Table of Contents
Study design and population
The CLSA is a large, national, long-term study with 51,338 respondents between the ages of 45 and 85 years at baseline who will be followed longitudinally. Baseline data were collected between 2011 and 2015. A large subset referred to as the Comprehensive Cohort (n = 30,097) provided detailed sociodemographic, lifestyle, and medical data, as well as genome-wide genotyping33. The details of the CLSA are provided elsewhere34. In rapid response to the pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the CLSA initiated three COVID-19 studies between 2020 and 2021 to explore the impacts of the pandemic on older adults (COVID-19 Questionnaire Study, COVID-19 Seroprevalence (Antibody) Study, and COVID-19 Brain Health Study)35. The current investigation utilized data from both the COVID-19 Questionnaire Study (comprised of self-reported data collected by questionnaires) and COVID-19 Seroprevalence (Antibody) Study (comprised of self-reported data collected by a separate questionnaire and serology from a blood sample). Genetic data from the CLSA Comprehensive Cohort were merged with these COVID-19 data sources to evaluate TAS2R38 diplotypes in analyses. Two common single nucleotide polymorphisms (SNPs) from TAS2R38 were available from the genome-wide assays: rs1726866 (V262A) and rs10246939 (I296V). Participants were categorized into three TAS2R38 diplotypes based on these SNPs: [P]AV/[P]AV (subsequently referred to as “[P]AV homozygotes”), [P]AV/[A]VI (subsequently referred to as “heterozygotes”), and [A]VI/[A]VI (subsequently referred to as “[A]VI homozygotes”). The bracketed letter indicates the inferred variant that was not available from the genetic data, and thus in the present study, [P]AV is analogous to the PAV haplotype and [A]VI is analogous to the AVI haplotype.
CLSA COVID-19 questionnaire study and seroprevalence study
A series of questionnaires were collected as part of the COVID-19 Questionnaire Study over a 9-month period (April 15, 2020 to December 29, 2020) with frequencies varying from weekly, biweekly, and monthly. Questionnaires were completed online or via telephone interview. The present investigation utilized data from the final COVID-19 questionnaire (Exit Questionnaire, n = 24,114), collected between September 29, 2020 to December 29, 2020. The Exit Questionnaire took approximately 30 min to complete and captured self-reported information on COVID-19 outcomes that occurred since the start of the pandemic (March 1, 2020), including participant infection status and symptoms, healthcare use, health behaviours, and psychosocial and economic impacts of the pandemic. In addition to the Exit Questionnaire, n = 19,334 CLSA participants participated in the CLSA COVID-19 Seroprevalence (Antibody) Study. These participants completed a separate questionnaire between November 2020 to July 2021 (internet-based or telephone interview) that collected self-reported information on COVID-19 infection status and symptoms, healthcare use, health behaviours, psychosocial and economic impacts of the pandemic since it began on March 1, 2020. In addition, these participants provided a blood sample either at a CLSA data collection site or using a self-collection kit at home. Blood samples were used to test for the presence of antibodies against SARS-CoV-2. Informed consent was obtained from all CLSA participants at baseline and prior to data collection for the COVID sub-studies. All methods conducted in this investigation were carried out in accordance with the principles outlined in the Declaration of Helsinki.
For the present investigation, data from the CLSA COVID-19 Questionnaire Study and the CLSA COVID-19 Seroprevalence (Antibody) Study were analyzed separately given that the samples were not identical, but also contained substantial overlap. Upon merging with genetic data for TAS2R38 diplotypes, a total of n = 13,825 respondents from the CLSA COVID-19 Questionnaire Study and n = 8791 respondents from the CLSA COVID-19 Seroprevalence (Antibody) Study were included in the current analyses (Supplementary Fig. S1 and S2). Ethics approval was obtained from the McGill University Faculty of Agricultural and Environmental Sciences Research Ethics Board.
Outcome measures
CLSA COVID-19 questionnaire study
Participants were considered to be cases of COVID-19 infection based on their responses to two questions in the self-reported Exit Questionnaire: “Have you ever had a positive test result?” and “Have you ever been told by a health care provider that you have COVID-19, but you did NOT have a test to confirm this?” Of the 13,825 participants with a response available, n = 42 reported a positive test result for COVID-19 and an additional n = 34 reported being told they had COVID-19 by a health care provider, providing a total of n = 76 respondents considered as COVID-19 cases for analysis.
The experiences of COVID-19 symptomatology were determined based on 12 questions in the Exit Questionnaire (detailed wording of the questions are presented in Supplementary Table S1). Common symptoms included fever, dry cough (no phlegm or mucus), wet cough (with phlegm or mucus), shortness of breath or difficulty breathing, decreased sense of smell, fatigue, sore/scratchy throat, muscle and/or joint aches/pains, headache, runny or stuffy nose, sinus pain and feeling generally unwell. For each symptom, participants self-reported the degree to which they experienced it: no symptom, mild, moderate, or severe. Due to the limited sample size of cases available, mild, moderate and severe responses were combined, and a binary outcome was analyzed for each symptom representing yes/no presence of the symptom.
CLSA COVID-19 seroprevalence (Antibody) study
From this study, participants with COVID-19 infection were determined based on the results of the serology variable: COVID-19 antibody result interpretation. Four groups were available based on antibody result: (1) results suggest prior SARS-CoV-2 infection (n = 177); (2) results suggest prior SARS-CoV-2 infection and/or vaccination (n = 4065); (3) results suggest prior SARS-CoV-2 infection OR prior SARS-CoV-2 infection and vaccination (n = 259); (4) no antibodies to SARS-CoV-2 detected (n = 4290). Only the first group was able to be considered as COVID-19 cases, as their results did not include the uncertainty in infection status due to vaccination.
The experiences of COVID-19 symptomatology were determined based on seven questions in the seroprevalence study questionnaire (detailed wording of the questions are presented in Supplementary Table S1). Common symptoms included fever, cough, shortness of breath, decreased sense of smell, sore throat, sore muscles and headache. The questionnaire asked only about the presence of the symptoms without severity level, so a binary outcome was analyzed for each symptom representing yes/no presence of the symptom.
Statistical analyses
Statistical analyses were conducted using SAS software (version 9.4, SAS Institute, Cary, North Caroline, USA). Sample means and frequencies for characteristics of the CLSA respondents according to TAS2R38 diplotypes were evaluated with analysis of variance and Pearson Chi-square tests to compare continuous and categorical characteristics, respectively. Characteristics of CLSA participants with data available vs. unavailable for the present investigation were also compared using the same statistical approach to assess the potential for attrition bias (Supplementary Table S2). Some variables assessed at CLSA’s baseline assessment were not reassessed in the COVID-19 Questionnaire Study (body mass index, self-reported ethnicity and highest level of education). In these cases, baseline data were used in these comparisons. Pearson Chi-square tests (or Fisher’s Exact test where appropriate) were used to assess differences in COVID-19 infection status, comorbidities, and symptomatology according to TAS2R38 diplotypes, with assessment of pairwise comparisons when overall exploratory (uncorrected) significant differences were observed. Furthermore, multivariable logistic regression analyses were used to assess associations between TAS2R38 diplotypes and COVID-19 infection status, as well as associations between TAS2R38 diplotypes and experiences of common COVID-19 symptomatology. The regression models were adjusted for age, sex, and the first five principal components of ancestry to account for population stratification. p-values are two-sided and due to multiple testing statistical significance was set at FDR corrected P < 0.0536. However, given the exploratory nature of this investigation, both uncorrected (i.e. exploratory) and FDR corrected p-values are reported.
