External validation of the French predictive model to estimate PAH survival: A REVEAL analysis

Background: The F rench P ulmonary Hypertension N etwork (FPHN) and R egistry to EV aluate E arly A nd L ong-term Pulmonary Arterial Hypertension (PAH) Disease Management (REVEAL) recently developed models to predict survival in PAH patients (pts). Aims and Objectives: The FPHN equation was tested in a REVEAL cohort to assess its generalizability in a different PAH population. Methods: The REVEAL validation cohort had 436 recently diagnosed (<1 year before enrollment), treatment-naive, ≥18-year-old patients (pts) with idiopathic, familial, or anorexigen-induced PAH, divided into subgroups with non-missing (n=292) and missing (n=144) data for all FPHN equation parameters. Results: The FPHN and REVEAL cohorts had similar characteristics. FPHN follow-up (f/u) was ≥3 years; REVEAL validation cohort and subgroup with non-missing data had a mean f/u from diagnosis of 29.2 and 30.1 months, respectively. FPHN model parameters applied to REVEAL showed a good correlation of estimated hazard ratios between the two studies and robust C index of 0.72 (95% CI, 0.64–0.80). Risk stratification based on the FPHN equation showed good discrimination in REVEAL between high and low-risk pts. Survival in REVEAL correlated with FPHN equation predictions and was slightly better than predicted (Figure). ![Figure][1] Conclusion: The FPHN equation accurately stratified a matched US population according to risk, suggesting its prognostic generalizability in PAH pts. [1]: pending:yes