ABSTRACT
This study aimed to assess whether a relationship exists between hyperglycaemia and outcome in a mixed cohort of critically ill patients with nosocomial bloodstream infection (BSI), and to evaluate patterns of blood glucose levels between survivors and non-survivors. A historical observational cohort study was conducted in the intensive care unit (ICU) of a tertiary care referral centre. One-hundred-and-thirty patients with a microbiologically documented ICU-acquired BSI (period 2003 to 2004) were included. For the study, morning blood glucose levels were evaluated from one day prior until five days after onset of BSI. The contribution of hyperglycaemia, divided in three subgroups (> or = 150 mg/dl, > or = 175 mg/dl and > or = 200 mg/dl), to in-hospital mortality was estimated by logistic regression. In-hospital mortality was 36.2%. Over the seven study days, no differences were found in daily morning blood glucose levels between survivors (n = 83) and non-survivors (n = 47). Nevertheless, the trend of blood glucose levels upon onset of BSI showed a remarkable increase in the non-survivors, whereas it decreased in the survivors. Hyperglycaemia (> or = 175 mg/dl and > or = 200 mg/dl) was observed more often among the non-survivors. Multivariate logistic regression showed that APACHE II (P = 0.002), antibiotic resistance (P = 0.004) and hyperglycaemia (> or = 175 mg/dl) upon onset of BSI (P = 0.017) were independently associated with in-hospital mortality, whereas a history of diabetes (P = 0.041) was associated with better outcome. Hyperglycaemia (> or = 175 mg/dl) upon onset of ICU-acquired BSI is associated with worse outcome in a heterogeneous ICU population. Patterns of morning blood glucose levels have only limited value in the prediction of the individual course.
