Logo image
Rates of In-Hospital SGLT2 Inhibitor Discontinuation After AKI: SA-OR016
Abstract   Peer reviewed

Rates of In-Hospital SGLT2 Inhibitor Discontinuation After AKI: SA-OR016

Benjamin R. Griffin, Mary Vaughan Sarrazin, Melissa L. Swee, Masaaki Yamada, Brian C. Lund and Diana I. Jalal
Journal of the American Society of Nephrology, Vol.37(10S), 10.1681/ASN.2026pbf2h7jn
10/2026
DOI: 10.1681/ASN.2026pbf2h7jn

View Online

Abstract

Background: Sodium glucose co-transporter 2 inhibitors (SGLT2i) are associated with decreased morbidity and mortality of cardiovascular and kidney events following acute kidney injury (AKI). However, clinical practices for discontinuing these agents after AKI in inpatient settings and restarting after discharge are not well described. Methods: We evaluated Veterans hospitalized across 120 VA facilities from 2019-2025 who had prior SGLT2i use and developed AKI during the hospital episode. The primary outcome was continued SGLT2i use after AKI, with secondary outcome SGLT2i use within 180 days after live discharge. Mixed-effects logistic regression models including demographics, comorbidities, AKI severity, year of admission, and VA facility random effects were used to estimate patient- and facility-level associations. Facility variation in SGLT2i use was summarized using the median odds ratio (MOR). Results: Among 63,957 hospitalized Veterans with prior SGLT2i use and in-hospital AKI, 34,310 patients (53.6%) did not receive an SGLT2i after AKI. Discontinuation varied significantly by AKI severity: 50.4% after stage 1 AKI, 63.7% after stage 2 AKI, and 71.7% after stage 3 AKI. In risk-adjusted models, SGLT2i use after AKI was less likely among patients with heart failure (OR 0.52 compared to no CHF, 95% CI 0.51-0.53), and decreased over time (OR 0.30 for 2025 versus 2019, 95% CI 0.28-0.32). Facility-level variation remained substantial after adjustment with an adjusted MOR of 1.48 (95% CI 1.38-1.57), a larger effect size than most clinical variables. Among 61,923 patients discharged alive, 18,084 (29.2%) had no SGLT2i use within 180 days. Post-discharge use was lower among patients with SGLT2i discontinued during hospitalization compared to those continued (60.4% vs. 82.5%), and this association persisted after adjustment for patient and facility factors (OR 0.33, 95% CI 0.32-0.34). Conclusion: More than half of patients with prior SGLT2i use did not receive an SGLT2i after AKI during a hospitalization. In-hospital SGLT2i discontinuation was strongly associated with decreased long-term use and represents a potentially modifiable target to improve post-AKI outcomes. Variations in continuation rates across VA facilities were substantial and more impactful than most clinical features.

Details

Metrics

1 Record Views
Logo image