Spinal manipulative therapy for acute low‐back pain

Sidney M Rubinstein; Caroline B Terwee; Willem JJ Assendelft; Michiel R de Boer; Maurits W van Tulder

doi:10.1002/14651858.CD008880.pub2

. 2012 Sep 12;2012(9):CD008880. doi: 10.1002/14651858.CD008880.pub2

Spinal manipulative therapy for acute low‐back pain

Sidney M Rubinstein ^1,^✉, Caroline B Terwee ², Willem JJ Assendelft ^3,⁴, Michiel R de Boer ⁵, Maurits W van Tulder ⁵

PMCID: PMC6885055 PMID: 22972127

Abstract

Background

Many therapies exist for the treatment of low‐back pain including spinal manipulative therapy (SMT), which is a worldwide, extensively practised intervention. This report is an update of the earlier Cochrane review, first published in January 2004 with the last search for studies up to January 2000.

Objectives

To examine the effects of SMT for acute low‐back pain, which is defined as pain of less than six weeks duration.

Search methods

A comprehensive search was conducted on 31 March 2011 in the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, CINAHL, PEDro, and the Index to Chiropractic Literature. Other search strategies were employed for completeness. No limitations were placed on language or publication status.

Selection criteria

Randomized controlled trials (RCTs) which examined the effectiveness of spinal manipulation or mobilization in adults with acute low‐back pain were included. In addition, studies were included if the pain was predominantly in the lower back but the study allowed mixed populations, including participants with radiation of pain into the buttocks and legs. Studies which exclusively evaluated sciatica were excluded. No other restrictions were placed on the setting nor the type of pain. The primary outcomes were back pain, back‐pain specific functional status, and perceived recovery. Secondary outcomes were return‐to‐work and quality of life. SMT was defined as any hands‐on therapy directed towards the spine, which includes both manipulation and mobilization, and includes studies from chiropractors, manual therapists, and osteopaths.

Data collection and analysis

Two review authors independently conducted the study selection and risk of bias (RoB) assessment. Data extraction was checked by the second review author. The effects were examined in the following comparisons: SMT versus 1) inert interventions, 2) sham SMT, 3) other interventions, and 4) SMT as an additional therapy. In addition, we examined the effects of different SMT techniques compared to one another. GRADE was used to assess the quality of the evidence. Authors were contacted, where possible, for missing or unclear data. Outcomes were evaluated at the following time intervals: short‐term (one week and one month), intermediate (three to six months), and long‐term (12 months or longer). Clinical relevance was defined as: 1) small, mean difference (MD) < 10% of the scale or standardized mean difference (SMD) < 0.4; 2) medium, MD = 10% to 20% of the scale or SMD = 0.41 to 0.7; and 3) large, MD > 20% of the scale or SMD > 0.7.

Main results

We identified 20 RCTs (total number of participants = 2674), 12 (60%) of which were not included in the previous review. Sample sizes ranged from 36 to 323 (median (IQR) = 108 (61 to 189)). In total, six trials (30% of all included studies) had a low RoB. At most, three RCTs could be identified per comparison, outcome, and time interval; therefore, the amount of data should not be considered robust. In general, for the primary outcomes, there is low to very low quality evidence suggesting no difference in effect for SMT when compared to inert interventions, sham SMT, or when added to another intervention. There was varying quality of evidence (from very low to moderate) suggesting no difference in effect for SMT when compared with other interventions, with the exception of low quality evidence from one trial demonstrating a significant and moderately clinically relevant short‐term effect of SMT on pain relief when compared to inert interventions, as well as low quality evidence demonstrating a significant short‐term and moderately clinically relevant effect of SMT on functional status when added to another intervention. In general, side‐lying and supine thrust SMT techniques demonstrate a short‐term significant difference when compared to non‐thrust SMT techniques for the outcomes of pain, functional status, and recovery.

Authors' conclusions

SMT is no more effective in participants with acute low‐back pain than inert interventions, sham SMT, or when added to another intervention. SMT also appears to be no better than other recommended therapies. Our evaluation is limited by the small number of studies per comparison, outcome, and time interval. Therefore, future research is likely to have an important impact on these estimates. The decision to refer patients for SMT should be based upon costs, preferences of the patients and providers, and relative safety of SMT compared to other treatment options. Future RCTs should examine specific subgroups and include an economic evaluation.

Keywords: Adult; Humans; Low Back Pain; Low Back Pain/therapy; Manipulation, Spinal; Manipulation, Spinal/methods; Randomized Controlled Trials as Topic

Spinal manipulative therapy for acute low‐back pain

Low‐back pain is a common and disabling disorder, representing a great burden both to the individual and society. It often results in reduced quality of life, time lost from work, and substantial medical expense. Spinal manipulative therapy (SMT) is widely practised by a variety of healthcare professionals worldwide and is a common choice for the treatment of low‐back pain. The effectiveness of this form of therapy for the management of acute low‐back pain is, however, not without dispute.

For this review, acute low‐back pain was defined as pain lasting less than six weeks. Only cases of low‐back pain not caused by a known underlying condition, for example, infection, tumour, or fracture, were included. Also included were patients whose pain was predominantly in the lower back but may also have radiated (spread) into the buttocks and legs.

SMT is known as a 'hands‐on' treatment directed towards the spine, which includes both manipulation and mobilization. The therapist applies manual mobilization by passively moving the spinal joints within the patient’s range of motion using slow, passive movements, beginning with a small range and gradually increasing to a larger range of motion. Manipulation is a passive technique whereby the therapist applies a specifically directed manual impulse, or thrust, to a joint at or near the end of the passive (or physiological) range of motion. This is often accompanied by an audible ‘crack’.

In this review, a total of 20 randomized controlled trials (RCTs) (representing 2674 participants) assessing the effects of SMT in patients with acute low‐back pain were identified. Treatment was delivered by a variety of practitioners, including chiropractors, manual therapists, and osteopaths. Approximately one‐third of the trials were considered to be of high methodological quality, meaning these studies provided a high level of confidence in the outcome of SMT.

Overall, we found generally low to very low quality evidence suggesting that SMT is no more effective in the treatment of patients with acute low‐back pain than inert interventions, sham (or fake) SMT, or when added to another treatment such as standard medical care. SMT also appears to be no more effective than other recommended therapies. SMT appears to be safe when compared to other treatment options but other considerations include costs of care.

Summary of findings

Summary of findings for the main comparison.

Spinal manipulative therapy compared to other interventions for acute low‐back pain

Spinal manipulative therapy compared to other interventions for acute low‐back pain
Patient or population: Patients with acute low‐back pain Settings: Primary or tertiary care Intervention: Spinal manipulative therapy Comparison: Other interventions (e.g. physiotherapy, exercise, back school)
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Other interventions	Spinal manipulative therapy
Pain at one week 0 (no pain) to 10 (worse pain)	The mean pain at one week ranged across control groups from 2.6 to 3.5 points	The mean pain at one week in the intervention groups was 0.1 higher (0.5 lower to 0.7 higher)		383 (3 studies)	⊕⊕⊝⊝ low^1,2	Small, not clinically‐relevant effect.
Pain at one month 0 (no pain) to 10 (worse pain)	The mean pain at one month ranged across control groups from 0.5 to 2.3 points	The mean pain at one month in the intervention groups was 0.2 lower (0.5 lower to 0.2 higher)		606 (3 studies)	⊕⊕⊕⊝ moderate¹	Small, not clinically‐relevant effect.
Functional status at one week Roland Morris Disability Questionnaire. Scale from: 0 (no dysfunction) to 24 (worse function)	The mean functional status at one week in the control groups was 7.2 points	The mean functional status at one week in the intervention groups was 0.1 standard deviations higher (0.2 lower to 0.3 higher)		241 (1 study)	⊕⊕⊝⊝ low^2,3	Small, not clinically‐relevant effect.
Functional status at one month Roland Morris Disability Questionnaire. Scale from: 0 (no dysfunction) to 24 (worse function)	The mean functional status at one month in the control groups was 4.1 points	The mean functional status at one month in the intervention groups was 0.5 points lower (1.2 lower to 0.2 higher)		681 (3 studies)	⊕⊕⊕⊝ moderate¹	Small, not clinically‐relevant effect. Based on pooled SMD: ‐0.11 (‐0.26 to 0.05).⁴
Recovery at one month	Study population		RR 1.06 (0.94 to 1.21)	117 (2 studies)	⊕⊕⊝⊝ low^1,5	Small, not clinically‐relevant effect.
	87 per 100	92 per 100 (81 to 100)
Serious adverse events	Study population		Not estimable	2 studies		Total 578 participants. No serious adverse events were observed in the SMT group.
CI: Confidence interval; RR: Risk ratio;⊕⊕⊝⊝ = these symbols indicate how many of the items were fulfilled (for each ⊕, one item was fulfilled and corresponds to the different levels of evidence).
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Spinal manipulative therapy plus another intervention compared to the intervention alone for acute low‐back pain
Patient or population: Patients with acute low‐back pain Settings: Primary or tertiary care Intervention: Spinal manipulative therapy plus another intervention Comparison: The intervention alone (e.g. usual care, exercise)
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	The intervention alone	Spinal manipulative therapy plus another intervention
Pain at one week Scale from: 0 (no pain) to 10 (worse pain)	The mean pain at one week in the control groups was 1.9 points	The mean pain at one week in the intervention groups was 0.8 points higher (0.04 lower to 1.7 higher)		102 (1 study)	⊕⊕⊝⊝ low¹	Small, not clinically‐relevant effect.
Pain at 3 to 6 months Scale from: 0 (no pain) to 10 (worse pain)	The mean pain at 3 to 6 months in the control groups was 1.5 points	The mean pain at 3 to 6 months in the intervention groups was 0.7 points higher (0.3 lower to 1.6 higher)		104 (1 study)	⊕⊕⊝⊝ low¹	Small, not clinically‐relevant effect.
Functional status at one week Oswestry Disability Index. Scale from: 0 (no dysfunction) to 100 (worse function).	The mean functional status at one week in the control groups was 33 points	The mean functional status at one week in the intervention groups was 5.7 points lower (10.1 to 1.4 lower)		225 (2 studies)	⊕⊕⊝⊝ low^2,3	Moderately clinically‐relevant effect. Based on pooled SMD: ‐0.41 (‐0.73 to ‐0.10).⁴
Functional status at 3 to 6 months Oswestry Disability Index. Scale from: 0 (no dysfunction) to 100 (worse function)	The mean functional status at 3 to 6 months in the control groups was 24.4 points	The mean functional status at 3 to 6 months in the intervention groups was 3.8 points lower (10.6 lower to 2.8 higher)		225 (2 studies)	⊕⊕⊝⊝ low^2,3	Small, not clinically‐relevant effect. Based on pooled SMD: ‐0.22 (‐0.61 to 0.16).⁴
Recovery at one week	Study population		RR 0.89 (0.32 to 2.47)	196 (2 studies)	⊕⊝⊝⊝ very low^2,5,6	Small, not clinically‐relevant effect. Based on pooled RR: 0.88 (0.36 to 2.19).
	16 per 100	14 per 100 (5 to 40)
Recovery at 3 to 6 months	Study population		RR 0.75 (0.51 to 1.1)	195 (2 studies)	⊕⊝⊝⊝ very low^2,6	Small, not clinically‐relevant effect. Based on pooled RR: 0.96 (0.71 to 1.31). I²=57%.
	64 per 100	48 per 100 (33 to 70)
Serious adverse events	Study population		Not estimable	2 studies		Total 199 participants. In one of the studies, two serious adverse events were observed in the SMT group; however, they "appeared not to be related to the treatment". An equal number of adverse events were seen in the control group (Juni 2009).
CI: Confidence interval; RR: Risk ratio;⊕⊕⊝⊝ = these symbols indicate how many of the items were fulfilled (for each ⊕, one item was fulfilled and corresponds to the different levels of evidence).
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Spinal manipulative therapy (SMT) compared to sham SMT for acute low‐back pain
Patient or population: Patients with acute low‐back pain Settings: Primary or tertiary care Intervention: Spinal manipulative therapy (SMT) Comparison: Sham SMT
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Sham SMT	Spinal manipulative therapy (SMT)
Pain at one month 0 (no pain) to 10 (worse pain)	The mean pain at one month in the control groups was 2.2 points	The mean pain at one month in the intervention groups was 0.5 lower (1.4 lower to 0.4 higher)		74 (1 study)	⊕⊝⊝⊝ very low^1,2	Small, not clinically‐relevant effect.
Functional status at one month Oswestry Disability Index. Scale from: 0 (no dysfunction) to 100 (worse function)	The mean functional status at one month in the control groups was 16.3 points	The mean functional status at one month in the intervention groups was 0.4 standard deviations lower (0.8 lower to 0.1 higher)		94 (1 study)	⊕⊝⊝⊝ very low^1,2	Small, not clinically‐relevant effect.
Recovery at one month	Study population		Not estimable	0 studies		No data were available.
Serious adverse events	Study population		Not estimable	0 studies		No data were available.
CI: Confidence interval; RR: Risk ratio;⊕⊕⊝⊝ = these symbols indicate how many of the items were fulfilled (for each ⊕, one item was fulfilled and corresponds to the different levels of evidence).
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Author	Presence/absence of radiating pain	Duration LBP (according to inclusion criteria)	Duration LBP (current episode)	Type of manipulator (n= # of manipulators); experience (if stated)	Type of manipulation	No. txs SMT allowed and duration
Bergquist‐Ullman 1977	No radiation below knee	< 8 wks	>50% less than 4 wks	Physiotherapist (n=?)	Manipulation/MOB according to Cyriax	4 tx's (mean), 10 (max)
Brennan 2006	Absence of nerve root compression	<3 mo.	Median (IQR): 16d (10, 41)	Physiotherapist (n=?)	Thrust manipulation or low‐amplitude mobilization	? median range: 6.5 to 7 sessions
Cherkin 1998	No sciatica	Duration was not listed in inclusion criteria	78% < 6 wks	Chiropractors (n=?); collectively 6 to14 yrs. experience	Short‐lever HVLA SMT	6.9 tx's (mean)
Childs 2004	Absence of nerve root compression	?	Median = 27 d	Physiotherapist (n=14)	HVLA SMT	?
Cleland 2009	Absence of nerve root compression	Duration was not per se an inclusion criteria	Median (IQR) = 45 (27 to 60)	Physiotherapists (n=17); collectively avg. 9 yrs. experience	HVLA SMT or low‐amplitude mobilization	Total 2 sessions
Cramer 1993	Without compressive neuropathy	< 2 wks	?	Chiropractors (n=?)	Side‐lying (short‐lever?) HVLA? SMT	3 to 5 times over a 10‐d period
Farrell 1982	Without neurological signs	< 3 wks	?	Physiotherapists (n=?)	Manipulation/MOB according to Maitland	3x/wk for 3 weeks. Tx was continued, prn
Glover 1974	Without neurological signs	?	52% <7 d	Osteopathic physician? (n=1)	Rotational manipulation	1 tx (followed by 4d of detuned diathermy)
Hadler 1987	With or without signs of radiculopathy	< 4 wks	?	Osteopathic physician? (n=1)	Long‐lever high‐velocity SMT	One visit?
Hallegraeff 2009	No symptoms distal to the knee	<16 d.	69% <3 wks	Manual therapists (n=?)	HVLA SMT	4 visits over 2 1/2 wks
Hancock 2007	Absence of nerve root compromise	< 6 wks	Mean = 9.13 d	Physiotherapists (n=15)	Most (97%) received low‐velocity mobilization; a small proportion (5%) also received high‐velocity thrusts	2 to 3x/wk for a max. of 12 txs. over 4 wks
Hoehler 1981	?	Duration was not listed in the inclusion criteria	52% of SMT grp & 48% of ctrl. grp < 1 mo	?	High‐velocity thrust	?
Hoiriis 2004	No neuropathy	2 to 6 wks	Total for all subjects = 3.7 wks	Chiropractors (n=?)	HVLA SMT	Most attended 7 chiropractic visits
Juni 2008	Absence of nerve root compression, no radiation below the knee	< 4 wks	54% of SMT grp & 75% of ctrl. grp <7d with LBP	Medical manipulator (n=2), osteopathy (n=1)	HVLA SMT	Median (IQR): 3 (2, 4)
MacDonald 1990	Absence of nerve root compromise	?	55% of both grps<14d LBP	Osteopathy (n=?)	HVLA SMT	4.7 tx's (mean), 87% were delivered within the first 2.5 wks.
Postacchini 1988	With or without radiation to the knee	grp.A="acute"	Mean duration: 15d & 17d	Chiropractor (n=?)	Manipulation	12 over 6 wks
Rasmussen 1979	Without signs of nerve root pressure	< 3 wks	?	Physiotherapist (n=1?) or medical manipulator (n=1?)	Rotational manipulation in the pain‐free direction	3x/wk for 2 wks
Seferlis 1998	With or without sciatica	< 2 wks	?	Physiotherapist (n=?)	"Manipulation of the lumbar facet and SI joint"	10 txs (mean)
Skagren 1997	Absence of nerve root signs	Duration was not listed in the inclusion criteria	55% of SMT grp. & 48% of control grp <4 wks with LBP	Chiropractors (n=6), collectively 9.9 yrs experience (range 1‐15)	HVLA SMT	4.9 txs (mean) in 4.1 wks
Sutlive 2009	Absence of nerve root compression	Duration was not a required item	62% w/ LBP <16d	Physiotherapist (n=1?)	HVLA SMT	1 tx only

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pain	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.1 Pain at 1 week	3	311	Mean Difference (IV, Random, 95% CI)	0.14 [‐0.69, 0.96]
1.2 Pain at 1 month	1	178	Mean Difference (IV, Random, 95% CI)	‐1.20 [‐2.01, ‐0.39]
1.3 pain at 3 months	1	181	Mean Difference (IV, Random, 95% CI)	‐1.20 [‐2.11, ‐0.29]
2 Functional status	2		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
2.1 Functional status at 1 week	2	205	Std. Mean Difference (IV, Random, 95% CI)	‐0.08 [‐0.37, 0.21]
2.2 Functional status at 1 month	1	178	Std. Mean Difference (IV, Random, 95% CI)	‐0.27 [‐0.58, 0.04]
2.3 Functional status at 3 months	1	181	Std. Mean Difference (IV, Random, 95% CI)	‐0.28 [‐0.59, 0.02]
3 Recovery	2		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only
3.1 Recovery at 1 week	2	263	Risk Ratio (M‐H, Random, 95% CI)	0.96 [0.50, 1.85]
3.2 Recovery at 1 month	1	239	Risk Ratio (M‐H, Random, 95% CI)	0.97 [0.85, 1.10]
3.3 Recovery at 3 months	1	239	Risk Ratio (M‐H, Random, 95% CI)	1.0 [0.98, 1.02]

Outcome or subgroup title	No. of studies	Statistical method	Effect size
1 Pain	2	Mean Difference (IV, Random, 95% CI)	Totals not selected
1.1 Pain at 1 week	2	Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.2 Pain at 1 month	1	Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
1.3 Pain at 3 to 6 months	1	Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
2 Functional status	3	Std. Mean Difference (IV, Random, 95% CI)	Totals not selected
2.1 Functional status at 1 week	3	Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
2.2 Functional status at 1 month	2	Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
2.3 Functional status at 3 to 6 months	1	Std. Mean Difference (IV, Random, 95% CI)	0.0 [0.0, 0.0]
3 Recovery	1	Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
3.1 Recovery at 1 week	1	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
3.2 Recovery at 1 month	1	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]
3.3 Recovery at 3 to 6 months	1	Risk Ratio (M‐H, Random, 95% CI)	0.0 [0.0, 0.0]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pain ‐ For funnel plot	8		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.1 Pain at 1 week	6	704	Mean Difference (IV, Random, 95% CI)	‐0.13 [‐0.82, 0.56]
1.2 Pain at 1 month	5	809	Mean Difference (IV, Random, 95% CI)	‐0.56 [‐1.07, ‐0.06]
1.3 pain at 3 to 6 months	3	676	Mean Difference (IV, Random, 95% CI)	‐0.42 [‐1.00, 0.17]
1.4 Pain at 12 months	1	314	Mean Difference (IV, Random, 95% CI)	0.40 [‐0.08, 0.88]
2 Functional status ‐ For funnel plot	10		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
2.1 Functional status at 1 week	6	683	Std. Mean Difference (IV, Random, 95% CI)	‐0.31 [‐0.59, ‐0.03]
2.2 Functional status at 1 month	9	1280	Std. Mean Difference (IV, Random, 95% CI)	‐0.23 [‐0.42, ‐0.03]
2.3 Functional status at 3 to 6 months	5	901	Std. Mean Difference (IV, Random, 95% CI)	‐0.26 [‐0.49, ‐0.02]
2.4 Functional status at 12 months	2	437	Std. Mean Difference (IV, Random, 95% CI)	0.06 [‐0.14, 0.25]

Methods	Method of sequence generation considered adequate. Allocation concealment unclear. Statistical analysis: Contingency tables and Chi² tests were used when comparing absence from work, number and length of recurrences in one year in the three groups of therapy. Differences in pain index were analysed using analysis of variance. A covariance analysis was used when comparing the duration of symptoms following the first treatment in the three groups (using logarithms of the values). No sample size determination was performed.
Participants	217 subjects; study setting: occupational (Automotive Division of AB Volvo, mainly manual workers (light industrial work with a predominance of assembly‐line work) and clerks and executives; country: Sweden Period and mode of recruitment: via the healthcare centre of the company Age: median 34.5 years (range 17‐64) Gender: 13% women Inclusion criteria: Acute or subacute back pain localized to the lumbosacral region with or without radiation to the thigh; duration of pain not longer than three months; a pain‐free year before the onset of the current episode. Exclusion criteria: chronic pain; rhizopathy; pregnancy; spondylolisthesis; infections; tumours; ankylosing spondylitis; senile osteoporosis, structural scoliosis.
Interventions	I) Manipulation and mobilization according to Cyriax, Kaltenborn, Lewit and Janda (n=72); postural advice and strengthening exercises were also allowed; avg. number of treatments = 4 (max. 10). C1) Back school (including instruction and exercise) (n=70); avg. number of treatments = 4 sessions of 45 minutes given during a 2 week period. C2) Short‐wave diathermy (considered a placebo treatment) (n=75); avg. number of treatments = 5 (max. 10).
Outcomes	1. Pain index (range: 0‐70) 2. Back specific functional status: 10 items, 4‐point scale 3. Recovery: not reported 4. Spinal mobility: via Schober's test 5. RTW: patient and insurance data based upon work absenteeism 6. Adverse events: not reported. Note: outcomes were not defined as primary or secondary Period of follow‐up: 10 days; 3, 6 weeks; 12 months
Notes	Authors' results and conclusions: 70% of the studied group recovered from the initial episode within two months and 86% within three months, regardless of the treatment given. The 95% confidence interval for the difference between the combined physiotherapy group (antilogarithms of the adjusted mean value 15.8 days) and the placebo group (28.7 days) was 0.59 ± 0.37. No difference was detected between the Back School group (14.8 days) and the combined physiotherapy group. There were significantly more patients with a shorter duration of sick‐leave in the Back School group compared to the placebo group (P<0.01). A similar decrease in the pain index was observed in the three groups. No relation was found between the type of treatment and the number of recurrences of pain or total duration of absence from work. "There is enough evidence in this study to conclude that Back School and combined physiotherapy are superior to "placebo" treatment in acute low back pain." Funded by: the Swedish Work Environment Fund (government) and AB Volvo (Industry)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Stratified randomisation (based on vocational and psychologic factors). "Separate tables of random numbers were used" (p. 39).
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	No mention of an attempt to blind the patients to therapy.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	Care provider was not blinded.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Patient was not blinded; therefore, outcomes assessor also not blinded.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	High risk	Percentage drop‐out. At 10 days: Back school ‐ 37% (n=26/70); SMT ‐ 31% (n=22/72); Diathermy ‐ 24% (n=18/74) At 3 weeks: Back school ‐ 64% (n=45/70); SMT ‐ 74% (n=53/72); Diathermy ‐ 57% (n=42/74) At 6 weeks: Back school ‐ 80% (n=14/70); SMT ‐ 78% (n=56/72); Diathermy ‐ 80% (n=59/74)
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	High risk	"22 patients were excluded from the analysis as they refused treatment. Four patients were initially randomly allocated to the physiotherapy group and 18 patients were allocated to the "placebo" group. Since a predominance of patients from the placebo group refused treatment, a separate comparison........ was carried out."
Selective reporting (reporting bias)	Unclear risk	No published protocol available
Similarity of baseline characteristics?	Unclear risk	Not described
Co‐interventions avoided or similar?	Unclear risk	Not described
Compliance acceptable?	High risk	Seemingly differences between groups: "All patients attended all four sessions of the back school"; "...four patients (6%) allocated to manual therapy did not attend a single session"; "....16 patients (21%) of those allocated to diathermy did not follow treatment."
Timing outcome assessments similar?	Low risk	1 year follow‐up
OVERALL RISK OF BIAS	High risk

Methods	Adequate sequence generation and allocation procedures. Statistical analysis: ANOVA for treatment group, and classification subgroup. ODI was the principal dependent variable. Last variable carried forward was used to impute missing data. Sample size calculation based upon determining a MCID (of 6 points) for the ODI.
Participants	123 subjects; study setting: physical therapy clinics; Country: USA Period and mode of recruitment: Primary recruitment occurred at one clinic between January 1, 2000 and July 1, 2003. Additional recruitment occurred at two other clinics between January 1, 2002 and September 1, 2002. Age (all) (mean (SD)): 37.7 (10.7) yrs Gender (all) (% female): 45% Inclusion criteria: Patients between 18 and 65 years with a primary complaint of LBP of less than 90 days, with or without referral into the lower extremity, and an ODI>25% were eligible. Exclusion criteria were a visible lateral shift or acute kyphotic deformity, signs of nerve root compression (positive straight leg raise test and reflex or strength deficits), any red flags indicating a serious pathology such as spinal neoplasm, infection, or fracture, an inability to reproduce any symptoms with lumbar spine active range of motion (AROM) or palpation, current pregnancy, or prior surgery to the lumbar and/or sacral region.
Interventions	I) Manipulation (n=40): manual therapy techniques that could include thrust manipulation, or low amplitude mobilization procedures directed to the lumbosacral region, along with instruction in a lumbar AROM exercise. The therapist performing the treatment was permitted to reexamine the patient and could choose one of two manual therapy techniques. The choice of which technique to use was left to the therapists’ discretion, but one of the two techniques had to be used. In the first technique, the patient was supine, with the lumbar spine placed into side‐bending and rotation to the opposite direction. The therapist delivered a force through the patient’s pelvis in a posterior and inferior direction. For the second technique, the patient was side‐lying. The lumbar spine was positioned in either flexion or extension followed by rotation in an attempt to isolate forces to a particular spinal level. The therapist delivered the force through the patient’s pelvis and trunk. The choice of technique was left to the discretion of the therapist. The AROM exercise was performed by instructing the patient to alternately flex and extend the lumbar spine while in a quadruped position. C1) Specific exercise (n=37): received instruction in repeated ROM exercises into either lumbar flexion or extension. All patients in this group had to be treated with directional exercises; however, the direction of the exercise was determined by the treating therapist based on a reassessment of the patient’s response to movement testing and symptom response to positions of sitting, standing, or walking. Flexion exercises were used for patients who centralized with or had a preference for flexion movements or positions (i.e., sitting), whereas extension exercises were used for patients who centralized or had a preference for extension (i.e., standing or walking). Either flexion or extension exercises were used, but not both. Flexion exercises were performed with the patient sitting, supine, or quadruped. Extension exercises were performed in prone, using prone on elbows or prone press‐up activities. C2) Stabilization (n=46): treated with a program of trunk strengthening and stabilization exercises. Patients were instructed to perform abdominal bracing exercises in supine and quadruped positions, progressing to more functional positions and activities. Patients were also instructed in alternating arm and leg extension exercises in quadruped to strengthen the lumbar extensor muscles. Strengthening for the oblique abdominals included curl‐up and side support exercises.
Outcomes	1. Pain: 11‐pt. NRS, current pain 2. Back specific functional status: modified ODI 3. Recovery: not reported 4. FABQ: including two subscales (work and physical activity) 5. Adverse events: not reported. Note: outcomes were not designated as primary or secondary. Follow‐up: at completion of treatment (˜4 wks. after baseline assessment), 1 year.
Notes	Authors results and conclusions: Patients receiving matched treatments experienced greater short‐ and long‐term reductions in disability than those receiving unmatched treatments. After 4 weeks, the difference favouring the matched treatment group was 6.6 ODI points (95% CI, 0.70–12.5), and at long‐term follow‐up the difference was 8.3 points (95% CI, 2.5–14.1). Compliers‐only analysis of long‐term outcomes yielded a similar result. Conclusions. Nonspecific low back pain should not be viewed as a homogenous condition. Outcomes can be improved when subgrouping is used to guide treatment decision‐making. Funded by Deseret Foundation (non‐profit).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"A random number generator was used to generate a randomization list before initiation of the study. The list was maintained by the secretarial staff of the participating clinics."
Allocation concealment (selection bias)	Low risk	"Before the first treatment session, the secretarial staff consulted the randomization list and assigned the patient to one of three groups."
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	No attempts were made to blind the patient.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	No attempts were conducted to blind the provider.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Patient was not blinded; therefore, outcome assessor is also not blinded. Physiological measures were examined, including lumbar active ROM, judgement of centralization or peripheralization, aberrant movements occurring during lumbar active ROM (indicating poss. instability), and mobility of each level of the lumbar spine was assessed.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	High risk	"66% (n=81) completed the long‐term follow‐up, with no differences in the median number of days between baseline and follow‐up or the proportions of patients with completed follow‐up between patients receiving the matched or unmatched treatments. There were no significant differences between those with complete or incomplete long‐term follow‐up with respect to age, sex, duration of symptoms, baseline pain, OSW, or FABQ scores (P>0.05). The proportion of matched versus unmatched patients did not differ between those with complete or incomplete long‐term follow‐up (P>0.05)." However, patients with complete long‐term follow‐up did have lower 4‐week OSW scores (17.3 vs. 25.0, P=0.02).
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	High risk	Both ITT and compliers‐only analyses were conducted and are presented in Table 2 and the ITT analysis (for matched and unmatched treatment) is presented in Fig.3. ITT analyses used the last available OSW score carried forward for missing data. A large proportion of patients were lost to the long‐term follow‐up, so a compliers‐only analysis was conducted, including only those patients completing the 1‐year OSW score. This item was scored negatively because data from T.3, which represents the compliers‐only data, were used for data extraction purposes in this review. These data represent randomization without matching. This point is discussed further in the review.
Selective reporting (reporting bias)	Unclear risk	No published protocol available. Outcomes were presented separately for functional status by type of randomized treatment group and by classification subgroup, but not for pain (scores were presented only for those receiving the matched and unmatched treatments).
Similarity of baseline characteristics?	Unclear risk	Data presented only by those receiving matched and unmatched treatments and not by allocation group. Baseline variables presented: age, gender, education level, prior history of LBP, symptoms distal to the knee, duration of current symptoms, missed work or school days related to the current LBP episode, FABQ ‐ work and physical activity sub‐scales, ODI, pain.
Co‐interventions avoided or similar?	Unclear risk	Note: This was not stated or not measured.
Compliance acceptable?	Low risk	"All patients were scheduled for treatment twice weekly for 4 weeks for a maximum of eight sessions." "Median number of sessions attended by the matched group was 6.5, and 80% attended at least 4 sessions. In the unmatched group median number of sessions was 7, and 77% attended at least 4 sessions".
Timing outcome assessments similar?	Low risk	At completion of treatment and 1‐year follow‐up.
OVERALL RISK OF BIAS	High risk

Methods	Sequence generation and allocation considered adequate. Statistical analysis: For dichotomous outcomes, logistic regression was used with adjustment for baseline values. Sample size calculation was based upon determining a MCID (2.5 points) for the RMDQ and (1.5 points) for the bothersomeness scale.
Participants	321 subjects; Study setting: Physical therapy clinics situated within a health maintenance organization (HMO) and SMT was performed in the private practices of chiropractors; Country: USA. Period and mode of recruitment: Patients were recruited from primary care clinics; November 1993 to September 1995. Age (mean (SD) (all): 40.7 (10.7) years (range across groups: 39.7 (9.4) to 41.8 (11.5) Gender (% female) (all): 48% (range across groups: 42% to 53%) Inclusion criteria: 20 to 64 years of age who saw their primary care physician for low back pain and who still had pain seven days later. Exclusion criteria: mild or no pain seven days following the visit to the physician, a history of back surgery, sciatica, systemtic or visceral causes of the pain, osteoporosis, a vertebral fracture or dislocation, severe neurologic signs, spondylolisthesis, coagulation disorders, or a severe concurrent illness. Subjects who had received corticosteroid therapy, were pregnant, were involved in claims for compensation or litigation because of the back injury, had received physical therapy or chiropractic or osteopathic manipulative treatment for their current back pain, or visited practitioners other than their primary care physicians were also excluded.
Interventions	I) SMT (n=122): The most common method of chiropractic manipulation was used: a short‐lever, high‐velocity thrust directed specifically at a “manipulable lesion.” This procedure is typically performed with the patient lying on his or her side on a segmental table. No other physical treatments were permitted. Chiropractors evaluated patients according to their usual procedures and were allowed to make the same recommendations about exercise and activity restrictions that they usually did. An exercise sheet was used that emphasized stretching and strengthening but excluded extension exercises, an important part of McKenzie therapy. C1) Physical therapy (n=133) following McKenzie principles: Patients were taught to perform exercises that centralize their symptoms and to avoid movements that peripheralize them. This method relies on patient‐generated forces and emphasizes self‐care. McKenzie Institute faculty trained the therapists before the study, and all but one therapist passed an advanced McKenzie credentialing examination. Subjects received McKenzie’s Treat Your Own Back book and a lumbar‐support cushion. Therapists were asked to avoid adjuncts such as heat, ice, transcutaneous electrical nerve stimulation, ultrasonography, and back classes. C2) Educational booklet (n=66): A minimal‐intervention control group received an educational booklet to minimize potential disappointment with not receiving a physical treatment. The booklet discussed causes of back pain, prognosis, appropriate use of imaging studies and specialists, and activities for promoting recovery and preventing recurrences. A previous trial found that the use of this booklet as a supplement to standard care was not associated with improved outcomes. This group was deemed to be similar in some respects to a no‐treatment control group.
Outcomes	1. Pain: "bothersomeness" of back pain, leg pain, and numbness or tingeling in the preceding 24 hours; 0‐10 pt. scale (unclear if this was a VAS or NRS) 2. Back specific functional status: RMDQ 3. Recovery: 5‐point scale, ranging from poor to excellent 4. RTW: number of days spent home from work or school 5. Other: number of days spent in bed or with reduced activity (specifically with reference to the back) 6. Adverse events: "No important adverse events of treatment were reported in any of the group" Reviewers note: outcomes were not defined as primary or secondary. Comment reviewers: The score for the most bothersome symptom was used. It is unclear if this was a different measure within the groups at each of the follow‐up measures nor whether it was different between the groups. Follow‐up: 1, 4, 12 weeks; 1 to 2 years.
Notes	Authors results and conclusions: The McKenzie method of physical therapy and chiropractic manipulation had similar effects (based upon pain and functional status) and costs and patients receiving these treatments had only marginally better outcomes than those receiving the minimal intervention of an educational booklet. Note reviewers: "At both one and four weeks, about 75 percent of the subjects in the physical‐therapy and chiropractic groups rated their care as “very good” to “excellent,” as compared with about 30 percent of the subjects in the booklet group (P<0.001). However, about one quarter of the subjects in the booklet group failed to answer this question, possibly because only 18 percent received care during this period." Funded by: Agency for Health Care Policy and Research (Governmental)
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Subjects were randomly assigned without stratification .... with the use of sealed, opaque envelopes."
Allocation concealment (selection bias)	Low risk	See above
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	There is no mention of attempts to blind the patients to other interventions or their perceptions of potential effectiveness of the different interventions.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	No mention if there were any attempts to blind the care providers to the other groups.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Patient was not blinded; therefore, this item was scored as "no".
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	Low risk	"Between 89 and 96 percent of the subjects responded to each of the follow‐up questionnaires." Reviewers note: There was minimal drop‐out and differences between groups, however, no reasons were offered for those who dropped out or were lost to follow‐up.
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	Low risk	"Two subjects were excluded after randomization (one because of a urinary tract infection and one because of pancreatic cancer)." Reviewers note: it was not stated to which group they belonged.
Selective reporting (reporting bias)	Unclear risk	No protocol available
Similarity of baseline characteristics?	Low risk	Variables examined: age, gender, employment, smoker (y/n), general health and mental health perceptions, history of LBP (prior episodes, prior chiropractic and PT care), duration and intensity of current LBP (including pain and functional status), medication usage and expectations of recovery.
Co‐interventions avoided or similar?	High risk	Eighteen per cent of the subjects in the booklet group visited a healthcare provider for back pain during the study month, and only 8% of the subjects in the chiropractic group and 9% of those in the physical‐therapy group visited providers other than those assigned. The reported use of exercise was almost identical in the three groups at baseline (about 57 percent) and one month (about 81%). During the month, the percentage of subjects who used back‐pain medication of any type decreased from 82% to 18% in the chiropractic group, from 84% to 27% in the physical‐therapy group, and from 77% to 32% in the booklet group (P<0.05 for the differences among the groups after adjustment for baseline use). Fewer than 2% of the subjects reported using corsets, braces, traction, transcutaneous electrical nerve stimulation, or injections.
Compliance acceptable?	Unclear risk	"96% of the chiropractic group and 97% of the physical therapy group visited their assigned provider at least once. The mean number of chiropractic visits exceeded the mean number of physical therapy visits by 50% (6.9 vs. 4.6). According to the subjects' reports, the total amount of time spent with the provider was virtually identical in the two groups (about 145 minutes)."
Timing outcome assessments similar?	Low risk	Follow‐up: 1, 4, 12 weeks; 1 to 2 years
OVERALL RISK OF BIAS	Low risk

Methods	Method of sequence generation and allocation unclear. Statistical analyses: Presented as mean differences (pre‐post testing); t‐Tests used to test for significance. No sample size calculation was performed.
Participants	36 subjects; Study setting: outpatient clinic; Country: USA. Period and mode of recruitment: those presenting to the clinic; otherwise not stated. Age (all subjects): 18 to 56 years of age Gender (% female; all subjects): 42% Inclusion criteria: Mechanical low‐back pain less than 2 weeks duration; ODI > 8; VAS > 33 mm; no litigation or workers' compensation; not pregnant. Exclusion criteria: Subjects with clinical evidence of a compressive neuropathy.
Interventions	I) SMT (n=17): included a side‐lying manipulation to the affected area of the lumbar spine; additional treatment consisted of electrical stimulation and cold‐packs to the lumbar spine (L3‐S1). C) control (n=19): included detuned ultrasound to the low back followed by cold packs (5‐10 min.) and very gentle soft tissue massage (15‐30 sec.). Treatments for both groups were delivered 3 to 5 times over a 10‐day period.
Outcomes	1. Pain: 0‐100 (VAS) 2. Back specific functional status: ODI 3. Recovery: not reported 4. Adverse events: not reported 5. H_max/M_max ratio were measured. No mention of which were primary or secondary; however, the basis of the report was the H_max/M_max ratio. Follow‐up: Pre‐ and post‐testing; Reviewers note: personal communication with the primary author: post‐treatment measurement represents the measurement following the 10‐day period.
Notes	Authors results and conclusions: The H/M ratio (reviewers: "an objective and clinically useful physiological measure for acute low back pain...") showed greater change in the group which received spinal manipulation, but the change was subtle. The results indicate that the H/M ratio may be of limited value in clinical practice. Funded by: Foundation for Chiropractic Education and Research (non‐profit).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Reviewers comment: No mention of the sequence generation procedure nor method of allocation. The abstract indicates that it is a randomized trial and a flow chart is provided suggesting that randomization occurred.
Allocation concealment (selection bias)	Unclear risk	See above
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	No mention of any attempts to blind the patient.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	No mention of any attempts to blind the provider.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Outcomes assessor was not blind to allocation.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	Low risk	Unclear, but the tables suggest that all subjects were retained in the study.
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	Low risk	Unclear, but presumably all subjects were analysed by allocation treatment and there were no drop‐outs.
Selective reporting (reporting bias)	Unclear risk	No protocol available. Recovery not reported, although selective reporting not likely.
Similarity of baseline characteristics?	Unclear risk	Not stated
Co‐interventions avoided or similar?	Unclear risk	Not stated
Compliance acceptable?	Unclear risk	Not stated
Timing outcome assessments similar?	Low risk	Collected post‐treatment over a 10‐day period.
OVERALL RISK OF BIAS	High risk

Methods	Unclear methods of sequence generation and allocation. Statistical analysis: ANCOVA; number of days required to reach symptom free status (assessed as number of subjects symptom‐free at fixed points (1, 2, 3, and 4 weeks following the first visit). No sample size calculation was performed.
Participants	48 subjects; Study setting: private clinics? of physiotherapists; Country: (Western) Australia. Period and mode of recruitment: family‐oriented general practice. Age, y (mean): SMT group = 43.4; comparison group = 41.8 Gender (% F): SMT group = 33%; comparison group = 42% Inclusion criteria: 20 to 65 years of age; pain with lumbar movement or straight leg raising; pain (intermittent or constant) centrally or para vertebrally between T12 and the gluteal folds; symptoms of 3 weeks duration or less; and experienced a pain‐free period of 6 months before the onset of the current episode. Exclusion criteria: other physical treatment for the current episode of LBP; pregnancy; signs of cauda‐equina pressure or altered sensation, reflexes, or muscle weakness in the lower extremity; previous surgery in the lumbar region; past history of fracture in the lower thoracic/lumbar region; evidence of systemic disease, such as rheumatoid arthritis, ankylosing spondylitis or carcinoma.
Interventions	I) SMT (n=24): passive mobilization and manipulation. The choice of technique included 1) central, posterior‐anterior pressures, 2) unilateral, posteroanterior pressures over the transverse processes, 3) transverse pressures on spinous processes, and 4) mobilization. These techniques have been described by Stoddart and Maitland. C) Physiotherapy (n=24): 1) 15 min. of microwave diathermy, 2) 10 repetitions of isometric abdominal exercises (which the subject was request to perform another 3 to 4 times per day), and 3) ergonomic instructions, including advice on lifting, sitting, standing, carrying objects and rest postures. Each subject was treated 3 times per week for up to 3 weeks. Treatment was discontinued if the subject met the criteria for discharge, which occurred for 8 subjects. Treatment was continued beyond 3 weeks, if necessary.
Outcomes	1. Pain: 0 to 10 (unclear if a VAS or NRS) 2. Back specific functional status: according to the questionnaire by Berquist‐Ullman and Larsen (a list of 10‐different functional activities) 3. Recovery: according to the following criteria: 1) subject could perform all functional activities without difficulty, 2) his subjective pain rating was very low (0 or 1 on the 11‐point scale), 3) the objective measures of lumbar movements and straight leg raising were pain‐free, with passive overpressure at the extreme of the patient's active range; Physiological measures: active range of motion and straight leg raising 4. Adverse events: not reported Reviewers note: outcomes were not defined as primary or secondary. Follow‐up: following the first and third treatment, and at 3 weeks.
Notes	Authors results and conclusions: The duration of LBP symptoms was significantly shorter for subjects receiving mobilization and manipulation. They also achieved symptom‐free status with fewer treatment sessions. Funded by: Spinal Pain Research Foundation of the Western Australian Manipulative Therapy Association (non‐profit).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"....subjects were placed at random into two groups .....". (comment: no other description was provided).
Allocation concealment (selection bias)	Unclear risk	See above
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	No mention of any attempt to blind the patients to the procedures.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	Provider not blinded.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Patients were not blinded; therefore, outcomes assessor was also not blinded.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	Low risk	Drop‐out seems minimal, although reasons for non‐reporting of data are not given: 3 subjects (13%) from the comparison group and 1 subject (4%) from the SMT group were not assessed for symptom‐free status.
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	Unclear risk	The number of drop‐outs seems minimal and presumably assessed according to allocated assignment, although remains unclear.
Selective reporting (reporting bias)	Unclear risk	No protocol available.
Similarity of baseline characteristics?	Low risk	Age, gender, and baseline pain seem roughly similar. No other baseline measures are presented.
Co‐interventions avoided or similar?	Unclear risk	Not stated. Treatment was continued beyond 3 weeks if necessary, but not recorded. Unclear if patients sought other therapies.
Compliance acceptable?	Unclear risk	The SMT group required 3.5 (1.6) treatments to reach pain‐free status, while the comparison group required 5.8 (2.3) treatments to achieve the same result.
Timing outcome assessments similar?	Low risk	Principal outcome measure ‐ days required to reach symptom‐free status.
OVERALL RISK OF BIAS	High risk

Methods	Adequate sequence generation, but unclear allocation procedure. Statistical analysis: Mann‐Whitney (non‐parametric test). No sample size calculation was performed.
Participants	84 subjects (total) (44 subjects w/ LBP <7d.); Study setting: industry ("medium‐sized engineering works") over a period of 15 months; country: UK Age (mean): 34 to 47y. (range for those w/ LBP <7d.) Gender (%F): 11% (of those w/ LBP <7d.) Inclusion criteria: Back pain between the inferior angle of the scapula and the lower end of the sacrum. Exclusion criteria: Bilateral pain and hyperaesthesia; those under treatment by another doctor; abnormal radiological or neurological signs.
Interventions	I) SMT (n=21 (w/ LBP <7d.)): one lumbar rotational manipulation session of 15 min. or less followed by four daily detuned short‐wave diathermy sessions of 15 min. C) Detuned short‐wave diathermy only (n=23 (w/ LBP <7d.): five 15 min. daily sessions of detuned short‐wave diathermy only.
Outcomes	1. Pain: recorded as a percentage of the original pain (ranging from 0% (no relief) to100% (complete relief)) 2. Back specific functional status: not reported 3. Recovery: 3 pt. scale (better, worse, same) ‐ measured but not reported 4. Other: skin hyperaesthesia, deep tenderness, restriction of straight leg raising, forward flexion of the spine. Reviewers note: these other outcomes were to be reported in a subsequent publication, however, no such publication was identified. 5. Adverse events: not reported Reviewers note: outcomes were not defined as primary or secondary. Follow‐up: Days 3 ,7; results from 1 month were not reported.
Notes	Authors results and conclusions: There was no demonstrable difference between the intervention (SMT) and control groups, except that at the 15‐minute stage, the relief from pain in the manipulated group was always greater than in the controls. Funded by: Nuffield Foundation, UK ("covered the salaries" of the principal researchers (non‐profit)).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were allocated to the manipulation or control series in accordance with a pre‐arranged sequence for that subgroup (reviewers note: four subgroups were made based upon duration of the pain (<7days, >7 days), based on random sampling numbers and contained in a set of sealed envelopes."
Allocation concealment (selection bias)	Unclear risk	Note: The person involved in the allocation was also involved in the treatment of the subjects, so it is uncertain to what extent he might have influence the allocation procedure.
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	No mention of attempting to blind patients to the intervention.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	Provider not blinded
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Subjects were not blinded; therefore, outcomes assessor was also not blinded.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	Unclear risk	Data presented for the first week only. Unclear what proportion of subjects dropped out beyond this interval because this data is not presented (although it was reportedly measured).
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	Unclear risk	Not explicitly stated and difficult to assess for follow‐up beyond one week.
Selective reporting (reporting bias)	High risk	Protocol available in which the outcomes are described and time of measurement. "All cases were followed up for one month or more." (Reviewers note: Yet these data are not reported; the author only reports the short‐term (1‐week) follow‐up data. In addition, the authors present only those outcomes related to pain and not the other outcomes, such as those related to hyperaesthesia, tenderness or range of motion).
Similarity of baseline characteristics?	Unclear risk	Baseline data reported: age, gender (only). No indication of severity of baseline pain, functional status, etc.
Co‐interventions avoided or similar?	Unclear risk	Not reported
Compliance acceptable?	Unclear risk	Not reported.
Timing outcome assessments similar?	Low risk	At 3 and 7 days f/u
OVERALL RISK OF BIAS	High risk

Methods	Adequate sequence generation, unclear allocation concealment. Statistical analysis: Descriptive statistics were used to analyse the baseline characteristics. A MANOVA was used for analysing the effects on pain, disability and mobility between the treatment groups. Sample size calculation was based upon determining a MCID (50% reduction) for pain.
Participants	64 subjects; study setting: physical therapy and manual therapy practices; Country: The Netherlands Period and mode of recruitment: patients presenting to PT clinics in the period 2005‐2006. Age: stratified by age, no mean (SD) is presented; the majority of subjects were between 20 and 50 years of age. Gender (%F): 45% Inclusion criteria: acute (<16d.), non‐specific low‐back pain, between 20 to 55 years of age, with or without previous complaints and no symptoms distal to the knee. Exclusion criteria: specific low back pain (e.g. with neurological signs, rheumatic disease, signs of osteoporotic fractures); inability to fill in the questionnaires.
Interventions	I) SMT + PT (n=31); C) PT (n=33). ‐SMT consisted of high‐velocity low‐amplitude SMT designed to cause cavitation of the joint and to improve overall joint function, decrease any restrictions in movement at segmental levels in the lumbar spine and sacroiliac joint, and reduce pain. No other technique was applied. The manual therapist chose the techniques on the basis of the location of the dysfunction and in each treatment session, only one manipulation was applied, with an added time investment of approximately four minutes. ‐Physiotherapy consisted of: standard PT care under the principle of increasing physical functioning. Participants participated in a low‐intensity, low‐load endurance exercise programme designed to train specific abdominal muscles, to be completed 5 minutes, two times per day. All participants were also given instructions in staying active and a pamphlet from the Royal Dutch Society for Physical Therapy.
Outcomes	1. Pain: 0‐100 VAS (last 24 hours) 2. Back‐specific functional status: ODI 3. Recovery: 2‐point scale, measured at the 4th visit (improved or not improved) 4. Mobility: evaluated using the Sit‐and‐Reach Test (designed to test the mobility of the lower‐body and spinal column) 5. Adverse effects: not reported Note: outcomes not defined as primary or secondary. Follow‐up: at the 4th visit (ca. 2 1/2 weeks post‐baseline).
Notes	Authors results and conclusions: The addition of SMT to PT care demonstrates a significant effect on improvement in functional status, but not on pain relief or improvement of mobility. This study does not support the efficacy of a clinical prediction rule in the treatment of acute, non‐specific low‐back pain. Funded by: not specified. Note: Primary author provided additional information regarding the RoB assessment, which resulted in a number of items being adjusted from "high" to "low" risk of bias.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"An independent employee, not involved in recruitment of participants, generated a random stratified list where by means of a computerized programme, ...."...patients were assigned to one of the two treatment groups.
Allocation concealment (selection bias)	Low risk	Unclear who actually conducted the allocation from the publication; however, we personally contacted the principal author and confirmed that an independent researcher conducted the allocation.
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	No attempt was made to blind patients to the intervention.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	Providers were not blinded.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Patients were not blinded; therefore, outcome assessor was also not blinded.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	Low risk	According to personal contact with the principal author, only one subject dropped out. No flow chart is provided in the publication.
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	Low risk	Stated that an ITT analysis is used; no alternative analyses are presented (e.g. per‐protocol). One patient discontinued care (from the experimental group) due to increasing pain ‐ based upon the numbers it would appear that this subject was removed from the analyses; however, this is not likely to have influenced the overall results.
Selective reporting (reporting bias)	Unclear risk	No protocol available, although recovery was measured, but not reported.
Similarity of baseline characteristics?	High risk	Baseline variables measured: age (stratified by group), gender, duration, onset of LBP (sudden, gradual), presence/absence of relapse, type of physical labour performed, use of medication, outcomes: pain, functional status, mobility. Note: baseline pain for the two groups: SMT + PT = 42.7 (18.4)) vs. PT = 54.0 (17.5).
Co‐interventions avoided or similar?	Unclear risk	According to the primary author, "no other technique was applied", but it is unclear if participants sought care elsewhere outside the constructs of the trial. This was not examined.
Compliance acceptable?	Low risk	According to the primary author, participants were provided a checklist which was to be completed every day. No differences were found between the groups and it appears that both groups adequately completed their exercises at home. Patients were required to complete 5min. twice per day and this was checked by the checklist.
Timing outcome assessments similar?	Low risk	Follow‐up was conducted at the 4th visit (2 1/2 weeks following baseline).
OVERALL RISK OF BIAS	Low risk

Methods	Unclear sequence generation and allocation. Statistical analysis: Mann‐Whitney U test for non‐parametric data; correlations were evaluated using Spearman rank‐order correlation coefficient. No sample size calculation was performed.
Participants	95 subjects; Study setting: medical clinic/department of physical medicine and rehabilitation; Country: USA. Period and mode of recruitment: subjects presenting to University of California, Irvine Medical Center Back clinic between June 1973 to June 1979. Age, y (mean (SD)): SMT group ‐ 30.1 (8.4); control group ‐ 32.1 (9.8) Gender (%F): SMT group ‐ 41%; control group ‐ 41% Inclusion criteria: Presence of palpatory cues indicating hyperalgesia or a restricted or painful range of vertebral motion; absence of any contraindications for vertebral manipulation; absence of any psychosocial problems. Exclusion criteria: Previous experience with manipulation; disability income; pending litigation; previous back surgery; obesity; drug or alcohol abuse; pain not treatable by manipulation of the lumbosacral area.
Interventions	I) SMT (n=58): rotational manipulation of the lumbar spine. "This was carried out as follows: the patient lies on his or her side on a table facing the manipulator. The inferior leg is extended and the superior leg is flexed, tilting the superior aspect of the pelvis toward the manipulator. The superior shoulder is rotated away from the manipulator and the spine is locked in extension. A short, high‐velocity thrust is then applied to the pelvis. This presumably has the effect of gapping the facet joints and stretching the paravertebral muscles of the lumbosacral area." C) Massage (n=39): soft‐tissue massage to the lumbosacral area only. The number of treatments was variable and left to the discretion of the practitioner.
Outcomes	1. Pain: 5‐point ordinal scale, ranging from none to very severe 2. Back specific functional status: listed as specific activities and ability to perform them, such as walking, bending or twisting, sitting down in a chair, sitting up in bed, etc. 3. Recovery: reported as those responding to the question whether the "treatment was effective" (or not) 4. Adverse events: not reported 5. Other: physiological signs: improvement in straight leg raising to pain or to pelvic rotation, and distance of the fingertips to the floor with maximum forward flexion Reviewers note: outcomes were not defined as primary or secondary. Follow‐up: following first treatment; discharge and 3 weeks following discharge.
Notes	Authors results and conclusions: Patients who received manipulative treatment were much more likely to report immediate relief after the first treatment; and at discharge, there was no significant difference between the 2 groups because both showed substantial improvement. Funded by: Medical Trust, Los Angeles, USA (non‐profit?).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"...each patient was randomly assigned to either the experimental or control group." Comment reviewers: no other information was provided.
Allocation concealment (selection bias)	Unclear risk	See above
Blinding (performance bias and detection bias) All outcomes ‐ patients?	Unclear risk	Results from the query post‐treatment (at 3 weeks follow‐up) is as follows (data available for 60% of the baseline sample (note: therefore, possible respondent bias)): 33 received OMT (57% of the original sample (n=33/58)): 21 (65%) perceived it to be OMT and 12 (35%) perceived it to be sham OMT; 25 received sham OMT (64% of the original sample (n=25/39)): 14 (58%) perceived it to be OMT and 11 (42%) perceived it to be sham OMT. The authors claimed the study was successful in comparing manipulation to an appropriate placebo treatment. Reviewers note: This comparison was rated unclear risk because it was unclear what was said to the subjects and whether the subjects were aware that they would receive either soft‐tissue massage or SMT, or whether they were told that they would receive treatment A or treatment B and which was most effective would be evaluated. In other words, if type of treatment was mentioned, the participating subjects could have determined if they thought they received an effective treatment or not, whereas if the type of treatment was not mentioned, the participating subjects would not have had a frame of reference to compare and thus, remain blinded to type of treatment modality.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	Providers were not blinded.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	Unclear risk	See reviewers comment above.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	High risk	Lost to follow‐up: Discharge: SMT: 29% (n=17/58); control: 28% (n=11/39) 3 wks following discharge: SMT: 43% (n=25/58); control: 36% (n=14/39)
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	Unclear risk	Not stated
Selective reporting (reporting bias)	Unclear risk	No protocol available
Similarity of baseline characteristics?	High risk	Baseline variables reported: age, gender, proportion with acute and chronic pain, proportion with moderate to very severe pain, proportion with impaired gait and abnormal lumbar curve (as reported by the physician). Difference in percentage of patients with chronic pain: 17 (SMT) versus 29 (control). Difference in percentage of patients with severe to very severe pain: 37 (SMT) versus 16 (control).
Co‐interventions avoided or similar?	Unclear risk	Not stated
Compliance acceptable?	Unclear risk	Not stated
Timing outcome assessments similar?	High risk	Discharge and 3 weeks following discharge, which would have been different for individual subjects. "The number of treatments received was variable, at the discretion of the treating physicians."
OVERALL RISK OF BIAS	High risk

Methods	Randomization procedure not described. Unclear if patients were aware that they were randomized. Statistical analysis: Results were analysed in three groups according to duration of the present episode: groups 1 up to and including 13 days; group 2: 14‐28 days inclusive; group 3: 29 days and over. t‐test was performed to detect statistically significant differences between groups. No sample size calculation was performed.
Participants	95 subjects; Study setting: general practice; Country: UK. Period and mode of recruitment: not stated. Age (all subjects): 16 to 70 years of age Gender (% female): 28% (SMT), 24% (control) Inclusion criteria: 16‐70 y/o; pain partly or wholly between the inferior angles of the scapulas and the buttock folds. Exclusion criteria: Patients suffering from inflammatory joint disease, skeletal metastases or infection, spondylolisthesis, neurological deficit in structures innervated by lumbar or sacral roots that could not be ascribed to a previous resolved episode or other pathology, osteomalacia or osteoporosis, visceral pathology that could refer pain to the low back, pregnancy. Also excluded were those who sought or intended to seek physical treatment outside the practice for their present episode, and transient patients.
Interventions	I) SMT (n=49): classical range of osteopathic manipulative manoeuvres of the type most likely to be delivered to a patient in the UK from a registered osteopath. The following elements were used in case (if could be applied without pain): direct pressure and stretching to involved musculature, low‐velocity high‐amplitude thrust techniques (HVT) to hypomobile vertebral motion segments, especially those from which the presenting pain was deemed to originate. Treatment continued twice weekly until either patients deemed themselves recovered or the manipulator decided that further treatment was unlikely to produce benefit. C) Control (n=46): controls were seen in the clinical as necessary for examination for incapacity certificates, reinforcement of postural advice, and reassurance. Control patients were told that was no treatment that had been shown to be superior to the program of rest and graded resumption of activities on which they were embarked. Both groups received advice on posture, exercise, and avoidance of occupational distress as appropriate to their situation.
Outcomes	Primary outcome: 1. Patient‐reported recovery (yes/no) Secondary outcomes: 2. Either patient‐reported recovery or a zero Disability Index (DI) score. Improvement in DI score from presentation to 1, 2 and 3 weeks into the trial were also examined Outcome measures: DI score, asking the subject to mark which of 12 activities were comfortable (scored from 0‐12); VAS pain indicating level of pain between pain‐free (0) and the maximum score of 75, midpoint being the level of pain during the worst 24 hours before presentation; Activity Loss Analog (ALA), a similar VAS ranging from "full normal activity" to "least possible activity"; Index of compliance = ALA divided by VAS pain (activity loss per unit pain); return to partial or full work (or home duties), and reported recovery (yes/no). Follow‐up: 2x/wk for 3 weeks following trial entry and then weekly until the patients deemed themselves recovered or for 3 months, if un recovered. Adverse events: "One control and one manipulated patient developed signs of neurologic deficit soon after trial entry....". No significant difference in development of excess lumbar lordosis or paraesthesias for the 2 treatment groups.
Notes	Authors results and conclusions: "Even with the small numbers appropriate to a pilot trial, we have confirmed a significant benefit from manipulation to identifiable group of back pain patients. The advantage to manipulated patients was maximal between 1 and 2 weeks after commending treatment, but was not discernable after 4 weeks." Funded by: Osteopathic Trust Ltd (Non‐profit), Rehabilitation Products Ltd (industry ‐ loaned the manipulation tables).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Patients were randomly allocated to one of two groups". No further statements on randomization.
Allocation concealment (selection bias)	Unclear risk	See above
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	Not stated. It is not clear whether patients were informed about the two treatment arms and the fact that they were randomized.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	No mention of any attempts to blind the provider.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Not stated. It is not clear whether patients were informed about the two treatment arms and the fact that they were randomized.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	Low risk	"3 control patients were lost to follow‐up less than 2 weeks into the trial and 2 developed complications shortly following trial entry."
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	High risk	5 were excluded from the analyses (2 that developed complications shortly following trial entry and 3 which were lost to follow‐up within the first 2 weeks (n=95). Table 3 shows 94 patients.
Selective reporting (reporting bias)	Unclear risk	No protocol.
Similarity of baseline characteristics?	Low risk	Variables examined: age, gender, excess lumbar lordosis, straight leg raising, erect pelvic tilt, leg length difference, periods of "pins and needles", periods of "numbness", duration of present episode, previous episodes, disability index, VAS pain, Activity Loss (ALA). The only significant difference was a higher prevalence of complaints of "pins and needles" in the SMT group (P<0.005), and also in the SMT group a higher proportion reported episodes of transient numbness in the leg (P<0.01).
Co‐interventions avoided or similar?	Unclear risk	No mention of co‐interventions or avoidance of co‐interventions.
Compliance acceptable?	Low risk	231 manipulative treatments were given: 87% of these in the first 2.5 weeks of the trial. It can be calculated that most patients received 2‐3 treatments.
Timing outcome assessments similar?	Low risk	Questionnaires were completed twice weekly for 3 weeks after trial entry and then weekly until the patients deemed themselves recovered, or for 3 months if not recovered.
OVERALL RISK OF BIAS	High risk

Methods	Sequence generation and allocation procedure unclear Statistical analysis was performed using Wilcoxon signed‐ranks test. No sample size calculation was performed.
Participants	159 randomly allocated to 6 treatment groups (of a total 459 who were either lost to follow‐up, changed treatment assignment or had chronic low‐back pain); Study setting: hospital outpatient department (university orthopaedic clinic and a "Static Center" of Rome); Country: Italy. Period and mode of recruitment: January 1985 ‐ October 1986 Age (mean (years)): group1B ‐ 38.4; group2B ‐ 39.5 Gender (%F): group1B ‐ 51% (39/77); group2B ‐ 49% (39/80) Inclusion criteria: low‐back pain, aged 17‐58 years. Pattern of pain radiation: with and without radiation below knee; 2 groups ‐ acute (< 4 weeks) and chronic (> 9 weeks) LBP. Duration of the current LBP (mean): group 1B ‐ 13 months; group 2B ‐ 9 months (all other groups are not relevant for this report). Exclusion criteria: Pregnancy or nursing women, serious general diseases, psychiatric disturbances, medico‐legal litigation.
Interventions	Two principal grps: group1 ‐ LBP only; group2 ‐ LBP radiating to the buttocks and/or thighs and no neurological changes. Subgroups were defined as: A ‐ LBP <4 wks. duration and no LBP in the preceding 6 months; B ‐ continuous or almost continuous LBP lasting more than 2 months; C ‐ chronic LBP with an episode of acute pain at the time of clinical observation. I) Manipulation by trained chiropractor [at follow‐up: n=87]; no. chronic patients: 12; at a rate of 2 treatment per week C1) Diclofenac "full dose" [at follow‐up: n=81]; duration treatment: 2 weeks C2) Physiotherapy: massage, electrotherapy, infrared, etc. [at follow‐up: n=78]; no. treatment: 15, daily for 3 weeks C3) Bed rest [at follow‐up: n=29]; duration treatment: 6 ‐ 8 days C4) Back school [at follow‐up: n=50]; no. treatment: 4 in 1 week C5) Placebo gel [at follow‐up: n=73]; duration 1 or 2 weeks
Outcomes	1. Pain: 4 point scale: ranging from none to most severe pain imaginable 2. Back specific functional status: 4‐point scale: extremely, moderately, slightly or not limited 3. Recovery: not reported. Overall evaluation was based upon a sum score including both subjective and objective measures. 4. Spinal mobility: forward flexion: fingertip to floor distance; Abdominal muscle strength: assessed by the leg‐lowering test, and isometric endurance 5. Adverse events: not reported Note: Outcomes not defined as primary or secondary by the authors. Follow‐up: 3 weeks, 2, 6 months.
Notes	Authors results and conclusions: In subgroup1B, the best results were obtained with physiotherapy at short‐term and low‐back school at the long‐term. For subgroup2B, physiotherapy gave the best results at both short‐ and long‐term follow‐up. Funded by: Centro Studi di Patologia Vertebrale, Rome (non‐profit). Unequal numbers for the intervention groups because not all interventions applied to the various groups (acute ‐ chronic).
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Patients in each group were randomly assigned to the following treatments.
Allocation concealment (selection bias)	Unclear risk	Note: No other information was provided on the sequence generation or allocation.
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	There is no mention of attempts to blind the patients to other interventions or their perceptions of potential effectiveness of the different interventions.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	No mention if there were any attempts to blind the care providers to the other groups.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Patient was not blinded; therefore, this item was scored as "no". No mention if there were any attempts to blind the outcome assessors to treatment allocation for the subjective or objective outcome measures.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	Low risk	5% (n=23/459) were lost to follow‐up.
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	High risk	13% of those randomized were either lost to follow‐up or changed their assigned treatment and subsequently not included in the analyses.
Selective reporting (reporting bias)	Unclear risk	No published protocol available; recovery not reported.
Similarity of baseline characteristics?	Unclear risk	Similar for the 2 groups with chronic LBP (based upon age, gender, and duration of symptoms), but unclear for the baseline scores for functional status.
Co‐interventions avoided or similar?	High risk	8% (38/459) of the subjects had interrupted or changed their assigned treatment.
Compliance acceptable?	Unclear risk	Not stated
Timing outcome assessments similar?	Low risk	At 3 weeks; 2, 6 months.
OVERALL RISK OF BIAS	High risk

Methods	Unclear methods of randomization and allocation. Statistical analysis: Mann‐Whitney rank‐sum test. No sample size calculation was performed.
Participants	76 subjects; Study setting: Dept. of physical medicine and rheumatology, Alborg Hosptial; Country: Denmark. Period and mode of recruitment: 1975, referred from the general practitioner to the rheumatologist. Age, y (all, mean (SD)): 34.9 (7.3) Gender (%F): 0% Inclusion criteria: males; 20‐50 years of age; LBP without signs of root pressure; duration less than 3 weeks; no treatment except analgesics before entering the trial. Exclusion criteria: contraindication to manipulation; no other exclusion criteria were stated.
Interventions	I) SMT (n=12): rotational manipulation in the pain‐free direction, delivered by either a physiotherapist or physician using the same technique. C) short‐wave diathermy (n=12): no further description was provided. Therapy was delivered for both groups 3 times per week for 2 weeks.
Outcomes	1. Pain: unclear how this was measured 2. Back specific functional status: not reported 3. Recovery: declared restored if subjects fulfilled the following criteria: no pain at all, normal function (according to Schober's test?), no objective signs of disease (unclear how this was determined), and fit to work (not reported) 4. Adverse events: not reported Follow‐up: 1, 2 weeks
Notes	Authors results and conclusions: There was a significant difference in recovery (92% of patients treated with manipulation were free of symptoms within 2 weeks compared to 25% in the short‐wave diathermy group). All patients in the manipulation group improved function/mobility whereas only half of the patients in the short‐wave group improved (P<0.01). Funded by: Not stated.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"... patients were randomized ....". Reviewers note: no other details were provided.
Allocation concealment (selection bias)	Unclear risk	See above.
Blinding (performance bias and detection bias) All outcomes ‐ patients?	High risk	No mention of an attempt to blind the patients to therapy.
Blinding (performance bias and detection bias) All outcomes ‐ providers?	High risk	Care provider was not blinded.
Blinding (performance bias and detection bias) All outcomes ‐ outcome assessors?	High risk	Patient was not blinded; therefore, outcomes assessor also not blinded.
Incomplete outcome data (attrition bias) All outcomes ‐ drop‐outs?	Low risk	"92% of the allocated subjects returned for final measurement at 2 weeks."
Incomplete outcome data (attrition bias) All outcomes ‐ ITT analysis?	High risk	Not explicitly stated. In addition, 2 patients (8% of the sample) were dropped from the analysis (one patient dropped out and another did not fulfil the inclusion criteria).
Selective reporting (reporting bias)	Unclear risk	No protocol is available.
Similarity of baseline characteristics?	Unclear risk	Baseline characteristics for the 2 groups are not presented.
Co‐interventions avoided or similar?	Unclear risk	Not stated
Compliance acceptable?	Unclear risk	Not stated
Timing outcome assessments similar?	Low risk	At 1, 2 weeks.
OVERALL RISK OF BIAS	High risk

Study	Reason for exclusion
Andersson 1999	Includes a population with primarily subacute and chronic LBP.
Beyerman 2006	Duration not specified.
Bishop 2010	Unique contribution of SMT cannot be determined from the study design. The intervention group (SMT) received "reassurance regarding the natural history of acute mechanical LBP; advice to avoid passive treatment approaches (e.g., bed rest, heat, or the use of back supports/corsets/braces); advice to carry out a progressive walking program (two walks a day, each with an initial duration of between 5 and 15 minutes depending on the patient’s tolerance increasing by 2 minutes each walk per week); acetaminophen, 650mg every 6 to 8 hours as required for 2 to 4 weeks, unless medically contraindicated (e.g., because of allergy, compromised liver function, or acute porphyria); and a maximum 4 weeks of lumbar spinal manipulative therapy using conventional side posture,high‐velocity, low‐amplitude techniques....". The control group "were advised of their diagnosis (i.e., mechanical low back pain) and referred back to their referring family physician with a letter explaining the protocol of the present study." Note: see also Table 3.
Blomberg 1994	Multi‐modal treatments were delivered in the various treatment arms, thus, making the assessment of the effects of SMT impossible. The intervention group consisted of: manipulation (thrust techniques) or "more gentle specific mobilization", muscle stretching (almost all received), auto‐traction (15% of the group received this) followed by "steroid injections, often in combination with needling and local anaesthetics" for those not responding after 1‐2 weeks of treatment (which represents 54% of the intervention group). The conventional care group consisted of: "drugs, low‐back pain school training, active back exercises, corsets, taping, short‐wave ultrasonic waves, transcutaneous nerve stimulation (TNS), transcutaneous electric muscle stimulation (TEMS), heat, cold, postural exercises, and in some cases, plunge‐bath training and massage".
Bronfort 1989	Proportion of subjects with acute low‐back pain is unclear (cut‐off reported in the study was 8 wks).
Coyer 1955	Pseudo‐RCT: alternate inclusion.
Delitto 1993	Unique contribution of SMT cannot be determined from the study design. The intervention group consisted of SMT (mobilization to affect the sacroiliac joint) followed by McKenzie extension‐oriented exercises. The comparison group consisted of flexion‐oriented exercises. "On the return visits, the major focus was to assess how the patient was performing the prescribed exercise regimen." Thus, the value of SMT is unclear and it is uncertain if SMT was delivered only once at the first visit.
Doran 1975	Proportion of subjects with acute low‐back pain is unclear (16% < 1 week; 56% > 1 month (of which 14% > 6 months)).
Erhard 1994	Unique contribution of SMT cannot be determined from the study design. The intervention group consisted of SMT immediately followed by "hand‐heel rocking" (consisting of an exercise designed to induce flexion and extension to the spine). The comparison group included an extension‐oriented treatment regimen as proposed by McKenzie.
Gemmell 1995	No clinical assessment beyond one day.
Godfrey 1984	Unique contribution of SMT cannot be determined from the study design. The experimental group received SMT (according to the technique described by Maigne) followed by soft tissue massage. The comparison group consisted of massage (including light effleurage) administered by a kinesiologist or electrical stimulation to either side of the lumbar spine.
Grunnesjo 2004	Unique contribution of SMT cannot be determined from the study design. Two groups were examined: the reference therapy consisting of the "stay‐active" concept (includes physical training, non‐specific traction, passive physiotherapy modalities) and the experimental group consisting of the "stay active" concept plus SMT and steroid injections (half of the patients) in combination with needling and local anaesthetics, where indicated. In a subsequent publication (Bogefeldt 2008) the groups were isolated into the original 4‐group factorial design so that theoretically the additional effect of SMT could be isolated. However, an examination of Table 1 (Grunnesjo 2004) suggests that many in the experimental group did not also receive passive and active physical training (in contrast to the study design, which examined the additional benefit of stretching, SMT and injections) to the "stay active" concept; therefore, it is our opinion that the effects of SMT cannot be isolated. Furthermore, in Table 2 (Bogefeldt 2008), there are some fundamental differences across the groups and most notably for the SMT group, such as percentage women included, percentage on sick‐leave at the time of inclusion and in the previous two years, and percentage of those with x‐rays due to a previous low‐back pain episode.
Helliwell 1987	No relevant outcome measure.
Hsieh 2002	Overall duration of back pain between 10.7 to 11.8 weeks (SD range from 6.6 to 7.2 weeks).
Hurley 2004	Mean duration of the LBP for the study population was 8 weeks (range 7.6 to 8.3 weeks for the three intervention groups).
Kinalski 1989	Duration not specified.
Mathews 1987	Evaluates subjects with exclusively sciatica.
Meade 1990	Proportion of subjects with acute low‐back pain is unclear (cut‐off that was reported in the study was 1 month: 59% of the chiropractic group and 60% of the hospital group had back pain longer than 1 month).
Nwuga 1982	Pseudo‐RCT ‐ alternate inclusion of subjects.
Pope 1994	Evaluates chronic LBP.
Rupert 1985	Mix of acute, subacute and chronic LBP, and although a subgroup analysis is presented separately for the acute patients, it is unclear what proportion these patients represent as no numbers of subjects are presented for any of the groups separately.
Sanders 1990	Outcome assessment not longer than 1 day.
Santilli 2006	Evaluates subjects with exclusively sciatica.
Sims‐Williams 1978	Duration not specified.
Sims‐Williams 1979	Duration not specified.
Terrett 1984	Asymptomatic subjects
Waterworth 1985	Unique contribution of SMT cannot be determined from the study design. Not all subjects in the SMT group received spinal manipulation: some received exercises (according to McKenzie) and some received SMT alone. The other two comparison groups examined medication (Diflunisal) and standard physiotherapy (heat, ultrasound, flexion and extension spinal exercises).
Williams 2003	Subjects recruited with 2‐12 weeks of spinal pain (of which 61% had low‐back pain only); unclear what percentage of subjects had acute or subacute low‐back pain. In addition, the effect of SMT could not be ascertained: "All patients in the trial continued to receive treatment as usual from their GPs.... The control group did not receive any additional intervention. ...The treatment package consisted mainly of osteopathic spinal manipulation....... Occasionally, if symptoms persisted despite osteopathy, tender ligaments or peripheral joints were injected with corticosteroid and local anaesthetic."
Wreje 1992	Proportion of subjects with acute low‐back pain is unclear (no data were presented regarding duration of the back pain). Exclusion criteria was pain of more than 3 months duration.
Zylbergold 1981	Duration not specified.

Trial name or title	http://ClinicalTrials.gov/show/NCT00497861 Title: Comparison of Mechanical Force, Manually Assisted Activator Manipulation Versus Manual Side Posture Manipulation in Patients With Low Back Pain: a Randomized Pilot Study Purpose: This study compared the treatment effect of Activator Methods Chiropractic Technique (AMCT) and manual Diversified type spinal manipulative therapy in a sample of patients with acute and sub‐acute low back pain.
Methods	Allocation: Randomized Endpoint Classification: Safety/Efficacy Study Intervention Model: Parallel Assignment Masking: Open Label Primary Purpose: Treatment
Participants	Inclusion Criteria: Being 18 years or older; Having current acute or subacute low back pain defined as pain that has not lasted more than 16 weeks; Minimum score of 30mm on a 100mm visual analog pain scale. Exclusion criteria consisted of the following: Have any of six possible un underlying causes of low back symptoms in their history (spinal osteomyelitis, spinal fracture, herniated disc, ankylosing spondylitis, cauda equina syndrome, or cancer, excluding nonmalignant skin cancer); Have undergone surgery involving the low back; Have received workers’ compensation benefits within the preceding year or were potentially involved in litigation relating to back problems; Pregnancy, because of possible need for exposure to diagnostic x‐rays; Have participated as a subject in research previously at the trial clinic site; Have received spinal manipulation within the preceding 3 months or on more than three occasions during the preceding year. Subjects with sciatica were excluded if they had any one of the following:Ankle dorsiflexion / plantar flexion weakness;Great toe extensor weakness;Absence of knee or ankle reflexes;Loss of light touch sensation in the medial, dorsal, and lateral aspects of the foot;Ipsilateral straight‐leg‐raising test (positive result: leg pain at <60°);Crossed straight‐leg‐raising test (positive result: reproduction of contralateral pain).These six neurologic tests allow detection of most clinically significant nerve root compromises resulting from L4‐L5 or L5‐S1 disc herniations, which together make up more than 90% of all clinically significant radiculopathies attributable to lumbar disc herniations (21‐25). Because approximately 12% of ambulatory patients with back pain h Have symptoms of sciatica or leg pain without neurologic compromise related to lumbar disc herniation,[5] investigators attempted to include such subjects in the trial. The criteria described above were intended to minimize the likelihood of including subjects with a lumbar disc herniation.
Interventions	Manually Assisted Activator Manipulation versus Manual Side Posture Manipulation
Outcomes	Primary outcomes measured include pain measurement with a VAS scale, the use of the ODI and the Bournemouth back pain scale questionnaire.
Starting date	Study completion date: April 2007
Contact information	Principal Investigator: Mark T. Pfefer, DC, RN; Cleveland Chiropractic College
Notes	SMR had contact with a colleague of the PI and stated that as of Sept. 2010, the study had not been submitted for publication; a search in PubMed by SMR in May 2011 did not reveal any listing.

Trial name or title	http://clinicaltrials.gov/show/NCT00632060 Title: The Efficacy of Manual and Manipulative Therapy for Low Back Pain in Military Active Duty Personnel: A Feasibility Study The specific aims of this research project are to determine feasibility of, and the comparative treatment effect size for, conducting a larger clinical trial of Manual/Manipulative Therapy (M/MT) in restoring peak performance in military personnel in operational environments and to evaluate the ability of the addition of M/MT to standard care to decrease pain and increase function for patients with low back pain. The following two hypotheses will guide the data collection: The primary hypothesis is that the addition of a course of M/MT to standard care for low back pain will decrease pain at 4 weeks when compared to standard care alone; In addition, the secondary hypothesis will be that the addition of a course of M/MT to standard care for low back pain will decrease pain and increase function over 2 and 4 weeks when compared to standard care alone.
Methods	Allocation: Randomized Endpoint Classification: Efficacy Study Intervention Model: Parallel Assignment Masking: Open Label Primary Purpose: Treatment
Participants	Criteria Inclusion Criteria: Active Duty Aged 18‐35 New episode of low back pain (LBP) or a reoccurrence of a past episode of low back pain Exclusion Criteria: LBP from other somatic tissues as determined by history, examination, and course (i.e. pain referred from visceral conditions) Radicular pain worse than back pain Co‐morbid pathology or poor health conditions that may directly impact spinal pain. Patients who have case histories and physical examination findings indicating other than average health will be excluded from the study Bone and joint pathology contraindicating patient for M/MT. Patients with spinal fracture, tumors, infections, inflammatory arthropathies and significant osteoporosis will be referred for appropriate care and will be excluded from the study Other contraindications for M/MT of the lumbar spine and pelvis (i.e. bleeding disorders or anticoagulant therapy) Pregnancy (all potential female participants will undergo pregnancy testing) Use of manipulative care for any reason within the past month Unable to follow course of care for four weeks Unable to give informed consent for any reason Unable to confirm that they will not be deployed during the course of the study: "Will you be deployed, receiving orders for a distant temporary active duty assignment, attending training at a distant sight, or otherwise absent from Ft. Bliss over the next 6 weeks?"
Interventions	1. No Intervention Standard Care Control Group ‐ Participants randomized to the standard care group will continue their use of non‐prescription or prescription medication and reduced duty loads, as prescribed by the credentialed medical provider. 2. Experimental Manual / Manipulative Therapy Group: Participants randomized to the M/MT group will receive a course of M/MT along with standard care. The patient will see the chiropractor twice a week for the entire course of the study, regardless of manipulation or not.Intervention: Procedure: Manual / Manipulative Therapy (M/MT).
Outcomes	Primary Outcome Measures: Decreased pain [ Time Frame: Baseline, 2 weeks, 4 weeks ] [ Designated as safety issue: No ] Secondary Outcome Measures: Increased function [ Time Frame: Baseline, 2 weeks, 4 weeks ] [ Designated as safety issue: No ]
Starting date	February 2008
Contact information	Maria Hondras, DC, MPH (maria.hondras@palmer.edu)
Notes	SMR had contact with one of the principal investigators in May 2011: The mean (SD) duration of LBP in the study is 11.5 (8.5) days and the median is 9 days. The investigators are currently in the final stages of manuscript preparation.

PERMALINK

Spinal manipulative therapy for acute low‐back pain

Sidney M Rubinstein

Caroline B Terwee

Willem JJ Assendelft

Michiel R de Boer

Maurits W van Tulder

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Spinal manipulative therapy for acute low‐back pain

Summary of findings

Summary of findings for the main comparison.

Summary of findings 2.

Summary of findings 3.

Summary of findings 4.

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Inclusion criteria

Exclusion criteria

Types of interventions

Experimental intervention

Types of comparisons

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Assessment of clinical relevance

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Results

Description of studies

Table 1.

Results of the search

Figure 1.

Included studies

Types of studies

Study population

Technique: type, practitioner, number and duration of treatments

Outcome measures: type, timing

Primary outcomes

Secondary outcomes

Other outcomes

Follow‐up

Safety

Excluded studies

Risk of bias in included studies

Figure 2.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Figure 3.

Figure 4.

Effects of interventions

Trial name or title	http://clinicaltrials.gov/show/NCT01211613 A Comparison of Chiropractic Manipulation Methods and Standard Medical Care for Low Back Pain. Purpose: The investigators will be comparing the effectiveness of two types of chiropractic manipulation and standard medical care for patients with a recent onset of low back pain. The two types of chiropractic treatments being compared will be hands‐on (manual) manipulation and mechanical‐assisted (Activator) manipulation. The standard medical care will consist of a medical examination and prescription for over‐the‐counter anti‐inflammatory medication.
Methods	Allocation: Randomized Endpoint Classification: Efficacy Study Intervention Model: Single Group Assignment Masking: Single Blind (Investigator) Primary Purpose: Treatment
Participants	Criteria Inclusion Criteria: Age: ≥ 18 years of age Ability to read and write English Experiencing a new episode of LBP with onset in the past 3 months ODI between 20‐70 points (0‐100 scale) Numeric pain rating score between 3‐8 points (0‐10 scale) Exclusion Criteria: Prior history of lumbar spine surgery History of unstable spondylolisthesis, spinal stenosis, or scoliosis > 20° Signs or symptoms suggestive of nerve root tension and/or neurological deficit in the lower extremity History of metastatic cancer, osteoporosis, long‐term corticosteroid use, or any other red flags of serious illness including the following: unexplained weight loss of >10% of body weight, spinal pain associated with fever, and severe night pain unrelieved by medication Receiving any physical therapy, chiropractic therapy, or any other manual therapy for this episode of LBP (within the past 3 months) Receiving any on‐going medical care for this episode of LBP Current use of opiate or other prescription medications for LBP
Interventions	Procedure: Manual Manipulation Doctor of chiropractic will apply manual high‐velocity low‐amplitude thrust to lumbar spine of research participants. Device: Mechanically‐assisted manipulation Doctor of chiropractic will use the Activator Instrument to apply a mechanically‐assisted thrust to the lumbar spine of research participants.Other Name: Activator IV Instrument: FDA approval# K003185. Other: Standard Medical Care Patients will receive an examination with a physician who is board certified in physical medicine and rehabilitation. Treatment will consist of medical monitoring of the patient's condition over 4 weeks (baseline and 2 follow up exams) and a prescription for over‐the‐counter anti‐inflammatory medications if indicated.
Outcomes	Primary Outcome Measures: ODI [ Time Frame: 4 weeks ] [ Designated as safety issue: No ] Questionnaire of level of self‐reported impairment of ADLs due to low back pain. Secondary Outcome Measures: Numeric Pain Rating Score [ Time Frame: 4 weeks ] [ Designated as safety issue: No ] Likert scale from 0‐10 measuring self‐reported level of low back pain.
Starting date	Nov. 2010; Estimated study completion date Nov. 2013
Contact information	Michael J Schneider, PhD, DC Tel. 412.383.6640, mjs5@pitt.edu Christine McFarland Tel. 412.623.6872, mcfarlandce@upmc.edu
Notes	http://clinicaltrials.gov/show/NCT01211613 SMR had contact with the PI in May 2011. At that time, 50 subjects had been recruited of which, 13 had acute LBP (<6 weeks), 20 had sub‐acute LBP (6‐12 weeks) and 17 had either acute or sub‐acute LBP (unclear).