Predicting Gastrointestinal Complications After Cardiac Surgery From Mean Arterial Blood Pressure Under a Data‐Driven Threshold for Hypotension: Protocol for an Advanced Database Study

Background Hypotension during cardiopulmonary bypass may increase the risk of gastrointestinal complications due to insufficient splanchnic autoregulatory capacity. We hypothesize that both duration and level of blood pressure below a data‐driven threshold during cardiopulmonary bypass are associated with the risk of gastrointestinal complications. Methods Patients who underwent cardiac surgery at Copenhagen University Hospital, Rigshospitalet, during 2018–2022 will be included if they underwent on‐pump valve surgery, coronary artery bypass grafting, and/or surgery on the ascending aorta. The primary outcome is major gastrointestinal complications within 30 days, defined as acute intestinal ischemia, abdominal surgery, bowel perforation, or gastrointestinal hemorrhage. Secondary endpoints are serious gastrointestinal complications, delayed recovery of bowel function, cholinesterase inhibitor administration, and 30‐day mortality. The exposure is the exposure to lower blood pressure, calculated as the product of the duration and level of blood pressure below a data‐driven threshold. To identify the optimal threshold, we split the population in a training (75%) and validation cohort (25%). In the training cohort we will identify the threshold with the highest area under the receiver operating curve for the primary outcome. The selected threshold will then be applied to quantify the exposure to lower blood pressure for each patient. Patients will be grouped using a modified Youden's index (weighing sensitivity 80%) as a cutoff for high versus low exposure to lower blood pressure, and association with the study endpoints will be determined in the validation cohort. Conclusion Under the assumption of pressure‐dependent perfusion outside of an autoregulatory plateau, the exposure to blood pressure below a threshold might reflect the risk of gastrointestinal complications more accurately than point measures or averages. The results from this study may lead to relevant hypotheses for prospective testing.