Usefulness of different risk factor associations in predicting admissions due to respiratory syncytial virus in premature newborns of 32 to 35 weeks gestation in Spain

Aim: The aim of the study was to evaluate the risk factor associations for respiratory syncytial virus (RSV) hospitalization in preterm infants from 32 to 35 weeks gestation, treated during two consecutive RSV seasons in Spain. Patients and methods: A database (FLIP-2) was used after excluding the infants who received prophylactic palivizumab. A total of 193 RSV+ admissions and 4568 non-hospitalized children were studied. The risk factors analyzed were: chronological age < = 10 weeks at start of RSV season or to be born during the first 10 weeks of the season; school-age siblings or daycare attendance; mother smoking during pregnancy; mate gender; breastfeeding < = 2 months; > = 4 adults at home; history of wheezing; small for gestational age; > = 2 smokers at home. Results: Logistic regression model included the first four previously mentioned risk factors as independently significant variables, with R-2 of 0.062 and area under curve of 0.687 (P<0.001). Predictive values for a child with the four risk factors were: sensitivity 6.2%, specificity 98.6%, predictive positive value 16.2%, negative predictive value 96.1%, accuracy 94.9%, positive likelihood ratio 4.581, and negative likelihood ratio 0.951. Positive likelihood ratio for a child with the two major risk factors is 2.657. Conclusions: Usefulness of different risk factor associations to predict hospitalization for respiratory syncytial virus infection in preterm infants 32 to 35 weeks gestation in Spain is tow, although similar to other models. (C) 2009 Asociacion Espanola de Pediatria. Published by Elsevier Espana, S.L. All rights reserved.