Does symptom profile modify revascularization benefit in ischemic LV dysfunction? A win ratio analysis pooling STICHES and REVIVED-BCIS2
G Tong, Y Zeng, A J Lansky, M Ryan, M C Petrie, E Kontopantelis, M Dodd, G Marquis-Gravel, K F Doherty, T Clayton, M A Mamas, J L Rouleau, D Perera, E J VelazquezAbstract
Background
In the Surgical Treatment for Ischemic Heart Failure Extension Study (STICHES), coronary artery bypass grafting (CABG) improved long-term survival in patients with ischemic left ventricular (LV) dysfunction compared with medication alone. In contrast, in the Revascularization for Ischemic Ventricular Dysfunction trial (REVIVED-BCIS2), percutaneous coronary intervention (PCI) did not reduce all-cause mortality compared with medical therapy. Whether baseline symptom profile modifies the treatment effect of revascularization across a broader spectrum of clinical end points remains unknown.
Purpose
To determine whether the clinical benefit of revascularization over medical therapy differs according to baseline heart failure (HF) severity and angina burden using a win ratio approach applied to pooled patient-level data from STICHES and REVIVED-BCIS2.
Methods
Patient-level data from both trials (N=1912; all with LV ejection fraction ≤35%) were pooled. The primary end point was a hierarchical composite evaluated by win ratio (WR) using time-to-first event: (1) all-cause death, (2) myocardial infarction, (3) hospitalization for HF, and (4) unplanned revascularization. WRs were computed within each trial (median follow-up 80.5 vs. 40.6 months) and combined via fixed-effects meta-analysis. Inverse probability of treatment weighting adjusting for age, sex, and comorbidities was applied as a sensitivity analysis. Participants were classified into 4 strata by NYHA class (I–II vs III–IV) and CCS angina grade (0–1 vs 2–4).
Results
Of 1905 patients (median age 63.6, 88% male, mean LVEF 28.2%; 67% NYHA I–II, 35% CCS ≥2), WRs were directionally consistent across all 4 strata: NYHA I–II/CCS 0–1 (47%; WR 1.12, 0.93–1.35), NYHA I–II/CCS 2–4 (20%; WR 1.21, 0.93–1.58), NYHA III–IV/CCS 0–1 (18%; WR 1.26, 0.96–1.66), NYHA III–IV/CCS 2–4 (15%; WR 1.09, 0.82–1.44) (Figure 1). The overall meta-analytic WR was 1.17 (95% CI, 1.04–1.32; p=.011; I²=34.2%); trial-specific WRs were 1.22 (1.06–1.41) in STICHES and 1.03 (0.81–1.30) in REVIVED-BCIS2. Of 902,496 pairwise comparisons, 64% were resolved at all-cause death, with unplanned revascularization showing the largest asymmetry favoring revascularization (24,807 vs 3216 wins). Cox models showed reduced all-cause death (HR 0.87; 0.76–0.98), HF hospitalization (HR 0.77; 0.64–0.92), and unplanned revascularization (HR 0.20; 0.14–0.29); myocardial infarction did not differ significantly (HR 0.82, 0.59–1.14). A 30-day landmark analysis confirmed consistent mortality benefit beyond the periprocedural period.
Conclusions
In pooled data from STICHES and REVIVED-BCIS2, the direction of revascularization benefit was consistent across all 4 symptom strata defined by NYHA class and CCS grade, though individual strata were likely underpowered. Overall, revascularization demonstrated significant clinical benefit over medical therapy.