Physiologic Pulmonary Phenotyping of Infants Born Preterm and Post-Discharge Respiratory Morbidity.

Abstract

To determine whether airway and parenchymal function identifies subgroups of infants born preterm according to the predominant pulmonary pathophysiology, and whether these subgroups have different risks for respiratory disease during infancy. We prospectively enrolled a cohort of 125 infants born preterm with planned clinical follow-up after neonatal intensive care unit discharge. The study included monthly questionnaires for wheeze and visits to a physician or care provider for any respiratory illness. In addition, infant lung function testing near 5 months corrected-age included measures of airways and parenchymal function using forced expiratory flows, alveolar volume (V_A), and the carbon monoxide transfer constant (diffusion capacity of lung [D_L]/V_A). Phenotypes were defined using 2 approaches: an a priori defined phenotypes based on forced expiratory flow 75% and D_L/V_A z-scores, and an unbiased approach to classifying infants using k-means clustering. We identified 4 pulmonary physiologic phenotypes that distinguished participants with predominantly decreased airway and/or parenchymal function. Although the worst physiologic phenotypes were associated with a lower gestational age at birth, these phenotypes had a better predictive value than gestational age, sex, and diagnosis of bronchopulmonary dysplasia for increased respiratory morbidity during infancy (area under the curve = 0.71 vs 0.63 for respiratory illness and 0.69 vs 0.63 for wheeze). Physiologic pulmonary phenotypes of infants born preterm were associated with differential risks for respiratory morbidities as infants, which may identify heterogeneous risks for long-term respiratory sequelae to individualize therapeutic strategies.