Dataset related to the article "Heart and Gut: Gastrointestinal Complications in Cardiac Surgery and Establishing a New Risk Score"

This record contains raw data related to the article “Heart and Gut: Gastrointestinal Complications in Cardiac Surgery and Establishing a New Risk Score”.

Introduction:

Gastrointestinal (GI) complications following cardiac surgery, though rare (0.4-3%), exhibit a disproportionately high mortality rate (13-63%). This study aims to investigate independent predictors of GI complications and mortality, proposing a risk classification score for GI complications.

Methods:

A retrospective case-control study involved 8544 cardiac surgery patients (2005-2019). GI complications, defined as bleeding, ischemia/perforation, occlusion, or pancreatitis, prompted propensity-matching (1:2) with a control group. Multivariable logistic regression assessed independent predictors of a combined post-operative complications endpoint. A risk score for GI complications was created and cross-validated, determining the optimal cutoff for identifying complications. 

Results:

After matching, 162 patients were analyzed; 54 experienced postoperative GI complications (GI group), and 108 did not (Control group). Independent predictors of combined endpoints included GI complications (OR 6.976, CI95% 3.005-16.192), NYHA>2 (OR 2.931, CI95% 1.233–6.972), and pre-operative acute myocardial ischemia (OR 1.958, CI95% 1.015–3.778). The risk score, incorporating mitral valve replacement, sex, coronary disease, diabetes, kidney disease, smoking habit, and cardiac rhythm, demonstrated an area under the ROC curve of 0.829. The optimal cutoff was 0.305, with 69.2% sensitivity and 73.5% specificity. 

Conclusions:

Mortality remains high in patients with GI complications, irrespective of treatment type. The proposed risk score offers a valuable tool for identifying individuals at risk for GI complications, emphasizing the need for vigilant monitoring and timely intervention in this critical subset of post-cardiac surgery patients.