Our study investigated the relationship between both spontaneous and iatrogenic hypoglycaemia and mortality.

A total of 200,859 patients from the eICU Collaborative Research Database were grouped by whether they had registered episodes of hypoglycaemia or not. Patients with hypoglycaemia were then further divided into subgroups according to type of hypoglycaemia—whether spontaneous or iatrogenic. Spontaneous hypoglycaemia during an ICU stay was defined as one or more registered blood glucose measurements < 70 mg/dL (3.9 mmol/L) with no evidence of insulin therapy.

Evidence of at least one hypoglycaemic episode during ICU hospitalization was associated with a significant increase in mortality: the observed mortality rate in such patients was 15.6% compared with 8% in patients without hypoglycaemia (P < 0.001). Multivariate logistic regression analysis was performed with APACHE Scores, hypoglycaemia and baseline data (age, gender, and ethnicity). Spontaneous hypoglycaemia remained a statistically significant predictor of mortality with an adjusted odds ratio (OR) of 1.61 (95% CI: 1.38–1.88; P < 0.001), whereas iatrogenic hypoglycaemia was not a significant predictor with an adjusted OR of 0.97 (95% CI: 0.82–1.14; P = 0.71).

Spontaneous hypoglycaemia observed in ICU patients was associated with increased mortality and increased length of ICU stay. Although the present study, given its observational design, cannot provide a definitive answer, the clear difference between spontaneous and iatrogenic hypoglycaemia does not support a causal relationship between (short-lasting) hypoglycaemia and adverse outcomes, but instead indicates that (short-lasting) hypoglycaemia may be a marker of illness severity.