Study protocol and procedures:
The study participants first completed MBW according to American
Thoracic Society (ATS) guidelines27. LCI was determined
according to standardized protocols28. Subsequently
spirometry followed by lung volumes (plethysmography) were performed
according to ATS criteria and standards29,30.
To measure airway
hyper-responsiveness27(AHR), spirometry was repeated 20 minutes after administration of 4
puffs of albuterol via MDI. For healthy controls and SCD patients
without a recent blood draw (3 months), whole blood samples were
collected for processing on the day of testing.
Definitions: Obstructive lung disease was
defined as 1 or more lung function measures (FEV1 or
FEV1/FVC) < lower limit of normal (LLN)
equivalent to a z-score of -1.645 based on Global Lung Initiative (GLI)
normative data for age, height, weight and race.
Restrictive lung disease was defined as Total Lung
Capacity (TLC) or Forced Vital Capacity (FVC) <LLN based on
GLI normative data for age, height, weight and race. AHRwas defined as having a positive bronchodilator response after repeat
spirometric measures with an increase in FEV1 and/or FVC
of ≥12% from baseline31.
Statistical Analysis: The sample size was based on results from
Zwitserloot et al.32where the comparison of LCI between patients with asthma and healthy
children (controls) showed a LCI difference of 0.27 with standard
deviations of 0.48 and 0.38 for the two groups respectively. There are
no defined normative LCI values in African American children. We assumed
a similar LCI difference and standard deviation between SCD and healthy
control children. In order to achieve 80% power and a 5% error rate,
we needed to enroll 42 children in each group. Demographics, clinical
events, lab results, and pulmonary function data were compared across
groups using Student’s t-tests for continuous variables. Wilcoxon tests
were performed for non-linear data. Chi-Square tests were performed for
categorical variables and Fisher’s Exact tests were used when cell
counts were low. Correlation analyses were performed to analyze the
associations between clinical and pulmonary function data. To check for
possible moderating effects due to participant age, interaction terms
were assessed with age groups 6-12 years of age and 13-18 years, as well
as performing analyses stratified by age groups. Multivariable models
were then used to determine if associations with LCI would become
attenuated in the larger model. Variables in the multivariable model
were chosen due to statistical significance at the bivariate level and
for clinical meaningfulness. All analytic assumptions were verified,
with non-parametric tests (i.e. Wilcoxon tests and Spearman
correlations) used when data were non-linear. Collinearity was assessed
in the multivariable model with variance inflation factors. All analyses
were performed using SAS v9.4 (SAS Institute, Cary, NC).