Incidence of hypoglycemia in NICE sugar not so nice after consideration of the treatment period
KeywordsTreatment Period Hypoglycemia Glycaemic Control Major Side Effect Hypoglycaemic Event
Hypoglycaemia is one of the major side effects when establishing glycaemic control in critically ill patients and may contribute to increased mortality rates. Until today, no comparable measure of hypoglycaemia frequency has been suggested. Current assessments depend either on the treatment period or the number of glucose measurements performed per patient. Therefore, we suggest an alternative method to present the hypoglycaemia incidence by number of episodes per patient normalized for a 24-hour treatment period.
Two major studies performing glycaemic control in critically ill patients were chosen for comparison: the Leuven study with 1,200 patients at a medical ICU (605 standard care (STD), 595 tight glycaemic control (TGC); ) and the NICE study with 6,030 patients in a mixed ICU population (3,014 STD, 3,016 TGC; ). For both studies, rates of hypoglycaemia per patient per 24 hours were calculated based on published data.
Mean duration of treatment was 12.5 days for the Leuven study and 4.2 days for the NICE study. APACHE scores were 23 for Leuven and 21 for NICE, respectively. Number of hypoglycaemic events per study population was reported as follows: 19 (3.1%) for STD and 111 (18.7%) for TGC for the Leuven study and 15 (0.5%) for STD and 206 (6.8%) for TGC in the NICE study. When applying the new evaluation method, rates of hypoglycaemia per patient per 24 hours were 0.25% (STD) vs 1.49% (TGC) for Leuven and 0.12% (STD) vs 1.63% (TGC) for NICE, respectively.
The results reveal that the treatment period may have a significant impact on the reported rate of hypoglycemia. We conclude that reporting hypoglycaemic episodes in critically ill patients on a perpatient and per-24-hour basis may be a more adequate tool to accurately describe hypoglycaemia incidence which facilitates the comparison between different studies and study protocols.