https://diabetes.acponline.org/archives/2017/02/10/6.htm

Review: In type 2 diabetes, sodium–glucose cotransporter 2 inhibitors do not increase major CV events or mortality

The meta-analysis included 37 trials that compared canagliflozin, dapagliflozin, or empagliflozin to placebo or other active antidiabetic treatments.


A meta-analysis found no increase in risk of mortality or adverse cardiovascular (CV) events from sodium–glucose cotransporter (SGLT) 2 inhibitors. The analysis included 37 trials with 29,859 patients that compared canagliflozin, dapagliflozin, or empagliflozin to placebo or other active antidiabetic treatments. Compared to placebo, only empagliflozin showed a significant association with the studied outcomes—a lower risk of all-cause mortality and major adverse CV events, with the effect largely driven by the EMPA-REG OUTCOME trial.

The study was published in the Dec. 1, 2016, American Journal of Cardiology. The following commentary by James Brophy, MD, PhD, was published in the ACP Journal Club section of the Jan. 17 Annals of Internal Medicine.

CV disease is the leading cause of death and complications in type 2 diabetes mellitus. Because of previous controversies, regulatory authorities now require CV safety of new therapies for type 2 diabetes to be established. Tang and colleagues performed a systematic random-effects network meta-analysis assessing the benefits and risks of SGLT2 inhibitors for CV outcomes. They conclude that the 3 common SGLT2 inhibitors are not associated with increased all-cause mortality or CV outcomes and that empagliflozin, largely driven by a single trial, may have a protective effect.

An intriguing question is whether evidence is better interpreted by assessing each study independently or by also examining indirect comparisons. The development of rigorous standards for network meta-analyses allows integration of more evidence and arguably leads to improved evidence synthesis and decision making. Tang and colleagues followed recommended guidelines and performed analyses to ensure transitivity (comparable groups) and consistency between direct and indirect comparisons, and assessed individual study bias, thereby enhancing the validity of their analysis. However, their conclusion, “Neither dapagliflozin nor canagliflozin was significantly associated with any harm”, is not supported by the data. The upper 95% CI for mortality is >2 for both drugs, meaning that the studies are underpowered to rule out a potentially important doubling of mortality. In addition, the mortality benefit with empagliflozin was almost entirely driven by 1 trial in patients with established coronary artery disease, and it would be inappropriate to extrapolate these results to primary prevention. As Tang and colleagues point out, more definitive conclusions will require the completion of ongoing confirmatory studies.