Methods
Participants
Participants were recruited from two medical centers in California. One
hundred and twelve children were recruited due to having had a positive
CF NBS and a designation of either CRMS/CFSPID or CF. As defined by the
IRT/DNA/DNA sequencing algorithm, participants had at least twoCFTR variants identified.8 Protocol was
approved by the Institutional Review Boards of Children’s Hospital Los
Angeles (CHLA) and Kaiser Permanente Los Angeles Medical Center (KP) and
informed consent was obtained from all participating families.
Procedures
As part of this research study, clinical data were collected
cross-sectionally and included height, weight, physical exam, and a
study-specific questionnaire related to relevant signs and symptoms in
the past 12 months (Supplemental Figure 1 ). Sweat chloride
concentration (sw[Cl-]) testing (ELItech
Macroduct® system) was performed following Cystic Fibrosis Foundation
guidelines.9 Spirometry was performed using the
CareFusion Vmax Encore device (Care-Fusion, Yorba Linda, California,
USA). Forced expiratory volume in one second reported in percent
predicted (ppFEV1), forced vital capacity (FVC), and
FEV1/ FVC measurements were converted to global lung
function initiative (GLI).10 Lung Clearance Index
(LCI2.5) derived from nitrogen multiple breath washout
testing (MBW) was offered to all subjects ≥3 years old, and used the
Exhalyzer D® device (Eco Medics AG, Duernten, Switzerland) with
Spiroware software (version 3.1.6). Device calibration was performed
daily prior to testing. MBW test was performed per ATS/ERS consensus
statement and as described previously.11-13 Cystic
Fibrosis Questionnaires-Revised (CFQ-R) were applied for parents and
children ≥6 years old.14
Electronic medical records (EMR) were reviewed from birth; data
collected included CFTR genotype, immunoreactive trypsinogen
(IRT) value, sw[Cl‑] results, microbiology data
(focusing on Pseudomonas aeruginosa [PSA]), fecal elastase
values, history of hospitalizations, history of lower respiratory tract
infections (Pneumonia) and pulmonary exacerbations, use of oral or
intravenous antibiotics in the past 12 months, and provider
documentation of crackles, chronic or recurrent respiratory symptoms,
constipation, signs of malabsorption, and relevant treatments.
Pancreatic insufficiency was defined as fecal elastase <200
µg/g. Chronic PSA was defined as two positive respiratory cultures in
the past 12 months.15 Both clinical research sites
used diagnostic criteria based on CF Foundation guidelines for CF and
CRMS/CFSPID.16 CRMS/CFSPID-to-CF reclassification
occurred when an individual’s sw[Cl-] rose to ≥60
mmol/L, one or more of an individual’s CFTR variants changed to a
CF-causing interpretation by CFTR2 (https://cftr2.org), and/or an
individual developed signs and symptoms of CF according to provider’s
discretion.
Nasal epithelial samples were obtained from a subgroup of subjects to
explore the utility of Human Nasal Epithelial (HNE) assays in defining
baseline CFTR function, which could potentially identify risk for
reclassification. Subjects with one CF-causing variant and a copy of
R117H;7T, 5T;TG12, or 5T;TG13 were prioritized. Sampling and Ussing
chamber testing of planar respiratory cultures were performed as
previously described.17, 18
Statistical analysis
Data are described in mean and standard deviation or median and
interquartile range for continuous variables, frequency and percentage
for categorical variables. Differences in key clinical variables between
sites, and demographic and clinical characteristics between diagnostic
groups (CF vs. CRMS/CFSPID) were examined using chi-squared or Fisher’s
exact tests for categorical variables and Wilcoxon rank-sum or two
sample t-tests for continuous variables. A non-additive effect was
assessed using Firth logistic regression and generalized linear model to
evaluate whether the difference in CF and CRMS/CFSPID varied by site on
ever having a positive culture for PSA (PSA-ever), fecal elastase,
sw[Cl-], and CFQ-R respiratory domain. Using longitudinal data
collected from EMR, sw[Cl-] progression was evaluated using a linear
median quantile mixed-effects model which time is modeled as age at
measurement and random effect at participant level. Individual
sw[Cl-] increases were estimated as a slope that was based on at
least three measurements (18 individuals were excluded for not having at
least 3 values). The association of clinical characteristics with the
risk of reclassification was assessed by Firth logistic regression model
to reduce bias due to the small number of reclassified patients.
Multivariate Firth logistic regression model further assessed the
association of clinical characteristics with the risk of
reclassification adjusting for site and demographic characteristics. All
statistical computations were performed in Stata 17 (StataCorp, College
Station, TX).