Role of preoperative laboratory values in predicting mortality after elective CABG: a single center report including 6,436 patients

Objective: Recent literature has pointed out the predictive role of many preoperative laboratory values on mortality after CABG. Despite the innovative contribution of those studies, literature lacks studies regarding the influence of those preoperative values when analyzed all together. In the current study we sought to investigate the influence of numerous preoperative laboratory values as independent predictors for in-hospital mortality after elective CABG.

Methods: Between January 2000 and January 2011 a total of 6436 patients (77.7% men, 22.3% women, aged 21-95 years with a mean ± SD of 67.1 ± 9.5 years) underwent elective CABG operation at our department. To identify a discriminative point for every preoperative value for in-hospital mortality Receiver-operating characteristic (ROC) was performed. Significant variables were dichotomized and entered in multivariante logistic regression model.

Results: Overall in-hospital mortality was 3.05%. Nineteen out off 33 available variables were significantly predictive of in-hospital mortality and were retained for the multivariate analysis. After dichotomization, multivariate regression analysis included the following laboratory values in the predictive model: Creatinine > 1.4 mg/dl (p < 0.025, OR 1.71, CI 1.07 to 2.71); Urea > 20 mg/dl (p = 0.042 OR 1.53 CI 1.02 to 2.30); Hemoglobin  < 13.5 g/dl (p < 0.001, OR 1.98, CI 1.34 to 2.96); LDH > 199 U/L (p = 0.045 OR 1.52 CI 1.01 to 2.26); CK  < 64 U/L (p = 0.002 OR 0.53 CI 0.34 to 0.80); GOT > 28 U/L (p = 0.008 OR 1.79 CI 1.16 to 2.72); total serum protein  < 7g/dl (p = 0,05, OR 1.45, CI 1.00 to 2.11).

Conclusion: Preoperative laboratory values represent strong independent predictors of in-hospital mortality after elective CABG. This constitutes a useful and easily accessible information for the surgeon in order to quantify the preoperative risk.