A putative placebo analysis of the effects of sacubitril/valsartan in heart failure across the full range of ejection fraction

Abstract

Abstract Aims The PARADIGM-HF and PARAGON-HF trials tested sacubitril/valsartan against active controls given renin–angiotensin system inhibitors (RASi) are ethically mandated in heart failure (HF) with reduced ejection fraction and are used in the vast majority of patients with HF with preserved ejection fraction. To estimate the effects of sacubitril/valsartan had it been tested against a placebo control, we made indirect comparisons of the effects of sacubitril/valsartan with putative placebos in HF across the full range of left ventricular ejection fraction (LVEF). Methods and results We analysed patient-level data from the PARADIGM-HF and PARAGON-HF trials (n = 13 194) and the CHARM-Alternative and CHARM-Preserved trials (n = 5050, candesartan vs. placebo). The rate ratio (RR) of sacubitril/valsartan vs. putative placebo was estimated by the product of the RR for sacubitril/valsartan vs. RASi and the RR for RASi vs. placebo. Total HF hospitalizations and cardiovascular death were analysed using the negative binomial method. Treatment effects were estimated using cubic spline methods by ejection fraction as a continuous measure. Across the range of LVEF, sacubitril/valsartan was associated with a RR 0.54 [95% confidence interval (CI) 0.45–0.65] for the recurrent primary endpoint compared with putative placebo (P < 0.001). Treatment benefits of sacubitril/valsartan vs. putative placebo varied non-linearly with LVEF with attenuation of effects observed at LVEF above 60%. When analyzing data from PARADIGM-HF and CHARM-Alternative, the estimated risk reduction of sacubitril/valsartan vs. putative placebo was 48% (95% CI 35–58%); P < 0.001. When analyzing data from PARAGON-HF and CHARM-Preserved (with LVEF ≥ 45%), the estimated risk reduction of sacubitril/valsartan vs. putative placebo was 29% (95% CI 7–46%); P = 0.013. Across the full range of LVEF, consistent effects were observed for time-to-first endpoints: first primary endpoint (RR 0.72, 95% CI 0.64–0.82), first HF hospitalization (RR 0.67, 95% CI 0.58–0.78), cardiovascular death (RR 0.76, 95% CI 0.64–0.89), and all-cause death (RR 0.83, 95% CI 0.71–0.96); all P < 0.02. Conclusion This putative placebo analysis reinforces the treatment benefits of sacubitril/valsartan on risk of adverse cardiovascular events across the full range of LVEF, with most pronounced effects observed at a LVEF up to 60%.