Respiratory viral infections continue to impose a large burden of hospitalization, organ failure and death, particularly among older adults and people with cardiometabolic disease. Whether prior semaglutide exposure is associated with lower post-infection severity across distinct respiratory viruses remains uncertain. We analyzed de-identified longitudinal electronic health record data from a federated U.S. network spanning more than 29 million patients. Adults with documented COVID-19 or influenza infection were classified by pre-infection semaglutide use versus metformin use without prior GLP-1 receptor agonist exposure and one-to-one propensity matched on demographics, BMI, HbA1c, vaccination status, prior infection status, antiviral use, healthcare utilization extent, and baseline comorbidities, yielding matched cohorts of N = 7092 for COVID-19 and N = 1920 for influenza (flu). In the COVID-19 cohort, semaglutide exposure was associated with significantly lower 30-day mortality (0.3% vs 0.8%, RR 0.39, P < .001), hospitalization (5.3% vs 7.4%, RR 0.72, P < .001), and the composite disease severity outcome (8.2% vs 10.6%, RR 0.77, P < .001). Differences in rates of intensive care admission (1.4% vs 1.8%, RR 0.79, P = .07) and mechanical ventilation (0.7% vs 0.9%, RR 0.76, P = .14) were not statistically significant. In the flu cohort, semaglutide use was also associated with significantly lower mortality (0.2% vs 0.7%, RR 0.29, P = .02), hospitalization (6.5% vs 9.6%, RR 0.68, P < .001), and the composite outcome (10.3% vs 14.6%, RR 0.70, P < .001); ICU admissions (1.9% vs 2.7%, RR 0.71, P = .11) and mechanical ventilation (0.7% vs 0.9%, RR 0.78, P = .48) were not statistically significant. Among organ-specific outcomes, for which P values were adjusted for multiple comparisons, a significant risk reduction was observed for acute kidney injury (COVID-19 RR 0.73, P = .02; flu RR 0.58, P = .03) in the 30-day observation period, whereas reductions in acute coronary syndrome (COVID-19 RR 0.67, P = .14; flu RR 1.00, P = 1.00) and acute respiratory failure (COVID-19 RR 0.84, P = .17; flu RR 0.74, P = .27) were not statistically significant against the metformin comparator. Within the COVID-19 cohort, prior semaglutide exposure was associated with lower composite COVID-19 severity across the full study period, reaching significance in both vaccinated (RR 0.81, P = .04) and unvaccinated (RR 0.76, P < .001) patients, and during the pandemic period ending May 11, 2023, reaching significance in unvaccinated patients (RR 0.76, P = .001) but not vaccinated patients (RR 0.85, P = .20). When restricting to those treated with standard-of-care therapeutics during the COVID-19 pandemic era, semaglutide users had significantly lower 30-day composite risk among those receiving Paxlovid within ±7 days of infection (9.6% vs 12.7%; RR 0.76, P = .02). Among those initiating corticosteroids ±7 days of infection, the composite outcome remained significantly lower (27.3% vs 36.3%; RR 0.75, P = .007). Among the influenza cohort, semaglutide was associated with a reduced risk of the composite outcome among unvaccinated patients (RR 0.69, P < .001), but this trend was not statistically significant among vaccinated patients (RR 0.74, P = .06). Notably, semaglutide use was associated with significantly lower composite risk among Medicare-eligible adults aged ≥65 years for both COVID-19 (11.2% vs 12.9%; RR 0.86, P = .04) and influenza (17.0% vs 23.7%; RR 0.72, P = .003). Finally, composite risk did not vary significantly across pre-infection weight loss strata of <5%, 5%-15%, and ≥15% for COVID (P = .89) or flu (P = .54), or across dosage strata of 0.25-0.5, 1.0, and 1.7-2.4 mg/week for COVID (P = .50) or flu (P = .32), suggesting semaglutide-associated lower COVID and influenza severity may not be related to substantial weight loss or maximal dosing. Taken together, this study motivates prospective evaluation of low-dose semaglutide to blunt the severity of respiratory viral infections.
