Association of pre-existing sodium-glucose cotransporter-2 inhibitor therapy with outcomes in patients presenting with cardiogenic shock: a retrospective cohort study
M Kunz, M Tjeerdsma, L Volken, S Wenger, J Wiedenhofer, B Kaufmann, C Mueller, F Mahfoud, J Sundermeyer, B Schrage, G Leibundgut, J BoeddinghausAbstract
Background and purpose
Sodium-glucose cotransporter-2 inhibitors (SGLT2-I) are associated with improved cardiovascular and renal outcomes in patients with chronic kidney disease (CKD) and heart failure (HF). Whether pre-existing SGLT2-I therapy confers a prognostic advantage in patients presenting with cardiogenic shock, despite being paused during hospital stay, remains largely unknown. This study aimed to investigate the association between pre-existing SGLT2-I therapy and outcomes in patients with cardiogenic shock.
Methods
In this retrospective, single-centre cohort study, consecutive patients with cardiogenic shock presenting to a tertiary care hospital between 09/2021-09/2025 were included. Patients were stratified by pre-existing SGLT2-I therapy. Two comparisons were performed: (i) SGLT2-I users versus non-users, and (ii) SGLT2-I users versus patients with a guideline-based indication for SGLT2-I due to known HF (irrespective of LVEF) and/or CKD (G3-G5) but without such therapy (“indication non-users”). The primary endpoint was in-hospital all cause death. Secondary endpoints included 30-day and 90-day all-cause mortality, need for organ support at any time during hospitalization (renal replacement therapy, mechanical circulatory support [MCS], vasoactive or inotropic therapy), and occurrence of acute kidney injury (AKI).
Results
The study population comprised 650 patients (median 74 years, 38% women). Shock aetiology was ischemic in 30.2%, and non-ischemic in 69.8%. In-hospital all-cause death occurred in 305 patients (46.9%). In the overall cohort comparison, pre-existing SGLT2-I use showed no significant association with endpoints. Among patients with an indication for SGLT2-I (n=347), baseline characteristics were well balanced between SGLT2-I users (n=152) and indication non-users (n=195). The primary outcome occurred in 40.8% of SGLT2-I users versus 49.7% of indication non-users. After adjustment for age, sex, renal function, and known HF or baseline LVEF, pre-existing SGLT2-I therapy was independently associated with reduced in-hospital mortality (adjusted OR 0.44, 95% CI 0.22-0.89; p=0.024). Model-based g-computation derived from the adjusted model estimated an in-hospital mortality of 45.3% without SGLT2-I versus 29.0% with SGLT2-I, corresponding to an ARR of 16.3% and a NNT of 6.1 to prevent one in-hospital death. SGLT2-I users also demonstrated improved 30-day (adjusted HR 0.52, 95% CI 0.31-0.89; p=0.017) and 90-day mortality (adjusted HR 0.57, 95% CI 0.35–0.93; p=0.023). No association was observed with the need for organ support or occurrence of AKI. Subgroup analyses suggested the strongest association in patients with known HF, baseline LVEF <50%, and shock due to acute HF or valvular disease.
Conclusion
Pre-existing SGLT2-I therapy in patients with cardiogenic shock and a guideline-based indication for SGLT2-I due to HF or CKD was associated with improved survival during hospitalization and up to 90 days.