Comment on Okabayashi et al. Intensive Versus Intermediate Glucose Control in Surgical Intensive Care Unit Patients. Diabetes Care 2014;37:1516–1524

Hyperglycemia is a major concern in critically ill patients, especially in surgical patients. The optimal target of glucose control in this population and the best method to achieve this goal in order to avoid hypoglycemia is unknown. Okabayashi et al. (1) demonstrated that intensive glucose control (glucose target 80–110 mg/dL) compared with intermediate glucose control (glucose target 140–180 mg/dL) in a selected group of patients who underwent hepato-biliary-pancreatic surgery was feasible using a closed-loop glycemic sensor, without hypoglycemic events and, more important, with a decreased number of surgical site infections and shorter hospitalization length. The result of this trial reinforces the need for new trials using novel technologies for intensive glucose control. However, there are some concerns about this study.

First, the patient eligibility criteria were very strict and preclude generalization of the findings. Patients with seriously impaired function of vital organs were excluded as well as those with body weight loss >10% and distant metastases, and all patients underwent surgery for curative removal of the tumor. The first three criteria are important markers of illness severity, and they might contribute to absence of hypoglycemia in the study. Hypoglycemia may hinder the benefit of tighter glycemic control in other studies (2).

Second, parenteral nutrition (PN) was started as soon as patient was admitted in the intensive care unit (ICU); however, no classic severe undernutrition criteria (weight loss >10% within 6 months, BMI <18 kg/m², or albumin <3 mg/dL) were met in the study to justify the early PN support that was adopted. PN causes hyperglycemia and tight glycemic control may be required as a consequence of this approach. Furthermore, the combined intervention of PN plus intensive glucose control might play an important role in the results as it may reduce glycemic variability and the risk of hypoglycemia. In fact, a meta-analysis of major clinical trials involving glycemic control in ICU suggested that only the subgroup of patients in PN would benefit of tight glycemic control (2). On the other hand, current data do not support use of early PN in well-nourished patients as late initiation of PN was associated with faster recovery and fewer complications, including lower rates of wound infections, in a multicenter trial (3). Of note, surgical site infection rates were higher in both groups in the study by Okabayashi et al. (intensive 4.1% vs. intermediate 9.8%) than wound infections in the multicenter trial (early PN 4.2% vs. late PN 2.7%). The question is when using tight glycemic control with these patients, are we improving outcome or simply treating a side effect of unnecessary treatment?

Last, the study by Okabayashi et al. (1) demonstrated a benefit in intensive insulin therapy in a very specific group of surgical patients in reducing surgical site infection using a closed-loop glycemic sensor. The peripheral blood sample technique used led to some loss of follow-up and may not be feasible in all patients in the ICU. Caution should be taken in applying this device and study results to other surgical patients and, more important, to critically ill patients.