Among nearly 24,000 propensity score-matched adults with type 2 diabetes who developed sepsis, those previously taking GLP-1 receptor agonists faced a 19% lower 90-day all-cause mortality risk compared with prior SGLT2 inhibitor users (HR 0.81; 95% CI 0.75–0.88; incidence rate 48.6 vs. 60.6 per 100 person-years). GLP-1RA users also showed reduced major adverse cardiovascular events (HR 0.77) and ICU admissions (HR 0.89). The mortality advantage persisted through 365 days and across subgroups. After Holm correction, only MACE and ICU admission survived multiple-comparison adjustment.
This is a genuinely intriguing hypothesis-generating finding, but requires careful interpretation. Sepsis is an acute crisis, not a chronic lifestyle condition — so what's being tested here is whether the *pre-illness metabolic state* shaped by these drugs influences survival, not direct treatment of infection. GLP-1 agonists have established anti-inflammatory, cardioprotective, and potentially organ-protective properties that could plausibly buffer the cytokine storm of sepsis. However, SGLT2 inhibitors are themselves cardio- and nephroprotective, making this comparison informative rather than straightforwardly pro-GLP-1.
Critical limitations: the retrospective design using TriNetX administrative data cannot rule out residual confounding (E-values of 1.4–1.7 are modest). Drug exposure is defined at class level, obscuring agent-specific differences like semaglutide versus liraglutide. For adults on either drug class, this is reassuring background evidence rather than actionable guidance. Prospective trials are needed before any clinical inference can be drawn.