(138) Defining and identifying early-onset lung disease in cystic fibrosis with cumulative clinical characteristics in the first 3 years of life

Background: Because of variations in cystic fibrosis (CF) lung disease among young children, a clinical method is sorely needed to identify children at increased risk for early-onset lung disease. As an essential component of a study applying whole genome sequencing to young children with CF, particularly to discover polygenic risk factors, we designed and developed a novel CF Early-onset Lung Disease (CFELD) scoring system by utilizing prospectively collected longitudinal data on pulmonary disease manifestations in the first 3 years of life.

Methods: The study population consists of 145 infants born during 2012-17, enrolled after NBS at age 1.8±1.01 months (mo) in FIRST (Feeding Infants Right… from the STart), a prospective longitudinal study being conducted at 6 CF Centers (Madison, Milwaukee, Boston, Indianapolis, Salt Lake City and Chicago), and completed 36 months of follow-up. Cough severity, pulmonary exacerbations (PEx), respiratory cultures and hospitalizations were collected on a pulmonary interval history form at each CF center visit every 1-2 months in infancy and quarterly thereafter. These longitudinal data were used to construct the novel CFELD system, and the overall CFELD scores were used to classify lung disease at 3 years of age into five categories: asymptomatic, minimal, mild, moderate, and severe. Multiple regression analyses were used to identify factors significantly predictive of the CFELD scores.

Results: The most frequent manifestation of CF early lung disease was MD-reported PEx episodes, PEx hospitalizations, and positive Pseudomonas aeruginosa cultures. Parent-reported cough occurred in ~60% of the FIRST cohort by age 3. A total of 266 respiratory hospitalization episodes were documented in 2469 pulmonary interval history forms (11%) in the first 3 years of life. Cough severity was correlated with the number of respiratory hospitalizations (r=0.48, p< 0.001). The distribution of CFELD categories was 10% asymptomatic, 17% minimal, 29% mild, 33% moderate, and 12% severe. The moderate and severe categories are early-onset phenotypes that occur 3-fold higher in pancreatic insufficient (PI, 49%) vs. sufficient subjects (16%), p< 0.001. Having GI/nutrition-related hospitalizations was also a predictor of worse CFELD scores. High plasma IL-6, which is a proinflammatory cytokine, was associated with worse CFELD scores. In contrast, high plasma IL-10, which is an anti-inflammatory cytokine, was associated with better CFELD scores. Regarding social determinants, ordinal logistic regression revealed that parental education of community college or above significantly decreased the likelihood of being in the severe CFELD category. Additional analyses on the relationship of breastfeeding to CFELD outcomes are being conducted.

Conclusions: The CFELD scoring system is novel, allows systematic evaluation of lung disease prognosis early, and may aid in therapeutic decision-making. The CFELD system should be especially useful for studies of epidemiologic and genetic factors that may play a role in the onset of lung disease and for clinical decisions regarding the initiation of CFTR modulator therapy.

Acknowledgements: We thank all investigators in the FIRST Study Group and for funding from the NIH (R01DK072126, R01DK109692), the CFF (LAI14A0, ALI15A0, LAI17A0), the Legacy of Angels Foundation, and the Shapiro Summer Research award from the UW-Madison School of Medicine and Public Health.

References: