Low serum sodium level during cardiopulmonary bypass predicts increased risk of postoperative stroke after coronary artery bypass graft surgery

Objective: Rapid decreases in serum sodium levels are associated with altered mental status, seizures, and coma. During cardiac surgery, serum sodium levels decrease rapidly when cardiopulmonary bypass is initiated because cardiopulmonary bypass causes hemodilution. However, whether this decrease influences neurologic outcome after cardiac surgery remains unclear. We investigated whether the average serum sodium level during cardiopulmonary bypass is independently predictive of postoperative stroke or 30-day all-cause mortality in patients who undergo primary coronary artery bypass grafting. Methods: In a single-institution, retrospective cohort of 2348 consecutive patients who underwent primary, isolated coronary artery bypass grafting, sequential multivariate logistic regression was performed to determine the threshold below which the average serum sodium level during cardiopulmonary bypass independently predicts postoperative stroke or early death. To further test the validity of this threshold and to control for selection bias, stepwise multivariate logistic regression was also performed on propensity score-matched patients (n = 924). Results: An average serum sodium level less than 130 mEq/ L during cardiopulmonary bypass was independently predictive of stroke, both in the entire study cohort (1.44% vs 2.92%; odds ratio, 2.09; 95% confidence interval, 1.1-4.1; P - .03) and in the propensity-matched patients (0.9% vs 3.0%; odds ratio, 4.1; 95% confidence interval, 1.3-13.0; P = .02). The average serum sodium level during cardiopulmonary bypass was not independently associated with early death, regardless of what threshold value was used. Conclusions: An average serum sodium level of less than 130 mEq/ L during cardiopulmonary bypass is independently associated with an increased risk of postoperative stroke in patients who undergo primary coronary artery bypass grafting.