GLP-1 vs. SGLT2: Which Cardiorenal Benefit Matters Most?

Two useful drug classes, no universal winner

If you have type 2 diabetes and risks that overlap—kidney disease, heart failure, obesity, a prior heart attack, or stroke—the question is rarely which drug class is "best." It is which outcome needs the most protection. Current evidence gives SGLT2 inhibitors the stronger track record for chronic kidney disease progression and heart failure. GLP-1 receptor agonists have the stronger comparative case for atherosclerotic outcomes, especially major adverse cardiovascular events (MACE) and stroke.

Using both classes may make clinical sense for selected people with several active risks. But the evidence supports a narrower claim than "the stack wins": benefits from one class appear to persist when the other is already being used. Randomized outcome trials have not yet directly shown that starting both drugs beats carefully selected, optimized monotherapy for every patient.

That distinction matters. "Not redundant" is useful evidence. "Proven synergy" would be a different and much bigger claim.

What the combination evidence actually shows

The best evidence comes from randomized placebo-controlled trials analyzed by background treatment status. In a 2024 SMART-C collaborative meta-analysis of 12 SGLT2 inhibitor trials, 3,065 of 73,238 participants (4.2%) were already taking a GLP-1 receptor agonist at baseline. SGLT2 inhibitors lowered chronic kidney disease progression with a hazard ratio of 0.65 among baseline GLP-1RA users and 0.67 among nonusers. The analysis found no heterogeneity signal for that difference.

The same analysis found a similar pattern for other outcomes. For MACE, the hazard ratio was 0.81 with background GLP-1RA use and 0.90 without it. For hospitalization for heart failure or cardiovascular death, the hazard ratios were 0.76 and 0.78, respectively. Serious adverse events were lower with SGLT2 inhibitors in both subgroups. These are randomized trial data, which gives the comparison within each trial strong protection against treatment-selection bias. Yet they remain subgroup data: they show that an SGLT2 inhibitor's usual signal did not disappear in people already using a GLP-1RA.

The reciprocal analysis reached the same practical conclusion. A 2024 meta-analysis of GLP-1RA cardiovascular outcome trials included 17,072 participants; 1,743 (10.2%) were taking an SGLT2 inhibitor at baseline across three trials. GLP-1RAs reduced MACE overall by 21% (hazard ratio 0.79). The estimate was 0.77 among participants already taking an SGLT2 inhibitor and 0.79 among those who were not. Hospitalization for heart failure also favored GLP-1RA treatment among background SGLT2 inhibitor users, with a hazard ratio of 0.58.

That analysis also reported favorable kidney composite outcomes and an overall improvement in estimated glomerular filtration rate (eGFR) slope of 0.78 mL/min/1.73 m² per year, without evidence that baseline SGLT2 inhibitor use changed those effects. It reported no heterogeneity in serious adverse events or severe hypoglycemia by background SGLT2 inhibitor use, although the published abstract did not provide exact subgroup estimates for those safety outcomes.

Adding the second class may preserve an outcome benefit. It does not prove that every person should take both classes.

The outcome map: kidney and heart failure versus atherosclerosis

Direct comparison data explain why sequencing should follow the patient's risk profile. In a Scandinavian new-user cohort study, investigators compared 87,525 SGLT2 inhibitor users with 63,921 GLP-1RA users. This observational design compared patients starting one class or the other; it did not randomly assign treatment.

GLP-1RAs were associated with slightly lower MACE risk than SGLT2 inhibitors, with a hazard ratio of 1.07 for SGLT2 inhibitors relative to GLP-1RAs, and lower stroke risk, with a corresponding hazard ratio of 1.14. Renal outcomes pointed the other way. Serious renal events favored SGLT2 inhibitors, with a hazard ratio of 0.76, and renal replacement therapy also favored SGLT2 inhibitors, with a hazard ratio of 0.74.

The study found no clear difference in cardiovascular death or all-cause mortality. Its heart-failure result was neutral in the intention-to-treat analysis, although the as-treated analysis showed a more favorable trend for SGLT2 inhibitors. Those details argue against a one-line rule such as "GLP-1 first" or "SGLT2 first." A class can lead on one endpoint without winning every endpoint that gets packed into a cardiometabolic headline.

Broader indirect comparisons generally reinforce this split. A 2023 network meta-analysis in people with chronic kidney disease and type 2 diabetes found stronger kidney protection with SGLT2 inhibitors than with GLP-1RAs. A 2024 American College of Physicians evidence review likewise concluded that newer glucose-lowering medicines do not have identical outcome profiles and that treatment should reflect the cardiorenal endpoint most relevant to the individual.

A newer network meta-analysis also placed SGLT2 inhibitors ahead on heart-failure and renal endpoints while GLP-1RAs compared well on MACE and atherosclerotic outcomes. Network meta-analysis is useful when head-to-head trials are scarce, but it is still indirect. Different trials enroll different populations, define endpoints differently, and come from different treatment eras. Rankings can therefore reflect study architecture as well as drug effects.

Why the biology fits, but does not settle the question

The pattern has a plausible physiologic explanation. Reviews link SGLT2 inhibitor kidney and heart-failure effects to natriuresis, hemodynamic unloading, and lower intraglomerular pressure. GLP-1RA benefits track more closely with weight loss, lower blood pressure, and anti-atherosclerotic biology. A 2025 meta-analysis of blood-pressure and weight mediation found that GLP-1RAs reduced MACE overall, with a pooled hazard ratio of 0.86, and examined the relationship between cardiovascular effects, blood pressure, and weight change.

Mechanistic fit can make a combination plausible because the classes act through partly different pathways. It cannot prove that the combined regimen produces additive protection against dialysis, hospitalization, stroke, heart attack, or death. Clinical outcome evidence still has to do that work.

What complicates the decision

GLP-1RAs deserve kidney credit. In a 2023 real-world comparison with basal insulin, GLP-1RA initiation was associated with slower kidney-function loss and less albuminuria progression in people with type 2 diabetes. That finding supports kidney benefit, but it does not overturn the broader pattern in which SGLT2 inhibitors have the stronger evidence for hard renal progression outcomes.

Observational comparisons can also change by population and analytic choices. The LEGEND-T2DM federated analysis and a Medicare-based comparison of empagliflozin and liraglutide in older adults add real-world context, but residual confounding remains possible. Clinicians may choose one class over another because of kidney function, frailty, body weight, heart failure, prior vascular events, or medication access. Even sophisticated matching cannot guarantee that these differences disappear.

The combination subgroup evidence has its own precision problem. Only 4.2% of SMART-C participants were taking a GLP-1RA at baseline. In the GLP-1RA meta-analysis, only 10.2% were taking an SGLT2 inhibitor, and just three trials contributed to that subgroup. These are meaningful data, but they are not the same as a large trial that randomizes patients with overlapping risk to combination therapy versus optimized single-class treatment.

Safety also stays individual. The subgroup analyses did not show a broad serious-adverse-event penalty from background use of the other class, but they do not fully quantify every tradeoff. SGLT2 inhibitors raise class-specific concerns such as genital infections, volume depletion, and euglycemic ketoacidosis. GLP-1RAs can cause gastrointestinal intolerance and raise questions about gallbladder events and other adverse effects in the appropriate clinical setting. Cost, access, kidney function, contraindications, and willingness to stay on treatment can matter as much as a modest relative-risk difference.

A practical verdict

For chronic kidney disease progression, albuminuria progression, renal replacement therapy, or heart-failure risk, SGLT2 inhibitors have the stronger overall case. For atherosclerotic risk—especially a profile dominated by prior stroke or MACE concern—GLP-1 receptor agonists have the stronger comparative case.

When one class is already in place and the patient still has a meaningful risk targeted by the other, current randomized subgroup evidence supports considering the second class with a clinician. It suggests the benefit can persist rather than cancel out. It does not establish routine combination treatment as the best next move for everyone. Blood-pressure control, lipid lowering, smoking cessation, kidney-protective therapy where indicated, and adherence still determine more than any medication-class rivalry.