Authors’ reply

We would like to thank Patoulias and colleagues for their observations, highlighting a possible association between proton pump inhibitors (PPIs) use and diabetes mellitus (DM). Besides the meta-analysis by Yuan et al,1 including 3 cohorts of health professionals showing that regular PPIs use was likely to be associated with an increased risk of type 2 DM compared with controls, the evidence on such association appears contradictory and inconclusive to date.

Although similar changes in the microbiota phenotype have been reported between PPI users and patients affected by DM, suggesting that gut microbiota may mediate this association, the exact mechanisms have not yet been clarified. On the contrary, it has been suggested that PPIs could be effective as antidiabetic agents, since they increase serum gastrin concentrations, whereby affecting glucose metabolism through promoting β-cell regeneration and also enhancing insulin secretion.2,3 Indeed, a meta-analysis on this topic found no significant effect on glucose metabolism and insulin resistance in patients with type 2 DM treated with PPIs4 and population-based cohort studies demonstrated a decreased risk of DM in patients with upper gastrointestinal disease taking PPIs.5

Nevertheless, we agree with Patoulias and colleagues that clinicians should prescribe PPIs with caution to those affected by DM, as they represent patients at a high risk for developing cardiovascular complications and, as reported in our review, long-term PPIs use is associated with an increased risk of cardiovascular events.6 Therefore, we also agree that further research is warranted to investigate the association, if any, between PPIs and DM, to clarify possible underlying mechanisms and their clinical implications.