(66) Restricted Mean Survival Time for Analysis of Real-World Cardiovascular Effect Associated with SGLT2 Inhibitor Use: A Population-Based Study

Saturday, August 27, 2022

8:00 AM – 6:00 PM CEST

Location: Hall C

Publication Number: 740

Presenting Author(s)

Zi-Yang Peng
– National Cheng Kung University, Tainan City, Taiwan

Co-Author(s)

Chun-ting Yang
– Institute of Clinical Pharmacy and Pharmaceutical Sciences, College of Medicine, National Cheng Kung University, Tainan, Taiwan, Taiwan (Republic of China)
Chih-Hsing Wu
– Department of Family Medicine, College of Medicine, National Cheng Kung University, Tainan, Taiwan, Taiwan (Republic of China)
Huang-Tz Ou
– Institute of Clinical Pharmacy and Pharmaceutical Sciences, College of Medicine, National Cheng Kung University, Tainan, Taiwan, Taiwan (Republic of China)

Background: Restricted mean survival time (RMST) has been emerged as an alternative measure to conventional hazard ratio for time-to-event outcomes owing to its robustness and clinical interpretability. Difference in the RMSTs (dRMSTs) can quantify the postponement (free) of an outcome event during a specific (restricted) interval and corresponds to the difference between the areas under the two survival curves of two comparison groups. A growing study has reported the cardiovascular effect of sodium-glucose cotransporter-2 inhibitors (SGLT2is) using the RMST analysis which was however mainly based on individual-level, time-to-event data reconstructed from available trials. Therefore, the applicability of trial-data based RMST estimates to support real-world decisions is uncertain.

Objectives: To report real-world cardiovascular effect of SGLT2i use using the RMST analysis.

Methods: New users of SGLT2i or dipeptidyl-peptidase 4 inhibitor (DPP4i) in 2017 were identified from Taiwan’s National Health Insurance Database. Propensity score (PS) matching was employed to ensure the between-group comparability. Hospitalization for heart failure (HHF) was the main outcome of interest and secondary outcomes included 3-points major adverse cardiovascular event (3P-MACE comprising myocardial infarction [MI], stroke, and CV death), 4P-MACE (including 3P-MACE and HHF), MI, stroke, CV death, and all-cause death. Each patient was followed until an outcome occurred, loss of follow-up, or end of 2018, whichever came first. For each outcome, we assessed the RMSTs with 95% CIs for SGLT2i and DPP4i groups based on direct integration of Kaplan-Meier survival curves and made inferences about between-group difference in the RMSTs (dRMSTs) up to a 2-year follow-up time.

Results: There were 21,144 PS-matched SGLT2i and DPP4i pairs identified for analyses. The RMSTs of SGLT2i and DPP4i users for HHF in 2 years after treatment initiation were 720.71 (719.80, 721.61) and 715.65 (714.52, 716.79) days, respectively. The dRMSTs between SGLT2i and DPP4i users for HHF was 4.99 days, implying that on average patients treated with SGLT2is for 2 years would be free of HHF for 4.99 (3.56, 6.42) days more than DPP4i users. For secondary outcomes, the postponements for 3P-MACE, 4P-MACE, MI, stroke, CV death, and all-cause death while using SGLT2is versus DPP4is for 2 years were 4.12 (2.72, 5.52), 7.72 (5.83, 9.61), 1.26 (0.47, 2.04), 2.70 (1.57, 3.82), 0.69 (0.28, 1.11), and 6.05 (4.89, 7.20) days, respectively.

Conclusions: Compared with DPP4is, using SGLT2is over 24 months in a real-world setting provided the favorable CV outcomes with a delay in the occurrence of CV events in a range of 1 to 8 days, which is clinically informative and interpretable to support real-world decisions.