Defining the optimal spirometry cut-off points for optimal bronchodilator response in prematurity associated lung disease

<bold>Introduction:</bold> Low lung function is observed in prematurity-associated lung disease (PLD). We determined optimal cut-off values for spirometry measures to identify preterm born children who have reversible airway disease which is most likely to respond to inhaled therapy. <bold>Methods:</bold> Children, from the Respiratory Health Outcomes in Neonates (RHiNO) study, aged 7-12 years, who were born ≤34 weeks’ gestation underwent spirometry before and after salbutamol. Receiver operator characteristic (ROC) curves were used to determine optimal cut-offs for sensitivity, specificity, positive and negative predictive values for all spirometry measures to identify optimal associations with bronchodilator responsiveness (≥10% predicted %FEV1 absolute change) utilising Youden’s J statistic. <bold>Results:</bold> Acceptable lung function data were available from 544 preterm-born children. Mean age and gestational age at birth were 9.6 years and 31 weeks, respectively. 51% were male. The area under the ROC curves (AUROC) were 0.758, 0.825, and 0.816 for %FEV1, percent predicted forced expiratory flows between 25 and 75% of forced vital capacity (%FEF25-75%), and FEV1/FVC ratio respectively but was 0.576 for %FVC. The optimal spirometry cut-off values for optimal Youden’s J statistic values were 85% for percent precited FEV1 (sensitivity: 0.789 and specificity: 0.591, Youden’s J: 0.38); 65% for %FEF25-75% (0.773 and 0.727, 0.50); and 0.80 for FEV1/FVC (0.816 and 0.697, 0.51). <bold>Conclusions:</bold> Since %FEF25-75% is associated with high variability in children, we conclude that 85% %FEV1 and 0.8 FEV1/FVC are optimal measures to identify bronchodilator response in preterm-born children.