A meta-analysis of 20 randomized trials and 83,004 patients found that GLP-1 receptor agonists cut composite renal events by 20 percent, the biggest effect in the bundle. Major adverse cardiovascular events (heart attacks, strokes, and cardiovascular deaths combined) fell 13 percent. Heart failure hospitalizations and coronary revascularizations did not reach statistical significance.
The paper, published online May 16 in the European Heart Journal Quality of Care and Clinical Outcomes ↗, pools the dedicated cardiovascular outcome trials that anchor every guideline recommendation the GLP-1 class currently carries. The conclusion the authors land on is the one that is now in every prescriber's pocket: "GLP-1RAs are associated with improved cardiovascular outcomes." That conclusion is correct. It also obscures the actual shape of the data.
Renal beats cardiac
The composite renal endpoint (some combination of macroalbuminuria, kidney function decline, end-stage kidney disease, and renal death across the 20 trials) carried a relative risk of 0.80 (95% CI 0.73 to 0.88). That is the largest effect size in the meta-analysis. Major adverse cardiovascular events, the headline endpoint for every dedicated cardiovascular outcome trial in this class, came in at RR 0.87 (95% CI 0.83 to 0.92). All-cause death, cardiovascular death, myocardial infarction, and stroke all clustered between RR 0.87 and 0.89.
In plain terms: across this pooled population of T2DM patients, the class prevents roughly twenty kidney events per hundred baseline kidney events, versus roughly thirteen heart attacks or cardiovascular deaths per hundred baseline cardiac events. The renal effect is about half again as large as the cardiac one.
This is not how the class was originally framed. LEADER (liraglutide, 2016) and SUSTAIN-6 (semaglutide, 2016) were designed and powered for MACE. Renal outcomes were secondary endpoints. The renal signal kept getting bigger across successive trials, FLOW (semaglutide in chronic kidney disease) eventually showed enough kidney benefit to land a dedicated label, and FDA approved semaglutide for kidney disease in T2DM in early 2025. The new meta-analysis confirms the renal effect at a larger population scale than any single trial has reached. It also makes the actual size of the cardiac effect harder to overstate: real, but smaller than the framing suggests.
Heart failure did not pass
This is the genuinely uncomfortable finding. Heart failure hospitalizations came in at RR 0.93 (95% CI 0.85 to 1.01). Coronary revascularization came in at RR 0.87 (95% CI 0.74 to 1.01). The point estimates favor GLP-1 receptor agonists. Both upper confidence bounds touch 1.0, the line where there is no effect.
The reason this matters: the field has been treating GLP-1 receptor agonists as heart-failure-favorable based on STEP-HFpEF (semaglutide in HFpEF plus obesity, 2023) and a handful of smaller dedicated readouts. Those trials selected specific patient populations (HFpEF, obesity) and showed real benefit in symptoms, biomarkers, and exercise tolerance. The pooled MACE-trial population is different. It is T2DM patients enrolled for cardiovascular outcome trials, not for heart failure. In that broader population, across 83,004 patients, the heart failure signal does not pass the conventional bar.
What this says, mechanistically, is plausible. The class may benefit HFpEF through weight loss and visceral adiposity reduction, not through a direct cardiac mechanism. In a T2DM cardiovascular-trial population that is not enriched for heart failure, the weight effect is smaller and the HFpEF benefit dilutes. The class story is still "cardiorenal," but the cardio part is more renal-than-cardiac, and the heart-failure part is more population-specific than class-wide.
Platform note
The 20 RCTs in the meta-analysis cover semaglutide ↗, liraglutide ↗, dulaglutide ↗, tirzepatide ↗, exenatide ↗, and lixisenatide ↗, plus albiglutide (withdrawn in 2018). The GLP-1R target page ↗ shows the full lineage. What the meta does not break out, and what per-agent dedicated trials are now starting to test, is whether the renal effect is class-wide or concentrated in one or two agents. FLOW was semaglutide-only. If the renal effect concentrates rather than distributing evenly, the cardiorenal label will need to start splitting per agent.
The other piece of the table that is not in this paper but sits next to it is the cancer story. Yesterday's coverage of the GLP-1 thyroid-cancer landmark analysis ↗ found the thyroid harm signal that anchors the boxed warning did not survive landmark analysis, while non-endocrine cancer protection held under active-comparator scrutiny. The prescriber-consent conversation for this class has shifted in 48 hours, in both directions.