BACKGROUND Therapies for transthyretin amyloidosis with cardiomyopathy (ATTR-CM), including transthyretin (TTR) stabilizers and silencers, have demonstrated mortality benefit in three randomized trials. However, the timing of this benefit-often appearing delayed-has been debated, and has broad implications for clinical use and trial design. OBJECTIVES The purpose of this study was to evaluate the time course of mortality benefit with TTR stabilizers and silencers in ATTR-CM by estimating time-varying treatment effects across three randomized trials. METHODS We extracted time-to-event mortality data from the published Kaplan-Meier curves of three ATTR-CM outcomes trials: ATTR-ACT (tafamidis), ATTRIBUTE-CM (acoramidis), and HELIOS-B (vutrisiran). Using flexible parametric survival models, we estimated instantaneous hazard ratios and assessed the time-varying treatment effects across trials. RESULTS Mortality curves in each ATTR-CM trial began to diverge between approximately 12-18 months after therapy initiation (Figure 1, Panel A). Instantaneous hazard ratios showed consistent time-varying treatment effects across trials (p=0.96), with a pooled model confirming a delayed but progressively strengthening benefit (p for treatment effect <0.001; p for time interaction < 0.001). No significant differences were found between the 3 trials in the instantaneous hazard ratios with widely overlapping confidence intervals. We estimate that the treatment effect hazard ratio for mortality drops below 0.80 around 15 months (95% CI, 10-19) after randomization and continues to strengthen throughout follow-up. CONCLUSIONS AND RELEVANCE TTR silencers and stabilizers in ATTR-CM confer a delayed but consistent mortality benefit, with no significant differences observed between the three major trials. This uniform pattern may reflect a shared mechanism of action-reducing new amyloid deposition rather than reversing established disease, and underscore the importance of early treatment initiation and adequate trial duration to capture delayed mortality effects.
Claggett et al. (Thu,) studied this question.