Integrative therapies for low back pain that include complementary and alternative medicine care: a systematic review.

STUDY
DESIGN: Systematic review of the literature.
OBJECTIVE: To evaluate whether an integrated approach that includes different Complementary and Alternative Medicine (CAM) therapies combined or CAM therapies combined with conventional medical care is more effective for the management of low back pain (LBP) than single modalities alone. SUMMARY OF BACKGROUND DATA: LBP is one of the leading causes of disability worldwide, yet its optimal management is still unresolved.
METHODS: The PRISMA Statement guidelines were followed. The Cochrane Back Review Group scale was used to rate the quality of the studies found.
RESULTS: Twenty-one studies were found that met the inclusion criteria. The CAM modalities used in the studies included spinal manipulative therapy, acupuncture, exercise therapy, physiotherapy, massage therapy, and a topical ointment. Twenty studies included acupuncture and/or spinal manipulative therapy. Nine high quality studies showed that integrative care was clinically effective for the management of LBP. Spinal manipulative therapy combined with exercise therapy and acupuncture combined with conventional medical care or with exercise therapy appears to be promising approaches to the management of chronic cases of LBP.
CONCLUSIONS: There is support in the literature for integrated CAM and conventional medical therapy for the management of chronic LBP. Further research into the integrated management of LBP is clearly needed to provide better guidance for patients and clinicians.

INTRODUCTION

Low-back pain (LBP) is a complex disorder and one of the most significant healthcare challenges affecting modern society. In the United States, LBP is the fifth most common reason for all physician visits[1,2] and is the single most common cause for chronic or permanent impairment in adults under the age of 65 years.[3] LBP is now considered the leading cause of disability worldwide.[4] While substantial heterogeneity exists among epidemiological studies of LBP, estimates of the recurrence of activity limiting LBP at 1 year range from 24% to 80%,[5] leading to a startling economic burden that appears to continue to increase.[6]

The number of interventions available to manage LBP is also extensive.[7,8] While there are more than 1000 randomized controlled trials (RCTs) published on the topic,[9] the most effective management approach remains unclear.[8,10] Conventional treatments such as anti-inflammatory medications have been shown to have limited benefit in improving patient outcomes.[11] Opioids appear to offer short-term benefits for chronic LBP, but long-term effects and safety remain unproven.[12] Stabilization exercises have been reported to help decrease pain and disability,[13] and post-treatment exercise programs have been shown to prevent recurrences of back pain, but Choi et al.[14] found conflicting evidence for exercise as a treatment. An RCT comparing fusion surgery with conservative treatment showed conflicting results as well.[15] These uncertainties in the literature and in conventional medical practice may help explain why LBP is the most common reason for patients to seek out care from a complementary and alternative medicine (CAM) provider.[16] Yet, the literature in support of CAM for management of LBP is also limited. A review published by the Agency for Healthcare Research and Quality found that CAM therapies for back pain (such as acupuncture, massage, and spinal manipulative therapy [SMT]) provide a greater, albeit a modest, benefit as compared to usual medical care.[17]

With so many treatment options available and insufficient evidence of efficacy,[15] it is no surprise that there is little consensus among healthcare practitioners across disciplines and even within disciplines with regard to what might be the most appropriate management intervention for LBP. Hirsh et al[18] reported that there does not seem to be a one-size-fits-all approach to LBP management, that no single treatment approach is the panacea for all patients.

Previous reviews[2,19-28] have evaluated the use of single CAM therapies for LBP, but in practice it is also common for patients to seek care from multiple practitioners and therefore combine different types of therapies (such as combining conventional medical care, SMT, exercise, and/or acupuncture).[8] National survey data suggest that more than half of US adults with LBP seek care from medical doctors and one or more types of CAM modalities.[16] The increasing use and acceptance of CAM[29] makes this combined approach to care likely to become even more prevalent in the coming years,[30] yet very little information has been collected regarding the actual practice of integrative medicine for the treatment of LBP. Studies have proposed the use of multi-modal therapies for LBP management[15] with Guzman et al[31] reporting a benefit for multidisciplinary biopsychosocial rehabilitation for LBP and Flor et al[32] finding multidisciplinary treatments for chronic LBP to be superior to single-discipline treatments such as medical treatment or physical therapy. However, Flor's[32] review did not include RCTs, and neither Flor et al[32] nor Guzman et al[31] considered CAM therapies. Rubinstein et al[27] included a review of SMT used as an adjunct therapy but limited their search to acute LBP and did not investigate any other CAM modalities. In Rubinstein's[27] review of SMT for chronic LBP, SMT as an adjunct therapy demonstrated varying quality of evidence, but again, the study did not investigate other CAM modalities.[20] No systematic reviews appear to have examined the studies that combine CAM modalities or CAM and conventional medical approaches for LBP. We therefore conducted a systematic review to address this gap in knowledge by summarizing efficacy evidence of an integrated approach to managing LBP.

Key Points:

Low back pain is a global disorder that causes significant disability.

People with low back pain often seek care from a variety of practitioners, both conventional medical practitioners and complementary and alternative medicine (CAM) providers. The optimal management for low back pain is still debated.

This systematic review found 21 articles that described the management of low back pain by integrated therapy of conventional medicine and CAM modalities or CAM modalities combined.

Spinal manipulative therapy combined with conventional medical care or with exercise therapy and acupuncture combined with conventional medical care or with exercise therapy appear to be promising approaches to the management of chronic cases of low back pain.

Our aim was to summarize evidence in relation to the following questions:

Is an integrated approach that includes CAM therapies and conventional medical care more effective for the management of LBP than either alone?

Is an integrated approach that combines CAM modalities more effective than each applied alone?

MATERIALS AND METHODS

The PRISMA statement for reporting systematic reviews was used as guide during the development of this project.[33]

Eligibility Criteria

To be eligible for this systematic review, articles had to meet the following criteria:

Reported on an RCT.

Included the treatment of LBP.

At least one treatment group received integrated therapy that included at least one CAM modality (ie, two or more CAM modalities used together, or one or more CAM modalities used with conventional medical treatment). For this study, a CAM modality was defined as a non-mainstream health-care intervention that is used either together with conventional medical care or in place of it.[34] Conventional medicine was defined as care provided by a medical or osteopathic physician or allied healthcare providers (eg, physical therapist). Exercise therapy was defined as conventional medical care as it is most often considered a mainstream therapy provided by physical therapists.

Published in English.

An original study and published in a peer-reviewed journal.

At least one outcome measure for pain or disability used (visual analog scale for pain [VAS], Oswestry Low Back Pain Disability Questionnaire [Oswestry], Roland-Morris Disability Questionnaire [Roland-Morris], Short Form-36 [SF-36], von Korff Scales, Dartmouth-Northern New England Primary Care Cooperative Information Project [CO-OP], Aberdeen Low Back Pain Scale (Aberdeen), McGill Pain Questionnaire [MPQ], Numeric Rating Scale for Pain [NRS], Rating of Perceived Capacity of Spine, self-generated pain or disability questionnaires).

Information Sources

The search engines used for this systematic review were PubMed, Medline, Index to Chiropractic Literature, Academic Search Premier, CINAHL, Cochrane Center Registry of Controlled Trials, and OVID. References of selected articles were searched for any additional studies that were missed by the initial search.

Search restrictions were human subjects, English language, peer-reviewed journal, RCTs and articles published up to February 2013.

Search

The search string used for this systemic review was: (low back pain OR back pain) AND (integrative care OR multidisciplinary care OR chiropractic OR spinal manipulation OR Chinese medicine OR traditional chinese medicine OR acupuncture OR homeopathy OR Ayurveda OR herbal medicine OR nutrition OR nutritional supplements OR mind-body medicine OR massage therapy OR yoga) AND (randomized controlled trial).

Study Selection

The three authors met together and reviewed each article to determine if each one met the predetermined inclusion criteria. Disagreements between reviewers were resolved through discussion and consensus and by performing a literature review on the meaning of integrated CAM therapies.

Data Collection Process

One author extracted the data from the articles into a spreadsheet. This was reviewed by a second author and disagreements between the authors resolved by consensus.

Data Items

Information that was extracted from the studies included:

Type of LBP

Sample size

Age of participants

Adverse events

Interventions used

Outcome measures

Results of interventions

Risk of Bias in Individual Studies

The Cochrane Back Review Group (CBRG) scale[35] was used for rating the quality of the studies found (Table 1). All three authors rated each article independently and then met to discuss their ratings. When there was a disagreement the original article was reread until a consensus was formed. Articles that scored six points or more on the 12-point scale were rated as high-quality. Articles that scored less than six points should be rated as low-quality.[35]

Table 1

The Cochrane Back Review Group Scale Questions

	Was the method of randomization adequate?
	Was the treatment allocation concealed?
Was knowledge of the allocated intervention adequately prevented during the study?	Was the patient blinded to the intervention?
	Was the care provider blinded to the intervention?
	Was the outcome assessor blinded to the intervention?
Were incomplete outcome data adequately addressed?	Was the dropout rate described and acceptable?
	Were all randomized participants analyzed in the group to which they were allocated?
	Are reports of the study free of suggestion of selective outcome reporting?
Other sources of potential bias	Were the groups similar at baseline regarding the most important prognostic indicators?
	Were co-interventions avoided or similar?
	Was the compliance acceptable in all groups?
	Was the timing of the outcome assessment similar in all groups?

Planned Methods of Analysis

Data were extracted for the pain and disability effect estimates (between-group and/or within group differences) and for measures of variability (eg, confidence intervals) where possible. Clinical significance was determined using very minimal estimates based on existing literature.[36-41] A statistician analyzed whether studies achieved clinically significant results. Statistical significance was reported for outcomes that did not have validated clinically significant cutoffs.

For this review, cutoffs were set at[36-41]:

> 10 mm on a 100-mm visual analog scale for pain (VAS)

> 10 points on Oswestry Low Back Pain Disability Questionnaire (Oswestry)

> 3 points on Roland-Morris Disability Questionnaire (Roland-Morris)

≥ 5 points on Short Form-36 (SF-36)

> 10 points on von Korff Scales (Pain and Disability)

> 5 points on Dartmouth Primary Care Cooperative chart system (COOP)

> 10 points on Aberdeen Low Back Pain Scale (Aberdeen)

> 2 points McGill Pain Questionnaire (MPQ)

> 1 on a 10-point Numeric Rating Scale for Pain (NRS)

No further pooling of study results could be performed because of the heterogenicity of the outcome measures used.

Additional Analyses

The studies were grouped according to the modalities that were used in the treatment arms. Subgroup analysis was performed to determine the efficacy of the different integrative treatments for LBP. SMT was considered to be manual therapy delivered to correct spinal biomechanics. It could be applied by chiropractors, osteopaths, or physical therapists. Acupuncture could be provided alone or in conjunction with electrical stimulation applied directly to the needles. The points chosen could be traditional acupuncture points or muscle motor points or other points using neurological levels. Physiotherapy was considered any passive modality applied to the patient while exercise therapy was considered any active movement therapy to be performed by the patient.

The strength of the evidence for each group was calculated by looking at the degree of agreement on the effectiveness and the quality of the studies as determined by the CBRG score. The strength of the evidence was rated using a five-level scale to determine their place in the best evidence synthesis (Table 2).[42]

Table 2

Best Evidence Synthesis Scale

Level	Description	Evidence Required
1	Strong evidence	More than 75% of high-quality RCTs report the same results
2	Moderate evidence	One high-quality RCT and/or multiple low-quality RCTs report the same results
3	Limited evidence	One low-quality RCT or one high-quality RCT and one low-quality RCT with opposite conclusion
4	Conflicting evidence	Contradictory results among multiple RCTs
5	No evidence	No RCTs

Abbreviation: RCT, randomized controlled trial.

RESULTS

Study Selections

The search of PubMed, Medline, Index to Chiropractic Literature, Academic Search Premier, CINAHL, Cochrane Center Registry of Controlled Trials, and OVID provided total of 622 citations. Five hundred and ninety one of these articles were excluded for not meeting the inclusion criteria. The full text of the remaining 31 articles was obtained. The references of these articles were examined for any additional relevant studies and none were found. After review, nine of the 31 articles were excluded for not including an integrative CAM treatment arm or not including pain or disability outcome measures. One article was excluded because it was a duplicate. A total of 21 studies were found that met the inclusion criteria (Figure 1).

Figure 1

Flow of citations through the retrieval and screening process.

Study Characteristics

A total of 4400 patients were included in the 21 articles. Individual study sample sizes ranged from 32 to 1334. Two studies (10%) had fewer than 50 patients; five (24%) had between 50 and 100 patients, and 14 (67%) had more than 100 patients. Five studies (24%) included only acute LBP, 13 (62%) only subacute or chronic LBP (one of which specified patients with osteoarthritis in the low back), and three (14%) did not specify any chronicity (Table 3).

Table 3

Study Characteristics

Authors	Type of LBP	Sample size	Age	Interventions	Outcome Measures
Beyerman et al, 2006[47]	Secondary to osteoarthritis (chronic)	252	Not reported	TG: SMT + Moist heat	Lumbar spine range of motion Oswestry VAS
				CG: Moist heat
Bronfort et al, 2008[45]	Chronic (≥6 weeks)	174	20-60 y	TG: SMT + Strengthening exercise	VAS Roland-Morris COOP
				CG: Conventional medical care + Strengthening exercise
				CG: SMT + stretching exercise
Childs et al, 2004[43]	Not reported	131	18-60 y	TG: SMT + Exercise	Oswestry
				CG: Exercise
Eisenberg et al, 2007[62]	Acute	444	18 y and over	TG: Conventional medical care + Choice of acupuncture, chiropractic or massage therapy	NRS Roland-Morris Self-generated satisfaction scale Cost
				CG: conventional medical care
Gunn et al, 1980[60]	Chronic	56	25-56 y	TG: Acupuncture (dry needling) + Conventional medical care	Self-generated pain and work status questionnaire
				CG: Conventional medical care
Hancock et al, 2007[51]	Acute	240	Not reported	TG: SMT + Conventional medical care	VAS for number of days to recovery
				CG: Conventional medical care + Placebo SMT
				CG: SMT + Placebo conventional medical care
				CG: Placebo SMT + Placebo conventional medical care
Hurley et al, 2004[48]	Sub-acute	240	18-65 y	TG: SMT + Electric muscle stimulation	Roland-Morris VAS MPQ EQ-5D SF-36 Self-generated questionnaire for LBP recurrence, work absenteeism, exercise participation, analgesic medication consumption, and additional healthcare use
				CG: SMT
				CG: Electric muscle stimulation
Hurwitz et al, 2002[49]	Not reported	341	18 years and overy	TG: SMT with physiotherapy modalities	NRS for pain Roland-Morris
				CG: SMT
Itoh et al, 2009[57]	Chronic	32	61-81 y	TG: Acupuncture + TENS	VAS Roland-Morris
				CG: Waiting list control
				CG: Acupuncture
				CG: TENS
Jüni et al, 2009[50]	Acute	104	20-55 y	TG: Conventional medical care + SMT	NRS Analgesic use
				CG: Conventional medical care
Leibing et al, 2002[56]	Chronic	131	18-65 y	TG: Acupuncture + Active physiotherapy	VAS Pain Disability Index
				CG: Active physiotherapy
				CG: Sham acupuncture + Active physiotherapy
Mayer et al, 2005[61]	Acute	122	18-55 y	TG: Heat-wrap therapy + Exercise	Rating of perceived capacity-spine
				CG: Heat-wrap therapy
				CG: Exercise
				CG: Control (booklet)
Meng et al, 2003[58]	Chronic	55	60 y and over	TG: Acupuncture + Conventional medical care	Roland-Morris
				CG: Conventional medical care
Mohseni-Bandpei et al, 2006[44]	Chronic	120	18-55 y	TG: SMT + Exercise	VAS Oswestry Lumbar spine range of motion Surface electromyography Lumbar extension muscle endurance
				CG: Ultrasound + Exercise
Molsberger et al, 2002[59]	Chronic	186	20-60 y	TG: Acupuncture + Conventional orthopedic therapy	VAS Self-generated effectiveness of treatment Lumbar spine range of motion
				CG: Conventional orthopedic therapy
				CG: Sham acupuncture + Conventional orthopedic therapy
Niemistö et al, 2003[53]	Chronic	204	24-46 y	TG: SMT + Stabilizing exercise + Conventional medical care	VAS Oswestry Frequency of LBP
				CG: Conventional medical care
Ongley et al, 1987[52]	Chronic	81	21-70 y	TG: Forceful SMT + Injection of dextrose-glycerine-phenol into soft tissues + High dose of anesthesia	VAS Analgesic use Roland-Morris Adverse events
				CG: Low dose of anesthesia + Less forceful SMT + Placebo injection
UK BEAM Trial 2004[46]	Not reported	1,334	18-65 y	TG: SMT + exercise	Roland-Morris Von Korff scales Back beliefs questionnaire Fear avoidance belief questionnaire SF -36 EQ-5D
				CG: Conventional medical care
				CG: Exercise
				CG: SMT
Weiner et al, 2008[63]	Chronic	65	65 y and over	TG: Acupuncture (percutaneous electrical nerve stimulation [PENS]) + General conditioning and aerobic exercise (GCAE)	MPQ Roland-Morris
				CG: PENS
				CG: Sham PENS
				CG: Sham PENS + GCAE
Yeung et al, 2003[55]	Chronic	52	18-75 y	TG: Exercise + Electro-acupuncture	NRS Aberdeen
				CG: Exercise
Zhang et al, 2008[54]	Acute	36	Not reported	TG: SMT + Biofreeze topical application	VAS Roland-Morris Low back muscle surface electromyography Heart rate variability
				CG: SMT

Abbreviations: CG, control group; COOP, Dartmouth Primary Care Cooperative chart system; EQ-5D, European Quality of Life; LBP, low back pain; MPQ, McGill Pain Questionairre; NRS, numeric rating scale; SF-12 & 36: Short Form 12 & 36; SMT, spinal manipulative therapy; TENS, transcutaneous electric nerve stimulation; TG, treatment group.

Eleven studies (52%) were limited to middle-aged participants. Three (14%) of the studies only included seniors, five studies (24%) included middle-aged and senior participants and two studies (10%) did not specify the age of the participants.

The modalities used in the studies included: SMT (13), conventional medical care (8), acupuncture (8), exercise (8), physiotherapy (5), massage therapy (1), topical application (1).

No studies were found that used integrated therapies that included homeopathy, Ayurveda, herbal medicine, nutrition, nutritional supplements, mind-body medicine, or yoga.

Risk of Bias Within Studies

The risk of bias within studies was determined using the CBRG scale (Table 4). Seventeen of the 21 studies (81%) were judged to be of high and four (19%) of low-quality. The most common problems encountered were lack of blinding of the participants, providers, and outcome assessors. Only eight of the studies reported on the use of any co-interventions by the participants and six on the participant's compliance with the treatment protocol.

Table 4

The Cochrane Back Review Group Scores

	Randomization Adequate	Treatment Allocation Concealed	Patient Blinded	Care Provider Blinded	Outcome Assessor Blinded	Drop-Out Rate	Intention to Treat	Selective Outcome Reporting	Groups Similar at Baseline	Co-interventions	Compliance	Timing of the Outcomes	Score
Beyerman, 2006[47]	?	–	–	–	–	+	–	+	+	?	?	+	4
Bronfort, 2008[45]	+	+	–	–	–	+	+	+	+	+	+	+	9
Childs, 2004[43]	+	+	–	–	–	–	+	+	+	?	+	+	7
Eisenberg, 2007[62]	+	+	–	–	–	+	+	+	+	+	?	+	8
Gunn, 1980[60]	+	+	–	–	?	+	+	+	?	?	?	+	6
Hancock, 2007[51]	+	+	+	?	+	+	+	+	+	+	+	+	11
Hurley, 2004[48]	+	+	–	–	–		+	+	+	?	+	+	8
Hurwitz, 2002[49]	+	+	–	–	–	+	+	+	+	+	?	+	7
Itoh, 2009[57]	+	+	–	–	–	+	?	+	+	?	?	+	6
Jüni, 2009[50]	+	+	–	–	–	+	+	+	?	+	+	+	8
Leibing, 2002[56]	–	–	?	–	?	–	+	+	+	?	?	+	4
Mayer, 2005[61]	+	+	–	–	–	+	–	+	+	?	?	+	6
Meng, 2003[58]	+	+	–	–	–	–	+	+	+	+	?	+	7
Mohseni-Bandpei, 2006[44]	+	+	–	–	+	–	+	?	+	+	?	+	7
Molsberger, 2002[59]	+	+	?	–	?	+	+	+	+	?	?	+	7
Niemistö, 2003[53]	+	+	–	–	–	+	+	+	+	+	?	+	8
Ongley, 1987[52]	+	+	+	–	+	?	–	+	+	?	?	+	7
UK Beam, 2004[46]	+	+	–	–	–	–	–	+	–	?	–	+	4
Weiner, 2008[63]	+	+	?	?	?	+	+	+	+	?	+	+	8
Yeung, 2003[55]	+	+	–	–	–	+	+	+	+	?	?	+	7
Zhang, 2008[54]	+	+	–	–	–	?	?	+	+	?	?	+	5

+, yes (1 point); –, no (0 points); ?, unclear (0 points).

Results of Individual Studies

A wide range of outcome measures were used in the studies included in this systematic review (Table 5). All of the outcome measures included were validated except Gunn et al,[60] who used a self-generated questionnaire. Some studies reported only the within-group changes over time while others only reported between-group differences. Seventeen studies (81%) used primary outcome measures for pain or disability due to LBP. Even this consistency was confounded by the use of many different instruments that purport to measure the same outcome. Other primary outcome measures used included lumbar range of motion, satisfaction with care, prescription medication use, surface electromyography (EMG), lumbar muscle endurance, quality of life, perceived capacity of spine, adverse events, and heart rate variability.

Table 5

Study Outcomes

Study Name		Group 1	Group 2	Group 3	Group 4
Beyerman, 2006[47]	Intervention	SMT + Moist Heat	Moist Heat
	N	143	109
	Oswestry mean ± SDb	8.56 ± 7.10a	12.82 ± 7.66a
	VAS mean ± SDb	2.55± 2.01a	3.99± 2.23a
Bronfort, 2008[45]	Intervention	SMT + Strengthening Exercise	Conventional Medical Care + Strengthening Exercise	SMT + Stretching Exercise
	N	71	52	51
	VAS mean ± SD	Mean not reporteda	Mean not reporteda	Mean not reporteda
	Roland-Morris	Mean not reporteda	Mean not reporteda	Mean not reporteda
	COOP	Mean not reporteda	Mean not reporteda	Mean not reporteda
Childs, 2004[43]	Intervention	SMT + Exercise	Exercise
	N	70	61
	Oswestry mean differences (95% CI)	8.3 (2.4, 14.2)[b,c]
Eisenberg, 2007 [62]	Intervention	Conventional Medical Care + Choice of Acupuncture, Chiropractic or Massage Therapy	Conventional Medical Care
	N	300	150
	NRS mean change after treatment (IQ)b	–5 (–7,–3)	–4 (–7,–2)
	Roland-Morris change	–9 (–15,–4)	–8 (–13,–2)
Gunn, 1980 [60]	Intervention	Acupuncture (Dry Needling) + Conventional Medical Care	Conventional Medical Care
	N	29	27
	Self-generated pain and work status b	Not reported	Not reported
Hancock, 2007 [51]	Intervention	Conventional Medical Care + SMT + Diclofenac	Conventional Medical Care + Placebo SMT + Diclofenac	Conventional Medical Care + SMT + Placebo Diclofenac	Conventional Medical Care + Placebo SMT + Placebo Diclofenac
	N	60	60	59	60
	Treatment Modality	Diclofenac	Placebo Diclofenac	SMT	Placebo SMT
	Days	13 (10-16)	16 (14-18)	15 (13-18)	15 (12-19)
	Number of days to a full day of zero on the VAS ( 95% CI)
Hurley, 2004 [48]	Intervention	SMT	Electric Muscle Stimulation	SMT + Electric Muscle Stimulation
	N	80	80	80
	Change in Roland-Morris mean (95% CI)	–4.53 (–5.7,–3.3)a	–3.56 (–4.8,–2.4)a	–4.65 (–5.8,–3.5)a
	Change in VAS (mm) mean (95% CI)	–19.88 (–26.1,–13.7)a	–21.38 (–27.5,–15.2)a	–24.69 (–30.8,–18.6)a
Hurwitz, 2002 [49]	Intervention	SMT	SMT + Physiotherapy
	N	169	172
	Clinically significant Improvement on NRS (2 points/10)	34.5%	45%
	Change in NRS (95% CI)	1.04 (0.74, 1.35)	1.35 (1.05, 1.66)
		–0.31 (–0.13, 0.75)
	Clinically significant Improvement on Roland-Morris	43.8%	51.5%
	Change in Roland-Morris (95% CI)	3.18 (2.48, 3.88)	3.16 (2.46, 3.86)
		0.02 (–1.02,0.97)
Itoh, 2009 [57]	Intervention	Acupuncture + TENS	No Specific Treatment	Acupuncture	TENS
	N	6	7	7	6
	VAS mean ± SD	36.6 ± 8.0a	53.1 ± 27.9	37.4 ± 25.8	53.2 ± 25.1
	Roland-Morris mean ± SD	7.3 ± 4.9a	9.8 ± 0.8	5.4 ± 3.4	6.2 ± 3.4
Jüni, 2009 [50]	Intervention	Conventional Medical Care + SMT	Conventional Medical Care
	N	52	52
	11-point Box scale for pain mean difference (95% CI)	0.6 (–0.1, 1.3)
	Analgesic dose mean difference (95% CI)	–13 (–42, 15)
Leibing, 2002 [56]	Intervention	Acupuncture + Active Physiotherapy (AG)	Active Physiotherapy (CG)	Sham Acupuncture + Active Physiotherapy (SG)
	N	40	46	45
		–2.7 ± 2.2	–1.0 ± 1.7	–2.1 ± 2.2
	Change in VAS mean ± SD	Contrast	Difference in Change in VAS Mean (95% CI)
		AG vs SG	–0.6 (–1.65, 0.45)
		AG vs CG[b,c]	–1.7 ( –2.71,–0.62)
		–13.9 ± 15.0	–2.6 ± 7.8	–9.7 ± 10.5
	Change in pain disability index (PDI) mean ± SD	Contrast	Difference in Change in PDI (95% CI)
		AG vs SG	–4.2 (–9.99, 1.71)
		AG vs CG[b,c]	–11.3 ( –17.01,–5.44)
Mayer, 2005 [61]	Intervention	Heat-Wrap Therapy + Exercise	Heat-Wrap Therapy	Exercise	Control (Booklet)
	N	24	25	25	26
	Rating of Perceived Capacity-Spine (RPC-S)	Treatment Contrast	Relative Increase
		Heat + Exercise vs Heatb	84%
		Heat + Exercise vs Exerciseb	95%
		Heat + Exercise vs Bookletb	175%
Meng, 2003 [58]	Intervention	Acupuncture + Conventional Medical Care	Conventional Medical Care
	N	31	24
	Change in Rolland Morris questionnaire [b,c]	4.1 ± 3.9a	0.7 ± 2.8
Mohseni-Bandpei, 2006 [44]	Intervention	SMT + Exercise	Ultrasound + Exercise
	N	60	60
	Change in VAS (95% CI)	41.6 (4.2, 49.6)a	25.1 (17.7, 32.5)a
		16.4 (6.1, 26.8)[b,c]
	Change in Oswestry (95% CI)	17.9 (14.0, 21.8 )a	10.1 (6.2, 13.9)a
		7.8 (2.4, 13.2)b
Molsberger, 2002 [59]	Intervention	Acupuncture + Conventional Orthopedic Therapy (AC)	Conventional Orthopedic Therapy (C)	Sham Acupuncture + Conventional Orthopedic Therapy (SC)
	N	65	60	61
	VAS mean ± SD	26 ± 21	39 ± 21	36 ± 19
		77% (62%, 88%)	14% (4%, 30%)	29% (16%, 46%)
	VAS 50% pain relief (95% CI)	Contrast
		AC vs SC[b,c]	Difference in means not reported
		AC vs Cb	Difference in means not reported
Niemist ӧ, 2003 [53]	Intervention	SMT + Stabilizing Exercise + Conventional Medical Care	Conventional Medical Care
	N	102	102
	VAS mean ± SD[b,c]	25.2 ± 23.3a	36.1 ± 23.3a
	Oswestry mean ± SDb	14.7 ± 11.6a	18.6 ± 11.6a
	% with daily LBP	37%a	39%a
Ongley, 1987 [52]	Intervention	Forceful SMT + Injection of Dextrose-Glycerine-Phenol into Soft Tissues + High-dose Anesthesia	Low Dose Anesthesia + Less Forceful SMT + Placebo Injection
	N	40	40
	VAS mean ± SD [b,c]	1.77 ± 0.22	2.93 ± 0.25
	Roland-Morris [b,c]	4.70 ± 0.73	8.49 ± 1.04
UK BEAM Trial, 2004 [46]	Intervention	General Practice (GP)	Exercise (E)	SMT (S)	SMT + Exercise (SE)
	N	338	310	353	333
	Roland-Morris mean difference	Contrast	Mean difference (95% CI)
		GP vs Eb	1.36 (0.63, 2.10)
		GP vs Sb	1.57 (0.82,2.32)
		GP vs SEb	1.87 (1.15,2.60)
	Von Korff scales mean difference	Contrast	Mean difference (95% CI)
		GP vs Eb	5.03 (1.02,9.05)
		GP vs S	.97 (-0.050,7.98)
		GP vs SEb	5.51 (1.75,9.28)
	Von Korff scales mean difference for pain	Contrast	Mean difference (95% CI)
		GP vs Eb	4.59 (0.43,8.75)
		GP vs Sb	8.90 (4.84,12.95)
		GP vs SEb	8.21 (4.20,12.21)
Weiner, 2008 [63]	Intervention	Acupuncture with PENS (A)	Acupuncture with PENS + Exercise (AE)	Acupuncture with sham PENS (S)	Acupuncture with Sham PENS + Exercise (SE)
	N	47	45	48	44
		–2.9 ± 9.2a	–4.1 ± 8.2a	–2.3 ± 6.3a	–3.1 ± 7.9a
	MPQ mean ± SD	Contrast	Mean ± SD
		A vs S	0.5 ± 1.4
		AE vs SE	–0.6 ± 1.5
		AE vs A	–1.4 ± 1.4
		SE vs S	–0.3 ± 1.4
		–2.6 ± 4.5a	–2.6 ± 4.6a	–2.7 ± 3.8a	–3.0 ± 4.7a
	Roland-Morris mean ± SD	Contrast	Mean ± SD
		A vs S	0.1 ± 0.9
		AE vs SE	0.2 ± 0.9
		AE vs A	–0.2 ± 0.9
		SE vs S	–0.2 ± 0.9
Yeung, 2003 [55]	Intervention	Exercise	Exercise + Electro Acupuncture
	N	26	26
	NRS mean ± SD [b,c]	5.12 ± 2.18	3.81 ± 2.10
	Aberdeen mean ± SD [b,c]	30.82 ± 13.03	20.02 ± 10.47
Zhang, 2008 [54]	Intervention	SMT	SMT + Topical Application
	N	18	18
	Roland- Morris questionnaire mean ± SD	3.600 ± 5.412	8.000 ± 3.807
	VAS mean ± SD	5.2 ± 2.167	1.333 ± 1.732a

Significant within-group difference at P=.05.

Significant between-groups difference at P=.05.

Clinically significant between-groups difference.

Abbreviations: CG, control group; COOP, Dartmouth Primary Care Cooperative chart system; EQ-5D, European Quality of Life; LBP, low back pain; MPQ, McGill Pain Questionairre; NRS, numeric rating scale; SF-12 & 36: Short Form 12 & 36; SMT, spinal manipulative therapy; TENS, transcutaneous electric nerve stimulation; TG, treatment group.

Some authors used multiple VAS for pain, such as current, average, and most severe pain. To simplify analysis, we used only average pain for this systematic review when multiple VAS scores for pain were reported.

When authors reported outcomes at multiple time points we chose to only include those taken at the conclusion of the intervention period.

Synthesis of Results

Studies were organized by the type of modalities used for treating LBP, the quality of the studies based on the CBRG score and whether the treatment was clinically effective. Levels of best evidence (Table 6) were calculated for each type of treatment using methods described above.

Table 6

Best Evidence Synthesis

Modalities Used	Clinically Effective		Clinically Not Effective		Level of Best Evidence
Quality of Study	High	Low	High	Low
SMT and Exercise	Childs, 2004[43] Mohseni-Bandpei, 2006[44]	UK BEAM, 2004[46]	Bronfort, 2008[45]		Clinically effective level 2
SMT and Physiotherapy		Beyerman, 2006[47]	Hurley, 2004[48] Hurwitz, 2002[49]		Clinically not effective level 2
SMT and Conventional Medical Care	Ongley, 1987[52]		Jüni, 2009[50]		Clinically not effective level 4
			Hancock, 2007[51]
SMT, Exercise and Conventional Medical Care	Niemistö, 2003[53]				Clinically effective level 2
SMT and Topical Ointment				Zhang, 2008[54]	Clinically not effective level 3
Acupuncture and Exercise	Yeung, 2003[55]	Leibing, 2002[56]	Weiner, 2008[63]		Clinically effective level 3
Acupuncture and Physiotherapy			Itoh, 2009[57]		Clinically not effective level 3
Acupuncture and Conventional Medical Care	Gunn, 1980[60] Meng, 2003[58] Molsberger, 2002[59]				Clinically effective level 1
Exercise and Physiotherapy	Mayer, 2005[61]				Clinically effective level 2
Conventional Medical Care with Choice of Acupuncture, Chiropractic, or Massage			Eisenberg, 2007[62]		Clinically not effective level 3

Quality as measured on Cochrane Back Review Group scale: High, ≥6; Low, < 6.

Abbreviation: SMT, spinal manipulative therapy.

Spinal Manipulative Therapy and Exercise

A total of four studies were identified in this category. All of these studies were pragmatic trials, and all or most participants received SMT that included high-velocity thrusts. Childs et al[43] and Mohseni-Bandpei et al[44] had physiotherapists performing the SMT, Bronfort et al[45] used chiropractors, and the UK BEAM trials[46] used chiropractors, osteopaths, and physiotherapists. The UK BEAM[46] study reported no adverse effects while Bronfort et al[45] reported adverse effects mainly from NSAIDs. Childs et al[43] and Mohseni-Bandpei et al[44] did not report if there were any adverse effects. All trials except for Bronfort et al[45] reported clinically effective results from the integrated therapy. Of the three high-quality studies,[43-45] Childs et al[43] and Mohseni-Bandpei et al[44] reported that integrative treatments were more effective than the comparison therapy, leading to a level 2 best evidence score for using the integrated therapy.

Spinal Manipulative Therapy and Physiotherapy

A total of three studies were identified in this category. The Beyerman et al[47] participants all had chronic LBP due to osteoarthritis. The Hurley et al[48] participants had subacute LBP and the Hurwitz et al[49] study did not specify the chronicity of the participants' LBP. The Hurley et al[48] and Hurwitz et al[49] studies were pragmatic, while the Beyerman et al[47] participants all received a prescribed treatment regime. All three studies included a mix of high-velocity, low-amplitude thrusts and mobilization. Beyerman et al[47] and Hurwitz et al[49] used chiropractors to deliver the SMT, while Hurley et al[48] used physiotherapists. Beyerman et al[47] used moist heat as the physiotherapy, Hurley et al[48] used electric muscle stimulation, and Hurwitz et al[49] used the above plus cold and ultrasound. No adverse events were reported by any of these studies. The Beyerman et al[47] study found that the integrated therapy was clinically effective; however, the two high-quality studies in this group, Hurley et al[48] and Hurwitz et al,[49] did not find a clinically significant effect for the integrative therapies, leading to a level 2 best evidence score against using the integrated therapy.

Spinal Manipulative Therapy and Conventional Medical Care

A total of three studies were identified in this category. The Jüni et al[50] and Hancock et al[51] participants had chronic LBP, while Ongley's[52] had acute LBP. Ongley et al[52] reported using only forceful SMT. The practitioner type was not reported. Jüni et al[50] and Hancock et al[51] reported using a mix of high-velocity, low-amplitude thrusts, and mobilization. The Hancock et al[51] study used SMT performed by physiotherapists, while Jüni et al[50] used a mix of different practitioner types. Hancock et al[51] and Jüni et al[50] reported adverse events. All three studies were high-quality. Ongley et al[52] reported clinically significant effectiveness from the integrative therapy while the other two did not, leading to a level 4 best evidence score (conflicting evidence).

Spinal Manipulative Therapy, Exercise, and Conventional Medical Care

Only the Niemesto et al[53] study was identified in this category. Participants all had chronic LBP. The SMT used was mobilization using a muscle energy technique. They did not report which profession performed the SMT. No adverse effects were reported by any of the participants. This high-quality study found that the integrated treatment was effective, leading to a level 2 best evidence score for using the integrated therapy.

Spinal Manipulative Therapy and Ointment

Only the Zhang et al[54] study was identified in this category. The participants had acute LBP. The type of SMT performed was a diversified high-velocity, low-amplitude thrust performed by a chiropractor. A topical menthol ointment (Biofreeze, The Hygenic Corporation, Akron, Ohio) was applied by the participants. The occurrence of adverse events was not reported. There were no clinically significant effects observed from the treatments. This was a low-quality study and led to a level 3 best evidence score against using the integrated therapy.

Acupuncture and Exercise

Three articles were identified in this category. Participants all had chronic LBP. Yeung et al[55] and Leibing et al[56] used standardized acupuncture points, while Weiner et al[63] used motor and other points according to a standardized program using neurological levels. All three studies applied electric stimulation to the needles. Yeung et al[55] prescribed back-strengthening and stretching exercise, Leibing et al[56] prescribed standardized active physiotherapy, and Weiner[63] prescribed general conditioning and aerobic exercises. Yeung et al[55] and Weiner et al[63] reported no adverse events from the therapy, while Leibing et al[56] reported pain and problems with circulation. Yeung et al[55] and Leibing et al[56] found this integrated therapy to be more effective than acupuncture alone, while Weiner et al found no difference between the integrated therapy and acupuncture or exercise therapy alone. The Yeung et al and Weiner et al studies were high-quality, while the Leibing et al study was low-quality, leading to a level 3 best evidence score (conflicting evidence).

Acupuncture and Physiotherapy

Only the Itoh et al[57] article was identified in this category. The participants all had chronic LBP. The authors used a standardized set of acupuncture points for needling and applied transcutaneous electric nerve stimulation (TENS) stimulation using pads separately from the needles. The only adverse event reported was deterioration of the symptoms in one participant. This study was high-quality and did not find the integrated therapy to be clinically effective. This led to a level 3 best evidence score against using this integrated therapy.

Acupuncture and Conventional Medical Care

Three studies that used integrated acupuncture and conventional medical treatments were found. All participants had chronic LBP. Meng et al[58] and Molsberger et al[59] used a standardized set of traditional acupuncture points with the option to choose additional points. Gunn et al[60] chose to needle muscle motor points instead of traditional acupuncture points. Meng et al[58] reported adverse events such as aches, bruises, light-headedness, and increased pain. Gunn et al[60] and Molsberger et al[59] did not report on the occurrence of adverse events. All three studies were high-quality and found that the integrated therapy was more effective than conventional medical care alone, leading to a level 1 best evidence score for using this integrated therapy.

Exercise and Physiotherapy

Only the Mayer et al[61] study was found in this category. The participants all had acute LBP and received heat wrap therapy and McKenzie protocol exercises. No adverse events from the therapy were reported. This high-quality study found that the integrative therapy was clinically effective, leading to a level 2 best evidence score for using this integrated therapy.

Conventional Medical Care With Choice of Acupuncture, Chiropractic, or Massage

Only the Eisenberg et al[62] study was found in this category. All patients had acute LBP. Of the participants in the treatment group, 51.4% chose massage therapy, 25.7% chose chiropractic, 19.6% chose acupuncture, and 3.4% declined additional therapy. No details on the conduction of CAM therapies were provided by the authors. Adverse events reported included minor discomfort among 5% of those receiving acupuncture care, 8% of those receiving chiropractic care, and 7% of those receiving massage. This study, deemed high-quality by the CBRG scale, found that the integrative care was not clinically effective, leading to a level 3 best evidence score against using this integrated therapy.

DISCUSSION

LBP remains a vexing problem throughout the world, causing significant pain and disability. Many years of research into the optimal approach to treating LBP have failed to find an ideal solution for all cases. Most likely, various combinations of modalities will be necessary for the different populations of LBP.[30,31] This systematic review set out to determine if an integrated approach including combinations of CAM therapies and conventional medical care would be more effective than single modality treatment for the management of LBP. Overall, the results of this systematic review indicate that integrated modalities seem to be effective for this condition.

Twenty of the 21 studies included acupuncture and/or SMT. This is probably due to these modalities being used commonly to treat LBP and the more advanced state of the research communities in the professions that use these interventions.

Complementary and Alternative Medicine Combined With Conventional Medical Care

The combination of conventional medical therapy and CAM modalities appears to be the most promising approach to the management of chronic cases of this complex condition. A consensus among three high-quality trials that used acupuncture and conventional medical integrated therapy found it to be more effective for the management of chronic LBP than conventional medical care alone.[58-60] However, none of these studies compared the integrated therapy with acupuncture alone. Future studies are required to determine if this integrated therapy is better than acupuncture alone. There were also no studies found on the use of this integrated therapy for acute LBP.

Combining conventional medical therapy and SMT led to conflicting evidence depending on the chronicity of the LBP. Two high-quality studies used participants with acute LBP and did not find that the integrated therapy was more effective than conventional medical care alone.[50,51] One high-quality study used participants with chronic LBP and found that the integrated therapy was more effective than conventional medical therapy alone.[52]

The optimum management of acute LBP requires further study. In addition to the above studies, Eisenberg et al[62] found that offering patients with acute LBP the choice of a CAM modality (acupuncture, chiropractic, or massage therapy) in addition to conventional medical care increased patient satisfaction and costs but did not improve outcomes. This may be due to the nature of acute LBP. In many cases it initially seems to resolve regardless of the therapy used, but there is high chance of recurrence. Long-term studies need to be performed to clarify which therapies are most effective in preventing future episodes.

Complementary and Alternative Medicine Combined With Active or Passive Care

Exercise has the advantages of being low risk, cost-effective approach that transfers responsibility for care to the patient. The addition of exercise to CAM therapies appears to be a promising approach to managing acute and chronic LBP. Five studies examined integrated therapy combining SMT and exercise. Of these, three high-quality[43-45] and one low-quality study[46] found that the integrated therapy was more effective than conventional medical therapy, physiotherapy, and exercise or exercise alone. One of these, Niemisto et al,[53] combined conventional medical care, SMT, and exercise therapy. The Bronfort et al study[45] had three groups which all received integrative therapy: SMT and strengthening exercise, SMT and stretching exercise, or conventional medical care with strengthening exercises. The three groups were all found to be equally effective, leading to the conclusion that SMT and strengthening exercise therapy was not more effective than the other two groups. However, the design of this study precludes coming to any conclusion about the effectiveness of integrated SMT and exercise compared to SMT or conventional medical therapy alone since all groups included some form of exercise.

Three studies[55,56,63] addressed the use of integrated therapy including acupuncture and exercise therapy for chronic LBP. One high-quality[55] and one low-quality[56] study found that this integrated therapy was more effective than exercise alone or with sham acupuncture. Weiner et al[63] did not find this integrated therapy to be more effective than acupuncture alone, but the participants in this study were all seniors aged 65 years or older, which may have impacted the outcomes.

While integrated CAM therapy with active care appears to be effective, combining passive physiotherapy modalities with CAM therapy was found to be generally ineffective.[48,49,56,57] This was a surprising finding considering that this combination is often used as a standard treatment for LBP. This reinforces the modern paradigm that passive care alone may not be helpful in the management of LBP. On the other hand, combining active care with passive physiotherapy may be more effective as demonstrated by the Mayer et al[61] study, which found that integrated therapy combining exercise with heat wrap was more effective than heat wrap or exercise alone or a control group. The authors described the participants of the study as having acute LBP, but the inclusion criteria specified that the condition had to have been present between 2 days and 3 months, which is commonly considered subacute.

Challenges of Integrated Care

There are some challenges to combining CAM modalities together or with conventional medical therapy. If multiple practitioners treat the same patients without coordinating care, there will be additional costs and time spent on each therapy, fragmentation of patient records, and duplication of services.[8] There may also be an inconvenience to the patient if more than one clinic needs to be visited. These challenges can partially be mitigated by the increasing prevalence of multidisciplinary clinics, interprofessional collaboration and practitioners who can offer multiple therapies. For example, contemporary chiropractic and acupuncture education includes exercise therapy as part of their curriculum. Increased short-term costs may be justified if the integrated therapy leads to reduced chronic disability.

LIMITATIONS

One limitation was the large number of CAM modalities that could potentially be used in an integrated approach for LBP and the limited number of studies found. For several combinations of modalities, no articles could be identified. Among the integrated therapies that were included in this study, for each there were only one to four studies identified. This makes it impossible to draw a definitive conclusion regarding the most effective treatment for LBP. For example, even though every study except one used SMT or acupuncture treatment, no study was found that integrated these two therapies. In addition, the control group of many trials only included one of the interventions that were part of the integrated therapy. For example, all studies of integrated acupuncture and conventional medical care compared it to conventional medical care alone, but not to acupuncture alone.

Another limitation of this study was that only articles published in English were selected. This may have led to missing articles published in other languages. This is a particular concern for articles about acupuncture, which are often published in Chinese language journals.

Most of the studies included only participants who were described as having chronic LBP. Studies on LBP are often inconsistent when describing the chronicity categories of acute, subacute, chronic, and recurrent or do not describe this aspect at all. This may lead to uncertainty when attempting to translate research findings to clinical care. To add to the confusion, recent studies have pointed out that many cases of LBP that are originally considered acute and then resolved instead lead to recurrent LBP. Interventions that do not seem effective in the short-term may prove to be superior in preventing chronic recurrences and disability. Future studies should also explore the effects of integrated therapies on acute, chronic, and recurrent LBP. It is important for future studies to include long-term follow-up to determine if cases of LBP have truly resolved.

A limitation to combining results of the studies was the heterogeneity of the treatment modalities and outcomes used. For example, SMT delivered by chiropractors may differ significantly from that delivered by physical therapists. Some acupuncture trials used traditional points, while others used motor points. Numerous variations of pain and disability scales were used by the studies described in this article. In addition, Gunn et al[60] used a self-generated LBP and disability questionnaire for the primary outcome measure, and Hancock et al[51] used a VAS for pain in an unconventional manner for the primary outcome measure (number of days until zero is recorded).

One common concern for all studies of CAM modalities is that blinding of the patient and practitioner is often impossible. This combined with the common use of subjective pain and disability scales also led to the outcomes not being blinded. Despite this, 17 of the 21 studies scored in the high-quality range as determined by the CBRG scale.

CONCLUSIONS

Patients often try an integrated approach for treatment of LBP, using a combination of conventional medical care and CAM modalities. Previous systematic reviews have examined the use of individual CAM modalities for LBP and found promising results. This systematic review sought to determine if an integrated approach that includes different CAM therapies or CAM therapies combined with conventional medical care is more effective for the management of LBP than either alone. The studies found support the conclusion that integrated therapy which includes SMT combined with exercise therapy and acupuncture combined with conventional medical care or with exercise therapy appears to be more effective than select single therapies alone for treating LBP, although many questions remain. More studies are needed as most of the articles included participants with chronic LBP and there is a lack of RCTs for many CAM modalities used in an integrated manner. Further research into the integrated management of LBP is clearly needed to provide better guidance for patients and clinicians, as is the development of researchers with expertise in CAM modalities. In particular, there is a need for long-term studies that use cost effectiveness in addition to pain and disability from LBP as outcomes.