Advertisement

Philadelphia Panel Evidence-Based Clinical Practice Guidelines on Selected Rehabilitation Interventions for Low Back Pain

John Albright, Richard Allman, Richard Paul Bonfiglio, Alicia Conill, Bruce Dobkin, Andrew A Guccione, Scott M Hasson, Randolph Russo, Paul Shekelle, Jeffrey L Susman, Lucie Brosseau, Peter Tugwell, George A Wells, Vivian A Robinson, Ian D Graham, Beverley J Shea, Jessie McGowan, Joan Peterson, Michel Tousignant, Lucie Poulin, Hélène Corriveau, Michelle Morin, Lucie Pelland, Lucie Laferrière, Lynn Casimiro, Louis E Tremblay

Abstract

Introduction. A structured and rigorous methodology was developed for the formulation of evidence-based clinical practice guidelines (EBCPGs), then was used to develop EBCPGs for selected rehabilitation interventions for the management of low back pain. Methods. Evidence from randomized controlled trials (RCTs) and observational studies was identified and synthesized using methods defined by the Cochrane Collaboration that minimize bias by using a systematic approach to literature search, study selection, data extraction, and data synthesis. Meta-analysis was conducted where possible. The strength of evidence was graded as level I for RCTs or level II for nonrandomized studies. Developing Recommendations. An expert panel was formed by inviting stakeholder professional organizations to nominate a representative. This panel developed a set of criteria for grading the strength of both the evidence and the recommendation. The panel decided that evidence of clinically important benefit (defined as 15% greater relative to a control based on panel expertise and empiric results) in patient-important outcomes was required for a recommendation. Statistical significance was also required, but was insufficient alone. Patient-important outcomes were decided by consensus as being pain, function, patient global assessment, quality of life, and return to work, providing that these outcomes were assessed with a scale for which measurement reliability and validity have been established. Validating the Recommendations. A feedback survey questionnaire was sent to 324 practitioners from 6 professional organizations. The response rate was 51%. Results. Four positive recommendations of clinical benefit were developed. Therapeutic exercises were found to be beneficial for chronic, subacute, and postsurgery low back pain. Continuation of normal activities was the only intervention with beneficial effects for acute low back pain. These recommendations were mainly in agreement with previous EBCPGs, although some were not covered by other EBCPGs. There was wide agreement with these recommendations from practitioners (greater than 85%). For several interventions and indications (eg, thermotherapy, therapeutic ultrasound, massage, electrical stimulation), there was a lack of evidence regarding efficacy. Conclusions. This methodology of developing EBCPGs provides a structured approach to assessing the literature and developing guidelines that incorporates clinicians' feedback and is widely acceptable to practicing clinicians. Further well-designed RCTs are warranted regarding the use of several interventions for patients with low back pain where evidence was insufficient to make recommendations.

INTRODUCTION

Low back pain (LBP) is the largest cause of workers' compensation in the United States and Canada. Sixty to 90% of the adult population is at risk of developing LBP at some point in their lifetime.14 Of those who develop acute LBP, 30% develop chronic LBP.5 Low back pain has a significant impact on functional ability, restricting occupational activities with marked socioeconomic repercussions.6,7

A number of different practitioners treat people with LBP. These include physicians, physical therapists, chiropractors, massage therapists, psychologists, kinesiologists, rehabilitation technicians, and others. The treatment goals are to relieve pain, reduce muscle spasm, improve range of motion (ROM) and strength, correct postural problems, and ultimately improve functional status.

A number of rehabilitation interventions are used in the management of people with LBP. Among current musculoskeletal interventions specific for LBP available to rehabilitation specialists, there are body mechanics and ergonomics training, posture awareness training, strengthening exerises, stretching exercises, activities of daily living (ADL) training, organized functional training programs, therapeutic massage, joint mobilizations and manipulations, mechanical traction, biofeedback, electrical muscle stimulation, transcutaneous electrical nerve stimulation (TENS), athermal modalities, cryotherapy, deep thermal modalities, superficial thermal modalities, and work hardening.8

The Philadelphia Panel (see article titled “Evidence-Based Clinical Practice Guidelines on Selected Rehabilitation Interventions: Overview and Methodology”) was convened to evaluate 9 selected rehabilitation interventions for LBP: thermotherapy, therapeutic massage, therapeutic exercises, electromyographic (EMG) biofeedback, mechanical traction, ultrasound, TENS, electrical stimulation, and combined rehabilitation interventions.

The purpose of this article is to describe the evidence-based clinical practice guidelines (EBCPGs) developed by the panel about rehabilitation interventions for LBP. The aim of developing the guidelines was to improve appropriate use of rehabilitation interventions for low back pain. The target users of these guidelines are physical therapists, physiatrists, orthopedic surgeons, rheumatologists, family physicians, and neurologists.

METHODS

The detailed methods of the EBCPGs development process are summarized in an accompanying article in this issue (“Evidence-Based Clinical Practice Guidelines on Selected Rehabilitation Interventions: Overview and Methodology”). Briefly, an a priori protocol was defined that was followed for the conduct of separate systematic reviews for each intervention.

Studies were eligible if they were randomized controlled trials (RCTs), nonrandomized controlled clinical trials (CCTs), or case control or cohort studies that evaluated the intervention of interest in a population with nonspecific LBP. Nonspecific LBP was defined as pain between the gluteal fold and the uppermost lumbar vertebrae and included postsurgery back pain. The outcomes of interest were functional status, pain, ability to work, patient global improvement, patient satisfaction, and quality of life. The interventions included massage, thermal therapy (hot or cold packs), electrical stimulation, EMG biofeedback, TENS, therapeutic ultrasound, therapeutic exercises, and combinations of these rehabilitation interventions. Studies whose control groups received “active” treatments were included. Concurrent treatments (eg, home exercises, educational booklets, advice on posture) were allowed if they were given in the same way to both the experimental and control groups. However, concurrent therapy that was given to one group but not to the other group was not accepted (eg, education by means of lectures for the control group were not accepted). No limitations based on methodological quality were imposed. Only English-, French-, and Spanish-language articles were accepted. Abstracts were not included.

A structured literature search was developed based on the sensitive search strategy for RCTs recommended by the Cochrane Collaboration9 and modifications proposed by Haynes et al.10 The search strategy was expanded to identify case control, cohort, and nonrandomized studies. The search was conducted in the electronic databases of MEDLINE, EMBASE, Current Contents, CINAHL, and the Cochrane Controlled Trials Register up to July 1, 2000. In addition, the registries of the Cochrane Field of Rehabilitation and Related Therapies and the Cochrane Musculoskeletal Group and the Physiotherapy Evidence Database (PEDro) were searched. The references of all included trials were searched for relevant studies. Content experts were contacted for additional studies.

Two independent reviewers (VAR, JP) appraised the titles and abstracts of the literature search, using a checklist with the a priori defined selection criteria. Relevant studies were retrieved and the full articles were assessed by 2 independent reviewers for inclusion. Data were extracted by 2 independent reviewers from included articles, using predetermined extraction forms regarding the population characteristics, details of the interventions, trial design, allocation concealment, and outcomes. Methodological quality was assessed by a 5-point validated scale that assigns 2 points each for randomization and double-blinding and 1 point for description of withdrawals.11,12 Differences in data extraction and quality assessment were resolved by consensus.

Data were analyzed at 3 approximate time points posttherapy: 1 month, 6 months, and 12 months. If outcomes were reported at different intervals, the closest time was used for these time points.

Because prognosis is widely thought to be dependent on disease duration, the analysis was conducted for 3 categories of LBP: acute (<4 weeks duration), subacute (4–12 weeks duration), and chronic (>12 weeks duration). If the population contained individuals with mixed acute and chronic disease durations, the study was excluded. If the population included individuals with mixed subacute and chronic disease durations, the study was classified as chronic.

STATISTICAL ANALYSIS

Data were analyzed using the Review Manager (RevMan) computer program, Version 4.1 for Windows.* Continuous data were analyzed using weighted mean differences (WMDs) between the treatment and control groups at the end of the study, where the weight is the inverse of the variance. Where an outcome was measured with different scales (eg, pain, functional status), the data were analyzed with standardized mean differences, calculated using the mean and standard deviation. Dichotomous data were analyzed using relative risks. Heterogeneity was tested using the chi-square statistic. When heterogeneity was not significant, fixed-effects models were used. With significant heterogeneity, random-effects models were used.

To calculate clinical improvement (defined as 15% improvement relative to a control), the absolute benefit and the relative difference in the change from baseline were calculated. Absolute benefit was calculated as the improvement in the treatment group less the improvement in the control group, in the original units. Relative difference in the change from baseline was calculated as the absolute benefit divided by the baseline mean (weighted for the treatment and control groups). For dichotomous data, the relative percentage of improvement was calculated as the difference in the percentage of improvement between the treatment and control groups.

The recommendations were graded by their level of evidence (I or II) and by the strength of evidence (A, B, or C). This grading system is shown in Table 1 and is described more fully elsewhere (see article titled “Evidence-Based Clinical Practice Guidelines on Selected Rehabilitation Interventions: Overview and Methodology”). Briefly, grade A recommendations indicate that a clinically important benefit was shown in one or more RCTs. Grade B recommendations were assigned for interventions with a clinically important benefit shown in nonrandomized trials. Because there is less confidence in the results of nonrandomized studies, grade B recommendations required that the study be assigned a quality score of 3 or more out of 5. Grade C recommendations were assigned to interventions that have been compared with a control and have shown no evidence of effect in controlled trials. A master grid showing each rehabilitation intervention assessed and the strength and level of evidence is shown in Table 2. The report follows the same order as this grid, from left to right, top to bottom.

Table 1.

Details of Philadelphia Panel Classification System

Table 2.

Master Grid of Low Back Pain (LBP) Guidelinesa

Clinically important benefit was shown for therapeutic exercise across subacute, chronic, and postoperative LBP as well as for the continuation of normal activities (Tab. 3). No clinically important benefit was demonstrated for 4 other interventions (Tab. 4). Insufficient data were available for 4 interventions (Tab. 5). The Philadelphia Panel guidelines are compared with other published guidelines in Appendix 1.

Table 3.

Grade A Guidelines: Clinically Important Benefit Demonstrateda

Table 4.

Grade C Rehabilitation Interventions: Clinically Important Benefit Not Demonstrateda

Table 5.

Rehabilitation Interventions With Insufficient Dataa

A survey questionnaire was sent to 324 practitioners for feedback on the 4 grade A or B recommendations. Their comments were reviewed by the Philadelphia Panel and were incorporated in this EBCPG document.

RESULTS

Literature Search

The literature search identified 4,981 articles related to LBP. Of these, 340 were considered potentially relevant based on the selection criteria checklist. Of these, 41 met the selection criteria and were included (Appendix 2). The number of included trials is shown for each of the interventions for nonspecific back pain in the “cityscape” shown in Figure 1.

Figure 1.

Cityscape, demonstrating number of included trials. EMG=electromyographic, TENS=transcutaneous electrical nerve stimulation.

Practitioner Feedback Survey

Of the 324 practitioners surveyed from the American Academy of Family Physicians (AAFP), American Academy of Orthopaedic Surgeons (AAOS), American College of Physicians (ACP), American Physical Therapy Association (APTA), American College of Rheumatology Health Professionals (ARHP), and Physiatric Association of Spine, Sports, and Occupational Rehabilitation (PASSOR), 9 were inappropriately sampled (wrong specialty) and 21 could not be reached due to incorrect addresses. Of the 294 practitioners who were appropriately sampled and received the questionnaire, 149 responded (51% response rate). Of these, 11 (4%) refused to participate and 138 (47%) completed the survey.

ACUTE LBP (<4 WEEKS)

Therapeutic Exercises for Acute LBP (<4 Weeks), Level I (RCT), Grade C for Pain, Function, and Return to Work (No Benefit Demonstrated)

Summary of Trials:

Four RCTs (N=1,035) of acute LBP with a control group were included.1316 One RCT was excluded due to active intervention in the control group, consisting of specific instructions on back care and biomechanics as well as pelvic tilt exercises not provided to the flexion exercise group.17 One RCT of endurance training of the trunk muscles was excluded because the population included a mix of subjects with acute and subacute LBP, with pain duration ranging from 1 to 7 weeks.18 One RCT was excluded because the exercise groups also received an educational booklet on posture, which was not given to the control group).

In the included RCTs, the exercises included McKenzie, back extension, Kendall flexion, and strengthening exercises (described in Appendix 2). The treatment schedule ranged from 1 to 3 sessions per week for 4 to 8 weeks.

Efficacy:

None demonstrated. Therapeutic exercises were no better than control therapy at improving function, ability to work, or reducing pain at 1 or 12 months. The pooled estimates at 1 month were not clinically important for pain (Fig. 2), function, or return to work. There was no difference in effect between types of exercise (McKenzie, Kendall, or strengthening). At 12 months after the start of therapy, there was no clinically important benefit on pain compared with untreated patients13,19 and no effect on sick leave days in 1 year (WMD=3 days, 95% confidence interval [CI] = −22 to 28 days).16

Figure 2.

Exercises for acute low back pain (<4 weeks): pain at 1 and 6 months. VAS=visual analog scale, CI=confidence interval.

Strength of Published Evidence in Comparison With Other Guidelines:

The Philadelphia Panel found level I (RCT) evidence that showed no clinically important benefit of stretching (including McKenzie and Kendall programs) or strengthening exercises (level I, grade C). This is consistent with the Québec Task Force on Spinal Disorders (QTF)20 and the Agency for Health Care Policy and Research (AHCPR),21 which found no scientific evidence for general exercises.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommended there is poor evidence to include or exclude stretching or strengthening exercises alone (grade C for pain, function, and return to work) as an intervention for acute LBP. This recommendation agrees with the AHCPR21 and BMJ22 guidelines. In contrast, the QTF20 recommended the prescription of general exercises as an option to increase strength, ROM, and endurance but did not discriminate between different types of exercise. The BMJ22 reported that increased stress from therapeutic exercises may be harmful in acute conditions based on the RCT by Malmivaara et al15 that was included in our study.

Continuation of Normal Activities Versus Enforced Bed Rest for Acute LBP (<4 Weeks), Level I (RCT), Grade A for Return to Work (Clinically Important Benefit), Grade C for Pain and Function (No Benefit Demonstrated)

Summary of Trials:

One RCT (N=186)15 of continuing normal activities versus 2 days of enforced bed rest was included.

Efficacy:

Clinically important benefit was found for return to work. Continuation of normal activities resulted in 49% fewer sick days after 3 weeks relative to the enforced bed rest group, with an absolute difference of 3.4 sick days (95% CI=1.6–5.2 days) (Tab. 6, Fig. 3). The relative change with continuing normal activities was 10% better for functional status (Fig. 4) and 5% lower for pain relative to bed rest (Fig. 5). After 3 months, the normal activities group had 51% fewer sick days (4.5 days less, 95% CI=3–6 days), better function by 10% on the Oswestry scale, and 5% less pain on a 10-cm visual analog scale (VAS).

Figure 3.

Normal activities versus bed rest for acute low back pain: return to work at 1 and 3 months. CI=confidence interval.

Figure 4.

Continue normal activity versus bed rest for acute low back pain: function at 1 and 3 months. CI=confidence interval.

Figure 5.

Continue normal activities versus bed rest for acute low back pain: pain at 1 and 3 months. CI=confidence interval.

Table 6.

Continue Normal Activities for Acute Low Back Pain: Pain, Function, and Return to Work at 3 Weeksa

Strength of Published Evidence in Comparison With Other Guidelines:

The Philadelphia Panel found good scientific evidence (level I) of important clinical benefit of functional activities on return to work, similar to the AHCPR21 evidence rating. The QTF20 found no scientific evidence for general exercises but did not evaluate continuation of normal activities as a separate intervention.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel concluded that there is good evidence to include continuation of normal activities (grade A for return to work, grade C for pain and function) as an intervention for people with acute LBP. This conclusion agrees with the AHCPR21 guidelines. The BMJ22 guidelines do not discuss normal activities as an intervention. The QTF20 did not discriminate between normal activities and stretching and strengthening programs.

Practitioner Agreement

  • Response rate for this EBCPG: 46%

  • Percentage of practitioners giving comments for this EBCPG: 41%

  • Agree with recommendation: 98%

  • Think a majority of my colleagues would agree: 98%

  • Will (or already) follow this recommendation: 98%

Practitioner Comments

  1. Guideline should differentiate acute herniated disk, which may benefit from bed rest.

  2. Amount of bed rest is important—more than 72 hours is unnecessary.

Panel's Response:

The trial on which this recommendation is based excluded patients with disk involvement; therefore, the effects of continuing normal activity in patients with acute herniated disk involvement cannot be assessed. The Philadelphia Panel did not assess whether bed rest is an effective therapy for patients with acute LBP, but a Cochrane review on bed rest has just been completed. The results suggest that 2 to 7 days of bed rest has no effect on pain or functional status in patients with acute LBP and that there is no difference between short (2 days) or long (7 days) bed rest.23

Mechanical Traction for Acute LBP (<4 Weeks), Level I (RCT), Grade C for Pain and Patient Global Assessment (No Benefit Demonstrated)

Summary of Trials:

Three RCTs (N=176) of intermittent mechanical traction versus placebo for acute LBP were included.2426 One RCT (N=16) compared vertical traction with bed rest.27 Disease duration was not well reported, but all patients were hospitalized. All trials included patients with back pain radiating below the knee.

Efficacy:

None demonstrated. For intermittent traction, there was no difference at 1 month between traction and placebo for number of patients with improved pain (relative risk [RR]=0.88, 95% CI=0.50−1.55) (Fig. 6) or pain (-3.4 mm on a 100-mm VAS, 95% CI = −21.2 to 14.5) (Fig. 7). For vertical traction at 1 month, there was no difference in pain (0.1 cm on a 10-cm VAS, 95% CI = −2.6 to 2.8) or patient global improvement (RR=1.14, 95% CI=0.88−1.49). No data were reported beyond 1 month.

Figure 6.

Traction versus placebo for acute low back pain: patient and clinician global assessment at 1 month. CI=confidence interval.

Figure 7.

Traction versus placebo for acute low back pain: pain at 1 month. CI=confidence interval.

Strength of Published Evidence in Comparison With Other Guidelines:

We found level I (RCT) evidence of no clinically important benefit of traction for acute LBP. In contrast, the QTF20 found no scientific evidence, and the AHCPR21 found moderate scientific evidence of lack of benefit. The AHCPR21 included the same trials except for one, used a different classification of the evidence, and included a trial involving clients with chronic LBP conditions.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is poor evidence to include or exclude mechanical traction alone (grade C for pain and patient global assessment) as an intervention for acute LBP. This recommendation is in accord with AHCPR21 and BMJ22 clinical recommendations compared with other guidelines. In contrast, the QTF20 recommended mechanical traction as an option to increase ROM. The BMJ22 reported potential harms, not validated in trials, including: (1) debilitation, (2) loss of muscle tone, (3) bone demineralization, and (4) thrombophlebitis.

Therapeutic Ultrasound for Acute LBP (<4 Weeks), Level II (CCT), Grade C for Pain (No Benefit Demonstrated)

Summary of Evidence:

One nonrandomized controlled trial (N=73) of continuous ultrasound versus placebo was included.28

Efficacy:

None demonstrated. There was no difference between continuous ultrasound and no treatment for pain improvement (Fig. 8). However, range of flexion and extension was improved after 1 month by 28 degrees (95% CI=26°−29°).28 No data were available for functional status, strength, quality of life, or return to work at 1 month, and no data were available beyond 1 month.

Figure 8.

Therapeutic ultrasound for acute low back pain: pain at 1 month. CI=confidence interval.

Strength of Published Evidence in Comparison With Other Guidelines:

We found level II scientific evidence that therapeutic ultrasound has no clinically important benefit. This finding agrees with the AHCPR21 rating. In contrast, the QTF20 found no scientific evidence.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is poor evidence to include or exclude therapeutic ultrasound alone (grade C for pain) as an intervention for acute LBP. This EBCPG agrees with AHCPR21 and BMJ22 guidelines, even though they make general statements to consider physical interventions, including therapeutic ultrasound. In contrast, the QTF20 recommended therapeutic ultrasound as an option to diminish muscle spasm and relieve symptomatic pain. However, ultrasound was classified as thermotherapy, which is misleading, because pulsed ultrasound does not produce thermal effects. There is insufficient information regarding adverse effects.22

TENS for Acute LBP (<4 Weeks), Level I (RCT), Grade C for Pain or Function (No Benefit Demonstrated)

Summary of Trials:

One RCT (N=58) compared TENS and placebo for the management of people with acute LBP.29 All patients were treated with mobility and strengthening exercises. The TENS consisted of 15 minutes of high-frequency TENS followed by 15 minutes of acupuncture-like TENS.

Efficacy:

None demonstrated. There was no difference in self-rated VAS pain (Fig. 9), functional status, strength, or ROM at 1 month.29

Figure 9.

Transcutaneous electrical nerve stimulation for acute low back pain: pain at 1 month. VAS=visual analog scale, TENS=transcutaneous electrical nerve stimulation, CI=confidence interval.

Strength of Published Evidence in Comparison With Other Guidelines:

We found level I evidence of no clinically important benefit of TENS on pain or function, similar to the AHCPR,21 which found limited research-based evidence. The AHCPR,21 however, included one study that we rejected because the intervention used was electroacupuncture, not TENS. In contrast, the QTF20 found no scientific evidence for TENS in acute LBP.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is poor evidence to include or exclude TENS alone (grade C for pain) as an intervention for acute LBP. This EBCPG agrees with the AHCPR21 and BMJ22 recommendations. In contrast, the QTF20 recommended TENS as a useful modality for symptomatic pain relief, but this may refer to electroanalgesia rather than TENS specifically. Insufficient information is available regarding adverse effects.22

Interventions With Insufficient Data for Acute LBP (<4 Weeks)

No evidence with acceptable research design, interventions, group comparisons, and outcomes were identified for thermotherapy, electrical stimulation, therapeutic massage, or EMG biofeedback. This lack of evidence concurs with both the QTF20 and AHCPR21 guidelines. However, the QTF20 recommended thermotherapy, massage, and EMG biofeedback as potential interventions for acute LBP.

Some trials of combinations of rehabilitation interventions for acute LBP were identified, but these trials were excluded due to poor definitions of the interventions, populations, or nonstandard outcomes. The Philadelphia Panel rated the evidence as insufficient for a recommendation. In contrast, both QTF20 and BMJ22 recommended that rehabilitation specialists use physical interventions at their own discretion to relieve spasm; reduce inflammation and pain; increase strength, ROM, and endurance; and improve functional status.

SUBACUTE LBP (4–12 Weeks)

Therapeutic Exercises for Subacute LBP (4–12 Weeks), Level I (RCT), Grade A for Pain, Function, and Patient Global Assessment (Clinically Important Benefit)

Summary of Trials:

Three RCTs (N=405) of therapeutic exercises versus a control were included.3032 The exercises consisted of McKenzie, Kendall, and strengthening exercises (described in Appendix 2). The treatment schedule was twice per week for 4 weeks.

Efficacy:

Clinically important benefit on pain relief and patient-assessed global condition was demonstrated by meta-analysis. Therapeutic exercises provided more pain relief relative to the control by 10% for strengthening exercises alone,30 11% for Kendall flexion,32 50% for McKenzie exercises,32 and 57% for McKenzie exercises31 after 1 month (P<.05) (Tab. 7). Functional status was improved relative to the control by 11% with McKenzie exercises31 and 15% with strengthening exercises30 (Fig. 10). Patient global improvement was 17% to 24% better with Kendall exercises relative to the control, but not statistically different32 (Tab. 8). We could not assess the use of these interventions in patients with neurological or radicular pain, as these diagnostic groups were excluded from the original trials.

Figure 10.

Therapeutic exercises for subacute low back pain: pain at 1 and 6 months. CI=confidence interval.

Table 7.

Therapeutic Exercises for Subacute Low Back Pain: Pain at 1 Montha

Table 8.

Therapeutic Exercises for Subacute Low Back Pain: Patient Global Assessment at 1 Montha

Strength of Published Evidence in Comparison With Other Guidelines:

There is level I evidence of clinically important benefit on pain relief and patient global assessment with therapeutic exercises consisting of extension, flexion, or strengthening exercises as therapy for subacute nonspecific LBP (level I). This disagrees with the QTF,20 which found no scientific evidence for general exercises, but 2 RCTs30,31 have been published since the QTF, which explains this difference. The AHCPR21 included 3 additional trials compared with our research team: a trial with clients with chronic LBP conditions, another with a psychological intervention, and another involving a back school approach.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is good evidence to include extension, flexion, and strengthening exercises as interventions for subacute LBP (grade A for pain, function, and patient global assessment). However, we did not assess the use of these interventions for patients with neurological or radicular pain, as these diagnostic groups were excluded from the original trials. This is in partial concordance with AHCPR,21 which recommended low-stress aerobic exercises within the first 4 weeks (acute LBP). The BMJ22 is also in agreement with our EBCPG concerning extension, flexion, and strengthening exercises. The QTF20 recommended the prescription of general exercises as an option to increase strength, ROM, and endurance. The BMJ reported that the increased stress of therapeutic exercise is potentially harmful in subacute conditions.

Practitioner Agreement

  • Response rate for this EBCPG: 49%

  • Percentage of practitioners giving comments for this EBCPG: 32%

  • Agree with recommendation: 90%

  • Think a majority of my colleagues would agree: 88%

  • Will (or already) follow this recommendation: 93%

Practitioner Comments

  1. Selection of exercises depends on clinical presentation; if there are neurological or sensory deficits, exercises could exacerbate the pain.

  2. Type of exercise (eg, Kendall, McKenzie) depends on patient.

  3. A combined approach with education is needed.

  4. Length of follow-up in Davies et al study32 is insufficient.

Panel's Response:

The guideline has been modified to specify that we did not assess the use of these interventions for patients with neurological/radicular pain. The individualized approach to exercise prescription is a widespread clinical opinion but has little empiric evidence. The effects of individualized approaches could not be assessed, because the trials report group outcomes. Education was not assessed by the Philadelphia Panel. An educational booklet on posture and biomechanics was provided to the patients in both groups of 1 trial,31 but not to the other 2 trials. When the Davies et al study32 was excluded from the analysis, the results remained the same; therefore, exclusion of this study would not change the recommendation.

Mechanical Traction for Subacute LBP (4–12 Weeks), Level 1 (RCT), Grade C for Patient Global Assessment and Return to Work (No Benefit Demonstrated)

Summary of Trials:

Two RCTs (N=212) of static traction compared with a placebo were included.33,34 One trial was excluded due to lack of a control group (traction was compared with heat).35 Both trials included patients with radiating pain.

Efficacy:

None demonstrated. There was no clinically important benefit for patient global assessment at 1 month (Fig. 11) or for return to work at 12 months (Fig. 12).34

Figure 11.

Mechanical traction for subacute low back pain: patient global assessment at 1 month. CI=confidence interval.

Figure 12.

Mechanical traction versus placebo for subacute low back pain: return to work at 12 months. CI=confidence interval.

Strength of Published Evidence in Comparison With Other Guidelines:

The Philadelphia Panel found level I evidence of no clinically important benefit of mechanical traction for subacute LBP. In contrast, the AHCPR21 found moderate scientific evidence of no benefit, and the QTF20 reported no scientific evidence. The AHCPR21 included the same trials as our research team except for one, did not use the same classification for scientific evidence, and included a trial involving clients with chronic LBP conditions.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is poor evidence to include or exclude mechanical traction alone (grade C for patient global assessment and return to work) as an intervention for subacute LBP. This EBCPG agrees with AHCPR21 and BMJ22 recommendations. The QTF20 recommended mechanical traction as an option to increase ROM. The BMJ22 reported the following potential harms of traction: (1) debilitation, (2) loss of muscle tone, (3) bone demineralization, and (4) thrombophlebitis.

CHRONIC LBP (>12 WEEKS)

Therapeutic Exercises for Chronic LBP (>12 Weeks), Level I (RCT), Grade A for Pain and Function (Clinically Important Benefit), Grade C for Return to Work (No Benefit Demonstrated)

Summary of Trials:

Seven trials were excluded due to lack of an appropriate control group.3642 One study could not be analyzed because standard deviations were not reported.43 Eight RCTs were included.4451 The exercises included flexion, extension,44 stretching, circuit training,45,49,50 and strength exercises with progressive increases in resistance.4648

Efficacy:

Clinically important benefit was demonstrated for pain relief and functional status. Five RCTs (N=361) demonstrated percentage reductions in pain relative to the control group of 7%,50 20%,44 23%,45 26%,46 and 60%47 (Tab. 9, Fig. 13). Two RCTs of nursing aides with back pain showed no difference in number of patients with pain improvement after 1 month (RR=1.45, 95% CI=0.59–3.56).48,49 Functional status was improved in 3 RCTs (N=209) relative to the control group by 7% with stretching exercises50; 17% with strengthening, stretching, and aerobics45; and 47% with strengthening exercises46 (Tab. 10). The pooled meta-analysis results were statistically significant for function (standardized mean difference [SMD]=0.36, 95% CI=0.1–0.6). There was no difference in ROM, strength, or return to work. One RCT (N=56)52 found no difference between flexion and extension exercises for pain or patient global assessment at 1 month posttherapy.

Figure 13.

Exercises for chronic low back pain: pain at 1 month. CI=confidence interval.

Table 9.

Therapeutic Exercises for Chronic Low Back Pain: Pain at 1 Montha

Table 10.

Therapeutic Exercises for Chronic Low Back Pain: Function After 1 Montha

At 6 to 12 months follow-up (not shown in the table), the relative reduction in pain was 60%47 and the relative improvements in function were 0% in one trial50 and 30% in another trial.45 Two RCTs were pooled and showed no difference in sick days over 12 months (WMD=0.5 days, 95% CI = −1.5 to 2.5 days).47,51

Strength of Published Evidence in Comparison With Other Guidelines:

The Philadelphia Panel found good scientific evidence (level I) of clinically important benefit on pain and function with stretching or strengthening exercises. In contrast, the QTF20 found no scientific evidence for general exercises. The QTF20 reviewed only one of the trials that was included in our meta-analysis.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is good evidence to include stretching, strengthening, and mobility exercises (grade A for pain and function, grade C for return to work) as interventions for chronic LBP. The BMJ22 is in agreement with this EBCPG concerning strengthening exercises. The QTF20 also recommended the prescription of general exercises as an option to increase strength, ROM, and endurance. The BMJ22 reported that exercise could have adverse effects due to increased stress on the spine.

Practitioner Agreement

  • Response rate for this EBCPG: 48%

  • Percentage of practitioners giving comments for this EBCPG: 38%

  • Agree with recommendation: 88%

  • Think a majority of my colleagues would agree: 91%

  • Will (or already) follow this recommendation: 81%

Practitioner Comments

  1. Evidence for functional status is not convincing.

  2. Be careful about lumping different types of exercise (eg, McKenzie and strengthening).

  3. McKenzie exercises are insufficient for chronic LBP, useful only for acute LBP.

  4. Article by O'Sullivan et al53 on abdominal muscle re-education for spondylolisthesis should be included.

  5. Neuromotor retraining should be included in this EBCPG.

Panel's Response:

The evidence for functional status is based on 3 RCTs, 2 of which demonstrated greater than 15% improvement in function relative to the control group. Different exercises were not lumped; each article is presented separately in Tables 9 and 10. The McKenzie approach was evaluated for chronic LBP by one trial, which showed a 20% relative improvement in pain compared with the control, thus meeting the Philadelphia Panel criteria for clinically important improvement.44 The study by O'Sullivan et al53 was excluded because neuromotor retraining was compared with a combined approach using exercises, heat, massage, and ultrasound, but not the control. The panel was not able to make any recommendations about neuromotor retraining, as there are no controlled studies evaluating the effectiveness of this intervention.

Mechanical Traction for Chronic LBP (>12 Weeks), Level I (RCT), Grade C for Pain, Function, Patient Global Assessment, and Return to Work (No Benefit Demonstrated)

Summary of Trials:

Four RCTs (N=176) of mechanical traction (2 intermittent and 2 static) versus placebo or untreated were included.5457 One RCT was excluded because it was impossible to separate the data for each treatment group.58 One RCT was excluded because the traction force was only 10% of body weight and considered a placebo therapy.47

Efficacy:

None demonstrated. There was no difference in pain, function, or patient global assessment. The point estimate for pain at 3 months favored the control group in the pooled analysis (WMD = −6.3, 95% CI=−15.8 to −3.1) (Fig. 14). However, work absence was shorter in the traction group after 6 months (35 days with traction versus 45 days without traction, but this was not statistically significant).

Figure 14.

Traction for chronic low back pain: pain at 1, 3, and 6 months. VAS=visual analog scale, CI=confidence interval.

Strength of Published Evidence in Comparison With Other Guidelines:

We found good evidence (level I) of no important benefit on pain, function, or patient-rated improvement of mechanical traction. In contrast, the QTF20 found no scientific evidence.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is poor evidence to include or exclude mechanical traction alone (grade C for pain, function, patient global assessment, and return to work) as an intervention for chronic LBP. This EBCPG is in accordance with BMJ22 clinical recommendations compared with other EBCPGs, but not the QTF,20 which recommended mechanical traction as an option to increase ROM. According to the BMJ,22 potential harms, not validated in trials, include (1) debilitation, (2) loss of muscle tone, (3) bone demineralization, and (4) thrombophlebitis.

Therapeutic Ultrasound for Chronic LBP (>12 Weeks), Level II (CCT), Grade C for Pain (No Benefit Demonstrated)

Summary of Trials:

One RCT (N=36) of continuous therapeutic ultrasound versus a placebo59 was included.

Efficacy:

None demonstrated. There was no difference in pain improvement between continuous therapeutic ultrasound and sham therapeutic ultrasound after 1 month of therapy (Fig. 15).59 No data were reported for functional status, ROM, strength, quality of life, or return to work, and no data were available beyond 1 month.

Figure 15.

Therapeutic ultrasound for chronic low back pain: pain (number improved at 1 month). CI=confidence interval.

Strength of Published Evidence in Comparison With Other Guidelines:

We found fair scientific evidence (level II) that showed no clinically important benefit on pain with therapeutic ultrasound. The QTF20 found no scientific evidence.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is poor evidence to include or exclude therapeutic ultrasound alone (grade C for pain) as an intervention for chronic LBP. This EBCPG is in concordance with BMJ22 guidelines, despite a general statement on physical interventions, including therapeutic ultrasound. In contrast, the QTF20 recommended the prescription of therapeutic ultrasound grouped with thermotherapy as an option to diminish muscle spasm and relieve symptomatic pain. There is insufficient information regarding adverse effects.22

TENS for Chronic LBP (>12 Weeks), Level I (RCT), Grade C for Pain and Function (No Benefit Demonstrated)

Summary of Trials:

Four CCTs (N=235) of TENS versus a placebo were included.50,6062 One observational study (N=78) was included, although some patients did not have LBP (but had other pain syndromes such as cervicalgia or phantom limb pain).63 This study also allowed patients to self-select TENS based on its perceived effectiveness during 2 weeks of in-hospital therapy. One CCT was excluded because, of the 16 patients assigned to the treatment group, only 6 had LBP.64 One RCT was excluded because the patient population was people with chronic myalgia and the etiology was not described.65 Two RCTs were excluded due to lack of an appropriate control group (both used a form of massage as the comparison intervention).66,67 Two crossover RCTs were excluded because both the treatment and placebo groups received acupuncture needles. Acupuncture is thought to have an effect of its own and was excluded from the scope of this review.68,69

The method of application was acupuncture-like in one RCT.50 Two trials used high-frequency (>10 Hz) TENS,60,61 2 trials used low-frequency (4 Hz) TENS,62 1 trial alternated between both low- and high-frequency TENS,50 and the other trial did not report the characteristics in sufficient detail.63

Efficacy:

None demonstrated (Fig. 16). There was no difference in the pooled estimate of patient-rated pain at 1 month posttherapy (SMD=−0.2, 95% CI=−0.4 to 0.1). This SMD is equivalent to a difference between treatment and control groups of 4 mm on a 100-mm VAS for pain. In addition, there were no differences between placebo and TENS for functional status, ROM, or strength at 1 month posttherapy. At 3 to 6 months posttherapy, there was no difference in self-rated pain or activity level in any study. The pooled results were not affected by the quality of the trial, the frequency of TENS, or whether acupuncture or traditional TENS was applied. No side effects were reported for TENS.

Figure 16.

Transcutaneous electrical nerve stimulation for chronic low back pain: pain at 1 and 6 months. VAS=visual analog scale, TENS=transcutaneous electrical nerve stimulation, CI=confidence interval.

Strength of Published Evidence in Comparison With Other Guidelines:

The Philadelphia Panel found good evidence (level I) of no clinically important benefit on pain with TENS. This finding disagrees with the QTF,20 which found weak scientific evidence based on a CCT.67 This trial was excluded from our analysis due to lack of placebo (the comparison intervention was massage).

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is poor evidence to include or exclude TENS alone (grade C for pain and function) as an intervention for chronic LBP. This EBCPG is in concordance with the BMJ22 recommendations. In contrast, the QTF20 recommended TENS as a rehabilitation modality for symptomatic pain relief, but this recommendation may include other forms of electroanalgesia. Insufficient information regarding adverse effects was reported by the BMJ.22

EMG Biofeedback for Chronic LBP (>12 Weeks), Level I (RCT), Grade C for Pain and Function (No Benefit Demonstrated)

Summary of Trials:

Five RCTs (N=162) of EMG biofeedback versus a control for chronic LBP were included.7276

Efficacy:

None demonstrated. Meta-analysis showed no effect on pain relief, functional status, or ROM after 1 month of therapy (Fig. 17). There were no data for quality of life or return to work. There were no data beyond 1 month.

Figure 17.

Electromyographic biofeedback for chronic low back pain: pain and function at 1 month. ADL=activities of daily living, VAS=visual analog scale, CI=confidence interval.

Strength of Published Evidence in Comparison With Other Guidelines:

The Philadelphia Panel found good scientific evidence (level I), which showed no clinically important benefit on pain or function with EMG biofeedback, but no scientific evidence. When no evidence was found, we rated it “insufficient data” regarding postural exercises. The QTF20 found no scientific evidence.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is poor evidence to include or exclude EMG biofeedback alone (grade C for pain and function) as intervention for chronic LBP. The BMJ22 made no recommendation due to conflicting evidence related to EMG biofeedback. The QTF20 recommended EMG biofeedback as an option to reduce muscle spasm. Postural exercises were not studied by the QTF.20 There is insufficient information regarding adverse effects for EMG biofeedback.22

Interventions for Chronic LBP With Insufficient Data

No eligible studies were found on which to base recommendations for thermotherapy, massage, or electrical stimulation (Tab. 5). This lack of evidence was also reported by the BMJ22 and QTF20 guidelines. However, both the QTF20 and BMJ22 recommend massage as an intervention for chronic LBP. Massage may have beneficial effects, as shown in an RCT published in abstract format.75

Combinations of rehabilitation interventions were classified by the Philadelphia Panel as having insufficient data to make a recommendation due to different combinations, unvalidated outcomes, and poor description of the actual interventions. This is in disagreement with the BMJ22 and the QTF,20 which both make general statements about the use of physical interventions in combination at the discretion of the rehabilitation specialist.

Deep abdominal stabilization exercises for patients with chronic spondylolisthesis improved pain and function relative to general exercises, heat, massage, and therapeutic ultrasound in one RCT (N=42), but no placebo comparison group was available.53

POSTSURGERY BACK PAIN

Therapeutic Exercises Post-Back Surgery, Level I (RCT), Grade A for Pain and Function (Clinically Important Benefit)

Summary of Trials:

One RCT (N=200) with 3 groups was included that compared strengthening exercises versus McKenzie exercise versus no therapy.79 One RCT of vigorous lumbar stabilizing exercises compared with mild exercises was excluded due to lack of a control group.77

Efficacy:

Clinically important benefit was shown on pain and function with both types of exercise versus no therapy. Both the resisted exercises and the McKenzie program improved functional status by 51% relative to the control group (Tab. 11). The exercise groups improved more on ROM (flexion increased by 45% and extension increased by 109%) and strength (by 97%) at 2 months. Both exercise programs also extended the time to re-enter treatment for LBP by 51 weeks (95% CI=50−52) and 89 weeks (95% CI=69−110), respectively.76

Table 11.

Exercises for Postsurgery Back Pain at 1 Montha

Strength of Published Evidence in Comparison With Other Guidelines:

The Philadelphia Panel found good scientific evidence (level I) of clinically important benefit on pain and function with back extension and strengthening exercises. In contrast, the QTF20 found no scientific evidence for general exercises postsurgery.

Clinical Recommendations Compared With Other Guidelines:

The Philadelphia Panel recommends that there is good evidence to include strengthening and extension exercises (grade A for pain and function) as an intervention for postsurgery LBP. This is in agreement with the BMJ,22 which recommends strengthening exercises, and the QTF,20 which recommends therapeutic exercises. The BMJ22 guidelines reported that increased stress on the spine is a potential risk of therapeutic exercises.

Practitioner Agreement

  • Response rate for this EBCPG: 46%

  • Percentage of practitioners giving comments for this EBCPG: 24%

  • Agree with recommendation: 90%

  • Think a majority of my colleagues would agree: 83%

  • Will (or already) follow this recommendation: 91%

Practitioner Comments

  1. High-technology equipment (isotonic and isokinetic) is not practical in a clinical situation.

Panel's Response.

The EBCPG recommends that either high-technology exercise or low-technology (traditional strengthening and McKenzie exercises) be used for postsurgery LBP.

DISCUSSION

Evidence-based practice is rapidly growing in the rehabilitation domain,78 especially for LBP.6 This systematic review demonstrates that there are a number of rehabilitation interventions for LBP that have been shown in one or more controlled trials to provide a clinically important benefit. Such evidence is still needed for the other interventions.

However, as with all such reviews, there are a number of limitations. Methodologic issues such as the potential for publication bias, variations in the methodologic quality of the included trials, and lack of standardized outcomes are discussed in the Philadelphia Panel article on methodology in this issue (“Evidence-Based Clinical Practice Guidelines on Selected Rehabilitation Interventions: Overview and Methodology”).

The effectiveness of conservative management of people with LBP is a complex issue.6,79 Physical rehabilitation is defined as a combination of physical agents. Rehabilitation specialists often use concomitant interventions in their daily practice.80 Each intervention is usually used as an adjunct. Certain interventions such as cryotherapy, hot pack application, and massage are used for pain relief or as a treatment preparation before the main intervention. The use of a single and specific intervention does not reflect the complexity of the global approach adopted by rehabilitation specialists in real-life clinical situations.

The effectiveness of physical rehabilitation interventions for LBP are thought to be influenced by a number of risk factors,4,8184 including biological,85 psychosocial,8689 and occupational9094 health indicators. A multidimensional clinical evaluation is recommended in LBP management.95,96 It was not possible to examine these risk factors in this review.

The Philadelphia Panel EBCPGs for the management of LBP are largely in agreement with previous and relatively recent EBCPGs21,22 for LBP shown in Appendix 1. The Philadelphia Panel EBCPGs for LBP were developed based on a systematic grading of the evidence determined by an expert panel, and the evidence was derived from systematic reviews and meta-analyses using the Cochrane Collaboration methodology. The finalized guidelines were circulated for feedback from practitioners to verify their applicability and ease of use for practicing clinicians. This additional procedure provides credibility for rehabilitation specialists who intend to use these EBCPGs for LBP management in their daily practice.

Exercises

Our recommendations are in agreement with those of the AHCPR guidelines that continuation of normal activities (such as walking) is more effective than bed rest for the management of acute LBP.15

Extension, Strengthening, or Flexion Exercises:

Our systematic review3032,39,4447,51,52,76,97101 also showed that extension, flexion, or strengthening exercises are effective for subacute and chronic LBP and for postsurgery LBP. The results for acute LBP are in full agreement with guidelines and other reviews6,102 concerning moderate effectiveness of stretching or strengthening exercises, but highly effective “advice to stay active.”103 Certain authors recommend return to functional and work activities as soon as possible after lumbar injury to avoid the negative effects of immobilization and bed rest prescription.104,105 Task-oriented activities are recognized in rehabilitation. Patients with LBP benefit from these activities as they improve ADL. For chronic LBP, our results are concordant with the summary by van Tulder et al6 on the effects of exercise therapy (flexion, extension, strengthening). Exercise therapy for subacute and for postsurgery LBP was not considered by van Tulder et al.6 For subacute and chronic LBP conditions, our conclusions are in full agreement with the results of Faas.106 In future studies, it will be important to look at clarifying types of exercises used, adequate exercise intensity, and progression according to patient-specific classification of physical dysfunction, needs, treatment goals, and outcomes.95,107110

Mechanical Traction

Eleven RCTs of static, intermittent, or vertical traction have been conducted in acute, subacute, and chronic LBP.2427,3335,54,56,58,111 None of these RCTs showed evidence of clinically important benefit56 on any of the outcomes identified as patient-important outcomes by the Philadelphia Panel; thus, neither static nor intermittent mechanical traction was shown to be of clinically important benefit. This lack of important benefit was consistent for acute, subacute, and chronic LBP as well as for static and intermittent traction. These results agree with previous systematic reviews for acute and chronic LBP management,6 even though those reviews did not clearly distinguish between manual and mechanical traction.

It would have been of great interest to pool the results of different subgroups in order to evaluate whether the presence or absence of neurological signs influenced the results. Unfortunately, subgroup analyses across the different trials according to specific clinical characteristics were not possible due to noncomparable outcome measures. Our systematic review included a mix of patients with and without neurological signs in 4 out of 11 trials.23,34,54,58 Therefore, our results appear to be robust across mixed conditions. However, as none of these 4 trials included patients with disk involvement, no conclusions can be drawn about effects in patients with disk involvement.

Some have claimed that the clinical indication for static or sustained traction is the presence of a nuclear disk protrusion.112,113 Of the 11 trials, 2 included populations in which all patients had disk involvement,25,26 and 2 included a mix of patients with and without nuclear disk protrusions.50,57 The results of these 3 trials were consistent with the others, showing no evidence of clinical benefit. This point shows the importance of identifying homogenous subgroups of patients with LBP based on precise differential physical dysfunction diagnostic classes, such as nerve root adhesion, hypomobility dysfunction, and sacroiliac hypermobility.114 However, current literature does not support the suggestion that lumbar traction has a beneficial effect on LBP111,115 by repairing disk herniations.

Therapeutic Ultrasound

Therapeutic ultrasound has not been shown to provide clinically important benefit for acute LBP28 or chronic LBP.59 The Philadelphia Panel recommendation (level II, grade C) agrees with the AHCPR21 and BMJ22 guidelines that evidence for the effectiveness of therapeutic ultrasound is lacking. Comparison with the results of van Tulder and colleagues'6 systematic review was not possible because they found no RCTs that met their criteria. However, the QTF20 recommended therapeutic ultrasound for muscle spasm and pain relief. This QTF recommendation was based on common practice rather than any evidence from clinical trials. Furthermore, this recommendation grouped therapeutic ultrasound with thermotherapy.

The only 2 trials available were both of low quality (0 out of 5 on the Jadad scale11,12). Furthermore, the type of therapeutic ultrasound was continuous in the study by Roman59 and not specified in the other trial.28 A pulsed therapeutic ultrasound type may be more effective than continuous therapeutic ultrasound in acute conditions because of a nonthermal effect.116 These results concur with a previous systematic review,80 even though it was conducted for all musculoskeletal conditions.

TENS

Despite several RCTs of TENS, no consistent benefit was shown on clinically relevant outcomes (eg, pain, functional status, patient global assessment) for acute, subacute, or chronic LBP.29,50,6064 Subgroup analyses revealed no significant differences between acute and chronic conditions, low- and high-quality studies, or conventional and acupuncture-like application or duration of the TENS session. These results are in concordance with previous systematic reviews for acute and chronic LBP.6,117 The Philadelphia Panel EBCPGs (level I, grade C) are in agreement with AHCPR21 and BMJ22 guidelines, which do not recommend TENS for LBP, but they are in conflict with the QTF20 guidelines, which recommend TENS for pain relief. However, the QTF20 guidelines do not differentiate between electroanalgesia and TENS. Specific therapeutic application of TENS is of key importance. Vibratory stimulation has been recommended as part of the TENS application, especially for chronic pain relief.65,118,119 None of the trials included in our review considered vibratory stimulation. There is a need for strict and rigorous RCTs of TENS using combined vibratory stimulation.

Therapeutic Massage

There were insufficient data for the Philadelphia Panel to make a recommendation for therapeutic massage as an intervention for acute, subacute, or chronic LBP. The current results are in full agreement with 2 recent systematic reviews.6,120 However, a recent, large unpublished trial suggests that massage may have a clinically important benefit.75

The Philadelphia Panel recommendation agrees with the AHCPR,21 which concluded that insufficient evidence regarding massage was available. However, the Philadelphia Panel disagrees with both the BMJ22 and QTF20 recommendations that massage should be considered as a therapeutic option in LBP, particularly for the relief of muscle spasm. The QTF20 and BMJ22 recommendations were made based on common practice rather than any evidence from controlled clinical trials.

Massage is a global intervention involving a number of confounding variables related to the therapeutic application. For example, the effectiveness of massage is influenced by the types of maneuvers used, the massage approach adopted, years of experience of the therapist, the number and the size of the muscles involved, the patient position used, the pressure exerted, the rhythm and progression, and the frequency and duration of the treatment sessions.120

Thermotherapy

There was insufficient evidence to make a recommendation regarding thermotherapy as an intervention for acute or chronic LBP. These results are in agreement with a recent systematic review6 for chronic LBP and with the evidence ratings of the AHCPR21 and QTF20 guidelines. However, the QTF20 recommended thermotherapy as an intervention for acute LBP.

Only comparative trials56,66,121 were identified for chronic LBP. In these trials, the separate effects of ice or heat could not be determined because they were used in conjunction with other interventions or compared with other interventions of unknown therapeutic benefit.

Physiological studies have shown significant effects of cryotherapy on circulatory and temperature responses and on muscle spasm and inflammation,122,123 but the mechanism of action has not yet been fully elucidated.123 It is unknown whether these physiological effects translate to important effects on clinical outcomes (such as pain and functional status). Back muscles may be too thick to benefit from the penetrating effects of superficial, local ice application for pain relief.123

Electrical Stimulation

The Philadelphia Panel was unable to make a recommendation regarding electrical stimulation due to insufficient evidence. There are some head-to-head studies,61,124 but in the absence of placebo controls, efficacy cannot be established. This is in agreement with AHCPR21 guidelines. The BMJ22 and QTF20 did not evaluate this modality.

Electrical stimulation is thought to increase functional activity if the peripheral nervous system is intact.125 Only Pope et al124 excluded patients with LBP who had neurological deficits. It has been recommended for physical rehabilitation in the upper and lower extremities in patients with LBP.125 To our knowledge, no other systematic reviews have studied this specific therapeutic intervention.

EMG Biofeedback

No consistent clinically important benefit was found for EMG biofeedback for either acute LBP126 or chronic LBP.70,7274 Two trials used EMG biofeedback for back muscle relaxation.72,74 The other trials70,73,126 used EMG biofeedback to improve muscle activation control. Only 2 trials70,126 excluded patients with root and peripheral nerve entrapment. Electromyographic biofeedback126 may be important in the relief of muscle spasms in people with acute LBP.127 Our results are in concordance with other guidelines and the review by van Tulder et al.6

Combined Rehabilitation Interventions

For both acute and chronic LBP, the Philadelphia Panel concluded that insufficient evidence was available regarding rehabilitation interventions when several are combined. This differs from the QTF20 and BMJ22 guidelines, which both recommend that rehabilitation specialists use interventions in combination at their own discretion to achieve treatment goals. The extent to which multiple interventions are redundant needs to be assessed in the same fashion as individual interventions.

The studies of combinations that were found used heterogeneous modalities, many of which were not described in sufficient detail to allow replication.41,42,56,128130 Furthermore, comparisons among studies are difficult because the investigators did not select the same combination of rehabilitation interventions. Methodological quality was low in these studies.

Overall

The main difficulty in determining the effectiveness of rehabilitation interventions is the lack of well-designed prospective RCTs. Future research in physical therapy should adopt rigorous methods such as the use of an appropriate placebo (and double-blind procedure), adequate randomization, homogeneous sample of patients based on rigorous selection and diagnosis criteria, and adequate sample size to detect clinically important differences with confidence.

CONCLUSION

There is evidence to support and recommend the use of continued normal activities for acute nonspecific LBP and therapeutic exercises for chronic, subacute, and postsurgery LBP. These EBCPGs were developed with a transdisciplinary team approach, using a structured methodology for guideline development that includes practitioner feedback. There is a lack of evidence at present regarding whether to include or exclude the use of thermotherapy, therapeutic massage, EMG biofeedback, mechanical traction, therapeutic ultrasound, TENS, electrical stimulation, and combined rehabilitation interventions in the daily practice of physical rehabilitation.

Appendix 1.

Appendix 1.

Strength of Published Evidence and Clinical Recommendations of Previous Evidence-Based Clinical Practice Guidelines (EBCPGs) for Low Back Paina

Footnotes

  • Address all correspondence and requests for reprints to: Peter Tugwell, MD, MSc, Chair, Centre for Global Health, Institute of Population Health, 1 Stewart St, Rm 312, Ottawa, Ontario, Canada K1N 6N5 (ptugwell{at}uottawa.ca).

    This study was financially supported by an unrestricted educational grant from the Cigna Foundation, Philadelphia, Pa, USA, the Ministry of Human Resources and Development, Government of Canada (Summer Students Program), and the Ontario Ministry of Health and Long-Term Care (Canada). Ian Graham is a Medical Research Council Scholar, Canadian Institutes of Health Research (Canada).

    Acknowledgments: Summer students: Sarah Milne, Michael Saginur, Marie-Josée Noël, Mélanie Brophy, Anne Mailhot

  • * Oxford, England: The Cochrane Collaboration, 2000.

  • Dettori JR, Bullock SH, Sutlive TG, et al. The effects of spinal flexion and extension exercises and their associated postures in patients with acute low back pain. Spine. 1995;20:2303–2312.

References

View Abstract