Real world prognostic modelling and trial simulations to inform phase 3 design amid improved standard of care in ATTR-CM
O Lairez, N D Amaral, N G FerreiraAbstract
Background
Recent improvements in standard of care (SoC) [1, 2], particularly following HELIOS-B [3], are expected to lower composite event rates in transthyretin amyloid cardiomyopathy (ATTR-CM), risking underpowered Phase 3 trials. We used real-world data (RWD) to identify prognostic factors and ran Monte Carlo–style simulations to assess event accrual under Phase 3-like design refinements.
Purpose
To quantify the impact of improved SoC on composite event rates, validate expected Phase 3 rates using RWD prognostic modelling, and project event rates under contemporary SoC scenarios to generate actionable protocol refinements that preserve power and event accrual.
Methods
We assembled a US RWD cohort (TriNetX; n=1562; 607 composite events) aligned to the planned Phase 3 population. Baseline was anchored to tafamidis initiation or ATTR-CM diagnosis proxy, with inclusion/exclusion (IE) approximations (N-terminal pro–B-type natriuretic peptide [NT-proBNP] ≥1000 pg/mL [sinus or atrial-fibrillation], age ≥18 years, bilirubin ≤3× upper limit of normal, glomerular filtration rate ≥15 mL/min/1.73 m², no hospitalization within 14 days pre-index, and no acute coronary syndrome/unstable angina/transient ischemic attack/coronary revascularization/pacemaker within 60 days). The composite endpoint was cardiovascular or heart failure hospitalizations plus all-cause mortality; censoring accounted for inactivity gaps ≥12 months. A multivariable Cox model identified prognostic factors. We ran 1000 Phase 3-like simulations (N=1030; 1:1 randomisation; assumed 30% treatment effect) to estimate event rates in different scenarios: (a) historical RWD, (b) contemporary SoC scenarios (21%, 25% and 36% risk reduction, Fig 2), and (c) a risk-enriched "sicker" population (higher NT-proBNP, prior hospitalisation, elevated urea/bilirubin) – see workflow in Fig 1.
Results
RWD-based modelling reproduced historical composite event rates for Phase 3 (~20%/year; 19.9%/year). When applying literature-based SoC risk reductions, simulations projected contemporary rates of ~16%/year (15.2%/year), quantifying implications for event accrual under current SoC. Risk-enrichment ("sicker" cohort) increased the composite rate to 27.1%/year, demonstrating that IE adjustments can restore event accrual despite SoC-driven reductions. Adverse prognostic factors were prior hospitalisation (14–180 and >180 days preindex), higher log(NT-proBNP), very high urea, high bilirubin, diuretic use, and low LVEF; SGLT2 inhibitor use was protective.
Conclusions
Prognostic modelling on RWD validated historical event rates and, under literature-based SoC scenarios, projected lower composite event rates in ATTR-CM, with implications for Phase 3 power. IE criteria could be amended to enrich for higher-risk participants while maintaining feasibility. This evidence-based framework supports adaptive protocol design to preserve event accrual and statistical power in the current SoC landscape.For image description, please refer to the figure legend and surrounding text.For image description, please refer to the figure legend and surrounding text.