Establishment of Predictive Models for Nonocclusive Mesenteric Ischemia Comparing 8,296 Control with 452 Study Patients

Abstract

The aim of this study was to develop clinical preoperative, intraoperative, and postoperative scores for early identification of patients who are at risk of nonocclusive mesenteric ischemia (NOMI). Design: A retrospective analysis. Setting: Single center. Participants: From January 2008 to December 2014, all patients from the Department of Thoracic and Cardiovascular Surgery were included on the basis of the hospital database. Interventions: All mesenteric angiographically identified NOMI patients were compared with non-NOMI patients. Measurements and Main Results: The study population of 8,748 patients was randomized into a cohort for developing the scores (non-NOMI 4,214 and NOMI 235) and a cohort for control (non-NOMI 4,082 and NOMI 217). Risk factors were identified using forward and backward Wald test and were included in the predictive scores for the occurrence of NOMI. C statistic showed that the scores had a high discrimination for the prediction of NOMI preoperatively (C statistic 0.79; p < 0.001), intraoperatively (C statistic 0.68; p < 0.001), and postoperatively (C statistic 0.85; p < 0.001). A combination of the preoperative, intraoperative, and postoperative risk scores demonstrated the highest discrimination (C statistic 0.87; p < 0.001). The combined score included the following risk factors: renal insufficiency (preoperative); use of cardiopulmonary bypass and intra-aortic balloon pump support (intraoperative); and reexploration for bleeding, renal replacement therapy, and packed red blood cells ≥ 4 units (postoperative). The results were similar in the control group. Conclusions: These scores could be useful to identify patients at risk for NOMI and promote a rapid diagnosis and therapy.