OBJECTIVES: Explore potential sources of heterogeneity among estimates of tolterodine (TOL) and fesoterodine (FESO) efficacy relative to placebo (PBO) in patients with overactive bladder and urgency urinary incontinence (UUI) from randomized clinical trials (RCTs) published from 2001 – 2010. METHODS: RCTs evaluating TOL 4mg, solifenacin 5mg and/or 10mg, or FESO 4mg or 8mg compared to PBO reporting mean reduction of UUI episodes/d from baseline to endpoint were identified.  Treatment effects (treatment response minus PBO response) and PBO responses were tested for heterogeneity using Cochran’s Q statistic.  Where heterogeneity was present, other study variables (baseline UUI, baseline micturitions, gender, age, diary evaluation days, publication year, and study duration) were evaluated for potential confounding using linear regression methods. RESULTS: Statistical heterogeneity was found among the 17 PBO responses (mean reduction of UUI) of the included studies. PBO response increased with publication year, which accounted for more than 27% of response variation.  Publication year (p<0.02), gender (p<0.003), and study duration (6-week vs. other) (p<0.006) were significant predictors of PBO response (adj. R2=0.6261). TOL and FESO 8mg treatment effect estimates were also heterogeneous.  Among the nine TOL trials, treatment responses remained constant over publication year while PBO responses increased, resulting in a net decline in TOL treatment effect (p=0.0928).  The majority of this decline was explained by publication year and study duration (adj. R2=0.7703).  The four FESO 8mg UUI responses also displayed a publication year-dependent decrease leading to a decreasing treatment effect relative to PBO.  However, this trend was almost fully predictable by differential baseline UUI episodes (adj. R2=0.9721). CONCLUSIONS: Publication year, gender, 6-week duration, and baseline UUI were found to be significant predictors of PBO response or treatment effect.  Additional research should be done to understand why PBO response has increased over time and future trials should compare to real-world treatments wherever possible to demonstrate clinically-meaningful differences.