Variability of tidal breathing parameters in preterm infants and associations with respiratory morbidity during the first year of life: a cohort-study

Objectives: Low variability of tidal volume (V T ) and capnographic indices in preterm infants is an expression of impaired lung function and gas exchange. We tested if these parameters, when assessed near term, are predictive of respiratory morbidity during the first year of life. Methods: We performed lung function measurements at 44 weeks postmenstrual age in a birth cohort of 133 unsedated preterm infants. We asked whether lung function measurements (variability of V T and capnographic indices) are associated with subsequent rehospitalization, episodes of wheeze, or inhalation therapy. Using logistic regression and area under the curve (AUC) analysis, we determined if the predictive power of models that included BPD classification is enhanced by adding variability of V T or capnographic indices. Results: Both coefficient of variation (CV) of V T (CV VT ) (range 4%-35%) and CV of expired CO 2 volume (CV VE,CO2 ) (range 5%-40%) were negatively associated with rehospitalization during infancy. Per interquartile range decrease in CV VT and CV VE , CO2 , the adjusted odds ratio (95% CI) for rehospitalization increased by 2.25 (1.21–4.20) and 2.31 (1.20–4.45), respectively. Adding CV VT or CV VE,CO2 to ROC models that included BPD improved prediction of rehospitalization (AUC increased from 0.56 to 0.66 in both models). There was no association of CV VT or capnographic indices with other outcomes. Conclusion: In comparison to BPD classification alone, including near term variability of tidal breathing parameters improves prediction of rehospitalization in infancy. These findings may impact upon parent counseling and monitoring strategies in preterm infants.