Glycaemic control in Australia and New Zealand before and after the NICE-SUGAR trial: a translational study

Abstract

Introduction: There is no information on the uptake of Intensive Insulin Therapy (IIT) before the Normoglycemia in Intensive Care Evaluation and Surviving Using Glucose Algorithm Regulation (NICE-SUGAR) trial in Australia and New Zealand (ANZ) and on the bi-national response to the trial, yet such data would provide important information on the evolution of ANZ practice in this field. We aimed to study ANZ glycaemic control before and after the publication of the results of the NICE-SUGAR trial.

Methods: We analysed glucose control in critically ill patients across Australia and New Zealand during a two-year period before and after the publication of the NICE-SUGAR study. We used the mean first day glucose (Glu1) (a validated surrogate of ICU glucose control) to define practice. The implementation of an IIT protocol was presumed if the median of Glu₁ measurements was <6.44 mmol/L for a given ICU. Hypoglycaemia was categorised as severe (glucose ≤2.2 mmol/L) or moderate (glucose ≤3.9 mmol/L).

Results: We studied 49 ICUs and 176,505 patients. No ICU practiced IIT before or after NICE-SUGAR. Overall, Glu1 increased from 7.96 (2.95) mmol/L to 8.03 (2.92) mmol/L (P <0.0001) after NICE-SUGAR. Similar increases were noted in all patient subgroups studied (surgical, medical, insulin dependent diabetes mellitus, ICU stay >48/<48 hours). The rate of severe and moderate hypoglycaemia before and after NICE-SUGAR study were 0.59% vs. 0.55% (P =0.33) and 6.62% vs. 5.68% (P <0.0001), respectively. Both crude and adjusted mortalities declined over the study period.

Conclusions: IIT had not been adopted in ANZ before the NICE-SUGAR study and glycaemic control corresponded to that delivered in the control arm of NICE-SUGAR trial. There were only minor changes in practice after the trial toward looser glycaemic control. The rate of moderate hypoglycaemia and mortality decreased along with such changes.

Figure 1

Glu₁, mortality and Glu₁ in subgroups over time before and after NICE-SUGAR. Mean first day glucose (SE, Glu₁, blue circles) in all patients with crude hospital mortality (SE, black squares) in four consecutive six-month periods before (periods 1 to 4) and after (periods 5 to 8) Normoglycemia in Intensive Care Evaluation and Surviving Using Glucose Algorithm Regulation (NICE-SUGAR) study (upper panel). Mean Glu₁ values in patient subgroups (surgical/medical, IDDM/no IDDM, ICU stay >48 h/ICU stay <48 h) in the same six-month periods (lower panel). IDDM, insulin-dependent diabetes mellitus.

Figure 2

Adjusted mortality over time before and after NICE-SUGAR. Odds ratios (95% confidence intervals) for hospital mortality in all patients during four consecutive six-month periods before (periods 1 to 4) and after (periods 5 to 8) NICE-SUGAR study (P <0.0001). NICE-SUGAR, Normoglycemia in Intensive Care Evaluation and Surviving Using Glucose Algorithm Regulation.

Figure 3

Adjusted mortality in quartiles of ICUs according to change in Glu₁. Odds ratios (95% confidence intervals) for mortality in quartiles of ICUs according to changes in mean first day glucose values (Q1 denotes ICUs with the smallest change, Q4 denotes ICUs with the highest increase). P <0.001 for Q1, P ns (>0.01) for Q2 to 4.

Figure 4

Interrupted time series analysis of mortality before and after NICE-SUGAR. Interrupted time series analysis of mortality before and after NICE-SUGAR study. There was no significant vertical shift between pre and post periods (P = 0.29). NICE-SUGAR, Normoglycemia in Intensive Care Evaluation and Surviving Using Glucose Algorithm Regulation.