoing research in this area.

There is no published protocol.

16 August 2012

16 August 2012