(165) Effectiveness of early versus delayed post-discharge sodium-glucose cotransporter-2 inhibitor initiation in acute heart failure with type 2 diabetes: a population-based cohort study

Background: Recent guidelines have recommended early sodium-glucose cotransporter-2 inhibitor (SGLT2i) initiation in clinically stable acute heart failure, based on results from placebo-controlled randomized trials.

Objectives: To compare the effectiveness of early versus delayed post-discharge SGLT2i initiation, in older adults with acute heart failure and type 2 diabetes, using real-world data.

Methods: This retrospective, population-based cohort study used health services databases housed at ICES. Residents of Ontario, Canada ≥66 years with type 2 diabetes who were discharged to community from acute care hospital admission for heart failure as the most responsible diagnosis, during July 1, 2016, to March 31, 2020, entered the cohort. We chose a cohort entry period in which in-hospital use of SGLT2i was rare, to minimize exposure misclassifications. People dispensed an SGLT2i in the year before admission were excluded for prevalent users. People with contraindication were also excluded. We compared 4 different post-discharge SGLT2i initiation timing: ≤3 days, 4-90 days, 91-180 days, and delay for ≥180 days (reference). With a target trial emulation framework, the “clone-censor-weight” approach was used to address time-related biases. The primary outcome was hospitalization for heart failure (HHF) or cardiovascular death as a composite over 1 year from discharge. Stabilized inverse probability weights were used to adjust selection bias due to artificial censoring. Risk ratios (RRs) and risk differences (RDs) were obtained from cumulative incidence functions estimated from a weighted pooled logistic regression standardized to baseline covariates. Percentile confidence intervals (CIs) were derived from 200 non-parametric bootstraps.

Results: Among 9,641 eligible people, after cloning and artificial censoring, there were 38,564 clones, 7,584 events, and 12,439 person years. Compared to delay for ≥180 days, SGLT2i initiation within 3 days post-discharge was associated with a 35% lower risk of HHF or cardiovascular death over 1 year (RR [95% CI] = 0.65 [0.45, 0.83], RD [95% CI] = -13.1% [-19.8%, -6.3%]), but initiation 4-90 days post-discharge (RR [95% CI] = 0.83 [0.72, 0.93], RD [95% CI] = -6.1% [-9.9%, -2.6%]) and initiation 91-180 days post-discharge (RR [95% CI] = 0.89 [0.79, 0.97], RD [95% CI] = -4.1% [-7.6%, -1.0%]) showed smaller risk reduction.

Conclusions: In older adults with acute heart failure and type 2 diabetes, early SGLT2i initiation within 3 days post-discharge, compared to delayed initiation, was associated with a lower risk of HHF or cardiovascular death. This real-world evidence study supports guidelines recommending early SGLT2i initiation post-hospitalization for acute heart failure.