Individualised exercise in chronic non-specific low back pain: a systematic review with meta-analysis on the effects of exercise alone or in combination with psychological interventions on pain and disability
Objective To investigate the effects of individualised exercise interventions consisting with or without combined psychological intervention on pain intensity and disability in patients with chronic non-specific low back pain. Design Systematic review with meta-analysis and meta-regression. Data sources Five databases (PubMed, Cochrane Central, EMBASE, Clarivate Web of Science, and Google Scholar) were searched up to 31 March 2021. Selection criteria Randomised controlled trials were eligible if they included participants with chronic non-specific low back pain, compared at least one individualised/personalised/stratified exercise intervention with or without psychological treatment to any control / comparator group, and if they assessed at least pain intensity or disability as outcome measure. Results Fifty-eight studies (n = 9099 patients, 44.3 years, 56% female) compared individualised to other types of exercise (n = 44; 62%), usual care (n = 16; 23%), advice to stay active, or true controls. The remaining studies had passive controls. At short-term follow-up, low-certainty evidence for pain intensity (SMD -0.33 [95%CI -0.47 to -0.18]) and very low-certainty evidence for disability (-0.16 [-0.30 to -0.02]) indicates effects of individualised exercise compared to other exercises. Very low-certainty evidence for pain intensity (-0.35; [-0.53 to -0.17])) and low-certainty evidence for disability (-0.12; [-0.22 to -0.02]) indicates effects compared to passive controls. At long-term follow-up, moderate-certainty evidence for pain intensity (-0.14 [-0.23 to -0.06]) and disability (-0.23 [-0.33 to -0.12]) indicates effects compared to passive controls exercises. All findings stayed below the threshold for minimal clinically important difference (MCID). Certainty of evidence was downgraded mainly due to evidence of risk of bias, publication bias and inconsistency that could not be explained. Sensitivity analyses indicated that the effects on pain, but not on disability (always short-term and versus active treatments) were robust. Sub-group analysis of pain outcomes suggested that individualised exercise treatment is probably more effective in combination with psychological interventions (-0.32 [-0.51 to -0.14]), a clinically important difference Conclusion We found very low to moderate-certainty evidence that individualised exercise is effective for treatment of chronic non-specific low back pain. Individualised exercise seems superior to other active treatments and sub-group analysis suggests that some forms of individualised exercise (especially motor-control based treatments) combined with behavioural therapy interventions enhances the treatment effect. Certainty of evidence was higher for long-term follow-up. In summary, individualised exercise can be recommended from a clinical point of view.