Clinical rule predicts patients likely to benefit from spinal manipulation

QuestionIs it possible to identify the low back pain patients who will respond to spinal manipulation?DesignRandomised controlled trial with pre-planned subgroup nalysis.SettingEight physical therapy clinics in USA.PatientsPatients aged 18–60, with a primary complaint of low back pain and an Oswestry Disability score of at least 30%. Exclusions were serious spinal pathology, nerve root compromise, pregnancy, and previous surgery. 543 patients were screened; 131 were eligible and were randomised using sealed envelopes to a manipulation group (70 patients) or an exercise group (61 patients).InterventionsPatients in both groups attended physiotherapy for 5 sessions over 3 weeks. The manipulation group received a high velocity thrust spinal manipulation during the first two sessions and then low stress aerobic and lumbar strengthening exercises. The exercise group received exercise alone. Two participants in the manipulation group and 9 in the exercise group discontinued treatment. An independent examiner assessed the patients and classified them as positive on the clinical prediction rule if they met 4 of the 5 following criteria: symptom duration < 16 days, no symptoms distal to knee, <19 on Fear Avoidance Beliefs Questionnaire, at least one hypomobile segment and at least one hip with > 35 degrees of internal rotation.OutcomesThe primary outcome was disability measured using the 0–100% Oswestry disability index, measured at baseline, 1 week, 4 weeks, and 6 months. Treatment success was defined as 50% reduction in disability. All participants completed the baseline assessment and 1 week follow-up, 130/131 the 4 week follow-up, and 92/131 the 6 month follow-up. Analysis was by intention to treat with pre-planned subgroup analysis.Results47/131 participants were positive to the rule. ANOVArevealed that the outcome depended upon the both the participant's treatment group and status on the rule. Pairwise disability mean (95%CI) differences at one week were: manipulation vs exercise 9.2 (4.4 to 14.1), manipulation (+ve on rule) vs manipulation (-ve on rule) 15.0 (8.5 to 21.5), manipulation (+ve on rule) vs exercise (+ve on rule) 20.4 (13.0 to 28.8) and exercise (+ve on rule) vs exercise (-ve on rule) -1.9 (-8.6 to 4.9). (+ve values signify greater improvement with the first named group in a pair). At 1 week 44% of the manipulation group had a successful outcome, however the success rate was 92% in the manipulation subgroup positive to the rule and only 7% in the subgroup who met less than 3 of the criteria.ConclusionPatients were more likely to benefit from spinal manipulation if they met the clinical prediction rule