Consecutive or selectively included high bleeding risk patients in the MASTER DAPT screening log and trial

Aims Screening logs have the potential to appraise the actual prevalence and distribution of predefined patient subsets, avoiding selection biases, which are inevitably and potentially present in randomised trials and real-world registries, respectively. We aimed to assess the prevalence of high bleeding risk (HBR) characteristics in the real world and the external validity of the MASTER DAPT trial. Methods and results All consecutive patients who underwent percutaneous coronary intervention (PCI) for at least two consecutive weeks across 65 sites participating in the trial were entered into a screening log. Of 2,847 consecutive patients, 1,098 (38.6 %) were HBR and 109 (9.9 %) consented for trial participation. PRECISE-DAPT score ≥ 25 was the most frequent HBR feature, followed by advanced age, use of oral anticoagulation (OAC) and anaemia. Compared with consecutive HBR patients, consenting patients were older (≥ 75 years: 69 % versus 62 %, absolute standardized difference [SD] 0.16), more frequently male (78 % versus 71 %, absolute SD 0.18), had higher use of OAC (38 % versus 20 %, absolute SD 0.39), treatment with steroids or nonsteroidal anti-inflammatory drugs (10 % versus 5 %, SD 0.16), and prior cerebrovascular events (10 % versus 6 %, absolute SD 0.18) but lower PRECISE DAPT score ≥ 25 (54 % versus 66 %, absolute SD 0.24). Conclusions The HBR criteria distribution differed between consecutive versus selectively included HBR patients, suggesting the existence of selection biases in the trial population.