“…In fact, although we can generally confirm previous results from meta-analysis published in 2011 25 , 2013 24 and 2014 23 , our meta-analytic approach has identified a different overall risk of AEs, provided specific data on the risk of SAEs, and reported a more selective characterization of AEs induced by roflumilast. The differences between our results and those of previous meta-analyses may be related with the current availability of results from recently published RCTs [40][41][42] , that together enrolled more that 2,500 COPD patients, and with the fact that our meta-analysis has been specifically designed for assessing the safety profile of roflumilast in COPD.…”