AIM: Evidence of comparative effectiveness of different treatment approaches is important for clinical decision-making, yet absent for most recommended treatments of knee osteoarthritis pain. The objective of this study was to estimate the comparative effectiveness of exercise versus orally administered analgesics for pain in patients with knee osteoarthritis.

METHODS: The Cochrane Database of systematic reviews was searched for meta-analyses of randomized controlled studies comparing exercise or analgesics with a control group (placebo or usual care) and with pain as an outcome. Individual study estimates were identified and effect sizes were calculated from group differences. We combined study-level effects on pain with a random effects meta-analysis and compared effect sizes between exercise trials and trials with analgesic interventions.

RESULTS: We included six Cochrane reviews (four pharmacology, two exercise). From these, 54 trials were eligible (20 pharmacology, 34 exercise), with 9806 participants (5627 pharmacology, 4179 exercise). The pooled effect size of pharmacological pain interventions was 0.41 (95% CI: 0.23-0.59) and for exercise 0.46 standardized mean difference (95% CI: 0.34-0.59). There was no statistically significant difference between the two types of intervention (difference: 0.06 standardized mean difference [95% CI: -0.28-0.16; p = 0.61]).

CONCLUSION: This meta-epidemiological study provides indirect evidence that for knee osteoarthritis pain, the effects from exercise and from oral analgesics are comparable. These results may support shared decision-making where a patient for some reason is unable to exercise or who consider exercise as unviable and analgesics as a more feasible choice. PROSPERO registration: CRD42013006924.