Methodological Considerations on the Optimal Duration of Dual Antiplatelet Therapy in Patients with High Residual Inflammatory Risk [Letter]

We read with great interest the article by Wang et al, 1 which addresses an important clinical question regarding the optimal duration of dual antiplatelet therapy in patients with high residual inflammatory risk after percutaneous coronary intervention.The authors are to be commended for their efforts in exploring this clinically relevant topic using a largescale prospective registry.Their findings suggesting potential benefits of prolonged DAPT are thought-provoking and offer valuable insights for clinical practice.We would like to offer a few methodological considerations that may serve as constructive suggestions for further strengthening the interpretation of these results and facilitating their translation into clinical decision-making.Our primary consideration relates to the timing of patient classification.The study stratified patients into prolonged (>12 months) and conventional (≤12 months) DAPT groups based on actual medication duration after the 12-month landmark.This classification approach requires that patients in the prolonged group survived and remained event-free through the first 12 months to be eligible for assignment, whereas patients who experienced early adverse events-which may have prompted DAPT discontinuation-were assigned to the conventional group.This phenomenon, known as immortal time bias, has been shown to substantially influence treatment effect estimates in observational cohort studies, as discussed by Lévesque et al in their methodological work. 2 We noticed that the authors mention "patients free of events at 12 months" in Table 3, yet the baseline characteristics (Table 1) and primary analyses appear to use the full cohort, which may create some ambiguity regarding the time origin.It might be worthwhile to consider a landmark analysis starting at 12 months, with follow-up beginning only among patients event-free at that time point, as this approach could offer a more balanced comparison between the two treatment strategies.Second, we would like to raise the issue of competing risks for further consideration.The primary composite endpoint included all-cause death, which precludes the occurrence of other components such as myocardial infarction or stroke.Standard Kaplan-Meier estimation treats death as censoring, which may affect the estimation of cumulative incidence of non-fatal events.Austin et al have pointed out that accounting for competing risks can substantially influence event incidence estimates in cardiovascular cohorts.3 The application of Fine-Gray competing risk regression might provide complementary cumulative incidence function estimates and could be considered as a supplementary analysis to enhance the robustness of the findings.Third, the hazard ratios warrant thoughtful interpretation.The adjusted hazard ratio of 0.182 for all-cause death or myocardial infarction suggests a substantial risk reduction-an effect size that appears quite large for an observational study.Notably, zero deaths were recorded in the prolonged DAPT group (0/2,440) versus 32 deaths (3.0%) in the conventional group.When one group has zero events in a Cox model, the hazard ratio estimate may exhibit certain