Preoperative plasma levels of soluble tumor necrosis factor receptor type I (sTNF-RI) predicts adverse events in cardiac surgery

Objective: The objective was to estimate the sTNF-RI preoperative measure in the identification of patients with bad outcome and death. Methods: We assessed prospectively sixty-two patients submitted electively to myocardial revascularization with ECC or heart valve surgery. The sTNF-RI levels were determined by the Sandwich-Type ELISA method before anesthetic induction. Clinical, surgical characteristics and sTNF-RI levels were compared among patients with good (group I, n = 46) or bad outcome (group II, n = 16-length of stay in the ICU for over 72 h or death). Results: No difference was found between the verified mortality (6.4%) and the predicted by EuroSCORE (3.0%), p = 0.48. The sTNF-RI levels were higher in group II (1322) than group I (748) p = 0.009 (levels >954, 69% sensitivity and 70% specificity for good outcome, 44% positive predicted value and 85% negative). The sTNF-RI levels were higher in patients who died (1556) versus (759) p = 0.029, (levels >1230, 79% sensitivity, 75% specificity, 20% positive predicted value and 98% negative). In the multivariate logistic regression model sTNF-RI (OR = 1.002, IC95% 1.000-1.005, p = 0.014) and age (OR = 1.083, IC95% 1.010-1.161, p = 0.025) were independently related to the risk of bad outcome. Conclusions: Basal levels of sTNF-RI yield prognostic information in patients who undergo heart surgery.