Weight loss not significantly linked to incretin heart benefit in meta-regression

The study

This is a systematic review with aggregate-data meta-regression, published in Diabetes, Obesity & Metabolism. The authors searched MEDLINE, EMBASE, CENTRAL and ClinicalTrials.gov for placebo-controlled randomised trials of incretin mimetic therapy. Trials had to report cardiovascular hazard ratios (HRs) and weight changes.

The main question was whether greater weight loss is associated with larger reductions in three-point major adverse cardiovascular events (3P-MACE), individual cardiovascular outcomes and mortality. The moderator was the placebo-adjusted percentage weight difference at the latest reported follow-up. A second aim was to estimate the overall cardiovascular effect of incretin mimetics against placebo, using standard meta-analysis.

Sixteen trials with 97 199 participants, with and without diabetes, were analysed.

What they found

Meta-regression showed no statistically significant association between greater weight loss and treatment HRs. This applied to 3P-MACE, cardiovascular mortality, all-cause mortality, myocardial infarction, stroke and hospitalisation for heart failure (HHF).

The standard meta-analysis showed significant reductions with incretin therapy compared with placebo:

• 3P-MACE: reduced by 15%
• Cardiovascular mortality: reduced by 13%
• All-cause mortality: reduced by 12%
• Myocardial infarction: reduced by 14%
• Stroke: reduced by 13%
• HHF: reduced by 10%

The abstract does not report confidence intervals, which drugs were included, the length of follow-up, heterogeneity between trials, risk of bias or funding.

How much weight to give it

The pooled benefit comes from randomised, placebo-controlled trials with a large total number of participants. The abstract gives few details to check it.

The meta-regression is different. It compares trials with each other, so it shows association at trial level only. The authors state that the findings do not establish an individual-level or causal relationship. They also state that the findings do not exclude an individual-level contribution of weight loss to cardiovascular benefit.

A result that is not statistically significant is not proof of no relationship. The abstract does not report how precise the estimates were, so we cannot tell how large an association could have been missed. The trials included participants with and without diabetes, and the abstract does not report heterogeneity or design details. The authors used only the latest reported weight difference as the moderator.

The result should therefore not be read as showing that weight loss is unimportant for heart outcomes. It shows that, at trial level, the amount of weight loss was not significantly associated with differences in treatment HRs.

For your clinic

The overall message is that incretin therapy was associated with significant reductions in cardiovascular and mortality outcomes compared with placebo in these trials. This analysis does not give a basis to predict an individual patient's heart benefit from the weight they lose.

The trial-level findings do not establish causality. They also do not exclude an individual-level contribution of weight loss to cardiovascular benefit.

Before changing practice, read the full paper. Check which drugs and populations were included, how precise the estimates were, and whether the trials differed in important ways. The authors state that individual-participant-level analyses are needed to clarify the relationship between weight loss and cardiovascular outcomes.