Supplemental Table S1: Spearman correlation between age and surfactant protein (SP) levels

MATERIALS AND METHODSStudy population: We enrolled 325 consecutive children under the age of 18 years who sought medical care at Penn State Children’s Hospital during the COVID-19 pandemic (early 2020 to late 2021) and were found to have a positive COVID-19 PCR test. Children with a lack of informed consent from the parents were excluded. Blood and plasma samples were available from 325 and 194 children, respectively.Classification of subjects: Controls were defined as those with mild disease if they did not require hospitalization due to COVID-19 or were hospitalized for a reason other than COVID-19. Cases were defined as subjects with severe disease if they were admitted to either general inpatient wards or the intensive care unit (ICU). Clinical and demographic data were collected from all subjects’ medical records, including age, weight, sex, race, ethnicity, comorbidities, exposure history, visit type, hospitalization status, environmental exposures, co-infections, antibiotic use, respiratory support requirements and duration, use of extracorporeal membrane oxygen (ECMO), and treatments used (shown in Table 1).Plasma surfactant protein concentration: Plasma levels of SP-A, SP-B, SP-C and SP-D were measured using Enzyme linked immunosorbent assays (ELISA) kits, following manufacturer’s recommendations (Novus Biologicals, LLC, Centennial, CO, USA for SP-A and SP-B, Biomatik Corporations, Wilmington, DE, USA; and Invitrogen, Life Technologies Corporation, Carlsbad, CA, USA for SP-D). The samples were tested in duplicates and accepted with a coefficient of variation of 5%. SP-C levels could not be measured due to their very low concentrations in the plasma samples.DNA isolation: DNA was extracted from blood samples using the QIAamp Blood Kit (Qiagen, Valencia, CA, USA) as described in the manufacturer’s instructions (L. C. Depicolzuane et al., 2022; DiAngelo et al., 1999).Genotyping: A multiplexed polymerase chain reaction (PCR) workflow of Ampliseq using custom designed panels from Illumina (Illumina, San Diego, CA) was used to analyze the SFTPA1, SFTPA2, SFTPB, SFTPC, and SFTPD genes (L. C. Depicolzuane et al., 2022). The data processing are described in detailed here (L. C. Depicolzuane et al., 2022). The genotypes of SP-A1 (6A, 6Am, m=0-13) and SP-A2 (1A, 1An, n=0-15) were assigned as indicated by DiAngelo et al. (1999).Statistical analysis: All variables were summarized prior to analysis to assess their distributions. Demographic variables were compared between mild and severe cases using a two-sample t-test for continuous variables and a Chi-square test for categorical variables. A quantile regression model was used to compare the median SP-A, SP-B, and SP-D between mild and severe cases unadjusted for covariates and adjusted for age, co-viral, and co-bacterial infections. The correlation between SP levels and age was tested using Spearman correlation both overall and within each severity group. A receiver operating curve (ROC) analysis was applied to determine the optimal cut point for SP-A as a predictor of severe disease. The Youden Index was used to find the cut point where the sensitivity and specificity were maximum simultaneously. A binary predictor for each SP level was created and included in a logistic regression model as a predictor of the severity while adjusting for sex and history of asthma. Odds ratios (ORs) are used to quantify the magnitude and direction of any significant associations. All analyses were performed using SAS software version 9.4 (SAS Instituted, Cary, NC) and a type 1 error rate of 0.05.Genetic association analysis: To investigate the association between SP genetic variants and COVID-19 severity, we conducted logistic regression analyses for each SNP of interest using PLINK 2.0, with COVID-19 severity as the primary outcome. The models were adjusted for key confounders, including age, sex, and race. In these models, sex was coded as a binary variable, with males serving as the reference group. Race was represented by a series of dummy variables for the following categories: Patient declined (0), White (1), Black (2), Asian (3), Mixed (4), Other (5), and Native Hawaiian or Pacific Islander (6).Results for each SNP were expressed as ORs with corresponding p-values. To ensure the robustness of our findings, statistical significance was determined using the Bonferroni correction to account for multiple comparisons across all SNPs and SP-A genotypes analyzed. This correction was applied using R version 4.3.3.Ethical consideration: The study was approved by the Human Subjects Protection Office of The Pennsylvania State University College of Medicine. Informed consent was obtained from the parent or guardian of each subject. All blood and plasma samples were assigned numbers upon arrival with no additional identifiers. To minimize bias, those measuring plasma SP levels and performing DNA extraction and genotyping were blinded to the sample identities.