SGLT2 vs GLP-1 Heart Protection Changes With Age

Ozempic is everywhere, but its cardiovascular effect may not look the same at 65 as it does at 40. A 2025 network meta-analysis found that age changed the relative cardiovascular effects of two leading type 2 diabetes drug classes in opposite directions: SGLT2 inhibitors became relatively more protective against major cardiovascular events with increasing age, while GLP-1 receptor agonists became relatively less protective.

The finding does not mean older adults should automatically stop a GLP-1 drug or that younger adults should avoid an SGLT2 inhibitor. Both classes can reduce cardiovascular risk in appropriate patients. It does mean HbA1c reduction, weight loss, and a drug class's general reputation do not answer every treatment question. Age may alter the balance.

Age moved heart and glucose effects in opposite directions

Hanlon et al.'s 2025 JAMA network meta-analysis combined 23 cardiovascular outcome trials, including 168,489 participants, with 592 HbA1c trials involving 309,503 participants. The cardiovascular-trial population had a mean age of 64.0 years. The authors used network meta-regression to test whether age and sex modified treatment effects.

For SGLT2 inhibitors, each 30-year increase in age was associated with a major adverse cardiovascular event, or MACE, hazard-ratio interaction of 0.76 (95% credible interval 0.62 to 0.93). Because the interval excludes 1.0, the analysis found statistically significant evidence that relative cardiovascular protection increased with age.

Yet SGLT2 inhibitors lowered HbA1c less effectively in older patients. Compared with younger patients, the age interaction corresponded to about 0.24 percentage points less HbA1c reduction with monotherapy, 0.17 points less with dual therapy, and 0.25 points less with triple therapy.

GLP-1 receptor agonists showed the reverse pattern. Their glucose-lowering effect increased with age: about 0.18 percentage points more HbA1c reduction with monotherapy and 0.24 points more with dual therapy per 30-year age increment. But their relative MACE benefit attenuated with age, with an interaction hazard ratio of 1.47 (95% credible interval 1.07 to 2.02).

These interaction estimates compare how treatment effects change across age; they do not say that either class causes more cardiovascular events at a particular age. They also do not establish a birthday at which one drug should replace the other. Still, the direction of the effects matters. In older adults, the drug that lowers HbA1c more may not be the drug with the greater relative cardiovascular advantage.

HbA1c remains useful for managing glycemia, but it cannot stand in for cardiovascular protection—especially when drug effects separate as patients age.

Why SGLT2 inhibitors might gain relative advantage with age

The result fits a plausible, but unproven, biological explanation. SGLT2 inhibitors promote natriuresis and osmotic diuresis, which can reduce preload and afterload. They may also affect myocardial energy use through ketone bodies. Those pathways differ from glucose lowering and may matter more in older cardiovascular systems, where ventricular stiffness, impaired diastolic filling, and vulnerability to volume overload become more common.

That is a hypothesis, not a demonstrated mechanism. The meta-analysis did not randomize people to be older or younger, and it did not include prospective age-stratified mechanistic trials. A 2021 systematic review by Karagiannis and colleagues nevertheless reached a compatible broad conclusion in older people with type 2 diabetes: both SGLT2 inhibitors and GLP-1 receptor agonists were safe and effective, while SGLT2 cardiovascular effects appeared disproportionate to their glycemic effects.

A large real-world comparison also supports the idea that SGLT2 cardiovascular benefits do not simply track glucose lowering. In a 2023 JAMA Internal Medicine comparative-effectiveness study of 144,614 commercially insured adults with type 2 diabetes, SGLT2 inhibitors retained cardiovascular advantages over DPP-4 inhibitors across baseline HbA1c strata. Observational evidence cannot prove the mechanism, but it weakens the simple claim that lower glucose alone explains SGLT2 cardiovascular benefit.

Less relative benefit is not zero benefit

The most important guardrail is absolute risk. The JAMA analysis reported relative age-interaction terms, not absolute MACE rates for specific age bands. A smaller relative risk reduction can still prevent many events when baseline cardiovascular risk is high, which is often true in older adults.

A systematic review and meta-analysis by Diallo et al. makes this point concrete. In pooled HARMONY and EXSCEL data, 2,390 participants older than 75 who received GLP-1 receptor agonists had a 22% lower MACE hazard than comparators (hazard ratio 0.78, 95% confidence interval 0.63 to 0.95). That result was statistically significant. It does not contradict Hanlon et al.; it shows that attenuation relative to younger patients does not mean GLP-1 cardiovascular benefit disappears in later life.

The caveat is precision and generalizability. The over-75 result came from only two trials and involved albiglutide and once-weekly exenatide. It cannot settle whether modern, higher-potency agents such as semaglutide or dulaglutide deliver the same effect in every very old population. Diallo et al. also found greater SGLT2 MACE benefit in people with diabetes for more than 10 years, a characteristic correlated with age but not identical to it.

Obesity and disease profile can change the equation

Age is one modifier among several. A 2025 meta-analysis by Zhang et al. found that GLP-1 receptor agonists and tirzepatide reduced heart-failure risk in obese, but not non-obese, people with cardiovascular or renal disease. The analysis included 18 randomized trials and 97,800 participants. It did not test the same age interaction as Hanlon et al., so it cannot explain that result. It does show why a one-variable rule would be reckless: age and body size may point in different directions for a given patient.

GLP-1 cardiovascular effects also extend beyond glucose reduction. In SELECT, a randomized trial of 17,604 adults with overweight or obesity and established cardiovascular disease but without diabetes, weekly semaglutide 2.4 mg reduced MACE versus placebo (hazard ratio 0.80, 95% confidence interval 0.72 to 0.90). The SELECT trial cannot answer which diabetes drug is best at a given age, but it shows that GLP-1 cardiovascular effects cannot be reduced to HbA1c.

Other patient characteristics may matter as well. A matched population-based cohort study in people with NAFLD and type 2 diabetes found lower cardiovascular-event and mortality risk among GLP-1 receptor agonist users than among users of other glucose-lowering drugs. That association supports individualized prescribing, but its observational design cannot prove that the medication caused the difference.

Sex did not meaningfully modify treatment effects

Hanlon et al. found no consistent evidence that sex modified MACE effects for SGLT2 inhibitors or GLP-1 receptor agonists. Diallo et al. reached a similar conclusion. A null interaction does not prove men and women respond identically in every setting, but current evidence does not support choosing between these classes primarily on the basis of sex.

What the analysis can—and cannot—change today

The JAMA analysis is large, but it has limits that matter in practice. Individual participant data were available for only 6 of the 23 MACE trials; the other 17 supplied aggregate data. That restricts adjustment for age-linked factors such as kidney function, heart-failure prevalence, frailty, polypharmacy, and differential dropout. The cardiovascular trials also skewed older, leaving relatively sparse evidence in people younger than 50. The estimated age trend may not be linear across adulthood.

The authors did not model sex-by-age interactions or BMI-by-age interactions. They also did not provide absolute MACE rates by practical age categories. An individual-patient-data analysis that adjusted fully for eGFR, heart failure, BMI, and diabetes duration could weaken, confirm, or eliminate the apparent age effect.

For now, age is a useful question to add to treatment selection, not a substitute for clinical judgment. In a younger person with type 2 diabetes and cardiovascular risk, the analysis suggests GLP-1 receptor agonists may have the greater relative cardiovascular advantage, alongside stronger age-related glucose lowering. In an older person, SGLT2 inhibitors may have the greater relative cardiovascular advantage despite a smaller HbA1c effect, particularly when heart-failure risk is part of the picture.

Kidney function, obesity, heart failure, established atherosclerotic disease, diabetes duration, tolerability, cost, and the patient's glucose targets still shape the decision. DPP-4 inhibitors remain glucose-lowering options but do not offer the class-level cardiovascular benefit associated with SGLT2 inhibitors or GLP-1 receptor agonists. The useful lesson is narrower than the hype: choose cardiovascular prevention by cardiovascular evidence, then make sure glycemic control is also covered.