Journal of the New Zealand Medical Association, 08-August-2003, Vol 116 No 1179
Chiropractic manipulation for non-spinal pain – a systematic review
Joint manipulation is frequently used by chiropractors, osteopaths, physicians, physiotherapists and other healthcare professionals to treat a wide range of conditions. Even though the domain of chiropractic manipulation is spinal pain, it is frequently also used for non-spinal syndromes. Recent survey data from the US and Canada, for instance, show that 6% of all patients seen by chiropractors have non-musculoskeletal problems.1 An Australian survey demonstrated that more than half of the responding chiropractors favoured the role of spinal adjustments in the management of patients with visceral conditions.2
The American Chiropractic Association defines manipulation as:
‘...a passive manual manoeuvre during which the three-joint complex is carried beyond the normal physiological range of movement without exceeding the boundaries of anatomical integrity. The essential characteristic is a thrust – a brief, sudden, and carefully administered ‘impulsion’ that is given at the end of the normal passive range of movement. The ‘dynamic thrust’ is the defining factor, which distinguishes manipulation from other forms of manual therapy. The thrust technique can be low or high velocity. The most common characteristics of the adjustive dynamic thrust are a controlled force delivered with high velocity, in a specific direction or line of drive, at a regulated magnitude and depth. In short, manipulation is a passive dynamic thrust that causes an audible release (cavitation) and attempts to increase the manipulated joint’s range of motion.’3
The one-year prevalence figures for manipulation in representative population samples range between 10% (1988, Austria) and 33% (1996, UK).4 In the US, the proportion of the general population using chiropractic has doubled during the last 20 years.1 At present, there are about 50 000 chiropractors in active practice in North America, and their number has tripled between 1970 and 1990.1 It has been estimated that by 2010 their number will have doubled again.5
Several authors have reviewed the evidence for spinal manipulation or mobilisation as a treatment of spinal pain.6–8 This paper aims at critically evaluating the evidence for or against the effectiveness of chiropractic manipulation as a treatment of non-spinal conditions.
The following databases were searched, each from their inception to February 2003: Medline, Embase, CISCOM, Amed and The Cochrane Library. Furthermore, other experts were consulted. The keywords used were: chiropractic, spinal manipulation, spinal adjustments, controlled clinical trials, manual therapy, pain. The bibliographies of all articles thus located and major chiropractic texts were screened for further relevant papers. No language restrictions were applied.
Trials of spinal manipulation for treatment of headache/migraine have recently been submitted to a systematic review9 and were thus excluded. Non-randomised studies, trials of disease prevention, studies of conditions not related to pain management,10,11 clinical trials of mobilisation (as distinct from manipulation), and studies in which the therapists were not chiropractors were also excluded.9,12–15 The last exclusion was deemed appropriate because the chiropractic approach differs profoundly from that of other manual therapists (‘The ‘dynamic thrust’ is the defining factor, which distinguishes manipulation from other forms of manual therapy.’)3 While manipulation usually involves high-velocity thrusts (see above), mobilisation ‘includes any manual therapy directed at joint dysfunction that does not involve a high velocity thrust’.16
All trials meeting the above-mentioned criteria were read in full. Information on trial methodology, patient population, treatment schedule, outcome measures, follow up and results was validated and extracted by the current author in a standardised way (Table 1, click here to view). Methodological quality was assessed by two independent evaluators using the Jadad score,17 which ranges between a minimum of 0 and a maximum of 5. It is a validated measure for methodological quality of clinical trials based on the absence or presence of design features such as randomisation, double blinding, description of withdrawals or dropouts, etc. Statistical pooling (ie, meta-analysis) of the data was anticipated but turned out to be impossible, mainly because of the heterogeneity of the primary studies.
Eight publications met the inclusion criteria.18–25 Their methodological quality ranged from poor19 to excellent.22,24,25 Key data from all studies are summarised in Table 1 (click here to view).
Kokjohn et al conducted a randomised controlled trial (RCT) with 45 women suffering from primary dysmenorrhoea.18 They were treated either with high-velocity, short-lever, low-amplitude thrusts to all clinically relevant vertebral levels or with a sham intervention consisting of thrusts at an irrelevant level. The study was aimed at determining acute effects only and thus only one treatment session took place. Abdominal pain was measured with a visual analogue scale and menstrual distress with the Menstrual Distress Questionnaire. In addition, levels of prostaglandin metabolites were quantified in the peripheral blood. The results show a remarkable drop of prostaglandin levels in both groups. Inter-group comparisons show that pain and distress were alleviated significantly more in the experimental compared with the control group.
Blunt et al conducted a cross-over study including 21 patients with fibromyalgia.19 These patients were treated for four weeks with spinal manipulation, soft-tissue manipulations and stretching. During the control phase, no such treatment was applied. Medication was similar in both phases and was continued unchanged throughout. The authors noted significant improvements in pain and range of motion when comparing pre-manipulation and post-manipulation measurements. No inter-group comparisons were made and the study design allowed no control of placebo effects. These and other methodological weaknesses render this pilot study uninterpretable in terms of therapeutic efficacy.
Davis and colleagues compared oral ibuprofen at decreasing doses (800 mg three times a day to twice daily) with a series of manipulations of the cervical spine and the upper extremities as a treatment for carpal tunnel syndrome.20 The study included 91 patients and was evaluator blind. There were considerable improvements in both groups but no statistically significant difference between them. The authors nevertheless concluded that carpal tunnel syndrome may be treated with conservative medical or chiropractic care.
Wiberg et al randomised 50 babies with infantile colic to receive either oral medication (as licensed in Denmark) or three to five sessions of spinal manipulation ‘with specific light pressure with the fingertips’.21 The main outcome measure, time of crying per day, improved significantly more in the experimental compared with the control group. It should be noted, however, that the former group was significantly younger than the latter.
Sawyer et al published a pilot study of spinal manipulation versus sham spinal manipulation for children with otitis media.22 The sham intervention involved manual handling of the spine for diagnostic purposes without the high-velocity thrust performed in the experimental group. The authors do not reveal any results that suggest real spinal manipulation to be superior to sham treatment. They do, however, state that controlled trials of sham spinal manipulation are feasible.
Hondras and colleagues randomised 138 patients with dysmenorrhoea in two groups.23 The experimental group received spinal manipulation in the form of high-velocity, short-lever, low-amplitude thrusts with a force greater than 750 N delivered at all clinically relevant vertebral levels and sacroiliac joints bilaterally. The control group received thrusts with a force of around 200 N delivered to the left L2/3 segment, which was deemed irrelevant for the condition in question. Treatments were administered on Day 1 of cycles 2, 3, and 4, and on Day 7 before cycles 3 and 4. The primary outcome measure was pain measured with a visual analogue scale. The results did not yield significant inter-group differences between real and sham treatments.
Olafsdottir and colleagues conducted an RCT testing spinal manipulation versus a simple sham procedure for infants suffering from typical infantile colic pain.24 The primary outcome measure was a verbal rating scale by parents. The results show some improvements in both groups but do not reveal significant differences between real and sham therapy. The authors make the following interesting point: ‘This study emphasises the need for placebo-controlled and blinded studies when investigating alternative methods to treat unpredictable conditions such as infantile colic.’
Hawk et al recruited 39 women with gynaecological pelvic pain and treated them with a series of either spinal adjustments or sham adjustments.25 The trial was set up as a multicentre pilot study, and the authors did not formally evaluate the results. They did, however, comment that ‘patients in both groups were satisfied with their care and blinding appeared to be successful. Pain Disability Index change scores were not consistent across sites’.
Perhaps the most important result of this systematic review is the fact that very few RCTs exist in this area. Four of the eight RCTs had pilot character.18,19,22,25 The range of conditions is large and seems arbitrary. For any single condition a maximum of two RCTs only are available. This paucity of data limits the conclusiveness of this systematic review but does in itself represent an important finding, not least in relation to the widespread use of chiropractic manipulation for non-spinal pain.1,2
Chiropractic manipulation has been tested in conditions as diverse as fibromyalgia,19 carpal tunnel syndrome,26,27 infantile colic,21,24 otitis media,22 dysmenorrhoea,23 and chronic pelvic pain.25 There are a lack of independent replications, and multiple weaknesses in the trial designs have to be noted. They include small sample size, lack of follow up, lack of control for placebo response, lack of blinding and use of non-validated outcome measures. Where independent replications have been published they failed to confirm the initially encouraging findings.18,21,23,24 None of the high-quality trials (ie, those with a Jadad score of 5) yielded a conclusively positive result. The only study that generated positive results (ie, suggesting the superiority of manipulation over medical treatment of infantile colics) suffered from significant methodological limitations, eg, no control of placebo effects, no blinding, no validated outcome measure.21 In this study, the observed effect is of questionable clinical relevance; it scored only 3 of 5 possible points on the Jadad score. Most importantly, a larger and more rigorous study (Jadad score = 5) of the same indication failed to confirm the positive result.24
The combination of paucity and often low-quality primary data seriously limits the validity of the findings and the ability to generalise them. Moreover, one cannot be sure that all relevant RCTs have been identified, and the influence of publication bias is difficult to assess. In this context, it is noteworthy that no unpublished trials were identified with the above-mentioned search strategy. It could be argued that pilot studies should be omitted from this review. On the other hand, one might stress that it is the very purpose of systematic reviews to summarise evidence from even the smallest and most preliminary trials as long as they meet the pre-defined entry criteria.
The above evidence does not demonstrate the ineffectiveness of chiropractic manipulation for non-spinal syndromes. The evidence does, however, show that any claims that chiropractic manipulation might be effective in the treatment of non-spinal syndromes are not based on data from well-designed clinical trials. Critics of chiropractic theory would add that the notion of a benefit from chiropractic manipulation for non-spinal problems is scientifically implausible; in other words, there is no compelling rationale why manual adjustment of spinal malalignment should reduce non-spinal symptoms. A further point to bear in mind is the fact that chiropractic manipulation has repeatedly been associated with serious complications;28 this has obvious implications for a risk-benefit analysis and is a further important reason for segregating chiropractic trials for this review. The risks of spinal manipulation can be considerable,29 while the benefits for non-spinal conditions have yet to be demonstrated. It follows that an analysis of risk versus benefit has to yield a negative result.
Others might argue that the popularity of chiropractic amounts to proof of its efficacy. The frequency of chiropractic visits by the general population has been evaluated at 100 visits per 100 person-years in the US and 140 visits per 100 person-years in the US and Canada.30 Six per cent of these visits are likely to be due to non-musculoskeletal problems.1 Interestingly, patients with such complaints are more satisfied than other chiropractic patients.1 The open question is whether this level of satisfaction is due to specific or non-specific therapeutic effects.26
An important finding of this review is that high-quality studies (with a Jadad score of 5) of manipulation are feasible. This is not to imply that the issue of placebos in controlled clinical trials of chiropractic has been fully resolved. In fact, one aim of this review is to stimulate further research into this and related areas. Rigorous RCTs should now be conducted by those who make (and profit from) claims of efficacy. These RCTs should be conducted patient blind and evaluator blind against a sham control intervention in order to allow for placebo effects that have repeatedly been demonstrated.26 Obviously, they should be of sufficient sample size (ideally based on a proper sample-size calculation with sufficient power avoiding type I and II errors) and employ validated outcome measures of clinical effectiveness.
In conclusion, the notion that chiropractic manipulation is an effective treatment of non-spinal pain syndromes is not based on conclusive evidence. Those who make claims of effectiveness should provide the evidence to back them up.
Author information: Edzard Ernst, Professor of Complementary Medicine, Complementary Medicine, Peninsula Medical School, Universities of Exeter and Plymouth, Exeter, UK
Correspondence: Professor Edzard Ernst, Complementary Medicine, Peninsula Medical School, Universities of Exeter and Plymouth, 25 Victoria Park Road, Exeter EX2 4NT, UK. Fax: +44 1392 424989; email: Edzard.Ernst@pms.ac.uk
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads)
Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals