Persistent COVID-19 symptoms in community-living older adults from the Canadian Longitudinal Study on Aging (CLSA)

The study investigates the persistence and severity of COVID-19-related symptoms (long COVID) among community-living older adults, a group underrepresented in prior research that has largely focused on hospitalized patients. It aims to determine how age, sex, and pre-pandemic physical, psychological, social, and functional factors are associated with the cumulative incidence and persistence of 23 symptoms. Understanding these factors in a population-based cohort with a non-COVID control group is important to attribute symptoms to SARS-CoV-2 and to identify subgroups at higher risk who may need targeted support and rehabilitation.

Most existing studies of long COVID have focused on hospitalized cohorts, reporting high rates of persistent symptoms (e.g., fatigue, sleep difficulties, anxiety/depression) months after infection. Community-based evidence is limited; many studies used convenience samples, administrative data lacking socio-demographic/contextual variables, or lacked appropriate control groups, limiting generalizability and causal attribution. Only one prior study included a control group, making it difficult to attribute long-term symptoms directly to SARS-CoV-2. Definitions of post-acute (≥1 month) and chronic (>3 months) symptom persistence are aligned with proposed frameworks. The literature shows mixed findings regarding age, with more consistent associations for female sex and multimorbidity; socioeconomic indicators have been less frequently examined, particularly subjective social status.

Design and setting: Cross-sectional analysis nested within the Canadian Longitudinal Study on Aging (CLSA), a population-based cohort. Data sources: Linked pre-pandemic CLSA baseline (2011–2015) and follow-up 1 (2015–2018) data with CLSA COVID-19 questionnaires, including an Exit questionnaire (September 29–December 29, 2020) that captured symptom persistence. Participants: Community-dwelling adults originally aged 45–85 at CLSA recruitment; analysis restricted to 23,757 adults aged 50+ who completed the COVID-19 Exit questionnaire. COVID-19 status: Self-reported positive COVID-19 test or physician diagnosis defined the COVID-19 group (n = 121); those reporting it was “very likely” they had COVID-19 without a positive test/diagnosis were excluded from primary analyses (n = 357). Non-COVID-19 group comprised respondents without a positive test/diagnosis (n = 23,636). Measures: Participants reported whether they experienced any of 23 COVID-related symptoms since March 1, 2020 and rated each as mild, moderate, or severe; an “other” category captured additional symptoms. For COVID-19 positive participants, duration of each symptom was reported: ≤2 weeks, ≥1 month, 2 months, >3 months, or ongoing. Persistent symptom definitions followed: post-acute (≥1 month) and chronic (>3 months). Symptom reporting across baseline, biweekly, and 3-monthly questionnaires was summed to validate persistence reports; missing persistence status (~2.1%) was imputed using prior responses when possible. Pre-pandemic covariates included sex, age group (<65, 65–74, 75+), multimorbidity (≥2 chronic conditions), depression/anxiety, mobility ADL limitations, and subjective social status. Analytic approach: Estimated cumulative incidence of symptoms (overall and by severity) since March 2020 for COVID-19 and non-COVID-19 groups, including sex and age stratifications and direct standardization. Among COVID-19 positive participants, modeled the number of persistent symptoms (>1 month; >3 months) using regression to estimate (adjusted) incidence rate ratios (IRR/aIRR) with covariates age, sex, multimorbidity, depression/anxiety, mobility limitations, and subjective social status. Sensitivity analyses included participants who reported they were “very likely” to have had COVID-19. Hospital care utilization among COVID-19 positive respondents (ED visits, hospitalizations, ICU) was also summarized.

- Sample: 23,757 community-living adults (50+ years); 121 with COVID-19 (self-reported positive test or physician diagnosis) and 23,636 without. Overall, 53.16% female; no significant differences by sex or age between COVID-19 and non-COVID-19 groups, but the COVID-19 group had higher multimorbidity (78.15% vs. 67.31%), more depression/anxiety, greater mobility limitations, and slightly lower subjective social status. Among COVID-19 positive participants, 14 visited the ED; 7 (5.8%) were hospitalized; 2 (2.5%) had ICU stays. - Symptom burden: The most commonly reported symptoms in both groups were fatigue, dry cough, muscle/joint pain, sore throat, headache, and runny nose (>25% in each group). Moderate/severe symptom cumulative incidence was consistently higher in the COVID-19 group; more than double that of the non-COVID-19 group, with absolute differences ranging from 16.8% (runny nose) to 37.8% (fatigue). - Sex differences: Female COVID-19 participants reported more moderate/severe symptoms than males, notably for fatigue (52.7% vs. 44.7%), muscle/joint pain (40.5% vs. 30.4%), headache (32.4% vs. 19.2%), and decreased smell (33.7% vs. 19.1%). - Age patterns: <65 reported higher cumulative incidence than older groups, but the 75+ group had a higher proportion of moderate/severe symptoms; age group was not significantly associated with the rate of persistent symptoms in regression analyses. - Persistence: Among COVID-19 positive participants, approximately 60% of males and 73–75% of females reported at least one symptom persisting >1 month. Over 15% reported moderate/severe fatigue, dry cough, and decreased smell persisting >1 month; >10% reported moderate/severe shortness of breath, muscle/joint pain, or runny nose >1 month. Over 10% reported mild fatigue persisting >3 months. - Regression results: Persistent symptoms >1 month were more frequent in females (aIRR ≈ 1.68; 95% CI 1.03–2.73) and in those with multimorbidity (aIRR ≈ 1.90; 95% CI 1.02–3.49). For >3 months, female sex was not significant after adjustment; each unit increase in subjective social status was associated with a 15% reduction in the rate of persistent symptoms (aIRR 0.85; 95% CI 0.73–0.99), adjusting for age, sex, and multimorbidity. Sensitivity analyses yielded similar estimates.

The study shows that community-living older adults with COVID-19, even without hospitalization, exhibit higher cumulative incidence and severity of symptoms compared with non-COVID peers, supporting the attribution of persistent symptoms to SARS-CoV-2 infection. Females and individuals with multimorbidity are at greater risk of having more persistent symptoms within the first month post-infection, indicating subgroups that may benefit from targeted monitoring and intervention. While younger participants reported more symptoms overall, older participants (75+) had a higher proportion of moderate/severe symptoms, yet age was not an independent predictor of persistence in adjusted models. Notably, lower subjective social status independently predicted greater persistence beyond 3 months, highlighting socioeconomic perceptions as relevant determinants beyond traditional SES measures. Clinically, many persistent symptoms (e.g., dyspnea, fatigue, pain) are amenable to rehabilitation, underscoring the need for accessible post-COVID rehabilitation services in the community.

In a large, population-based Canadian cohort of community-living older adults, non-hospitalized individuals with COVID-19 commonly experienced persistent symptoms at 1 and 3 months post-infection, with higher burden among females, those with multimorbidity, and those with lower subjective social status. The findings emphasize the need for targeted supports, including access to rehabilitation, to facilitate recovery. Future research should incorporate detailed socio-demographic and contextual measures (including subjective social status), improve case ascertainment and timing, and follow participants beyond 3 months to characterize long-term trajectories and inform interventions.

- Participation bias due to a 67% response rate to the COVID-19 study; potential differences between respondents and non-respondents. - Small number of COVID-19 positive participants (n = 121) limits precision (wide CIs) and subgroup analyses, though consistent with contemporaneous community prevalence. - Exclusion of participants who were “very likely” to have had COVID-19 without confirmation may have omitted true cases; sensitivity analyses mitigated this concern. - COVID-19 status based on self-report (positive test or physician diagnosis), potentially subject to recall or misclassification, though recall bias likely minimal early in the pandemic; antibody-based alternatives also have limitations. - Lack of exact infection dates may underestimate persistence if infections occurred shortly before the Exit questionnaire. - Persistence beyond 3 months could not be assessed. - Generalizability to populations not represented in CLSA (e.g., institutionalized individuals, those on First Nations reserves, non-English/French speakers) may be limited.