Practice changes in blood glucose management following open heart surgery: from a prospective randomized study to everyday practice.

Abstract

Targeted blood glucose (BG) levels following cardiac surgery continues to be debated. According to the Society of Thoracic Surgeons (STS) guidelines, BG should be kept <180 mg/dl following cardiac surgery. However, our practice and others shifted to a stricter BG control (90-110 mg/dl) based on data suggesting an association with improved outcome. Recently, we conducted a randomized control study that demonstrated no added value to stricter control over liberal control (120-180 mg/dl). As a result, we shifted our management accordingly. The purpose of this study was to evaluate the impact that this change to a more liberal BG management (BGM) had on patient outcomes at our centre. BGM was changed in June 2011 from strict (90-110 mg/dl) to liberal (120-180 mg/dl). Insulin drips, managed through a computerized algorithm, controlled BG for the first 72 h post surgery. Consecutive cardiac surgery patients operated on throughout 1 year prior to BGM change (n = 934) were propensity score matched to patients operated on throughout 1 year after the change (n = 927). After matching, there were 846 patient pairs. There was no difference between cohorts for length of stay and perioperative complications, and both cohorts achieved acceptable outcomes. Incidence of perioperative renal failure (P = 0.02) and renal failure requiring dialysis (P = 0.004) were better for the cohort with liberal BGM. One-year cumulative survival did not differ between cohorts (log-rank = 0.70, P = 0.40). Implementation of glycaemic control of 120-180 mg/dl into clinical practice was not associated with increased morbidity. The present results confirm our prior findings that a more liberal glycaemic control strategy to maintain BG is equal to a stricter target range. These findings are important for patient care and demonstrate the safety and efficacy of practice change for all patients following a successful randomized controlled trial.