Kim Annink

78 Chapter 4 Statistical analyses SPSS Version 21 was used for statistical analysis (IBM corp., Armonk NY, United States). Differences in baseline characteristics between the two cohorts were calculated using the independent t-test or Mann-Whitney-U test for continuous variables and the χ 2 test for categorical variables. Univariate logistic regression was performed with the CUS items as independent variables and outcome as dependent variable. Non-significant items were excluded from further analysis. The sum of white matter involvement, of deep grey matter involvement and a Doppler ultrasound resistance index (RI) of a cerebral artery ≤0.55 were calculated and included in backward multivariable logistic regression analysis. Variables with a p-value <0.05 were entered in the model and those with a p-value ≥0.1 deleted. The inter-observer variability and the correlation with MRI were calculated using Spearman rank correlation test. Predictive values and ROC curves were determined for cohorts I and II per cut-off value. A p-value <0.05 was considered statistically significant. RESULTS Study population Between January 2008 and July 2014, 145 infants with HIE were treated with hypothermia in the University Medical Center Utrecht. In total, 83 infants were included in cohort I. Infants were excluded because of genetic or congenital abnormalities (n=9), preterm birth <36 weeks (n=12), because one or both of the CUS were not present (n=27), only a few images of an examination were saved (n=2), the quality was insufficient because of suboptimal settings (n=6) or because of missing outcome data (n=6). In the Erasmus Medical Center, 69 infants with HIE were treated with hypothermia in this period and 35 newborns were included in cohort II. Infants were excluded because of genetic or congenital abnormalities (n=9), preterm birth <36 weeks (n=2), diagnosis of arterial ischemic stroke (n=2), because the CUS between day three and seven was not present (n=15) or because no follow-up data were available (n=6). The incidence of death due to redirection of care between infants who were included and infants who were excluded from the study did not differ significantly.