Clusters of lung clearance index trajectories and survival in people with cystic fibrosis

Progressive muco-obstructive lung disease impairs lung function, i.e. ventilation in small airways, and leads to premature death in people with Cystic Fibrosis (pwCF). Serial measurements of lung clearance index (LCI), a biomarker of ventilation inhomogeneity, may predict survival in pwCF. We hypothesize that there are distinct characteristics shared within several underlying groups of pwCF. We aim to examine LCI clusters and the associated risk of death or the need of lung transplantation (LTx). We performed a retrospective, longitudinal analysis in a cohort of pwCF aged >4 years and at least one LCI measurement. LCI was measured using a nitrogen multiple-breath washout (MBW) setup during clinical routine visits in Bern (Switzerland) between 1986 and 2006. Preliminary analyses include descriptive statistics. We aim to apply latent classes analysis (LCA) from unobserved patterns of the indicator variables FEV1 and LCI and link these classes to the compound outcome risk of death or LTx in pwCF until 12/2018. In total, 268 pwCF were eligible and 259 (135 females) were included contributing 3158 person-years and 2787 LCI measurements in total. Mean age was 15.7 ± 0.9 years at study entry. The median of visits per patient was 11 (interquartile range (IQR) = 5.0-19.0). Median LCI was 18.4 (IQR = 13.8-24.6) units. LCA models with the fewest number of classes that best fits the data remain to be selected. In 106 (40.9%) pwCF the primary outcome occured: 59 patients received LTx and 47 died. Our preliminary analyses suggest that the dataset is correctly setup and sufficiently large for LCA. We will explore if pwCF who died or received LTx belong to specific LCI clusters.