Future Directions
In this review we have discussed both retrospective and prospective studies that have looked at HbA1c as predictor of various postoperative adverse outcomes. However, the majority of these studies had small sample sizes, thus from onset limited ability to draw statistically meaningful conclusions about adverse outcomes of inherently low event rates in any cohort; reduced statistical power and possible type II statistical error. Moreover, studies have used different cut-off values of preoperative HbA1c levels, such as 6.5%, 7%, 7.5%, and 8%. Further research should be directed at determining a preoperative cut-off of ‘suboptimal glycaemic control’ for pre‐operative optimisation clinical guidelines of the surgical patient.
A recent retrospective study by Kim et al. of 703 patients with diabetes mellitus who underwent off-pump coronary artery bypass surgery( OPCAB) provides strongest evidence to date of prognostic role of HbA1c 25. The use of composite of postoperative morbidity and mortality (CMM) endpoints (permanent stroke, prolonged ventilation, deep sternal wound infection, renal failure, reoperation, and 30-day mortality) attenuates the prospect of a misleading statistical conclusion by combining adverse events of low incidence. Kim et al. found that incidence of CMM endpoints was greater in patients with HbA1c ≥7.0% (21% vs 15%, P  = .041). Moreover, receiver operator-characteristic curve analysis revealed HbA1c 7.85% as the optimal threshold for CMM endpoints (area under the curve; 0.556, 95% CI, 0.501-0.611, P  = .048). This study has provided rationale for future prospective studies with sufficient power to examine whether postponing cardiac surgery in patients with high preoperative HbA1c levels would improve postoperative outcomes.
Moreover, Kim et al. indicated that high preoperative HbA1c (≥7.0%) level alone, and not the variables related to perioperative glycemic control, was independently associated with adverse outcome in diabetic patients undergoing OPCAB, although high HbA1c levels contributed to greater perioperative glycemic variability25. However, a randomised controlled trial conducted by Bláha et al. suggested that it is cardiac surgery patients with previously undiagnosed diabetes who have the worst prognosis65. Comparable conclusion was suggested in non-cardiac studies 66,67. Recent studies have shown that perioperative intravenous insulin infusion is more frequently administered in known diabetics due to more frequent monitoring of their capillary glucose concentrations 25,68. However, despite more frequently administered insulin in the high HbA1c group, adverse outcomes remained more prevalent in this group compared to normal HbA1c group, thereby further attenuating prognostic role of HbA1c. Nevertheless, optimisation of pre‐operative HbA1c concentrations with a combined intravenous and subcutaneous insulin glucose has been shown to reduce surgical mortality and morbidity in diabetic patients undergoing cardiac surgery69.
Future research should be directed at the determining the optimal level of perioperative glycaemic management and the crucial perioperative period to maintain this HbA1c level. Although there are current ongoing outcome studies currently in this area (e.g. the Optimising Cardiac Surgery outcomes in People with diabetes (OCTOPUS) trial – protocol number HTA16/25/12), there remain few data on the outcomes and effects of intervention on those not known to have diabetes70.