Study objective: To measure the association between the development of air leak (pneumothorax or pulmonary interstitial emphysema) during the first 27 postnatal days and neonatal death or chronic lung disease, as determined on day 28, among very low birth weight infants who required mechanical ventilation from the first day of life.
Design: Prospective, multicenter cohort study.
Patients: Two hundred sixty inborn, very low birth weight (501 to 1500 gm) infants given ventilatory support from the first day of life.
Results: The risk of an adverse outcome (death or chronic lung disease) changed with postnatal age at the time of diagnosis of the air leak. The association between air leak and an adverse outcome, as measured by gestational age-adjusted odds ratio (95% confidence interval), was 13.9 (1.7 to 114.6) for those in whom an air leak developed on day 0 or 1 (early), decreased to 1.7 (0.7 to 4.1) for those whose air leak developed on day 2 or 3 (intermediate), and increased to 16.6 (2.1 to 130.4) for those whose air leak developed on days 4 to 27 (late). The association with neonatal death showed even more striking fluctuations with postnatal age at occurrence of an air leak, ranging from an odds ratio of 40.3 (3.5 to 464.8) for the early group to 7.5 (2.3 to 25.0) for the intermediate group and 78.3 (6.9 to 889.5) for the late group.
Conclusions: Air leak in newborn infants requiring mechanical ventilation is associated with increased risks of death or future morbidity, but the magnitude of these risks changes with postnatal age at the time of diagnosis of the air leak. Failure to consider the age at which the air leak is detected may miss changes in its prognostic implications and may partly explain inconsistent results in previous studies.