Background: Tight glycemic control (TGC) may improve outcomes in hyperglycemic neurosurgical patients. The adoption of TGC has been limited by a lack of adequate data on optimal insulin delivery protocols and serum glucose concentration and by concerns about the risks of hypoglycemia.
Objective: This study was designed as a meta-analysis of outcomes to compare intensive insulin therapy and TGC with conventional insulin therapy and conventional glucose control. The secondary objective was to determine retrospectively whether a particular glucose range correlates with better outcomes.
Methods: Using electronic databases, we retrieved all English language studies published between January 1997 and December 2010 reporting outcomes in neurological and neurosurgical patients as a function of glucose levels and insulin protocols. We conducted a meta-analysis around 4 outcome measures: infection, neurological outcome, hypoglycemia, and mortality. Effect sizes in each study were individually correlated with target intensive insulin therapy glucose levels. Individual studies were assessed for quality by use of the Jadad scale.
Results: Nine studies reporting on 1459 patients met the inclusion criteria. Five were restricted to neurosurgical patients. Four included neurological patients. Compared with conventional glucose control, TGC lowered infection rates (odds ratio, 0.59; 95% confidence interval, 0.47-0.76; P < .001) and yielded better neurological outcomes (odds ratio, 1.72; 95% confidence interval, 1.36-2.16; P < .001). Beneficial effects increased as glucose limits tightened and study quality improved (R > 0.9 for both). TGC resulted in a higher rate of hypoglycemic events (odds ratio, 8.04; 95% confidence interval, 4.85-13.31; P < .001). Mortality was not affected.
Conclusion: TGC reduced infection risk and improved neurological outcome despite increased rates of hypoglycemic events. An optimal target for serum glucose concentrations could not be determined.