Abstract
Background:
Persistent symptoms after SARS-COV-2 (long-COVID) occur in 10–55% of individuals, but the impact on daily functioning and disability remains unquantified.
Methods:
To characterize disability associated with long-COVID, we analyzed baseline data from an online, US-based cohort study. Adult participants included those reporting a history of COVID-19 (n=8,874) or never having COVID-19 (n=633) without prior disability. The main outcomes were self-reported physical mobility, instrumental activities of daily living (IADL), and mental fatigue disability, assessed by measuring five disability components: difficulty walking ¼ mile and/or climbing 10 stairs (mobility), difficulty doing light or heavy housework (IADL), and Wood Mental Fatigue Inventory score (mental fatigue).
Results:
Of 7,926 participants with long-COVID, 65% were classified with at least one disability as compared to 6% and 14% for resolved-COVID and no-COVID, respectively. Additionally, 22% were classified as disabled in all three categories. Age, prior comorbidity, increased BMI, female gender, COVID-19 hospitalization, non-white/multi-race were associated with higher disability burden. Dizziness and heavy limbs at infection were associated with disability regardless of hospitalization. Dyspnea and tremors were associated with disability in non-hospitalized individuals. Vaccination was protective against disability.
Conclusions:
We observed a high burden of new disability associated with long-COVID which has serious implications for individual and societal health. Longitudinal evaluation of COVID-19 patients is necessary to identify patterns of recovery and treatment options.
Introduction
The public health response to COVID-19 has focused on the acute phase of infection, from curtailing transmission to preventing and treating severe disease.1–3 However, post-acute sequelae after SARS-CoV-2 infection, or long-COVID, occurs in 10 to 55%4–6 of individuals and has the potential to overwhelm health systems and economies.7,8 While long-COVID disproportionately affects those hospitalized during acute infection, it is also recognized that persistent symptoms affect those with mild/moderate infections.9–11 Moreover, long-COVID affects all ages, genders, and races and occurs among individuals with and without pre-existing comorbidity.12–14 Symptoms reported by individuals with long-COVID are heterogeneous and affect multiple organ systems, including fatigue, anosmia, shortness of breath, and general cognitive impairment.15,16
The breadth of symptoms that persist or newly occur in long-COVID is well characterized,15,16 but there are fewer assessments on disability. In a study of critically ill hospitalized COVID-19 patients, 48% had decreased functional status and 10% reported severe limitations in their daily life six months post-infection.17 Similarly, in a Norwegian cohort of older hospitalized patients, 35% reported impaired ability to perform daily activities, 33% reduced mobility, and 43% decreased cognitive function six months after hospitalization.18 While there have been fewer community-based assessments, one study of 328 participants reported 4.9% had symptoms post-infection that constrained daily activities.19 This functional restriction directly affects employment, caregiving, and independent living.19–22
There is a need for understanding the breadth of disability in a larger study spanning both hospitalized and non-hospitalized individuals. We characterize physical disability and mental fatigue in those with and without COVID-19 in a large sample of community dwelling adults who participated in a nationwide cohort study, the Johns Hopkins COVID Long Study (JHCLS).
METHODS
Study design and sample
The JHCLS is a prospective virtual cohort study that launched on February 2, 2021, for participants with a self-reported positive SARS-CoV-2 test or symptoms of COVID-19, and expanded on March 11, 2022, for those without a history of COVID-19. Inclusion was limited to participants 18 years of age or older with residence in the United States (US). The study was approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board. All participants provide informed consent.
Participants were recruited primarily through social media posts, Facebook advertisements, and direct messaging (i.e., email). Additional recruitment was through the Johns Hopkins Opportunities for Participant Engagement (HOPE) Registry matching individuals interested in open COVID-19 research studies.
Long-COVID was defined using self-reported baseline survey data and the World Health Organization (WHO) definition23,24 of a symptom lasting at least two months but occurring 12 weeks or more from the onset of COVID-19 with no other explanation. Participants who reported a SARS-CoV-2 infection, had at least 12 weeks between infection and survey completion, but did not meet the WHO definition were classified as resolved-COVID (Figure S.1.1). Individuals with no history of SARS-CoV-2 infection were classified as no-COVID. Individuals without at least 12 weeks between SARS-CoV-2 infection and survey completion were excluded.
Survey data
The survey collected data on socio-demographics, COVID-19 history, health conditions, general health history, current health status, and vaccines. We included individuals with a self-reported positive SARS-CoV-2 test and/or COVID-19 symptoms only because of the initial limited test availability and during peaks. For questions regarding symptoms, comorbidities, and physician-diagnosed conditions, we included a comprehensive list. Participants were asked to report symptoms experienced during acute infection and new/persistent symptoms post-acute infection. COVID-19 associated hospitalizations included length of hospitalization and the need for supplemental oxygen and ventilator support. All comorbidities were self-reported except for obesity; body mass index (BMI) was calculated using self-reported height and weight. COVID-19 vaccine history was recorded including the number of doses, date, and series type per dose. All data were collected in REDCap.25,26
Disability case definitions
Three disability outcomes were considered: mobility, instrumental activities of daily living (IADL), and mental fatigue. For mobility and IADL, we used the classification by the Baltimore Longitudinal Study of Aging.27 Mobility disability was defined as having some or greater level of difficulty walking ¼ mile and/or climbing 10 stairs. IADL disability was defined as having at least some or greater level of difficulty with heavy housework. Critical physical disability was defined as being unable to do mobility and/or IADL tasks.
To assess mental fatigue, we utilized the Wood Mental Fatigue Inventory (WMFI).28 Participants were assigned a score for each of nine domains: 0(not at all), 1(a little), 2(somewhat), 3(quite a lot), and 4(very much). The higher summed scores represent worse mental fatigue with a cutoff of 20 based on the 90th percentile of the no-COVID sample (Supplement Methods). Severe and critically severe mental fatigue were scores of 30–34 and 35–36, respectively.
Statistical analysis
We used baseline data from participants with long-COVID (n=7,926), resolved-COVID (n=948), and no-COVID (n=633) who had complete data on the three disability outcomes and did not report disability prior to infection (Figure S.1.1). Individuals missing any of the three disability outcomes or reported a disability prior to their SARS-CoV-2 infection were not included. Compared to those who had complete data, those missing data were no different with respect to age, gender, race/ethnicity, region, education, employment, occupation, or income.
We used inverse odds weights to ensure the distribution of variables within the resolved-COVID and no-COVID sample matched the long-COVID sample.29 Logistic regression was used to examine outcomes in individual models among those with long-COVID and symptoms from the acute phase of infection with each disability outcome. Questions added to the baseline survey after individuals enrolled were accounted for by single stochastic imputation since missingness was completely random. As the list of presenting symptoms was large and comprehensive, we used a sequential logistic model with relaxed LASSO in variable selection.
The effect of vaccination status at the time of infection on the development of long-COVID disability was evaluated in a weighted analysis to be representative of the US population through the COVID-19 Symptom Survey (Supplement Methods).31 For the vaccine analysis, individuals infected prior to vaccine availability (12/14/20) were excluded (n=5,386). Two approaches were used: a regression-based model informed by directed acyclic graphs and an instrumental variable (IV). Vaccination was defined in three categories: unvaccinated/one mRNA dose, two mRNA doses/one adenoviral dose, or additional vaccine doses. Due to sample size in the IV analyses, we used a binary category (≥2 mRNA doses or ≥1 dose of adenoviral vector vaccine vs. unvaccinated/one mRNA dose).
RESULTS
Among 7,926 individuals with long-COVID, the median age was 45 years, 84% were female, 89% self-reported white race, and 7.4% self-reported Hispanic/Latino ethnicity (Table 1). Pre-existing comorbidities were heterogeneous with participants reporting hypertension (15%), diabetes (4.0%), and asthma/other chronic pulmonary conditions (17%). Based on calculated BMI, 36% were classified as obese.
Table 1.
Characteristics of participants in the Johns Hopkins COVID Long Study*
| Long-COVID (n= 7,926) | Resolved-COVID (n= 948) | No-COVID (n=633) | |
|---|---|---|---|
|
| |||
| US region** | |||
| Northeast | 1,549 (20%) | 176 (16%) | 148 (23%) |
| Midwest | 1,842 (23%) | 224 (24%) | 117 (18%) |
| South | 2,693 (34%) | 385 (41%) | 211 (33%) |
| West | 1,830 (23%) | 163 (17%) | 157 (25%) |
| Median age, IQR** | 45 (36, 55) | 37 (30, 48) | 39 (30, 50) |
| Gender** | |||
| Cisgender man | 1,137 (14%) | 205 (22%) | 82 (13%) |
| Cisgender woman | 6,681 (84%) | 738 (78%) | 533 (84%) |
| Transgender man | 15 (0.2%) | 0 (0%) | 1 (0.2%) |
| Transgender woman | 9 (0.1%) | 0 (0%) | 0 (0%) |
| Genderqueer/nonbinary | 74 (0.9%) | 4 (0.4%) | 13 (2.1%) |
| Race** | |||
| White | 7,092 (89%) | 830 (88%) | 514 (82%) |
| Black | 177 (2.2%) | 26 (2.8%) | 9 (1.4%) |
| Asian/Pacific Islander/Native Hawaiian | 142 (1.8%) | 29 (3.1%) | 64 (10%) |
| Native American/Alaska Native | 44 (0.6%) | 1 (0.1%) | 2 (0.3%) |
| Other | 153 (1.9%) | 24 (2.6%) | 19 (2.7%) |
| Mixed race | 242 (3.1%) | 29 (3.1%) | 21 (3.3%) |
| Hispanic, Latino or Spanish origin | 583 (7.4%) | 59 (6.2%) | 36 (5.7%) |
| Educational attainment | |||
| High school, GED or less | 447 (5.6%) | 32 (3.4%) | 19 (3.0%) |
| Some college, Associates/technical degree | 2,124 (27%) | 148 (16%) | 75 (12%) |
| Bachelor’s degree | 2,429 (31%) | 350 (37%) | 196 (31%) |
| Post-graduate degree | 2,906 (37%) | 418 (44%) | 343 (54%) |
| Annual household income** | |||
| <$25,000 | 709 (8.9%) | 87 (10%) | 69 (12%) |
| $25,000 - $34,999 | 445 (5.6%) | 57 (6.6%) | 26 (4.7%) |
| $35,000 - $49,000 | 783 (9.9%) | 72 (8.3%) | 46 (8.2%) |
| $50,000 - $74,999 | 1,463 (18%) | 133 (15%) | 95 (17%) |
| $75,000 or greater | 3,810 (48%) | 520 (60%) | 323 (58%) |
| Body mass index (kg/m2) | |||
| Underweight (<18.5) | 129 (1.6%) | 18 (1.9%) | 12 (1.9%) |
| Normal weight (18.5 – 24.9) | 2,615 (33%) | 370 (39%) | 275 (44%) |
| Overweight (25 – 29.9) | 2,260 (29%) | 301 (32%) | 167 (26%) |
| Obese (30 and above) | 2,827 (36%) | 249 (27%) | 177 (28%) |
| Comorbid conditions* | |||
| Diabetes | 315 (4.0%) | 23 (2.4%) | 20 (3.2%) |
| Cardiovascular disease/Heart failure | 151 (1.9%) | 14 (1.5%) | 13 (2.1%) |
| Hypertension | 1,201 (15%) | 104 (11%) | 62 (9.8%) |
| Chronic Kidney disease | 55 (0.7%) | 2 (0.2%) | 2 (0.3%) |
| Cancer | 184 (2.3%) | 14 (1.5%) | 7 (1.1%) |
| Asthma/COPD/Chronic lung disease | 1,322 (17%) | 82 (8.6%) | 78 (12%) |
| Obesity | 2,338 (29%) | 210 (22%) | 143 (23%) |
| Autoimmune disorder | 743 (9.4%) | 40 (4.2%) | 42 (6.6%) |
| Neurologic condition | 37 (0.5%) | 2 (0.2%) | 4 (0.6%) |
| Depression/anxiety/other mental health | 2,640 (33%) | 230 (24%) | 229 (36%) |
| Self-rated health status prior to COVID-19 | |||
| Excellent | 2,896 (37%) | 538 (57%) | 257 (41%) |
| Very good | 3,146 (40%) | 293 (31%) | 263 (42%) |
| Good | 1,432 (8%) | 97 (10%) | 90 (14%) |
| Fair | 430 (5.4%) | 20 (2.1%) | 22 (3.5%) |
| Poor | 21 (0.3%) | 0 (0%) | 1 (0.2%) |
| Vaccination status at the time of enrollment | |||
| None | 3,295 (42%) | 308 (32%) | 20 (3.2%) |
| Partial vaccination | 95 (6.2%) | 39 (4.1%) | 5 (0.8%) |
| Complete first series | 2,175 (27%) | 284 (30%) | 82 (13%) |
| ≥1 Booster | 1,947 (25%) | 317 (33%) | 525 (83%) |
| Timing of initial SARS_CoV_2/COVID-19 | |||
| January-June 2020 | 2,883 (36%) | 189 (20%) | NA |
| July-December 2020 | 3,219 (41%) | 378 (40%) | NA |
| January-June 2021 | 826 (10%) | 136 (14%) | NA |
| July-December 2021 | 856 (11%) | 203 (21%) | NA |
| Time between initial infection and survey completion (median [IQR] in days) | 325 (162,465) | 217 (127,383) | NA |
| Presenting symptoms at initial COVID-19 illness | |||
| Cardiopulmonary | 7,240 (91%) | 771 (81%) | NA |
| Neuropsychiatric | 7,090 (89%) | 714 (75%) | NA |
| Systemic | 7,325 (92%) | 774 (82%) | NA |
| Gastrointestinal | 4,827 (61%) | 346 (36%) | NA |
| Hospitalization status at initial COVID-19 illness | |||
| Not hospitalized | 6,870 (87%) | 925 (98%) | NA |
| Hospitalized | 1,049 (l3%) | 23 (2.4%) | NA |
| Medications at initial COVID-19 illness | |||
| Paxlovid/Molnupiravir | 21 (0.3%) | 0 (0%) | NA |
| Monoclonal antibodies | 237 (3.0%) | 9 (0.9%) | NA |
| Remdesivir | 310 (3.9%) | 2 (0.2%) | NA |
| Steroids | 655 (8.3%) | 18 (1.9%) | NA |
| Azithromycin | 122 (1.5%) | 4 (0.4%) | NA |
| Convalescent Plasma | 96 (1.2%) | 2 (0.2%) | NA |
| Vitamin C | 1,948 (25%) | 92 (9.7%) | NA |
| Vitamin D | 2,210 (28%) | 95 (10%) | NA |
| Zinc | 1,797 (23%) | 92 (9.7%) | NA |
Percentages do not add to 100% because of missing data.
As compared to the resolved-COVID sample, the long-COVID sample was more from the Northeast and West, less Black and Asian/Pacific Islander/Native Hawaiian, lower income, older age, and contained less cisgender men (p<0.05). As compared to the no-COVID sample, the long-COVID sample was more from the Midwest, more white, lower income, and older age (p<0.05). As compared to the no-COVID sample, the resolved-COVID sample contained more cisgender men and fewer Genderqueer/nonbinary individuals (p<0.05).
The majority (77%) of infections occurred before January 2021 and vaccine availability. At initial infection, 89% of those with long-COVID and 73% with resolved-COVID had not been vaccinated. However, at enrollment, 58% of those with long-COVID and 67% of those with resolved-COVID had completed at least one (adenovirus) or two (mRNA) vaccine doses. Among long-COVID the five most reported infection symptoms were lack of energy (86%), headache (71%), muscle aches (61%), fever (54%), and dyspnea (52%). While the types of symptoms were comparable for those with resolved-COVID, only 2.4% with resolved-COVID required hospitalization compared to 13% with long-COVID (Table 1).
Prior to COVID-19, 77% of long-COVID participants reported that their general health status was ‘very good’ or ‘excellent’, yet only 6.6% of individuals in this group reported the same health status as before their infection, compared to 81% of those with resolved-COVID (Figure 1A). Most individuals reported being physically active before infection (71% of long-COVID, 66% of resolved-COVID). However, 50% of individuals with long-COVID experienced declines in their physical activity, including 33% moving from ‘active’ to ‘sedentary’ compared to 3% among those with resolved-COVID (Figure 1B).
Figure 1.
Self-reported changes in health status and physical activity since COVID-19 illness among persons with long-COVID and resolved-COVID. A: Self-reported health relative to pre-COVID health among persons with long-COVID vs. resolved-covid. B: Self-reported exercise level before and after COVID-19 illness among persons with long-COVID vs. resolved-COVID.
COVID-19 status and disability
Of long-COVID participants, 65% were classified with at least one disability compared to 6% of those with resolved-COVID and 14% with no-COVID. Those requiring hospitalization had higher prevalence of disability (68% mobility, 80% IADL, 47% mental fatigue) but the prevalence was also elevated in those not hospitalized (37% mobility, 53% IADL, 33% mental fatigue). Overall, 22% of those with long-COVID had all three disabilities (hospitalized 37%, non-hospitalized 19%) (Figure S.2).
Of those with long-COVID, 41% had mobility disability, compared to 2.7% of resolved-COVID and 4.5% of no-COVID (Figure 2). In the long-COVID group, 6.1% indicated they could not walk ¼ mile or climb 10 stairs, and 0.9% were incapable of doing both. For IADL, 57% of those with long-COVID had at least some IADL difficulty with housework compared to 2.7% and 10% for resolved-COVID and no-COVID, respectively (Figure 2). Of those with long-COVID, 12% were unable to do either light or heavy housework and 1.4% could not complete either. The median WMFI score was 15 (IQR:7–24), 2 (IQR:0–6), and 5 (IQR:2–12) for long-COVID, resolved-COVID, and no-COVID, respectively (Figure 2). Of those with long-COVID, 6.8% were classified with severe mental fatigue (WMFI 30–34) and 4.7% critically severe (WMFI 35–36), compared to 0.2% (severe) and 0% (critically severe) for resolved-COVID and 0.9% (severe) and 0.3% (critically severe) for no-COVID. Across the severity spectrum, new physician diagnosis post-infection included tachycardia, postural-orthostatic tachycardia syndrome (POTS), myalgic encephalomyelitis, and irritable bowel syndrome (Figure S.10.1–S10.3) with the highest burden among those classified as critically disabled. Several non-symptom factors increased the odds of all three disabilities, including age, prior comorbidity, increased BMI, female gender, COVID-19 hospitalization, lower education, and lower income.
Figure 2.
Disability components by COVID-19 status: long-COVID, resolved-COVID, and no-COVID. A: Level of difficulty in walking ¼ mile. B: Level of difficulty in climbing ten stairs; C: Level of difficulty in doing heavy housework. D: Level of difficulty in doing light housework. E: Level of mental fatigue (WMFI score) by age. All estimates are weighted in order to standardize covariates of the group without long-COVID and the group with no history of COVID-19 to that of the long-COVID sample. Factors that were standardized included: age, gender, BMI, whether individuals were hospitalized during their acute phase of infection, and prior co-morbidity, specifically diabetes, cardiovascular disease, history of heart attack, congestive heart failure, high blood pressure, high cholesterol, history of stroke, autoimmune disorders, Hepatitis C, asthma, chronic lung disease, chronic kidney disease, cancer, depression, being pregnant, overweight or obese, anxiety or other mental health condition, and chronic or acute Lyme disease.
Symptoms during acute infection and disability
We examined associations between symptoms reported during initial infection with five components of disability (walking ¼ mile, climbing 10 stairs, light housework, heavy housework, WMFI score) stratified by hospitalization status (Figure 3). Dizziness (33% non-hospitalized individuals, 39% hospitalized individuals) was associated with all disability components in both groups. Heavy limbs (17%), dyspnea (49%), and tremors (8.5%) were associated with 4/5 disability components in the non-hospitalized group, and heavy limbs (21%) with 4/5 disability components in the hospitalized group (Tables S.2–S.11).
Figure 3.
Association between symptoms and symptom severity at the time of initial COVID-19 illness and A: difficulty in walking ¼ mile; B: difficulty in climbing ten stairs; C: difficulty in doing light housework; D: difficulty in doing heavy housework; and E: level of mental fatigue. In each plot, the five symptoms with the strongest overall associations are shown. All models are adjusted for age, gender, body mass index, calendar period of infection, activity level prior to infection and vaccination status.
Vaccination status
The effect of vaccination on long-COVID disability was restricted to vaccine status prior to infection to reduce reverse causality from those unable to get a vaccine due to disability/health concerns. Individuals who received two mRNA doses/one adenoviral dose had a median time of six months (IQR: 3.5, 8.5 months) between last dose and infection, while individuals who had 3+ mRNA/two adenoviral doses had two months (IQR: 1, 3 months). Vaccination was protective against development of mobility, IADL, and mental fatigue disability as compared to unvaccinated/one mRNA dose (Table 2: Model 1; Table S.5.1–S.5.2). We were not sufficiently powered to evaluate the interaction of time between vaccination and infection to account for potential waning vaccine effectiveness. In separate models, we adjusted for hospitalization (probably on the causal pathway between vaccination and disability); those who had 3+ mRNA/two adenoviral doses prior to their infection had a significantly decreased mobility, IADL, and mental fatigue disability compared to those who were unvaccinated/one mRNA dose (Table 2: Model 2; Table S.5.1–S.5.2). The additional IV analyses provided evidence of a protective effect of vaccination on mobility disability (Table 2: Model 3; Table S.5.3–S.5.5). While there was some protection afforded with respect to IADL disability, the estimate lacked precision, and there was no evidence of effect for mental fatigue.
Table 2.
Effect of vaccination status prior to COVID-19 illness on disability*
| Vaccination Status prior to illness** | Mobility Disability | IADL Disability | Mental Fatigue | |||
|---|---|---|---|---|---|---|
|
|
|
|
|
|||
| Odds Ratio | 95% CI | Odds Ratio | 95 % CI | Odds Ratio | 95% CI | |
|
| ||||||
| Model 1: adjusted for date of infection, age, gender, race, BMI, prior exercise, education, income, and prior comorbidity | ||||||
| Unvaccinated or 1 dose mRNA | 1 | 1 | 1 | |||
| 2 doses mRNA or 1 dose adenovirus | 0.57 | 0.42, 0.77 | 0.67 | 0.51, 0.87 | 0.72 | 0.52, 0.96 |
| 3+ doses mRNA or 2 doses adenovirus | 0.38 | 0.24, 0.63 | 0.51 | 0.33, 0.79 | 0.44 | 0.25, 0.67 |
|
| ||||||
| Model 2: adjusted for date of infection, age, gender, race, BMI, prior exercise, education, income, prior comorbidity and hospitalization | ||||||
| Unvaccinated or 1 dose mRNA | 1 | 1 | 1 | |||
| 2 doses mRNA or 1 dose adenovirus | 0.67 | 0.50, 0.91 | 0.77 | 0.59, 1.02 | 0.75 | 0.55, 1.00 |
| 3+ doses mRNA or 2 doses adenovirus | 0.48 | 0.29, 0.78 | 0.62 | 0.40, 0.96 | 0.46 | 0.27, 0.71 |
|
| ||||||
| Model 3: Instrumental variable analysis ** | ||||||
| Unvaccinated | 1 | 1 | 1 | |||
| 2+ doses mRNA or 1+ dose adenovirus | 0.22 | 0.06, 0.81 | 0.53 | 0.20, 1.38 | 1.12 | 0.38, 2.90 |
Sample is weighted to represent the COVID-19 Symptom Survey with survey weights to generalize to the US. Unweighted results are in Tables S.5.1 and S.5.4. Sensitivity analyses of the needed strength of non-causal pathways are shown in Tables S.5.2 and S5.5. Directed acyclic graph that guided analyses is shown in Figure S.5.1. Analysis restricted to those whose COVID-19 illness occurred after vaccines were available in the US. This resulted in a reduced sample size of 2,540 persons with evidence of long-COVID in models 1 and models 2. There were 2,464 persons in the instrumental variable analysis as 76 persons were missing length of time between vaccination and infection. Sample size was too small to separate 2 doses mRNA or 1 dose adenovirus and 3+ doses mRNA or 2 doses adenovirus for the instrumental variable analysis.
Age, gender, co-morbidity, education, income, ethnicity, race, and date of infection were included in both the first and second stage of the instrumental variable analyses. Strength of state as an instrumental variable is provided in Table S.5.
Discussion
In this large and heterogeneous sample from across the US, nearly two-thirds of individuals with long-COVID were classified with some disability, a burden that was nearly eleven times that of those with resolved-COVID. Moreover, 1% were critically physically disabled or unable to function and 5% had critically severe mental fatigue. Considering that 30–165 million people are expected to have long-COVID (10–55% of cases),4–6 we might expect 300,000–1.65 million people (1% of long-COVID) to be physically incapacitated, and 900,000–4.95 million (3% of long-COVID) to have critically severe mental fatigue. While it is reassuring that those with critically severe disabilities appear to seek necessary medical care (high levels of self-reported new physician diagnoses), these disabilities will likely compromise the ability of many to work or care for themselves and others, extending the long-term societal and economic impacts of COVID-19.7,8,19–22
Hospitalization during acute illness is the strongest driver of COVID-19-associated disability, and is consistent with prior studies.32 However, in our study sample, 60% of non-hospitalized individuals experienced at least one long-COVID associated disability and 5.5% were classified with at least one critical disability. Although we observed some associations with age, gender, race, and pre-existing comorbidities with long-COVID disabilities, they contrasted with the risk factors of male gender, age, and comorbidities that have been consistently associated with COVID-19 severity and hospitalization.33,34 Essentially, these data suggest that factors predictive of hospitalization or severity of initial infection cannot be used to predict who will develop long-COVID-associated disability among those not requiring hospitalization.
While there was substantial heterogeneity in the presenting symptoms of acute COVID-19 in those with long-COVID, there were some patterns that emerged among those with incident disabilities. Among non-hospitalized individuals, dizziness, tremors, heavy limbs, and dyspnea were consistently associated with increased occurrence and sometimes severity of disability. Even among those hospitalized, dizziness and heavy limbs were related to all types of disability and severity. Dizziness and tremors are neurologic outcomes that may represent multiple disease pathways including disruption of the vestibular system and dysautonomia, and POTS.35,36 Associations of dizziness and tremors with imbalance may also explain observed associations with mobility/IADL disability.37,38 Additionally, the sensation of heavy limbs is often attributable to venous insufficiency or peripheral arterial disease.39,40 This incident symptom may represent the effect of SARS-CoV-2 on the cardiovascular system with alterations to microvascular perfusion on the peripheral limbs, thromboembolism, and endothelial injury. While the high burden of physician diagnoses, particularly among the most severely disabled group, suggests that many are seeking clinical care, our data reinforces the need for those with early lingering symptoms to seek medical care that considers careful neurologic, vestibular, cardiovascular, and/or pulmonary evaluations. Additionally, there may be opportunities for the early administration of medications like corticosteroids,38 although it remains unclear whether early intervention will alleviate long-term impacts.
Additional clarity is needed on whether vaccination reduces disability risk associated with long-COVID. We observed over a 50% reduction in odds for all three disabilities among those with 3+ mRNA/two adenoviral doses and a modest reduction in odds for those with two mRNA/one adenoviral dose. Some of this effect may be mediated through protection from severe COVID-19 resulting in hospitalization leading to disability. The strong reduction of mobility disability persisted in the 2+ mRNA doses/1+ adenoviral dose IV analysis that is robust to measured or unmeasured confounding (Supplement). With increasing viral positivity in the US, the window of complete vaccination (initial doses + booster) prior to infection is closing. However, these data support that vaccines may afford some protection from disabilities, as well as the intended purpose of attenuation of severe infection and hospitalization.
The cross-sectional design prevented evaluation of the impact of vaccination after infection on the resolution/exacerbation of COVID-19 disabilities. Comprehensive longitudinal studies are needed to explore the association of timing and dose of vaccination as well as number and timing of repeat infections. In this sample, we observed a high burden of all disabilities among those with long-COVID. While our sample may be skewed towards individuals reporting long-COVID, it was beneficial to identify common patterns in the long-COVID group, and especially those with disability. We included those with resolved-COVID and no-COVID for comparisons, acknowledging that the pandemic, outside of infection, was stressful and traumatic for some, and may contribute independently to disability. We lack representation in demographic groups that were especially impacted by the pandemic, and thus may underestimate the true burden of long-COVID disability across race and socio-economic factors; however, this was minimized in the vaccine analysis by weighting the study sample to the COVID-19 Symptom Survey sample weights increasing the generalizability of these results to the US population. Nevertheless, representativeness depends on the differences in distribution of effect modifiers in our sample as compared to the general population (Supplement Methods and Figures S.6.1–S.6.3). The use of a study survey provided the opportunity to capture the heterogeneity in initial and new/persistent symptoms but is subject to recall bias especially when the time of infection was far from enrollment in the study. However, this approach is more inclusive of the spectrum of COVID-19 and long-COVID symptoms than studies that use electronic health records that rely on information from those seeking medical care only.
In conclusion, we observed a high burden of physical and mental fatigue disability among individuals with long-COVID, which may have long-term consequences for these individuals and society-at-large. Long-term follow-up of those with long-COVID is necessary. In the meantime, specialized care is needed for the growing population of individuals with long-COVID to reduce the burden and impact of disability.
Supplementary Material
Clinical Significance Bullets.
65% with long-COVID had at least one new onset disability and 22% were disabled based on all three definitions: basic mobility, activities of daily living and mental fatigue
6% of long-COVID individuals are classified as critically disabled, and of those, 77% were never hospitalized for their infection.
4 symptoms (dizziness, tremors, dyspnea, heavy limbs) occurring during the acute phase of infection are associated with new onset disabilities
Funding/Support:
This study was supported by the Johns Hopkins University COVID-19 Research Response Program and in part by Johns Hopkins University Center for AIDS Research (P30AI094189), which is supported by the following NIH Co-Funding and Participating Institutes and Centers: NIAID, NCI, NICHD, NHLBI, NIDA, NIA, NIGMS, NIDDK, NIMHD. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Footnotes
Conflict of Interest Disclosures: We have no conflicts to disclose.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Paul P, El-Naas A, Hamad O, et al. Effectiveness of the pre-Omicron COVID-19 vaccines against Omicron in reducing infection, hospitalization, severity, and mortality compared to Delta and other variants: A systematic review. Hum Vaccin Immunother. 2023;19(1):2167410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.van der Ley PA, Zariri A, van Riet E, Oosterhoff D, Kruiswijk CP. An Intranasal OMV-Based Vaccine Induces High Mucosal and Systemic Protecting Immunity Against a SARS-CoV-2 Infection. Front Immunol. 2021;12:781280. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wang X, Du Z, James E, et al. The effectiveness of COVID-19 testing and contact tracing in a US city. Proc Natl Acad Sci U S A. 2022;119(34):e2200652119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Domingo FR, Waddell LA, Cheung AM, et al. Prevalence of long-term effects in individuals diagnosed with COVID-19: an updated living systematic review. doi: 10.1101/2021.06.03.21258317 [DOI] [Google Scholar]
- 5.van Kessel SAM, Olde Hartman TC, Lucassen PLBJ, van Jaarsveld CHM. Post-acute and long-COVID-19 symptoms in patients with mild diseases: a systematic review. Fam Pract. 2022;39(1):159–167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Perlis RH, Santillana M, Ognyanova K, et al. Prevalence and Correlates of Long COVID Symptoms Among US Adults. JAMA Netw Open. 2022;5(10):e2238804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tabacof L, Nicolau E, Rivera A, Putrino D. Post-COVID Conditions and Burden of Disease. Phys Med Rehabil Clin N Am. 2023;34(3):499–511. [DOI] [PubMed] [Google Scholar]
- 8.Katz GM, Bach K, Bobos P, et al. Understanding How Post-COVID-19 Condition Affects Adults and Health Care Systems. JAMA Health Forum. 2023;4(7):e231933. [DOI] [PubMed] [Google Scholar]
- 9.Bliddal S, Banasik K, Pedersen OB, et al. Acute and persistent symptoms in non-hospitalized PCR-confirmed COVID-19 patients. Sci Rep. 2021;11(1):13153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Peghin M, Palese A, Venturini M, et al. Post-COVID-19 symptoms 6 months after acute infection among hospitalized and non-hospitalized patients. Clin Microbiol Infect. 2021;27(10):1507–1513. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Carvalho-Schneider C, Laurent E, Lemaignen A, et al. Follow-up of adults with noncritical COVID-19 two months after symptom onset. Clin Microbiol Infect. 2021;27(2):258–263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hastie CE, Lowe DJ, McAuley A, et al. Outcomes among confirmed cases and a matched comparison group in the Long-COVID in Scotland study. Nat Commun. 2022;13(1):5663. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Al-Aly Z, Xie Y, Bowe B. High-dimensional characterization of post-acute sequelae of COVID-19. Nature. 2021;594(7862):259–264. [DOI] [PubMed] [Google Scholar]
- 14.Xie Y, Bowe B, Al-Aly Z. Burdens of post-acute sequelae of COVID-19 by severity of acute infection, demographics and health status. Nat Commun. 2021;12(1):6571. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Groff D, Sun A, Ssentongo AE, et al. Short-term and Long-term Rates of Postacute Sequelae of SARS-CoV-2 Infection. JAMA Network Open. 2021;4(10):e2128568. doi: 10.1001/jamanetworkopen.2021.28568 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Subramanian A, Nirantharakumar K, Hughes S, et al. Symptoms and risk factors for long COVID in non-hospitalized adults. Nat Med. 2022;28(8):1706–1714. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Taboada M, Moreno E, Cariñena A, et al. Quality of life, functional status, and persistent symptoms after intensive care of COVID-19 patients. Br J Anaesth. 2021;126(3):e110–e113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Walle-Hansen MM, Ranhoff AH, Mellingsæter M, Wang-Hansen MS, Myrstad M. Health-related quality of life, functional decline, and long-term mortality in older patients following hospitalisation due to COVID-19. BMC Geriatr. 2021;21(1):199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Duggal P, Penson T, Manley HN, et al. Post-sequelae symptoms and comorbidities after COVID-19. J Med Virol. 2022;94(5):2060–2066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sandmann FG, Tessier E, Lacy J, et al. Long-Term Health-Related Quality of Life in Non-Hospitalized Coronavirus Disease 2019 (COVID-19) Cases With Confirmed Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Infection in England: Longitudinal Analysis and Cross-Sectional Comparison With Controls. Clin Infect Dis. 2022;75(1):e962–e973. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kokolevich ZM, Crowe M, Mendez D, Biros E, Reznik JE. Most Common Long COVID Physical Symptoms in Working Age Adults Who Experienced Mild COVID-19 Infection: A Scoping Review. Healthcare (Basel). 2022;10(12). doi: 10.3390/healthcare10122577 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Vaes AW, Machado FVC, Meys R, et al. Care Dependency in Non-Hospitalized Patients with COVID-19. J Clin Med Res. 2020;9(9). doi: 10.3390/jcm9092946 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Health Organization W. A clinical case definition of post COVID-19 condition by a Delphi consensus. Accessed November 2, 2022. https://apps.who.int/iris/bitstream/handle/10665/345824/WHO-2019-nCoV-Post-COVID-19-condition-Clinical-case-definition-2021.1-eng.pdf?utm_source=yxnews&utm_medium=mobile [DOI] [PMC free article] [PubMed]
- 24.Soriano JB, Murthy S, Marshall JC, Relan P, Diaz JV, WHO Clinical Case Definition Working Group on Post-COVID-19 Condition. A clinical case definition of post-COVID-19 condition by a Delphi consensus. Lancet Infect Dis. 2022;22(4):e102–e107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: Building an international community of software platform partners. J Biomed Inform. 2019;95:103208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Stone JL, Norris AH. Activities and attitudes of participants in the Baltimore longitudinal study. J Gerontol. 1966;21(4):575–580. [DOI] [PubMed] [Google Scholar]
- 28.Bentall RP, Wood GC, Marrinan T, Deans C, Edwards RH. A brief mental fatigue questionnaire. Br J Clin Psychol. 1993;32(3):375–379. [DOI] [PubMed] [Google Scholar]
- 29.Westreich D, Edwards JK, Lesko CR, Stuart E, Cole SR. Transportability of Trial Results Using Inverse Odds of Sampling Weights. Am J Epidemiol. 2017;186(8):1010–1014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ishwaran H, Lu M. Standard errors and confidence intervals for variable importance in random forest regression, classification, and survival. Stat Med. 2019;38(4):558–582. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kreuter F, Barkay N, Bilinski A, et al. Partnering with a global platform to inform research and public policy making. In: Survey Research Methods. Vol 14. ; 2020:159–163. [Google Scholar]
- 32.Tsampasian V, Elghazaly H, Chattopadhyay R, et al. Risk Factors Associated With Post-COVID-19 Condition: A Systematic Review and Meta-analysis. JAMA Intern Med. 2023;183(6):566–580. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Vahey GM, McDonald E, Marshall K, et al. Risk factors for hospitalization among persons with COVID-19-Colorado. PLoS One. 2021;16(9):e0256917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Ioannou GN, Locke E, Green P, et al. Risk Factors for Hospitalization, Mechanical Ventilation, or Death Among 10 131 US Veterans With SARS-CoV-2 Infection. JAMA Netw Open. 2020;3(9):e2022310. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Raj SR. Postural tachycardia syndrome (POTS). Circulation. 2013;127(23):2336–2342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Carew S, Connor MO, Cooke J, et al. A review of postural orthostatic tachycardia syndrome. Europace. 2009;11(1):18–25. [DOI] [PubMed] [Google Scholar]
- 37.Sugaya N, Arai M, Goto F. Changes in cognitive function in patients with intractable dizziness following vestibular rehabilitation. Sci Rep. 2018;8(1):9984. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Kim G, Seo JH, Lee SJ, Lee DH. Therapeutic effect of steroids on vestibular neuritis: Systematic review and meta-analysis. Clin Otolaryngol. 2022;47(1):34–43. [DOI] [PubMed] [Google Scholar]
- 39.Blättler W, Thomae HJ, Amsler F. Venous leg symptoms in healthy subjects assessed during prolonged standing. J Vasc Surg Venous Lymphat Disord. 2016;4(4):455–462. [DOI] [PubMed] [Google Scholar]
- 40.Halperin JL. Evaluation of patients with peripheral vascular disease. Thromb Res. 2002;106(6):V303–11. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.