Advertisement

MS1 VALIDATION OF MODELED 5-YEAR SURVIVAL OUTCOMES AMONG PATIENTS WITH CYSTIC FIBROSIS (CF) TREATED WITH THE CF TRANSMEMBRANE CONDUCTANCE REGULATOR MODULATOR (CFTRM) IVACAFTOR USING US CF FOUNDATION PATIENT REGISTRY (USCFFPR) DATA

      Objectives

      CF is a rare genetic disease characterized by life-shortening lung function decline. Ivacaftor, a highly effective CFTRm, was introduced in 2012. A model was developed to project the impact of ivacaftor on survival based on known relationships between CF patient characteristics and mortality. This study validates the model methodology and survival projections using data from a recent real-world, postapproval, long-term safety study (LTSS) (Volkova 2019) that followed ivacaftor patients and matched comparators for 5 years (2012-2016).

      Methods

      The patient-level simulation model assesses the impact of initiating ivacaftor at age ≥6 years vs best supportive care (BSC) on outcomes in US CF patients with CFTR gating mutations. Lifetime survival was estimated using parametric equations fitted to historical USCFFPR survival data to estimate age-specific baseline mortality risk for each patient. Mortality hazards were estimated as the age-specific hazard adjusted for fixed and time-varying patient-level characteristics corresponding to published risk factors (Liou 2001). BSC disease progression was derived from published studies using the USCFFPR, and the expected impact of ivacaftor on ppFEV1, pulmonary exacerbations, and weight-for-age derived from clinical trials. LTSS patient characteristics were entered into the model at baseline and outcomes simulated; 5-year model-projected mortality with bootstrapped credible intervals (CrI) was compared to LTSS mortality.

      Results

      Modeled 5-year mortality projections closely approximate real-world data in both the BSC (6.4% [95% CrI: 5.3%-7.6%] modeled vs 6.0% observed) and ivacaftor-treated (3.5% [2.8%-4.5%] vs 3.1%) populations. The model also predicts that ivacaftor-treated patients have a 5-year relative risk of mortality of 0.54 ([0.48-0.61] modeled, 0.51 observed) vs untreated patients.

      Conclusions

      Modeled 5-year survival projections for CF patients initiating ivacaftor vs BSC track closely to observed registry data. Findings support the validity of modeling CF using the approach described herein to predict long-term survival and estimate clinical and economic outcomes of CFTRm. Sponsor: Vertex Pharmaceuticals.