Perioperative glycaemic control for diabetic patients undergoing surgery.
Patients with diabetes mellitus are at increased risk of postoperative complications. Data from randomised clinical trials and meta-analyses point to a potential benefit of intensive glycaemic control, targeting near-normal blood glucose, in patients with hyperglycaemia (with and without diabetes mellitus) being submitted to surgical procedures. However, there is limited evidence concerning this question in patients with diabetes mellitus undergoing surgery. To assess the effects of perioperative glycaemic control for diabetic patients undergoing surgery. Trials were obtained from searches of The Cochrane Library, MEDLINE, EMBASE, LILACS, CINAHL and ISIS (all up to February 2012). We included randomised controlled clinical trials that prespecified different targets of perioperative glycaemic control (intensive versus conventional or standard care) Two authors independently extracted data and assessed risk of bias. We summarised studies using meta-analysis or descriptive methods. Twelve trials randomised 694 diabetic participants to intensive control and 709 diabetic participants to conventional glycaemic control. The duration of the intervention ranged from just the duration of the surgical procedure up to 90 days. The number of participants ranged from 13 to 421, and the mean age was 64 years. Comparison of intensive with conventional glycaemic control demonstrated the following results for our predefined primary outcomes: analysis restricted to studies with low or unclear detection or attrition bias for infectious complications showed a risk ratio (RR) of 0.46 (95% confidence interval (CI) 0.18 to 1.18), P = 0.11, 627 participants, eight trials, moderate quality of the evidence (grading of recommendations assessment, development and evaluation - (GRADE)). Evaluation of death from any cause revealed a RR of 1.19 (95% CI 0.89 to 1.59), P = 0.24, 1365 participants, 11 trials, high quality of the evidence (GRADE).On the basis of a posthoc analysis, there is the hypothesis that intensive glycaemic control may increase the risk of hypoglycaemic episodes if longer-term outcome measures are analysed (RR 6.92, 95% CI 2.04 to 23.41), P = 0.002, 724 patients, three trials, low quality of the evidence (GRADE). Analysis of our predefined secondary outcomes revealed the following findings: cardiovascular events had a RR of 1.03 (95% CI 0.21 to 5.13), P = 0.97, 682 participants, six trials, moderate quality of the evidence (GRADE) when comparing the two treatment modalities; and renal failure also did not show significant differences between intensive and regular glucose control (RR 0.61, 95% CI 0.34 to 1.08), P = 0.09, 434 participants, two trials, moderate quality of the evidence (GRADE). We did not meta-analyse length of hospital stay and intensive care unit (ICU) stay due to substantial unexplained heterogeneity. Mean differences between intensive and regular glucose control groups ranged from -1.7 days to 2.1 days for ICU stay and between -8 days to 3.7 days for hospital stay (moderate quality of the evidence (GRADE)). One trial assessed health-related quality of life in 12/37 (32.4%) of participants in the intervention group and 13/44 (29.5%) of participants in the control group, and did not show an important difference (low quality of the evidence (GRADE)) in the measured physical health composite score of the short-form 12-item health survey (SF-12). None of the trials examined the effects of the interventions in terms of costs. The included trials did not demonstrate significant differences for most of the outcomes when targeting intensive perioperative glycaemic control compared with conventional glycaemic control in patients with diabetes mellitus. However, posthoc analysis indicated that intensive glycaemic control was associated with an increased number of patients experiencing hypoglycaemic episodes. Intensive glycaemic control protocols with near-normal blood glucose targets for patients with diabetes mellitus undergoing surgical procedures are currently not supported by an adequate scientific basis. We suggest that insulin treatment regimens, patient- and health-system relevant outcomes, and time points for outcome measures should be defined in a thorough and uniform way in future studies.