Spinal Manipulation and Other Conservative Treatments for Acute and Chronic Low Back Pain Compared with Natural History and Sham: A Bayesian Network Meta-Analysis
Year: 2026
Type: preprint
Abstract: Background Patients with low back pain (LBP) often improve during treatment. But some of that improvement would have happened anyway (natural recovery). Some comes from the experience of receiving care (contextual response). Only the rest is due to the treatment itself. We evaluated spinal manipulative therapy (SMT) and common alternatives in two ways: against each other, and against an independent estimate of how LBP improves without treatment (natural history). Methods We conducted a Bayesian hierarchical network meta-analysis (NMA) of 133 retained analytic units. Acute and chronic LBP were analyzed separately. A companion paper (Paper I) estimated natural history for each phase from separate prognosis studies. We imported those estimates unchanged. No treatment-trial data were used to build them. They served as an external “what would have happened anyway” comparison. The NMA compared SMT, exercise, medical care, physical modalities, self-care, and sham. We kept two comparisons separate: treatment minus natural history, and treatment minus sham. Uncertainty from both analyses was carried through to the final comparisons. Results In acute LBP, natural history predicted that 75.2% of patients would recover by 26 weeks. SMT recovery was 63.2%. That is 12.0 fewer recoveries per 100 than natural history (95% interval −23.2 to −0.7) and 3.1 more per 100 than sham (−6.3 to +12.8). In chronic LBP, where recovery means a favorable course, natural-history recovery at 54 weeks was 18.1%. SMT recovery was 31.8%. That is 13.8 more recoveries per 100 than natural history (+4.3 to +22.6) and 13.4 more than sham (+4.3 to +21.9). Exercise was similar (+12.5 per 100 versus sham; +5.5 to +20.4). Chronic sham and natural history were nearly identical (+0.4 per 100; −4.2 to +5.7). Conclusions Whether an LBP treatment looks effective depends on its comparator. Our linked design keeps the prognosis evidence and the treatment evidence apart until the final comparison. In acute LBP, recovery rates were high, and no treatment added a detectable benefit beyond sham or beyond expected natural recovery. In chronic LBP, SMT and exercise showed some benefit beyond natural history and sham. Even so, only about one in three chronic patients was modeled to recover with either treatment.
Source: medRxiv
Institution: University of Arizona
Language: English
Field: Health Professions
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7214062278