Recurrent Event Analysis: Andersen–Gill and Lin–Wei–Yang–Ying Models for Capturing Total Disease Burden in Cardiovascular Trials (Motivated by the JACC State-of-the-Art Review on Recurrent Events in Cardiovascular Research by Gregson et al.)

Time-to-first event analysis remains the dominant analytical framework in cardiovascular clinical trials, despite the fact that many chronic cardiovascular conditions are characterized by frequent recurrent nonfatal events such as hospitalizations. Restricting inference to the first event may therefore underestimate treatment effects and fail to reflect total disease burden experienced by patients. Recurrent event methods, particularly the Andersen–Gill and Lin–Wei–Yang–Ying models, extend traditional survival analysis by incorporating all qualifying events over follow-up while preserving the randomized comparison and familiar hazard ratio interpretation.

Motivated by a recent JACC state-of-the-art review, this report provides a structured overview of the rationale, assumptions, and practical implementation of recurrent event analysis in cardiovascular trials. We discuss when recurrent event methods are most appropriate, outline essential dataset characteristics required for valid inference, and illustrate their application using large heart failure trial examples. Particular emphasis is placed on the role of robust variance estimation in addressing within-patient event clustering and on the distinction between first-event and total-event estimands.

Through detailed tables and figures, we demonstrate how recurrent event analyses offer complementary clinical insights beyond conventional time-to-first event approaches, especially in trials targeting reductions in ongoing morbidity. While recurrent event methods do not consistently improve statistical power, they provide a more clinically aligned assessment of treatment benefit in chronic disease settings. Thoughtful application of these methods can enhance interpretation, reporting, and decision-making in cardiovascular research.