OBJECTIVES: Quantitative analysis of spontaneous adverse drug reaction reports is increasingly used in drug safety research. Signals are detected by disproportionality measures (DM). Different types of DMs are available: a debate is ongoing on which performs better. The aim was to evaluate the sensitivity on identification of signals with known safety profiles of incretin drugs. METHODS: Adverse Events (AE) reported to FDA Adverse Event Reporting System (FAERS) between 2005 and 2014 were included. To evaluate the impact of warning actions, two separated analysis were conducted: restricted to the time before and after a regulatory action. We selected 20 AEs for each drug, half as “positive controls” (with a known causal association from literature) and half as “negative controls” (with no evidence in literature). Proportional Reporting Ratio (PRR), Reporting Odds Ratio (ROR), Bayesian Confidence Propagation Neural Network Analysis (BCPNN) and Gamma-Poisson Shrinker (GPS) were calculated on 120 combinations of AE-drugs. Correlation between sensitivity and number of AEs per year reported for each drug was calculated to evaluate the effect of number of reports on sensitivity. RESULTS: The number of reports analyzed in 2014 was 1,934,607. After warning action, PRR showed a sensitivity of 0.42 (0.29; 0.55), ROR 0.55 (0.42; 0.68), BCPNN 0.53 (0.40; 0.66) and GPS 0.23 (0.13; 0.36). Analog findings were observed before warning actions. The concordance of signals identification was good for all pairwise comparison between DMs (>0.56). The correlation varied among 0.49 (for PRR) and 0.82 (for ROR and BCPNN) after warning action. CONCLUSIONS:  The sensitivity of measures was low (<0.6), without impact of warning actions. ROR and BCPNN showed the most elevated values of sensitivity not allowing to determine a clear superiority of neither frequentist nor bayesian DMs. As expected, the positive correlation suggests the presence of a strong impact on sensitivity of higher number of AEs reported.