Composite endpoints are widely used in randomized clinical trials to capture multiple aspects of treatment effect. However, conventional time-to-first-event analyses ignore outcome severity and subsequent events, potentially leading to misleading conclusions. The Win Ratio, introduced by Pocock et al. in 2012 and inspired by the Finkelstein-Schoenfeld test, addresses these limitations by hierarchically prioritizing outcomes according to clinical importance. Despite its increasing use, particularly in cardiovascular trials, guidance on its practical implementation remains fragmented. Motivated by methodological challenges encountered in the design of clinical trials using the Win Ratio as a primary endpoint, this work aims to provide a methodological overview of the Win Ratio and to summarize existing approaches for power and sample size determination in randomized clinical trials.

A broad literature search was conducted without restriction on the publication year, to capture all relevant methodological contributions. Searches were performed in major biomedical and statistical databases and publisher platforms, complemented by institutional library resources, targeted web searches, and manual citation tracking from seminal articles. Search queries were constructed using combinations of the following keywords : “Win Ratio”, “composite endpoint”, “hierarchical endpoint”, “sample size”, “power calculation”, “endpoint”, “Finkelstein-Schoenfeld”, “clinical trial”, “endpoint prioritization”, and “S. J. Pocock”. Articles were eligible if they were peer-reviewed publications describing the methodology, application, or statistical properties of the Win Ratio or hierarchical composite endpoints in randomized clinical trials. Studies lacking sufficient methodological detail, not conducted in a clinical trial context, or not published in English were excluded. In total, 25 articles were included in the review.

The Win Ratio is defined as the ratio of total wins to losses with values greater than one favoring the treatment group. This measure is derived from pairwise comparisons between treatment and control patients, in which hierarchical outcomes are evaluated sequentially in descending order of clinical importance until a win, loss, or tie is determined. Two main methodological variants have been proposed, the unmatched and the matched approaches, each with specific advantages and limitations. Statistical inference is well established for matched designs but remains challenging for the unmatched approach due to the dependence between pairs. Furthermore, sample size determination represents a major methodological challenge for both approaches. Due to the composite and hierarchical nature of the Win Ratio, variance estimation is complex and closed-form solutions are rarely available. Most existing approaches rely on simulation-based methods, which are often computationally demanding, incompletely described, or limited to specific endpoint structures.

The Win Ratio provides a clinically meaningful alternative to traditional composite endpoint analyses by aligning statistical comparisons with outcome severity. However, methods for power and sample size calculation remain heterogeneous and limited in scope. Further methodological development and clearer practical guidance are needed to support robust trial design using the Win Ratio.