Factors associated with poor prognosis in very-lowbirth- weight infants

Abstract

Objective: To evaluate predictors of poor outcome, including the CRIB (Clinical Risk Index for Babies) score, in a local population of very-low-birth-weight (VLBW) infants, in order to provide guidelines for selection of these babies for expensive tertiary care. Subjects: Two hundred and thirty-one neonates bom at less than 31 weeks' gestation and/or weighing between 1 001 g and 1 500 g, enrolled prospectively as part of a multicentre study evaluating the CRIB score. Design: Univariate analysis (chi-square/t-tests) and muttivariate analysis (stepwise logistic regression) on the above sample to detennine predictors of poor outcome. Setting: NeonataJ Unit, Johannesburg Hospital. Outcome measures: Death or impainnent (namely oxygen therapy> 28 days, grade 3 or 4 intraventricular haemorrhage, or ventricular enlargement). Results: Poor outcome was predicted by birth weight, lowest oxygen requirement in the first 12 hours (which are two components of the CRIB score), and maximum partial arterial carbon dioxide pressure (PaCO₂) in the first 72 hours. Other factors, including the full CRIB score, were not predictive of outcome. Conclusions: One method of selection of infants for expensive tertiary care is on the basis of predicted outcome. Birth weight remains a reasonable basis for this selection, but the inclusion of other factors, such as oxygen requirement, would improve accuracy. The CRIB score was not a suitable means to select infants in the local context, but may be of value in international comparisons.

S Afr Med J 1996; 86: 1457-1460