We would like to thank Hernández-Vaquero et al. for their interest on our investigation. Indeed, as we acknowledged,1 the most important limitations of our study concern the synthesis of data from heterogeneous trials, which included patients who differed widely in terms of their ischemic and bleeding risk profiles. Indeed, acute coronary syndrome patients ranged from the total population in the REDUCE trial2 to far less than 50%, or complete exclusion of ST-segment elevation myocardial infarction patients.

Moreover, the definition of the study endpoints differed among the included trials, leading Verdoia et al.1 to consider mortality, rather than the composite of “major cardiovascular events” as the primary study endpoint. In contrast, bleeding definition was not consistent across the studies. BARC 2-5 bleeding events were used in 3 studies and BARC 3-5 events were considered in 1 trial, while the STOPDAPT-2 applied the more stringent thrombolysis in myocardial infarction criteria, potentially explaining the greater benefits observed in the present study, which considered only severe bleedings.

Figures 2 and 3 clearly demonstrate that the included studies consistently showed a similar trend for benefit in the reduction of bleeding events with shorter dual antiplatelet therapy, with those events being associated with larger heterogeneity. In contrast, an opposite increase or reduction of deaths was reported in the REDUCE1 and other trials, although resulting in an I2=36%, far lower that the threshold of 50% suggested by the Cochrane guidelines and reported by Hernandez-Vaquero et al.

As for publication bias, the same issues could certainly refer to the large number of meta-analyses that have appeared in the literature in the last few years, reaching similar conclusions to our own.

Therefore, while awaiting large scale dedicated randomized controlled trials, the possibility of pooling together the data from different studies, despite some potential limitations, should certainly be considered in order to broaden the spectrum of included patients and increase statistical power for clearly underpowered endpoints.