Purpose Postoperative infection, particularly in cardiac surgery, results in significant morbidity, mortality, and healthcare cost. Identification of novel predictors of postoperative infection can target high-risk populations for prophylactic intervention. Methods Steroids in cardiac surgery (SIRS) was a multicentre randomized-controlled trial assessing intraoperative administration of methylprednisone during cardiac surgery, which enrolled 7,507 patients across 80 centres in 18 countries. It demonstrated that administration of steroids had no effect on mortality or major morbidity after cardiac surgery. Our primary objective was to identify risk factors for postoperative surgical site infections using SIRS participants as a cohort. We excluded patients who did not undergo surgery, died intraoperatively, or died within 48 hr of the operation.Patients were assessed for development of ''surgical site infection'' over the first 30 days postoperatively. Using theoretical and previously identified risk factors, we used forward stepwise entry to create a binary logistic regression model. Results Follow-up at 30 days was complete for all patients; 7,406 were included in the cohort. Surgical site infection occurred in 180 (4.8%) and 184 (5.0%) of patients in the placebo and steroid arms respectively. Significant risk factors (P \ 0.05 level) included: diabetes managed with insulin (adjusted odds ratio [aOR]: 1.55; 95% confidence interval [CI] 1.13 to 2.12), oral hypoglycemics (aOR 1.60; 95% CI 1.18 to 2.16), or diet (aOR 1.81; 95% CI 1.16 to 2.83), female sex (aOR 1.34; 95% CI 1.05 to 1.71), renal failure with (aOR 2.03; 95% CI 1.06 to 3.91), and without (aOR 1.50; 95% CI 1.04 to 2.14) dialysis, [ 96 min cardiopulmonary bypass (CPB) time (aOR 1.84; 95% CI 1.44 to 2.35), body mass index