Abstract
KEY POINTS: In a multicenter retrospective cohort study of adults hospitalized with acute heart failure, exposure to sodium-glucose cotransporter-2 inhibitor during AKI was associated with lower risk of 30-day mortality. Exposure to sodium-glucose cotransporter-2 inhibitor during acute heart failure–associated AKI was associated with no difference in time to renal recovery. The findings were reproducible in inverse probability-weighted analysis. BACKGROUND: Although sodium-glucose cotransporter-2 inhibitor (SGLT2i) use during acute heart failure (AHF) hospitalizations is associated with symptomatic improvement, reduction in rehospitalizations, and mortality, these medications are often withheld during AKI because of concerns about worsening GFR. We aimed to investigate the safety of SGLT2i exposure during AKI among patients hospitalized with AHF. We hypothesized that SGLT2i exposure would not worsen mortality but may prolong return of creatinine to baseline. METHODS: This was a retrospective study of adults hospitalized across five Yale New Haven Health System hospitals between January 2020 and May 2022 with AHF complicated by Kidney Disease Improving Global Outcomes–defined AKI. Patients with stage 5 CKD and those with potential contraindications to SGLT2i were excluded. We tested the association of SGLT2i use with kidney function recovery at 14 days and death at 30 days using time-varying, multivariable Cox-regression analyses. RESULTS: Of 3305 individuals hospitalized with AHF and AKI, 356 received SGLT2i after AKI diagnosis either as initiation or continuation. The rate of renal recovery was not significantly different among those exposed and unexposed to SGLT2i after AKI (adjusted hazard ratio, 0.94; 95% confidence interval, 0.79 to 1.11; P = 0.46). SGLT2i exposure was associated with lower risk of 30-day mortality (adjusted hazard ratio, 0.45; 95% confidence interval, 0.23 to 0.87; P = 0.02). Sensitivity analyses using an inverse probability-weighted time-varying Cox regression analysis and using alternate definitions of AHF with different NT-proBNP cutoffs yielded similar results. Rates of renal recovery were similar between the exposed and unexposed cohorts regardless of the proximity of SGLT2i exposure to AKI diagnosis. CONCLUSION: In adults experiencing AHF-associated AKI, exposure to SGLT2i was associated with decreased mortality and no delay in renal recovery. Prospective studies are needed to elucidate the effect of SGLT2i exposure during AKI, particularly during heart failure hospitalizations.