Continuous postoperative insulin infusion reduces deep sternal wound infection in patients with diabetes undergoing coronary artery bypass grafting using bilateral internal mammary artery grafts: a propensity-matched analysis

Deep sternal wound infection (DSWI), especially in patients with diabetes mellitus (DM), is a major concern after coronary artery bypass grafting (CABG) with bilateral internal mammary artery (BIMA) grafts. We evaluated the risk of DSWI and other clinical outcomes between continuous insulin infusion therapy (CIT) and insulin sliding scale therapy (IST) in a cohort of DM patients who underwent CABG with BIMA. The clinical records of DM patients who underwent isolated CABG with BIMA were retrospectively reviewed. The study population consisted of 95 patients who received CIT and 126 patients who received IST. Furthermore, a one-to-one matched analysis based on estimated propensity scores for patients who received CIT or IST yielded two groups comprising 58 patients each. The proportion of patients with DSWI, overall survival rates and major adverse cardiac events were compared between the two groups in the overall and the propensity-matching cohort. The prevalence of DSWI requiring debridement and closure was significantly reduced in the CIT group compared with that in the IST group [1/95 (1.1%) vs 9/126 (7.1%), P = 0.031]; these results were not attenuated even after propensity-matching analysis [0/58 (0%) vs 6/58 (10.3%), P = 0.031]. The mean preoperative glucose levels were similar between the two groups (157.5 +/- 54.6 vs 176.1 +/- +/- 70 mg/dl, P = 0.063), whereas the mean glucose values were significantly lower on the first and second operative days in the CIT group than in the IST group (132.9 +/- 44.1 vs 197.8 +/- 78.6 mg/dl, P < 0.0001; 153.5 +/- 58.8 vs 199.6 +/- 89.1 mg/dl, P < 0.0001, respectively). The glucose variability levels within 24 h postoperatively were significantly higher in the IST group (46.1 +/- 19.4 vs 66.4 +/- 26.8 mg/dl, P < 0.0001). The 30-day and 1-year survival rates were similar between the two groups (100 vs 99.2%, P = 0.384; 96.6 vs 94.4%, P = 0.454). No results were changed in the propensity-matching models. The CIT approach reduced the variability in glucose concentration and resulted in fewer instances of DSWI after CABG with BIMA grafts.