Post-Discharge COVID-19 Symptoms Predict 1-Year Functional Decline, Falls, and Emergency Department Visits: A Cohort Study

Abstract

Background: Hospitalization frequently results in persistent symptoms among older adults, raising concerns about the long-term impacts of acute events-a problem amplified by COVID-19. We investigated the effects of persistent symptoms on functional decline and unplanned events over 1 year in older patients recovering from COVID-19 hospitalization.

Methods: This prospective cohort included patients aged ≥ 50 years who survived COVID-19 hospitalization between March and December 2020 as part of the CO-FRAIL study at Brazil's largest academic medical center. Persistent symptoms were defined as those reported at admission and continuously present at one-, three-, six-, nine-, and 12-month post-discharge, covering 16 symptoms. Outcomes included functional decline in basic activities of daily living (ADL), mobility activities, instrumental activities of daily living (IADL), number of falls, emergency department (ED) visits, and hospital readmissions. Associations between persistent symptoms and outcomes were examined using mixed-effects negative binomial regression models adjusted for sociodemographic, clinical, hospitalization-related factors, and post-discharge rehabilitation.

Results: Among 1019 patients (mean age = 65 ± 10 years; women = 45%; White = 62%), 324 (32%) experienced persistent symptoms throughout the year. Fatigue (28%), myalgia (19%), and dyspnea (13%) were the most common. Patients with ≥ 2 symptoms had an increased risk of functional decline in mobility activities (IRR = 2.11; 95% CI = 1.50-2.96), IADL (IRR = 2.00; 95% CI = 1.44-2.79), falls (IRR = 2.56; 95% CI = 1.14-5.75), and ED visits (IRR = 2.69; 95% CI = 1.27-5.70), but not readmissions. Among women, ≥ 1 persistent symptom was associated with a twofold increased risk of ADL decline. One year after discharge, patients with ≥ 2 persistent symptoms developed 1.27 more new disabilities (3.26 vs. 1.99 on a 15-point functional scale) and had 31 more unplanned events per 100 person-years (54.3 vs. 23.2) than those without symptoms.

Conclusions: Persistent symptoms after hospitalization are common and contribute to functional decline, falls, and ED visits in older COVID-19 survivors. These findings suggest that greater attention to symptom burden may support risk identification and improve post-discharge care planning.

Keywords: COVID‐19; functional decline; hospital; persistent symptoms; post‐acute care.

FIGURE 1

Functional decline and unplanned events within 1 year after discharge by number of persistent symptoms at follow‐up (n = 1019). Persistent symptoms were defined as those reported at admission and consistently present in sequential follow‐up assessments across 16 symptoms. Functional decline was defined as any new disability in ADL, mobility, or IADL compared to the patient's status 2–4 weeks before admission. ADL decline may or may not co‐occur with decline in mobility and IADL; mobility decline may or may not co‐occur with IADL decline. All comparisons across levels of persistent COVID‐19 symptoms for functional disability and unplanned events yielded p‐values < 0.01, except for unplanned events at 1 month (p‐value = 0.20). ADL, basic activities of daily living; ED, emergency department; IADL, instrumental activities of daily living.

FIGURE 2

Incidence rate ratios for functional decline and unplanned events over 1 year after discharge by number of persistent symptoms (n = 1019). Estimates were computed using mixed‐effects negative binomial regression models with random intercepts for each participant and random slopes for time. The primary exposure was the number of persistent symptoms, defined as those reported at admission and consistently present in sequential assessments across 16 symptoms. Functional decline was defined as new disabilities (“worsening independence”) in ADL, mobility, and/or IADL at 1, 3, 6, 9, and 12 months post‐discharge, compared to the patient's status 2–4 weeks before admission. Models were adjusted for demographics (age, sex, race/ethnicity, education, marital status), Charlson Comorbidity Index, depression, obesity, hospitalization‐related factors (delirium, dialysis, length of stay), post‐discharge rehabilitation, type of informant, follow‐up time, and interactions between persistent symptoms with sex and time. Twenty‐six patients were excluded from the mobility disability analysis (final n = 993), and 84 from the IADL disability analysis (final n = 935), due to complete dependence on others for these activities before admission. 95% CI, 95% confidence interval; ADL, basic activities of daily living; IADL, instrumental activities of daily living.

FIGURE 3

Incidence rate ratios at multiple time points for functional decline (panel A) and unplanned events (panel B) by number of persistent symptoms after discharge (n = 1019). Estimates at each follow‐up time point were computed using mixed‐effects negative binomial regression models with random intercepts for each participant and random slopes for time. The primary exposure was the number of persistent symptoms, defined as those reported at admission and consistently present in sequential assessments across 16 symptoms. Functional decline was defined as new disabilities (“worsening independence”) in ADL, mobility, or IADL at 1, 3, 6, 9, and 12 months post‐discharge. Models were adjusted for demographics (age, sex, race/ethnicity, education, marital status), Charlson Comorbidity Index, depression, obesity, hospitalization‐related factors (delirium, dialysis, length of stay), post‐discharge rehabilitation, type of informant, follow‐up time, and interactions between persistent symptoms with sex and time. 95% CI, 95% confidence interval; ADL, basic activities of daily living; ED, emergency department; IADL, instrumental activities of daily living.

FIGURE 4

Incidence rate ratios for functional decline in basic activities of daily living by number of persistent symptoms, stratified by sex, over 1 year after discharge. Estimates were computed using mixed‐effects negative binomial regression models, which included random intercepts for each participant and random slopes for time, and were stratified by sex. The primary exposure was the number of persistent symptoms, defined as those reported at admission and consistently present across 16 symptoms in sequential assessments. Functional decline was defined as worsening independence in ADL at 1, 3, 6, 9, and 12 months post‐discharge compared to the patient's status 2–4 weeks before admission. Models were adjusted for demographics (age, sex, race/ethnicity, education, marital status), Charlson Comorbidity Index, depression, obesity, hospitalization‐related factors (delirium, dialysis, length of stay), post‐discharge rehabilitation, type of informant, follow‐up time, and interactions between persistent symptoms and time. 95% CI, 95% confidence interval; ADL, basic activities of daily living.