Background and aims
Guidelines recommend a moderate glucose control target of 7.8-10mmol/L for patients with hyperglycemia recovering in the intensive care unit(ICU) after cardiac surgery. It is unclear whether the same target can be used for hyperglycemic patients with and without diabetes. 

Methods
We studied 9,987 patients including 5,830 patients without diabetes and 4,157 patients with diabetes who underwent coronary artery bypass grafting (CABG) at a single hospital in China from 2011-2014. Patients were stratified into three groups according to their mean blood glucose (MBG) levels during the ICU admission: 1) strict-control, MBG <7.8 mmol/L; 2) moderate-control, MBG 7.8-10.0 mmol/L; and 3) liberal-control, MBG ≥10mmol/L. The relationship between glycemic control and  the composite outcome which including in-hospital mortality and major complication (sum of acute myocardial infarction, stroke and renal failure during the hospitalization after CABG surgery) was investigated by using multivariate logistic regression model adjusting for cardiovascular risk factors.

Results
Compared with patients without diabetes, patients with diabetes had higher systolic blood pressure, fasting blood glucose at admission, and HbA1c; and more frequently had a history of hypertension, dyslipidemia, myocardial infarction, and three-vessel disease (P<0.01 for all comparisons) (Table 1).The rate of composite outcome was 6.1% in the whole population, and occurred more frequently in patients with diabetes than in those without diabetes (7.6% vs 5.0%; P<0.001). The in-hospital mortality and major complication were also more common for patients with diabetes.
The rate of composite outcome was significantly higher in liberal-control group, while significantly lower in strict-control group, when compared with moderate-control group among patients without diabetes. In diabetes cohort, the rate of composite outcome was significantly higher in both liberal-control and strict-control groups. Among patients without diabetes, in-hospital mortality in liberal-control group was significantly higher than moderate-control group, but patients with diabetes had higher rate in strict-control group. The rate of major complication was consistent with composite outcome in the patients without diabetes, while in the diabetes group, liberal-control group showed significantly higher rate of complication comparing with moderate-control group (Figure 1) .
After adjustment of the multiple risk factors, patients with diabetes in strict-control group had higher risk of in-hospital mortality compared with patients in moderate-control group (odds ratio [OR], 11.04; 95% confidence interval [CI], 4.98-24.46; P<0.001); but among patients without diabetes, liberal-control group had higher risk of death (OR, 7.14; 95% CI, 2.34-21.80; P<0.001). The risk of major complication was significantly lower in strict-control group comparing with moderate-control group in patients without diabetes (OR, 0.72; 95% CI, 0.53 - 0.99; P=0.048), in contrast to patients with diabetes, there was no difference in glycemic control groups (OR, 1.17; 95% CI, 0.68-2.00; P=0.57) (Figure 2).

Conclusions
Hyperglycemia was independently associated with in-hospital mortality and major complication among CABG patients during the postoperative hospitalization. Adverse outcomes were often lower for those with moderate glycemic control (7.8-10.0mmol/L). However, strict glucose control targeted to <7.8mmol/L was associated with increased in-hospital mortality among patients with diabetes, but decreased occurrence of a major complication among patients without diabetes. The optimal glycemic control target after CABG may differ by diabetes status.

Impact of perioperative glycemic control on post heart surgery mortality and complications in 10000 Chinese patients