Objective: To determine the association between mean glycemia and its variability with perinatal mortality in preterm newborns hospitalized in an intensive care unit (ICU).
Methods: Patients admitted to the ICU within the first 12 hours of life, with birth weight <1500 g, at least three blood glucose measurements/day and lack of insulin treatment were evaluated. Association of mean glycemia and its standard deviation (SD) with death during initial 7 days of life was evaluated. Multivariate logistic regression analysis was performed twice, using continuous glucose concentrations and by means of a quintile-based approach correcting for nonnormal distribution and nonlinear effects.
Results: A total of 95 newborns were enrolled. Eleven patients (11.5%) died during the initial 7 days of life, overall mortality equaled 22%. Multivariate analysis showed that 5 minute Apgar score and SD of glucose concentrations were significantly associated with increased mortality in both models. Odds ratios (ORs) equaled 0.44; 95% confidence interval (95% CI) 0.27-0.74 and OR 1.34; 95% CI 1.03-2.03 for the continuous model and 0.50 95% CI 0.34-0.75 and OR 1.82 95% CI 1.07-3.11 for the quintile-based model. In both cases, mean glycemia was removed during the stepwise model-building procedure.
Conclusions: Higher glycemic variability may be associated with greater odds of perinatal mortality.