Metformin Is Associated with Reduced Risk of Sepsis: Analysis of a U.S. Cohort from 2009-2019

Speaker(s)

Bandy S1, Black CA2, Leung K3, Lipscomb J2, Koeller J2, Benavides R2, Lee GC2
1The University of Texas at Austin, San Antonio, TX, USA, 2The University of Texas at Austin, Ausitn, TX, USA, 3AllazoHealth, San Antonio, TX, USA

Presentation Documents

OBJECTIVES:

The global rates of sepsis and sepsis-related mortality have not significantly declined. Metformin, a first-line medication for type two diabetes mellitus (T2DM), has been found to have immunomodulatory effects in aging, cancer, and in high inflammatory states (e.g. severe infections/sepsis), highlighting its potential role in preventative-medicine; however, real-world studies are lacking. This study evaluated whether metformin-use is associated with lower risks for all-cause sepsis.

METHODS:

This was a retrospective U.S. cohort analysis of third-party medical and pharmacy claims databases from 2009 to 2019. Adults 18 years of age and older who received a routine pneumococcal vaccine (index date) and were continuously enrolled for at least one year before and after were included. Diagnosis of all-cause sepsis during hospitalization was determined using ICD-9/10-CM codes within one year from the index date. Metformin-use was defined as having at least a 90-day supply prior to index date. The primary outcome of all-cause sepsis was analyzed using multivariable logistic-regression models stratified by T2DM status.

RESULTS:

Overall, 600,132 adults met inclusion; the median age was 65 years, 45.1% were males, and 25.9% had T2DM. Among the 44,346 metformin-users (median dose=1,000 mg/day), 84.4% (n=37,420) had T2DM and 15.6% did not have T2DM (n=6,929); there were no differences in baseline demographics. The overall rate of sepsis within 1 year was 0.43%. Overall, metformin-use was associated with lower risks of sepsis (aOR 0.847, 95% CI 0.721-0.995) after adjusting for age, sex, year, and co-morbidities. This protective effect of metformin-use was observed when stratified among metformin-users without T2DM (aOR 0.549, 95% CI 0.327-0.920), but not among those with T2DM (aOR 0.920, 95% CI 0.768-1.103).

CONCLUSIONS:

Metformin-use was associated with reduced risk of sepsis among persons without T2DM, underscoring the need for future investigations to identify populations who might benefit from metformin.

Code

CO196

Topic

Clinical Outcomes, Study Approaches

Topic Subcategory

Clinical Outcomes Assessment

Disease

Diabetes/Endocrine/Metabolic Disorders (including obesity)