The effect of insulin therapy algorithms on blood glucose levels in patients following cardiac surgery: A systematic review.

BACKGROUND A number of insulin infusion algorithms, with varying methods and effects, have been developed to guide the management of postoperative hyperglycemia in critically ill patients. OBJECTIVES The objective of this study was to investigate the effect of insulin therapy algorithms on blood glucose levels in patients in critical care environments following cardiac surgery. INCLUSION CRITERIA Types of participants: Adult patients aged 18 or older who were admitted to a critical care environment after cardiac surgery and who received insulin therapy for glycemic control during the acute postoperative phase of their admission. Types of intervention: The intervention of interest was continuous intravenous insulin therapy. Types of studies: Experimental study designs including randomized controlled trials, non-randomized controlled trials and controlled before and after studies published in the English language were included in this review. Types of outcomes: Primary outcomes of interest included objective measures of glycemic control and secondary outcomes of interest included the incidence of adverse events. SEARCH STRATEGY The search aimed to find both published and unpublished studies through electronic databases, reference lists, key reports and the World Wide Web. An extensive search was undertaken for the following databases: Medline, CINAHL, PubMed, Embase, Scopus, the Cochrane Library, Dare, Social Science Index, ProQuest, and MedNar. Databases were searched up to March 2014. METHODOLOGICAL QUALITY All studies selected were assessed by two reviewers independently for methodological quality prior to inclusion in the review using the standardized Joanna Briggs Institute Meta-Analysis of Statistics Assessment and Review Instrument critical appraisal tool. Any disagreements that arose between the reviewers were resolved through discussion. DATA COLLECTION Quantitative data was extracted from papers included in the review using the standardized data extraction tool from Joanna Briggs Institute Meta-Analysis of Statistics Assessment and Review Instrument. All results were subject to double data entry. DATA SYNTHESIS Statistical pooling of the data in meta-analysis was carried out using Review Manager meta-analysis software where possible. Odds Ratios, Relative Risks, Weighted Means Differences and their 95% Confidence Interval were calculated where appropriate. Where statistical pooling was not possible due to the heterogeneity of the studies the findings are presented in narrative form including tables and figures. RESULTS Thirteen studies were included in the final review. Pooled data demonstrated significantly improved mean blood glucose levels (Weighted Means Differences -27.24, 95% Confidence Interval: -27.77 - -26.72), p < 0.00001 and achievement of target blood glucose levels range (Relative Risks 1.43, 95% Confidence Interval: 1.18 - 1.72), p = 0.0002, among participants who received the paper nomogram continuous intravenous insulin method of glucose control compared to the bolus regime group. Studies that investigated paper nomogram directed continuous intravenous insulin compared to computer calculator directed continuous intravenous insulin demonstrated a statistically significant improved mean blood glucose levels (Weighted Means Differences -23.74, 95% Confidence Interval: -24.45 - -23.02), p <0.00001 and higher percentage of time in which glucose levels were within the target range in the computer calculator group. A significantly lower incidence (p < 0.05) of hyperglycemia was observed in the computerized calculator directed CII group (1.3 ± 1.2%) compared to the paper nomogram CII (6.5 ± 2%). CONCLUSIONS There is evidence of a benefit of continuous insulin infusions for patients following cardiac surgery with the computer calculator CII method achieving optimal glycemic control. IMPLICATIONS FOR PRACTICE The implementation of a computer calculator CII algorithm allows for individualized patient treatment. However, given the limited availability of the computer based algorithm, the implementation of a paper based algorithm would be justified. IMPLICATIONS FOR RESEARCH Future studies to compare the methods of computer calculator directed CII to paper nomogram directed CII should be undertaken using more rigorous research designs such as randomized controlled trials with sufficiently powered sample sizes. Definitions for both hypoglycemia and hyperglycemia should be standardized.