A systematic review found extensive gaps in heart failure (HF) data capture in cardiovascular outcome trials of glucose-lowering therapies in patients with type 2 diabetes, according to results presented by Stephen J. Greene, MD, on Saturday, March 10, at ACC.18 in Orlando, FL. The study was simultaneously published in the Journal of the American College of Cardiology.

This systematic review evaluated the ascertainment of baseline HF, incident HF and reporting of HF-related clinical events within published clinical trials of glucose-lowering therapies. An extensive literature search was conducted of several databases from their inception through June 2017 to identify publications from large (n > 1,000) phase III or IV randomized clinical trials with primary clinical event endpoints evaluating glucose-lowering therapies in adults older than 18 years. The main outcome variables were the ascertainment and definitions of HF at baseline and during follow-up. Data were included on baseline HF assessment, incident HF during follow-up as well as resource utilization and outcomes.

After excluding 4,457 articles that did not meet eligibility criteria, 21 articles including 152,737 patients were selected for analysis. Baseline HF was reported in 67 percent of the studies and one excluded patients with HF. All but one trial providing baseline HF prevalence were published prior to 2010. Among trials with baseline HF documentation, one defined baseline HF.  Six trials tracked incident HF during follow-up. No trial reported ejection fraction (EF) data, natriuretic peptide level, or HF therapy among patients with baseline or incident HF. Baseline EF data were provided in three studies but not specific to patients with HF. Data on NYHA class were inconsistently reported. HF hospitalization data were reported in 15 trials. Only two studies included an HF-related event in the composite primary endpoint.

The authors highlighted the following findings from the review: limited characterization of baseline prevalence, severity or treatment optimization of HF; variable and low HF representation (<15 percent of enrolled sample); and description of new-onset HF in only 30 percent of trials. They concluded that even recent large cardiovascular outcome trials of novel glucose-lowering agents lack sufficient data to fully appraise treatment effects on an HF endpoint or relative safety in patients with prevalent HF.

They recommend enrollment of patients with baseline HF that is proportional to the HF prevalence in the general type 2 diabetes population in glucose-lowering trials and improved data collection to better inform HF prevention strategies, better define the safety profile of the drugs in the HF population, and better inform possible trials to examine the efficacy of these drugs as potential treatments for HF.

Keywords: ACC18, ACC Annual Scientific Session, Diabetes Mellitus, Type 2, Prevalence, Glucose, Stroke Volume, Follow-Up Studies, Natriuretic Peptides, Heart Failure, Hospitalization

< Back to Listings