Perioperative glycemic control and postoperative cognitive dysfunction: an expression of caution

Abstract

To Editor: If we are to try to decrease incidence of postop- erative cognitive dysfunction (POCD) by aggressively treating hyperglycemia perioperatively, we first need to be sure we are not harming our patients while we try to attain this laudable goal. In his informative edi- torial about hyperglycemia and POCD, Dr. Grocott asserts that the best clinical studies in this field of research to date are observational and thus define only an association of hyperglycemia to POCD. 1 Assum- ing that stroke is accepted as a form of POCD, a recent prospective, randomized, controlled trial (not observational data) has demonstrated a higher rate of stroke in patients treated with intensive insulin therapy during cardiac surgery. 2 In this study, 4% of patients in intensive insulin arm had a stroke, vs 1% in conventional treatment arm. Rather than highlighting some potential flaws in these links between hypergly - cemia and other adverse outcomes, 1 this study raises disturbing questions about whether anesthesiologists are too quickly adopting a treatment that we do not yet fully understand. In an editorial concerning this trial, Van den Berghe writes … we should regard tight glucose control during cardiac surgery as experi- mental and confine its use to clinical trials. 3 Although it is true that Van den Berghe's trial in 2001 4 (in mostly cardiac surgical patients) demon- strated a mortality reduction in group of patients treated with tight glucose control in intensive care unit (ICU), four subsequent prospective randomized controlled trials in critically ill patients have not shown any mortality benefit of intensive insulin therapy in ICU, but have demonstrated a fourfold to tenfold higher incidence of hypoglycemic episodes. 5-8 Two of these trials were stopped early because of safety con- cerns in intensive insulin arms, 6,7 and use of tight glucose control in critical care literature is currently being actively debated. 9 (Hopefully, with publication of NICE-SUGAR trial, A expected to enroll 6,100 patients in over 35 ICUs throughout Australia, New Zealand, Canada, and USA, we will have more information - at least in ICU setting). Extrapolating data from critical care domain into intraoperative domain must be done with due caution. Anesthesiologists should be committed to providing evidence that any novel therapies proposed as part of intraoperative/perioperative management will ensure a measurable benefit to our patients, with an acceptable risk profile. The animal data on glycemic control are promising, and while pathophysiologic link is compelling, clinical data is not incontrovert- ible in this particular case. Further research will clarify role of commencing tight glycemic control intra- operatively.