Effect of blood thiamine concentrations on mortality: Influence of nutritional status.

Abstract

To test the hypothesis that low blood thiamine concentrations in malnourished critically ill children are associated with higher risk of 30-d mortality. Prospective cohort study in 202 consecutively admitted children who had whole blood thiamine concentrations assessed on admission and on days 5 and 10 of intensive care unit (ICU) stay. The primary outcome variable was 30-d mortality. Mean blood thiamine concentrations within the first 10 d of ICU stay, age, sex, malnutrition, C-reactive protein concentration, Pediatric Index of Mortality 2 score, and severe sepsis/septic shock were the main potential exposure variables for outcome. Thiamine deficiency was detected in 61 patients within the first 10 d of ICU stay, 57 cases being diagnosed on admission and 4 new cases on the 5th d. C-reactive protein concentration during ICU stay was independently associated with decreased blood thiamine concentrations (P = 0.003). There was a significant statistical interaction between mean blood thiamine concentrations and malnutrition on the risk of 30-d mortality (P = 0.002). In an adjusted analysis, mean blood thiamine concentrations were associated with a decrease in the mortality risk in malnourished patients (odds ratio = 0.85; 95% confidence interval [CI]: 0.73-0.98; P = 0.029), whereas no effect was noted for well-nourished patients (odds ratio: 1.03; 95% CI: 0.94-1.13; P = 0.46). Blood thiamine concentration probably has a protective effect on the risk of 30-d mortality in malnourished patients but not in those who were well nourished.