Background: For dichotomous outcomes, inadequate or unclear concealment of allocation is associated with an overestimation of treatment effects when subjective outcomes are assessed. Objectives: We determined whether inadequate or unclear concealment is associated with biased estimates of treatment effects on pain intensity in osteoarthritis trials. Methods: We searched for meta-analyses of at least two randomized controlled trials comparing an intervention with placebo or no intervention in patients with osteoarthritis. We calculated effect sizes (ES) from between-group differences in means of pain intensity divided by the pooled standard deviation. Within each meta-analysis, we calculated the difference in ES between trials with adequate and trials with inadequate or unclear concealment, and calculated a weighted average across all meta-analyses using a random-effects model. Then, we determined the impact of restricting random-effects meta-analyses to trials with adequate concealment. Results: Fourteen meta-analyses including 163 trials and 40,436 patients contributed. Forty-eight trials reported adequate concealment (29%), for 115 trials concealment was unclear (71%). Trials with unclear concealment tended to be more beneficial than trials with adequate analyses (difference in ES, -0.16, 95%-CI -0.33 to 0.02). The association was evident in meta-analyses with a large degree of between-trial heterogeneity (difference in ES, -0.40, 95%-CI -0.72 to -0.07), but not in meta-analyses with small to moderate heterogeneity (0.00, 95%-CI -0.10 to 0.10, p-value for interaction, 0.042). Treatment effects became less beneficial in nine meta-analyses and more beneficial in five after restriction to trials with adequate concealment (range 0.71 ES less to 0.32 ES more beneficial). Results became less heterogeneous in nine meta-analyses and more heterogeneous in five. In four meta-analyses, results lost statistical significance at the 5% level. Conclusions: In osteoarthritis trials using continuous outcomes, there is evidence for selection bias associated with unclear concealment of allocation. The association of estimates of treatment effects with adequacy of concealment of allocation was only found in meta-analyses with a high degree of between-trial heterogeneity.