Subgroup effects of ticagrelor treatment strategies: credibility considerations in an individual patient data meta-analysis

This commentary refers to ‘Ticagrelor monotherapy for acute coronary syndrome: an individual patient data meta-analysis of TICO and T-PASS trials’, by Y.-J. Lee et al., https://doi.org/10.1093/eurheartj/ehae249 and the discussion piece ‘Subgroup effects of ticagrelor monotherapy for acute coronary syndrome using individual patient data from the TICO and T-PASS trials’, by Y.-J. Lee et al., https://doi.org/10.1093/eurheartj/ehae685. In an individual patient data meta-analysis, Lee and colleagues1 found that in patients with acute coronary syndrome (ACS), monotherapy with ticagrelor after short-term dual antiplatelet therapy (DAPT) reduced major bleeding events without increasing the risk of ischaemic events compared with 12 months of ticagrelor-based DAPT. Although this finding has potential clinical significance, we are concerned about the credibility of the subgroup effects reported, particularly for patient groups with diabetes, chronic kidney disease, and multivessel disease. Identifying diabetes as a factor for potential treatment effect heterogeneity is meaningful. In both diabetic and non-diabetic patients, ticagrelor monotherapy may offer similar ischaemic protection as DAPT. However, we remain sceptical about the credibility of this subgroup effect. The Instrument for assessing the Credibility of Effect Modification Analysis (ICEMAN) is an assessment tool specifically designed for evaluating effect modification claims in randomized controlled trials and meta-analysis. ICEMAN aids researchers, systematic reviewers, and journal editors in systematically evaluating the credibility of effect modification claims through a series of standardized questions and a guidance manual, ultimately providing an overall rating.2 We used this tool to assess the credibility of the ischemic protection effect in the diabetes subgroup. Our evaluation indicates that the credibility of the ischaemic protection effect in the diabetes subgroup is very low (Supplementary data). This conclusion is based on several key factors: first, the analysis of effect modifications was based on comparisons between trials. Second, the analysis was based on a limited number of trials, only two, which could increase the risk of trial-level bias. Furthermore, the study did not pre-specify the direction of effect modification. Lastly, the study tested a large number of effect modifiers, increasing the risk of finding spurious effects. These concerns apply not only to the diabetes subgroup but also to the analysis of other subgroups in the study. Therefore, we believe there is an urgent need to address the credibility of subgroup effects. Subgroup analyses can have significant implications for clinical practice, and the credibility of their conclusions must be grounded in rigorous methodology and transparent reporting.3,4 Supplementary data are available at European Heart Journal online.