RESUMO
OBJECTIVE: To test whether low variability of tidal volume (VT) and capnographic indices are predictive of subsequent respiratory morbidity in preterm infants. STUDY DESIGN: In a birth cohort of 133 preterm infants, lung function was performed at 44 weeks postmenstrual age. Associations between the coefficient of variation (CV) of VT (CVVT) and of expired CO2 volume per breath (CVVE,CO2) with rehospitalization, wheeze, and inhalation therapy during infancy were assessed using logistic regression. Area under the curve (AUC) analysis was used to assess whether outcome prediction using bronchopulmonary dysplasia (BPD) classification was enhanced by CVVT or CVVE,CO2. RESULTS: For each IQR decrease in CVVT (range, 4%-35%) and CVVE,CO2 (range, 5%-40%), the OR for rehospitalization increased by 2.25 (95% CI, 1.21-4.20) and 2.31 (95% CI, 1.20-4.45), respectively. The predictive value of BPD for rehospitalization was improved when CVVT or CVVE,CO2 was added to the model, with the AUC increasing from 0.56 to 0.66 in both models. No association was found for the other outcomes. CONCLUSIONS: Compared with BPD classification alone, including near-term variability of tidal breathing parameters improves the prediction of rehospitalization in infancy. These findings may inform parent counseling and monitoring strategies in preterm infants.