GLP-1 Drugs Outperform SGLT2 Inhibitors in Survival After Sepsis in Diabetics
In 24,000 matched diabetic patients with sepsis, prior GLP-1 receptor agonist use cut 90-day mortality by 19% versus SGLT2 inhibitors.
Summary
A large propensity score-matched cohort study compared outcomes in over 23,000 type 2 diabetic patients hospitalized with sepsis, based on whether they had previously been using GLP-1 receptor agonists (like semaglutide or liraglutide) or SGLT2 inhibitors (like empagliflozin). Patients on GLP-1 drugs prior to sepsis had a 19% lower 90-day mortality risk, fewer major cardiovascular events, fewer heart attacks, and lower ICU admission rates. The mortality benefit held at one year and across all subgroups. However, several secondary outcomes — including septic shock, respiratory failure, and kidney events — showed no statistically significant difference. The authors caution that residual confounding cannot be ruled out, and these findings are hypothesis-generating, requiring prospective validation.
Detailed Summary
Sepsis remains one of the leading causes of death in hospitalized patients, and type 2 diabetes dramatically worsens outcomes. As GLP-1 receptor agonists (GLP-1RAs) and SGLT2 inhibitors (SGLT2is) become cornerstones of diabetes management, a critical but unanswered question is whether the choice of glucose-lowering therapy prior to a sepsis event meaningfully affects survival — a question with direct relevance to metabolic health and longevity.
This retrospective cohort study drew on the TriNetX US Collaborative Network to identify adults with type 2 diabetes who developed sepsis and had received either a GLP-1RA or an SGLT2i in the three months prior. After 1:1 propensity score matching, 11,969 patients were included in each group, creating a well-balanced comparison population of nearly 24,000 individuals.
The primary finding was striking: prior GLP-1RA use was associated with a 19% reduction in 90-day all-cause mortality compared to prior SGLT2i use (HR 0.81; 95% CI 0.75–0.88). GLP-1RA users also had lower rates of major adverse cardiovascular events (MACE), myocardial infarction, and ICU admission. The mortality benefit remained consistent at 365 days and across all prespecified subgroups. Notably, no significant differences were found for septic shock, acute respiratory failure, or major adverse kidney events.
After applying Holm correction for multiple comparisons, only MACE and ICU admission associations remained statistically significant among the secondary endpoints. This important statistical caveat tempers enthusiasm for some of the secondary findings.
The implications are significant: GLP-1 receptor agonists may confer protective effects beyond glycemic control — potentially through anti-inflammatory, cardioprotective, or immune-modulating pathways — that translate into better sepsis survival. For clinicians managing diabetic patients at risk of infection or critical illness, drug class selection may matter beyond metabolic targets. These hypothesis-generating findings now urgently warrant prospective randomized confirmation.
Key Findings
- Prior GLP-1RA use reduced 90-day all-cause mortality by 19% versus SGLT2 inhibitors in diabetic sepsis patients.
- GLP-1RA users had 23% lower risk of major adverse cardiovascular events (MACE) after sepsis.
- ICU admission rates were significantly lower in GLP-1RA users (HR 0.89), surviving multiple comparison correction.
- No significant difference was seen for septic shock, acute respiratory failure, or kidney events between drug classes.
- The mortality benefit was consistent at 365 days and across all patient subgroups.
Methodology
Retrospective propensity score-matched cohort study using the TriNetX US Collaborative Network. Adults with type 2 diabetes who developed sepsis and had received a GLP-1RA or SGLT2i within three months prior were matched 1:1, yielding 11,969 patients per group. Holm correction was applied to account for multiple secondary outcome comparisons.
Study Limitations
As a retrospective observational study, residual confounding cannot be excluded despite propensity matching; the authors note modest E-values and baseline imbalances before matching. Exposure was classified at the drug-class level, not by individual agent, potentially masking differences within each class. The summary is based on the abstract only, as the full text was not available for review.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: