Review: Dipeptidyl peptidase-4 inhibitors do not increase overall adverse events in type 2 diabetes

A meta-analysis included 67 randomized controlled trials (RCTs) comparing dipeptidyl peptidase-4 (DPP-4) inhibitors to placebo or other oral antidiabetic agents. The rates of adverse events, serious adverse events, and mortality were no higher in patients taking DPP-4 inhibitors than in the overall group. However, compared with placebo, DPP-4 inhibitors increased risk for hypoglycemia when insulin or a sulphonylurea was coadministered.

The study was published in Diabetes, Obesity and Metabolism in December 2012. The following commentary by Saurav Chatterjee, MD, and Sanjay Chatterjee, MD, was published in the ACP Journal Club section of the April 16 Annals of Internal Medicine.

The meta-analysis by Gooßen and Gräber has important implications for clinical medicine as it indicates that the gliptins, currently widely used in clinical practice, do not increase risk for such infections as nasopharyngitis, upper respiratory infections, or urinary tract infections. It thereby refutes prior publications that have indicated such a possibility.

The statistically significant findings for specific types of adverse effects may be robust considering the large sample size of the review and longer follow-up compared with previously reported meta-analyses. However, weaknesses are similar to previous meta-analyses of gliptins in not adjusting for multiple testing and not stringently defining and specifying the outcomes evaluated, including cardiac and vascular disorders, and hypoglycemia; therefore, findings of increased vascular and cardiac disorders in patients treated with gliptins need to be interpreted cautiously. However, these data do provide some evidence for continuing gliptins in patients with such minor infections as upper respiratory infections, urinary tract infections, and skin infections, while the pragmatic practitioner will probably switch to insulin in patients with more severe infections.

This review could be used as a basis for generating the hypotheses for adequately powered RCTs to rigorously and conclusively corroborate or refute the signals noted in this meta-analysis.