Limited SR demonstrates SGLT2 inhibitors improve outcomes in adults with heart failure

Reference

Vaduganathan M, Docherty KF, Claggett BL, et al. SGLT-2 inhibitors in patients with heart failure: a comprehensive meta-analysis of five randomised controlled trials. Lancet 2022;400(10354):757-767.

Synopsis

These authors conducted a prespecified meta-analysis of randomized trials of sodium-glucose cotransporter 2 (SGLT2) inhibitors versus placebo in adults with heart failure. The trials had to have at least 1000 participants and report clinical outcomes. To be clear, this was not a systematic review of all trials of SGLT2 inhibitors in patients with heart failure. The authors included 5 trials of patients with reduced ejection fraction (DAPA-HF, EMPEROR-Reduced), preserved or mildly reduced ejection fraction (DELIVER, EMPEROR-Preserved), and recent worsening heart failure (SOLOIST-WHF). The trials, all at low risk of bias, included mostly ambulatory patients, but also included some hospitalized patients who were followed up for a median of 9 months. The follow-up for the nonhospitalized patients ranged from 16 months to 28 months. Since the trials used variations of similar endpoints, the authors report performing “data harmonization” across the studies, but only did that for potential benefits and not for harms. In the studies that included adults with either mildly reduced or preserved ejection fraction, patients taking SGLT2 inhibitors had fewer cardiovascular deaths (6.8% vs 7.7%; number needed to treat [NNT] = 107; 95% CI 54 - 6115) and hospitalizations for heart failure (9.6% vs 12.6%; NNT = 34; 25-54), but no difference in all-cause mortality (15% vs.15.6%). In the studies that enrolled adults with reduced ejection fraction, those taking SGLT2 inhibitors had fewer cardiovascular deaths (9.8% vs 11.2%; NNT = 70; 37 - 773), fewer hospitalizations (11.3% vs 15.6%; NNT = 24; 18 - 35), and fewer deaths from all causes (12.4% vs 14.0%; NNT = 61; 33 - 492). There was no heterogeneity in these data. Among hospitalized patients, there was no difference in deaths from cardiovascular causes or from all causes. The authors report that outcomes were improved across 14 different subgroups. The authors did not make any statistical adjustments for doing so many analyses. The authors note differences in definitions used for harms (amputation, diabetic ketoacidosis, hypoglycemia, renal events, and so forth), but report that only 2 trials reported harms and that the proportion of serious adverse events was similar (44% - 50%) for each treatment group. Finally, although this analysis had no funding, every one of its authors has financial ties to industry. This doesn’t invalidate the work, but it raises suspicion about some choices they made (eg, including only mega-trials, using data harmonization only for benefits and not for harms, etc.).