Abstract
Background
Clinical details of long COVID are still not well understood because of potential confounding with a wide range of pre-existing comorbidities.
Methods
The present study used datasets from a nation-wide, cross-sectional, online survey. We determined which prolonged symptoms were more likely to be associated with post-COVID condition after adjusting for a wide range of comorbidities and baseline characteristics. This study also used EuroQol 5 dimensions 5-levels (EQ-5D-5L) and Somatic Symptom Scale-8 to assess health-related quality of life (QOL) and somatic symptoms in individuals with a previous history of COVID-19, defined as the diagnosis of COVID-19 made at least 2 months prior to the online survey.
Results
In total, 19,784 respondents were included for analysis; of these, 2,397 (12.1%) had a previous history of COVID-19. The absolute difference of adjusted prevalence of symptoms attributed to prolonged symptoms after COVID-19 ranged from -0.4 % to +2.0 %. Headache (adjusted odds ratio [aOR]: 1.22; 95% confidence interval [95% CI]:1.07-1.39), chest discomfort (aOR:1.34, 95% CI:1.01-1.77), dysgeusia (aOR: 2.05, 95% CI: 1.39-3.04), and dysosmia (aOR: 1.96, 95% CI: 1.35-2.84) were independently associated with a previous history of COVID-19. Individuals with a history of COVID-19 had lower health-related QOL scores.
Conclusion
After adjusting for potential comorbidities and confounders, clinical symptoms, such as headache, chest discomfort, dysgeusia, and dysosmia, were found to be independently associated with a history of COVID-19, which was diagnosed two or more months previously. These protracted symptoms may have impacted QOL and the overall somatic symptom burden in subjects with a previous history of COVID-19.
Keywords: Long COVID, prolonged symptoms, quality of life, somatic symptoms
Introduction
Long coronavirus disease (long COVID) is considered one of the most devastating sequelae of COVID-19. 1 Long COVID is clinically defined as the continuation or development of new symptoms three months after the initial SARS-CoV-2 infection which lasts at least two months without a discernible reason.2 Given the global extent of COVID-19, many people with a history of COVID-19 may have long COVID. The reported prevalence of long COVID varies substantially, ranging from 10 % to 70 %, depending on the severity of acute SARS-CoV2 infection.3 , 4 A recent, cross-sectional study in the United States revealed that up to 15% of patients with COVID-19 later had symptoms of long COVID.5 A meta-analysis of 54 studies demonstrated that 6.2% of patients with COVID-19 had persistent symptoms suggestive of long COVID.6 However, estimating true prevalence of long COVID is challenging primarily because of the ambiguity of its case definition, and accordingly, differences in survey methodologies.7
Although some studies assessed risk factors for long COVID, clinical symptoms associated with the long COVID is frequently confounded by a wide range of pre-existing comorbidities.8 Moreover, the strength of the association between long COVID and its various clinical symptoms is also not clearly understood.
Using data culled from an online survey conducted in Japan in 2022, the present study aimed to assess the prevalence of prolonged symptoms after COVID-19, the strength of the association of COVID-19 with each alleged symptom, the impact on the patients’ quality of life (QOL), and the presence of severe somatic symptoms.
Methods
Study participants
The present study used datasets from a nationwide, cross-sectional, online survey (Japan COVID-19 Society Internet Survey: JACSIS).9 JACSIS consists of pooled panels of participants for the survey since the beginning of the COVID-19 pandemic which is managed by an internet research company (Rakuten Insight Co.).10 Participants received a monetary incentive of JPY 100 (about USD 0.7 in September 2022) for responding to the survey. The present study was performed in accordance with the STROBE statement and was conducted following the 1964 Helsinki Declaration and its later amendments. The ethics committee at Osaka International Cancer Institute approved the study protocol (June 19, 2020; approval number 20084).
Data collection
An online survey was done from September 15th through October 15th, 2022. We asked each respondent about a previous history of COVID-19 and whether its onset was more than 2 months prior to the online survey date. We collected information on demographic characteristics, pre-existing comorbidities, and the presence of nineteen current symptoms which would indicate long COVID, lasting at least two months from the JACSIS respondents. The present study excluded participants with inconsistent responses and those who satisfied at least one of the following conditions: (1) did not provide a correct response to an attention check question included in the middle of the questionnaire asking respondents to “select the penultimate option from five given options”; (2) responded that they had used all ten categories of drugs listed (nine categories of drugs in the JACSIS 2022); (3) had all 20 health conditions listed; (4) had family members > 15.
We also obtained the information on COVID-19 vaccination status: 1) no vaccination or incomplete initial vaccination series (i.e., only received one dose of COVID-19 vaccine); 2) completion of the initial vaccination series without booster vaccination; and 3) completion of the initial vaccination series and at least one booster vaccination. The type of vaccine was not recorded.
The EuroQol 5 dimensions 5-levels (EQ-5D-5L) and Somatic Symptom Scale-8 (SSS-8) scores were employed to assess health-related QOL and somatic symptoms. The EQ-5D-5L consists of mobility, self-care, usual activities, pain/discomfort, and anxiety/depression and was shown to be effective for assessing health-related QOL.11 The EQ-5D-5L ranges from 0 to 1, with diminishing scores indicating an increasingly negative impact on health-related QOL. The SSS-8 consists of eight items assessing the severity of common somatic symptoms and has also demonstrated good reliability and validity.12 Increasing SSS-8 scores indicate increasing severity of somatic symptoms.
The respondents were classified into two categories: “no history of COVID-19” or previous history of COVID-19” group. Further, respondents with no history of COVID-19 were subcategorized into a suspected COVID-19 group if they had a history of close contact with an individual(s) known to have COVID-19 (i.e., they were potentially asymptomatic COVID-19 individuals) to minimize the misclassification of these respondents to be truly negative COVID-19 group.
Statistical analysis
We employed multivariate logistic regression model to assess the association between the previous history of COVID-19 and each symptom attributed to long COVID to elucidate which symptoms were more likely to be associated with long COVID. On univariate analysis, we compared categorical variables with the χ2 test or Fisher's exact test as appropriate and assessed continuous variables using the Mann-Whitney U test. A previous history of COVID-19 (the variable of interest) was forced into the multivariate model. Potential determinants associated with long COVID, including gender, age, and a wide range of comorbidities, were also included in the final model.5 , 8 , 13 , 14 The primary analysis was done for respondents with a previous history of COVID-19 (i.e., the diagnosis of COVID-19 made at least 2 months prior to the online survey). and those with no COVID-19 after excluding the suspected COVID-19 group. We also performed sensitivity analysis for all the respondents (i.e., including suspected individuals), and respondents with history of COVID-19 between two to 12 months prior to the survey date. We estimated crude and adjusted prevalence and adjusted odds ratio by assessing the association between respondents with a previous history of COVID-19 and the prevalence of symptoms lasting longer than two months.
We also employed multivariate adjusted linear regression model to estimate the impact of a previous history of COVID-19 on EQ-5D-5L and SSS-8 scores in the entire cohort of survey respondents. Sensitivity analysis was also done after including the suspected individual group. Two-sided P < 0.05 was considered to indicate statistical significance. All statistical analyses were performed using STATA version 16 (STATA Corp., College Station, NC, USA).
Results
In total, 32,000 respondents participated in the online survey in September 2022, and 3,370 (10.5%) were excluded after giving inconsistent responses. Of the remaining 28,630 respondents, 1,909 (6.7%) with a history of COVID-19 within the last two months from the survey date were excluded. Of the 26,721 respondents who remained, 6,937 (25.9%) with no previous history of COVID-19 but having close contact with a person with a diagnosis of COVID-19 (suspected individuals) were excluded, leaving 19,784 (74.0%) respondents who were eligible for the primary analysis (Figure 1 ).
Figure 1.
Study population of the nationwide, cross-sectional, online survey (Japan COVID-19 Society Internet Survey: JACSIS)
Table 1 shows the baseline characteristics of the respondents; 2,397 (12.1%) had a previous history of COVID-19, and 17,387 (87.9%) had neither had a previous history of COVID-19 nor close contact with a person with a diagnosis of COVID-19. Male and younger respondents, in particular, those without booster vaccination, tended to have a previous history of COVID-19 infection more frequently. The respondents with a previous history of COVID-19 had a higher proportion of a wide range of comorbidities than those without a history of COVID-19.
Table 1.
Baseline characteristics of respondents with and without COVID-19 (N = 19,784)
| Total (n = 19,784) | Respondents with COVID-19 (n = 2,397) | Respondents without COVID-19 (n = 17,387) | P-value | |
|---|---|---|---|---|
| Age, median (IQR) years | 49 (34-67) | 35 (27-44) | 51 (36-68) | < 0.001 |
| Male gender | 9,585 (48.5) | 1,274 (53.2) | 8,311 (47.8) | < 0.001 |
| Education | < 0.001 | |||
| High school or lower | 7,994 (40.4) | 833 (34.8) | 7,161 (41.2) | |
| College or higher | 11,790 (59.6) | 1,564 (65.3) | 10,226 (58.8) | |
| Income [Japanese yen] | < 0.001 | |||
| < 4 million | 6,154 (31.11) | 553 (23.1) | 5,601 (32.2) | |
| ≧ 4 million | 9,374 (47.4) | 1,485 (62.0) | 7,889 (45.4) | |
| N/A | 4,256 (21.5) | 359 (15.0) | 3,897 (22.4) | |
| Partner | 12,182 (61.6) | 1,479 (61.7) | 10,703 (61.6) | 0.89 |
| Employment status | < 0.001 | |||
| Executive, employer, or self-employed | 2,127 (10.8) | 293 (12.2) | 1,834 (10.6) | |
| Employee | 9,057 (45.8) | 1,582 (66.0) | 7,475 (43.0) | |
| Homewife or house-husband | 4,033 (20.4) | 246 (10.3) | 3,787 (21.8) | |
| Unemployed, retired, or student | 4,567 (23.1) | 276 (11.5) | 4,291 (24.7) | |
| Smoking | 7,126 (36.0) | 1,040 (43.4) | 6,086 (35.0) | < 0.001 |
| Alcohol | 11,275 (57.0) | 1,525 (63.6) | 9,750 (56.1) | < 0.001 |
| Comorbidities | ||||
| Hypertension | 5,681 (28.7) | 704 (29.4) | 4,977 (28.6) | 0.45 |
| Diabetes mellitus | 1,947 (9.8) | 462 (19.3) | 1,485 (8.5) | < 0.001 |
| Dyslipidemia | 3,996 (20.2) | 567 (23.7) | 3,429 (19.7) | < 0.001 |
| Asthma | 2,625 (13.3) | 608 (25.4) | 2,017 (11.6) | < 0.001 |
| Cardiovascular disease | 1,061 (5.4) | 341 (14.2) | 720 (4.1) | < 0.001 |
| Cerebral vascular disease | 837 (4.2) | 334 (13.9) | 503 (2.9) | < 0.001 |
| COPD | 590 (3.0) | 320 (13.4) | 270 (1.6) | < 0.001 |
| CKD | 845 (4.3) | 328 (13.7) | 517 (3.0) | < 0.001 |
| Chronic liver disease | 753 (3.8) | 328 (13.7) | 425 (2.4) | < 0.001 |
| Immunocompromised status | 951 (4.8) | 346 (14.4) | 605 (3.5) | < 0.001 |
| Malignancy | 1,726 (8.7) | 353 (14.7) | 1,373 (7.9) | < 0.001 |
| Number of COVID-19 vaccines | < 0.001 | |||
| 0 | 2,562 (13.0) | 399 (16.7) | 2,163 (12.4) | |
| 1 | 120 (0.6) | 52 (2.2) | 68 (0.4) | |
| 2 | 2,507 (12.7) | 554 (23.1) | 1,953 (11.2) | |
| 3 or 4 | 14,595 (73.8) | 1,392 (58.1) | 13,203 (75.9) |
NOTE:
All categorical data are presented as a number (percentage, %) except age, which is presented as the median (interquartile range, IQR).
Abbreviations: COVID-19, coronavirus disease 2019; IQR, interquartile range; N/A, data not available; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease.
Table 2 shows the results of multivariate logistic regression analysis on the association between respondents with a previous history of COVID-19 and the prevalence of symptoms lasting longer than two months. The absolute difference of the adjusted prevalence of symptoms attributed to post-COVID conditions was + 2.0 % for back pain, +1.9 % for headache, +0.8 % for dysgeusia, +1.3 % for dysosmia, and +1.7 % for hair loss. After adjusting for demographic characteristics, a wide range of comorbidities, and COVID-19 vaccination status, a previous history of COVID-19 was independently associated with back pain (adjusted odds ratio [aOR]:1.13, 95% confidence interval [95% CI]:1.01-1.26), headache (aOR:1.22, 95% CI: 1.07-1.39), chest discomfort (aOR:1.34, 95% CI: 1.01-1.77), dysgeusia (aOR: 2.05, 95% CI: 1.39-3.04), dysosmia (aOR: 1.96, 95% CI: 1.35-2.84), and hair loss (aOR:1.29, 95% CI: 1.09-1.50). Sensitivity analysis of the cohort, including the suspected cases, suggested that a previous history of COVID-19 was independently associated with dysgeusia and dysosmia (Supplementary Table 1). Another sensitivity analysis only including respondents with a previous history of COVID-19 two to 12 months prior to the survey date and those who never had COVID-19 revealed that a previous history of COVID-19 was independently associated with back pain, headache, dysosmia, memory impairment, hair loss, erectile dysfunction, fatigue, and cough (Supplementary Table 2).
Table 2.
Association between respondents with history of COVID-19 and prevalence of symptoms lasting for more than two months (N = 19,784)
| Sample, n | Prevalence, n (%) | Adjusted rate, % (95% CI) | Absolute difference of adjusted prevalence (%) | Adjusted OR (95% CI) | |
|---|---|---|---|---|---|
| GI upset | |||||
| Respondents with COVID-19 | 2,397 | 331 (13.8) | 13.3 (11.8-14.7) | +1.0 | 1.10 (0.96-1.25) |
| Respondents without COVID-19 | 17,387 | 2,121 (12.2) | 12.3 (11.8-12.8) | Reference | |
| Back pain | |||||
| Respondents with COVID-19 | 2,397 | 536 (22.4) | 24.4 (22.6-26.3) | +2.0 | 1.13 (1.01-1.26) |
| Respondents without COVID-19 | 17,387 | 3,936 (22.6) | 22.4 (21.8-23.0) | Reference | |
| Arm/Leg/Joint pain | |||||
| Respondents with COVID-19 | 2,397 | 312 (13.0) | 17.7 (15.9-19.5) | -0.1 | 0.99 (0.86-1.14) |
| Respondents without COVID-19 | 17,387 | 3,199 (18.4) | 17.8 (17.2-18.3) | Reference | |
| Headache | |||||
| Respondents with COVID-19 | 2,397 | 369 (15.4) | 12.1 (10.9-13.3) | +1.9 | 1.22 (1.07-1.39) |
| Respondents without COVID-19 | 17,387 | 1,712 (9.8) | 10.2 (9.8-10.7) | Reference | |
| Chest discomfort | |||||
| Respondents with COVID-19 | 2,397 | 96 (4.0) | 2.5 (1.9-3.1) | +0.6 | 1.34 (1.01-1.77) |
| Respondents without COVID-19 | 17,387 | 305 (1.8) | 1.9 (1.7-2.1) | Reference | |
| Dyspnea | |||||
| Respondents with COVID-19 | 2,397 | 171 (7.1) | 5.7 (4.7-6.6) | +0.9 | 1.20 (0.97-1.48) |
| Respondents without COVID-19 | 17,387 | 807 (4.6) | 4.8 (4.5-5.1) | Reference | |
| Dizziness | |||||
| Respondents with COVID-19 | 2,397 | 191 (8.0) | 5.9 (5.0-6.8) | -0.4 | 0.93 (0.77-1.11) |
| Respondents without COVID-19 | 17,387 | 1,040 (6.0) | 6.3 (5.9-6.7) | Reference | |
| Sleep disturbance | |||||
| Respondents with COVID-19 | 2,397 | 418 (17.4) | 17.5 (15.9-19.1) | -1.0 | 0.93 (0.83-1.06) |
| Respondents without COVID-19 | 17,387 | 3,215 (18.5) | 18.5 (17.9-19.1) | Reference | |
| Hearing loss | |||||
| Respondents with COVID-19 | 2,397 | 158 (6.6) | 8.8 (7.4-10.2) | +1.2 | 1.18 (0.96-1.44) |
| Respondents without COVID-19 | 17,387 | 1,373 (7.9) | 7.6 (7.2-8.0) | Reference | |
| Dysgeusia | |||||
| Respondents with COVID-19 | 2,397 | 70 (2.9) | 1.7 (1.2-2.3) | +0.8 | 2.05 (1.39-3.04) |
| Respondents without COVID-19 | 17,387 | 136 (0.8) | 0.9 (0.7-1.0) | Reference | |
| Dysosmia | |||||
| Respondents with COVID-19 | 2,397 | 64 (2.7) | 2.3 (1.6-3.0) | +1.3 | 1.96 (1.35-2.84) |
| Respondents without COVID-19 | 17,387 | 200 (1.2) | 1.2 (1.0-1.3) | Reference | |
| Memory impairment | |||||
| Respondents with COVID-19 | 2,397 | 134 (5.6) | 5.9 (4.9-7.1) | +0.5 | 1.11 (0.90-1.37) |
| Respondents without COVID-19 | 17,387 | 947 (5.4) | 5.4 (5.1-5.7) | Reference | |
| Poor concentration | |||||
| Respondents with COVID-19 | 2,397 | 248 (10.3) | 9.6 (8.4-10.8) | -0.6 | 0.94 (0.80-1.09) |
| Respondents without COVID-19 | 17,387 | 1,752 (10.1) | 10.2 (9.7-10.6) | Reference | |
| Hair loss | |||||
| Respondents with COVID-19 | 2,397 | 202 (8.4) | 8.5 (7.3-9.7) | +1.7 | 1.29 (1.09-1.53) |
| Respondents without COVID-19 | 17,387 | 1,176 (6.8) | 6.8 (6.4-7.1) | Reference | |
| Decreased libido | |||||
| Respondents with COVID-19 | 2,397 | 164 (6.8) | 6.8 (5.7-7.8) | +1.0 | 1.18 (0.97-1.43) |
| Respondents without COVID-19 | 17,387 | 1,014 (5.8) | 5.8 (5.5-6.2) | Reference | |
| Erectile dysfunction a | |||||
| Respondents with COVID-19 | 1,274 | 110 (8.6) | 12.1 (9.9-14.4) | +2.0 | 1.25 (0.98-1.59) |
| Respondents without COVID-19 | 8,311 | 877 (10.6) | 10.1 (9.5-10.7) | Reference | |
| Fatigue | |||||
| Respondents with COVID-19 | 2,397 | 564 (23.5) | 12.1 (9.9-14.4) | +2.0 | 1.25 (0.98-1.59) |
| Respondents without COVID-19 | 17,387 | 3,507 (20.2) | 10.1 (9.5-10.7) | Reference | |
| Cough | |||||
| Respondents with COVID-19 | 2,397 | 83 (3.5) | 4.0 (3.1-4.9) | +0.6 | 1.21 (0.93-1.59) |
| Respondents without COVID-19 | 17,387 | 595 (3.4) | 3.4 (3.1-3.6) | Reference | |
| Fever | |||||
| Respondents with COVID-19 | 2,397 | 18 (0.8) | 0.4 (0.2-0.7) | -0.2 | 0.70 (0.39-1.25) |
| Respondents without COVID-19 | 17,387 | 101 (0.6) | 0.6 (0.5-0.8) | Reference |
NOTE:
Respondents who first contracted COVID-19 within the last two month and whose family members or coworkers contracted COVID-19 were excluded.
Adjusted for age (-29, 30-39, 40-49, 50-59, 60-69, 70-), sex, smoking, alcohol, hypertension, diabetes mellitus, dyslipidemia, asthma, cardiovascular diseases, cerebrovascular diseases, COPD, CKD, chronic liver diseases, immunocompromised status, malignancies, and COVID-19 vaccine status (0, 1, 2, 3 or 4 doses).
Abbreviations: COVID-19, coronavirus disease 2019; CI, confidence interval; aOR, adjusted odds ratio; GI, gastrointestinal; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease.
Female respondents were excluded.
Table 3 represents the health-related quality of life scores of EQ-5D-5L and SSS-8 scores of the respondents. The median SSS-8 score was eight and five among those with a previous history of COVID-19 and without a history of COVID-19, respectively. The prevalence of each somatic symptoms on the SSS-8 was higher in the respondents with a previous history of COVID-19. The median EQ-5D-5L score in the respondents with a previous history of COVID-19 was 0.9 (interquartile range [IQR]: 0.8-1.0) and 1.0 (IQR: 0.8-1.0) in those without any history of COVID-19. The respondents who received more than 2 doses of COVID-19 vaccination tended to have a lower prevalence of each domain in the EQ-5D-5L and each somatic symptom in SSS-8.
Table 3.
Symptoms and QOL of respondents by COVID-19 and vaccination status (N = 19,784)
| Total (n = 19,784) | COVID-19 infection |
Number of COVID-19 vaccines |
|||||
|---|---|---|---|---|---|---|---|
| Respondents with COVID-19 (n = 2,397) | Respondents without COVID-19 (n = 17,387) | 0 (n = 2,562) | 1 (n = 120) | 2 (n = 2,507) | 3 or 4 (n = 14,595) | ||
| SSS-8 [IQR] | 5 (2-10) | 8 (3-14) | 5 (2-9) | 6 (1-12) | 10 (4.5-15.5) | 6 (2-11) | 5 (2-9) |
| Stomach or bowel problem | 9,119 (46.1) | 1,411 (58.9) | 7,708 (44.3) | 1,207 (47.1) | 85 (70.8) | 1,170 (46.7) | 6,657 (45.6) |
| Back pain | 11,448 (57.9) | 1,589 (66.3) | 9,859 (56.7) | 1,410 (55.0) | 84 (70.0) | 1,450 (57.8) | 8,504 (58.3) |
| Pain in arms, legs or joints | 9,853 (49.8) | 1,320 (55.1) | 8,533 (49.1) | 1,198 (46.8) | 75 (62.5) | 1,117 (44.6) | 7,463 (51.1) |
| Headache | 8,242 (41.7) | 1,436 (59.9) | 6,806 (39.1) | 1,171 (45.7) | 79 (65.8) | 1,268 (50.6) | 5,724 (39.2) |
| Chest pain or shortness of breath | 6,167 (31.2) | 1,063 (44.4) | 5,104 (29.4) | 914 (35.7) | 67 (55.8) | 822 (32.8) | 4,364 (29.9) |
| Dizziness | 5,674 (28.7) | 1,002 (41.8) | 4,672 (26.9) | 863 (33.7) | 69 (57.5) | 848 (33.8) | 3,894 (26.7) |
| Feeling tired or having low energy | 11,918 (60.2) | 1,703 (71.1) | 10,215 (58.8) | 1,519 (59.3) | 78 (65.0) | 1,611 (64.3) | 8,710 (59.7) |
| Trouble sleeping | 10,104 (51.1) | 1,424 (59.4) | 8,680 (49.9) | 1,350 (52.7) | 80 (66.7) | 1,299 (51.8) | 7,375 (50.5) |
| EQ-5D-5L [IQR] | 1.0 (0.8-1.0) | 0.9 (0.8-1.0) | 1.0 (0.8-1.0) | 1.0 (0.8-1.0) | 0.9 (0.5-1.0) | 1.0 (0.8-1.0) | 1.0 (0.8-1.0) |
| Mobility | 2,427 (12.3) | 549 (22.9) | 1,878 (10.8) | 442 (17.3) | 48 (40.0) | 278 (11.1) | 1,659 (11.4) |
| Self-care | 1,327 (6.7) | 469 (19.6) | 858 (4.9) | 320 (12.5) | 45 (37.5) | 203 (8.1) | 759 (5.2) |
| Usual activity | 2,035 (10.3) | 529 (22.1) | 1,506 (8.7) | 410 (16.0) | 46 (38.3) | 288 (11.5) | 1,291 (8.9) |
| Pain/Discomfort | 7,281 (36.8) | 928 (38.7) | 6,353 (36.5) | 973 (38.0) | 61 (50.8) | 818 (32.6) | 5,429 (37.2) |
| Anxiety/Depression | 5,408 (27.3) | 890 (37.1) | 4,518 (26.0) | 871 (34.0) | 58 (48.3) | 812 (32.4) | 3,667 (25.1) |
NOTE:
All categorical data are presented as a number (percentage, %) except SSS-8 and EQ-5D-5L, which are presented as the median (interquartile range, IQR).
Abbreviations: QOL, quality of life; COVID-19, coronavirus disease 2019; SSS-8, Somatic Symptom Scale-8; IQR, interquartile range; EQ-5D-5L, EuroQol 5-dimensions 5-levels.
Table 4 shows the results of a multivariate linear regression analysis of factors associated with the EQ-5D-5L and SSS-8 scores. A previous history of COVD-19 was independently associated with a lower EQ-5D-5L score (β = -0.0230; 95% CI: -0.0295, -0.0164; P<0.001) and a higher SSS-8 score (β = +0.54; 95% CI: 0.28, 0.79; P<0.001). Sensitivity analysis by including suspected cases, did not change the findings in Table 4 substantially (Supplementary table 3).
Table 4.
Multivariate linear regression for EQ-5D-5L and SSS-8 (N = 19,784)
| Variables | EQ-5D-5L |
SSS-8 |
||||||
|---|---|---|---|---|---|---|---|---|
| β | P-value | 95% | CI | β | P-value | 95% | CI | |
| Age | 0.0010 | < 0.001 | 0.0008 | 0.0011 | -0.07 | < 0.001 | -0.07 | -0.06 |
| Sex | ||||||||
| Female sex (ref) | ||||||||
| Male sex | 0.0157 | < 0.001 | 0.0109 | 0.0205 | -1.64 | < 0.001 | -1.82 | -1.45 |
| Education | ||||||||
| High school or lower (ref) | ||||||||
| College or higher | 0.0070 | 0.001 | 0.0028 | 0.0111 | -0.25 | 0.002 | -0.41 | -0.09 |
| Income | ||||||||
| < 4 million (ref) | ||||||||
| ≧ 4 million | 0.0218 | < 0.001 | 0.0168 | 0.0268 | -0.70 | < 0.001 | -0.89 | -0.50 |
| N/A | 0.0108 | < 0.001 | 0.0052 | 0.0164 | -0.53 | < 0.001 | -0.75 | -0.31 |
| Partner | 0.0177 | < 0.001 | 0.0129 | 0.0224 | -0.03 | 0.783 | -0.21 | 0.16 |
| Employment | ||||||||
| Executive, employer, or self-employed (ref) | ||||||||
| Employee | 0.0144 | < 0.001 | 0.0076 | 0.0213 | -0.06 | 0.676 | -0.32 | 0.21 |
| Homewife or house-husband | -0.0028 | 0.508 | -0.0113 | 0.0056 | -0.53 | 0.002 | -0.85 | -0.20 |
| Unemployed, retired, or student | -0.0086 | 0.025 | -0.0161 | -0.0011 | -0.77 | < 0.001 | -1.06 | -0.48 |
| Smoking | -0.0279 | < 0.001 | -0.0325 | -0.0232 | 1.00 | < 0.001 | 0.82 | 1.18 |
| Alcohol | 0.0075 | < 0.001 | 0.0033 | 0.0117 | -0.11 | 0.179 | -0.27 | 0.05 |
| Comorbidities | ||||||||
| Hypertension | -0.0231 | < 0.001 | -0.0283 | -0.0179 | 1.38 | < 0.001 | 1.18 | 1.58 |
| Diabetes mellitus | -0.0237 | < 0.001 | -0.0313 | -0.0160 | 0.62 | < 0.001 | 0.32 | 0.91 |
| Dyslipidemia | -0.0215 | < 0.001 | -0.0270 | -0.160 | 1.21 | < 0.001 | 1.00 | 1.42 |
| Asthma | -0.0320 | < 0.001 | -0.0381 | -0.0258 | 1.78 | < 0.001 | 1.54 | 2.02 |
| Cardiovascular disease | -0.0340 | < 0.001 | -0.0446 | -0.0234 | 1.43 | < 0.001 | 1.02 | 1.84 |
| Cerebral vascular disease | -0.0535 | < 0.001 | -0.0659 | -0.0411 | 0.07 | 0.777 | -0.41 | 0.55 |
| COPD | -0.0264 | 0.001 | -0.0427 | -0.0101 | 0.62 | 0.054 | -0.01 | 1.25 |
| CKD | -0.0398 | < 0.001 | -0.0522 | -0.0273 | 1.31 | < 0.001 | 0.83 | 1.79 |
| Chronic liver disease | -0.0176 | 0.010 | -0.0310 | -0.0042 | 0.93 | < 0.001 | 0.41 | 1.45 |
| Immunocompromised status | -0.0575 | < 0.001 | -0.0687 | -0.0462 | 2.05 | < 0.001 | 1.62 | 2.49 |
| Malignancy | -0.0142 | < 0.001 | -0.0221 | -0.0064 | 0.34 | 0.029 | 0.03 | 0.64 |
| Number of COVID-19 vaccines | ||||||||
| 0 (ref) | ||||||||
| 1 | -0.0208 | 0.120 | -0.0470 | 0.0054 | -0.60 | 0.251 | -1.61 | 0.42 |
| 2 | 0.0236 | < 0.001 | 0.0158 | 0.0315 | -0.20 | 0.196 | -0.51 | 0.10 |
| 3 or 4 | 0.0286 | < 0.001 | 0.0224 | 0.0347 | -0.25 | 0.038 | -0.49 | -0.01 |
| Remote history of COVID-19 | -0.0230 | < 0.001 | -0.0295 | -0.0164 | 0.54 | < 0.001 | 0.28 | 0.79 |
Abbreviations: EQ-5D-5L, EuroQol 5-dimensions 5-levels; SSS-8, Somatic Symptom Scale-8; 95%CI, 95% confidence interval; N/A, data not available; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; COVID-19, coronavirus disease 2019.
Discussion
The present study showed the prevalence of symptoms attributed to prolonged symptoms after COVID-19 using a nationwide, relatively large sample of approximately 20,000 respondents in Japan. After adjusting for pre-existing comorbidities and confounders, not all alleged prolonged symptoms were associated with a previous history of COVID-19. Back pain, headache, chest discomfort, dysgeusia, dysosmia, and hair loss were independently associated with prolonged symptoms after COVID-19, suggesting a post-COVID condition after adjusting potential confounders, including demographic characteristics, social factors, underlying illnesses, or other risk factors previously noted in the literature. Moreover, individuals with a previous history of COVID-19 experienced a lower quality of life and greater severity of somatic symptoms than those without COVID-19. Respondents who received the COVID-19 booster vaccine were associated with higher EQ-5D-5L and lower SSS-8 scores, suggesting the COVID-19 booster vaccination would mitigate some prolonged symptoms of after COVID-19.
Although there are a number of studies on long COVID, only a few of these have investigated the strength of the association between each alleged clinical symptom and long COVID.15 Among symptoms with a statistically significant association, back pain and headache were more prevalent than the other symptoms (22.4% and 15.4%, respectively). The prevalence of these physical symptoms in long COVID varied substantially in previous studies; fatigue, chest discomfort, dyspnea and cough might be more common symptoms associated with long COVID.6 , 16 , 17 It remains unclear why the prevalence of these symptoms differs across studies. Reporting bias in the respondents may play a role in the difference in the prevalence of these symptoms. Moreover, despite the WHO's standard definition of long COVID, a significant variation in the diagnostic process would exist because of a wide range of symptoms linked to long COVD,3 the lack of a clear definition of the symptoms and the severity of symptoms description,6 , 18 and inadequacy of understanding the whole clinical picture of long COVID.19 Access to individual-level data (e.g., medical record))regarding symptoms suggestive of long COVID might have influenced the prevalence of each symptom.6
Although the prevalence of dysgeusia and dysosmia is not as high as that of other physical symptoms, the adjusted odds ratio for these symptoms was the highest among those of all the other symptoms. These clinical symptoms, representing neurological symptoms, are key clinical features of post-COVID conditions.3 , 17 Likewise, the adjusted odds ratio for symptoms such as poor concentration, sleep disturbance, dyspnea, and fatigue did not attain statistical significance, but these symptoms may nonetheless be associated with post-COVID conditions. As seen in Table 2 and supplementary table 2, the prevalence of most symptoms in respondents with a previous history of COVID-19 tended to be higher than in those without any history of COVID-19.
It should be noted, however, that these symptoms are subjective and differences in the threshold for reporting them might have influenced the results. A previous study suggested that 62 symptoms might be associated with long COVID,8 and some of these claims may be implausible if judged on strictly pathophysiological terms.20 The pre-pandemic status of general physical and mental health may also influence the prevalence of long COVID.14 Moreover, symptoms associated with long COVID are potentially overestimated because a wide range of comorbidities can also lead to similar symptoms.8 , 21 In the present study, the median age and the prevalence of various comorbidities were greater in respondents without COVID-19. Because older age and various comorbidities were considered risk factors of long COVID, these differences may have impacted the subsequent analyses. As the COVID-19 vaccination acceptance rate was higher in elderly individuals, the latter were less likely to develop COVID-19, leading to the older median age and high prevalence of each comorbidity in respondents without COVID-19. The major advantage of the present study is that we managed to adjust for various comorbidities and age to avoid the confounding of underlying illnesses and respondents’ age on the symptom of post-COVID conditions. While post COVID condition is a multisystemic illness causing a wide variety of symptoms,3 understanding which clinical symptoms are more likely associated with it is essential for assessing the disease burden and diagnosing the condition clinically, especially amid the changes occurring in the transition from the pandemic era to the endemic era of COVID-19.21
A previous history of COVID-19 also affected both overall and somatic health-related quality of life scores.22 Although the effect size of the previous history of COVID-19 on the health-related quality of life scores and somatic symptoms score appears to be modest, sensitivity analysis, including suspected cases, did not change the results substantially, which ensures the effectiveness of these scores for determining the severity of overall and somatic health-related quality of life associated with long COVID.
A few studies demonstrated that patients with long COVID had a low EQ-5D-5L score and validated the usefulness of the EQ-5D-5L score to assess health-related QOL for patients with long COVID.23 , 24 A study suggested that changes in the EQ-5D-5L score in a certain time interval (e.g., 6 months) enable us to assess their QOL in a long-term follow up precisely.25 In contrast, very few studies using somatic score to understand the severity of somatic symptoms due to long COVID were published. Somatic symptoms were assessed by each symptom rather than overall symptom severity such as SSS-8 score. Since each domain of somatic clinical symptoms in SSS-8 was also common symptoms suggesting long COVID, the finding in the present study suggests the usefulness of SSS-8 for assessing the severity of somatic symptoms in long COVID.
The impact of COVID-19 vaccination on the reduction of long COVID remains controversial.3 The present study demonstrated that COVID-19 booster vaccination was associated with improvements in the EQ-5D-5L and SSS-8 scores and showed a dose-response relationship between the number of vaccination and the improvement of these scores. In line with previous studies, the present study may indicate that COVID-19 vaccination provided partial protection against prolonged symptoms after COVID-19, reducing its prevalence and considerably ameliorating the severity of its symptoms when they occurred.15 , 26 Given the efficacy of the COVID-19 booster vaccinations, promoting booster vaccination may be a reasonable strategy to prevent the post COVID condition.. This is especially important as symptoms suggestive of long COVID may persist for up to one year.27
The present study has some limitations. First, selection bias may occur because of the study design to invite respondents who were willing to participate in the online survey. Second, the study method was based on a self-reported online survey, and there is a potential reporting bias, which may over- or under-estimate the true prevalence of symptoms. This may be further confounded by the subjective nature of reporting symptoms as no specific criteria for each symptom were provided, and the respondents' misclassification might occur. Third, some respondents did not appropriately respond to the online survey, which reduced the sample size because we excluded 3,370 respondents with inconsistent responses. Fourth, the present study was primarily dependent on the respondents’ self-reporting. Moreover, because the subjects were asked generally whether the symptoms had been present for at least two months, the exact duration of these prolonged symptoms was unable to be ascertained. Nevertheless, a nationwide, large sample population and the exclusion of respondents with inconsistent responses further ensure the finding of the present study. Fifth, this type of study is largely dependent on the size of sample population. Fifth, this type of study is largely dependent on the size of the sample population. Whereas a large number of respondents increases the robustness of the study, clinical variables of negligible significance may become statistically significant owing to the increased statistical power. Conversely, if the respondent pool is too small, important differences that are statistically significant may be overlooked. Lastly, we investigated the relationship between a previous history of COVID-19 and nineteen common symptoms indicating long COVID, but we may have missed other less common, yet relevant clinical symptoms potentially associated with long COVID. Despite these limitations, the present study successfully elucidated the prevalence of common prolonged symptoms after COVID-19 and the strength of association between each symptom and post-COVID conditions. This study also validated the usefulness of overall and somatic health-related scores in the post COVID condition, using a large sample of the population.
The COVID-19 pandemic is now shifting towards endemic, given accumulated immunity against SARS-CoV2 and lower virulence of SARS-CoV2. While the widespread of SARS-CoV2 in the community occurs, exploring the clinical characteristics suggesting post-COVID conditions is vital for effectively diagnosing post-COVID conditions. The finding in the study would help establish clinical criteria and estimate the clinical burden of post-COVID conditions.
Contributors
AT, TM, HH, TT, and YT designed the study. TT obtained the data. TM, AT, and HH performed the data analysis. AT, TM, HH, TT, and YT interpreted the data. HH drafted the manuscript. KT and KS secured the funding and performed the critical review. HH revised the manuscript. All the authors contributed to the final version of manuscript.
Data sharing statement
The datasets generated and/or analyzed during the current study are not publicly available due to ethical considerations but are available from the corresponding author on reasonable request.
Ethics statement
Ethical approval was granted by the Ethics Committee of Osaka International Cancer Institute under authorization number 20084. Respondents provided their consent to participate before proceeding to the questionnaire response page.
Declaration of Competing Interests
The authors declare no conflicts of interest in this study.
Acknowledgements
None
References
- 1.Carfi A, Bernabei R, Landi F. Gemelli Against C-P-ACSG. Persistent Symptoms in Patients After Acute COVID-19. JAMA. 2020;324(6):603–605. doi: 10.1001/jama.2020.12603. Aug 11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.World Health Organization. Post COVID-19 condition (Long COVID). https://www.who.int/europe/news-room/fact-sheets/item/post-covid-19-condition Last accessed 02052023
- 3.Davis HE, McCorkell L, Vogel JM, Topol EJ. Long COVID: major findings, mechanisms and recommendations. Nat Rev Microbiol. Jan 13 2023:1–14. doi: 10.1038/s41579-022-00846-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Chen C, Haupert SR, Zimmermann L, Shi X, Fritsche LG, Mukherjee B. Global Prevalence of Post-Coronavirus Disease 2019 (COVID-19) Condition or Long COVID: A Meta-Analysis and Systematic Review. J Infect Dis. Nov 1 2022;226(9):1593–1607. doi: 10.1093/infdis/jiac136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Perlis RH, Santillana M, Ognyanova K, et al. Prevalence and Correlates of Long COVID Symptoms Among US Adults. JAMA Netw Open. Oct 3 2022;5(10) doi: 10.1001/jamanetworkopen.2022.38804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Global Burden of Disease Long CC. Wulf Hanson S, Abbafati C, et al. Estimated Global Proportions of Individuals With Persistent Fatigue, Cognitive, and Respiratory Symptom Clusters Following Symptomatic COVID-19 in 2020 and 2021. JAMA. Oct 25 2022;328(16):1604–1615. doi: 10.1001/jama.2022.18931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Alwan NA, Johnson L. Defining long COVID: Going back to the start. Med (N Y) May 14 2021;2(5):501–504. doi: 10.1016/j.medj.2021.03.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Subramanian A, Nirantharakumar K, Hughes S, et al. Symptoms and risk factors for long COVID in non-hospitalized adults. Nat Med. Aug 2022;28(8):1706–1714. doi: 10.1038/s41591-022-01909-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.JACSIS study. JACSIS study. https://jacsis-study.jp/about/Last accessed 02052023
- 10.Rakuten Insight. Rakuten Insight internet search https://insight.rakuten.co.jp/en/aboutus.html. Last accessed 02052023
- 11.Lubetkin EI, Long D, Haagsma JA, Janssen MF, Bonsel GJ. Health inequities as measured by the EQ-5D-5L during COVID-19: Results from New York in healthy and diseased persons. PLoS One. 2022;17(7) doi: 10.1371/journal.pone.0272252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gierk B, Kohlmann S, Kroenke K, et al. The somatic symptom scale-8 (SSS-8): a brief measure of somatic symptom burden. JAMA Intern Med. Mar 2014;174(3):399–407. doi: 10.1001/jamainternmed.2013.12179. [DOI] [PubMed] [Google Scholar]
- 13.Tenforde MW, Kim SS, Lindsell CJ, et al. Symptom Duration and Risk Factors for Delayed Return to Usual Health Among Outpatients with COVID-19 in a Multistate Health Care Systems Network - United States, March-June 2020. MMWR Morb Mortal Wkly Rep. 2020;69(30):993–998. doi: 10.15585/mmwr.mm6930e1. Jul 31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Thompson EJ, Williams DM, Walker AJ, et al. Long COVID burden and risk factors in 10 UK longitudinal studies and electronic health records. Nat Commun. Jun 28 2022;13(1):3528. doi: 10.1038/s41467-022-30836-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Al-Aly Z, Bowe B, Xie Y. Long COVID after breakthrough SARS-CoV-2 infection. Nat Med. Jul 2022;28(7):1461–1467. doi: 10.1038/s41591-022-01840-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Crook H, Raza S, Nowell J, Young M, Edison P. Long covid-mechanisms, risk factors, and management. BMJ. Jul 26 2021;374:n1648. doi: 10.1136/bmj.n1648. [DOI] [PubMed] [Google Scholar]
- 17.Nehme M, Braillard O, Alcoba G, et al. COVID-19 Symptoms: Longitudinal Evolution and Persistence in Outpatient Settings. Ann Intern Med. May 2021;174(5):723–725. doi: 10.7326/M20-5926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.H L. How common is long COVID? Why studies give different answers. 2023 doi: 10.1038/d41586-022-01702-2. https://www.nature.com/articles/d41586-022-01702-2 Last accessed. [DOI] [PubMed] [Google Scholar]
- 19.Prevention CfDCa. Post-COVID Conditions: Information for Healthcare Providers. https://www.cdc.gov/coronavirus/2019-ncov/hcp/clinical-care/post-covid-conditions.html Last accessed 02052023.
- 20.Saunders C, Sperling S, Bendstrup E. A new paradigm is needed to explain long COVID. Lancet Respir Med. Jan 5 2023 doi: 10.1016/S2213-2600(22)00501-X. [DOI] [PubMed] [Google Scholar]
- 21.Wu Q, Ailshire JA, Crimmins EM. Long COVID and symptom trajectory in a representative sample of Americans in the first year of the pandemic. Sci Rep. Jul 8 2022;12(1):11647. doi: 10.1038/s41598-022-15727-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Wong AW, Shah AS, Johnston JC, Carlsten C, Ryerson CJ. Patient-reported outcome measures after COVID-19: a prospective cohort study. Eur Respir J. Nov 2020;56(5) doi: 10.1183/13993003.03276-2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Soh HS, Cho B. Long COVID-19 and Health-Related Quality of Life of Mild Cases in Korea: 3-Months Follow-up of a Single Community Treatment Center. J Korean Med Sci. Nov 28 2022;37(46):e326. doi: 10.3346/jkms.2022.37.e326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Walle-Hansen MM, Ranhoff AH, Mellingsaeter 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. Mar 22 2021;21(1):199. doi: 10.1186/s12877-021-02140-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Fernandez-de-Las-Penas C, Rodriguez-Jimenez J, Moro-Lopez-Menchero P, et al. Psychometric properties of the Spanish version of the EuroQol-5D-5L in previously hospitalized COVID-19 survivors with long COVID. Sci Rep. Jul 23 2022;12(1):12605. doi: 10.1038/s41598-022-17033-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ayoubkhani D, Bosworth ML, King S, et al. Risk of Long COVID in People Infected With Severe Acute Respiratory Syndrome Coronavirus 2 After 2 Doses of a Coronavirus Disease 2019 Vaccine: Community-Based, Matched Cohort Study. Open Forum Infect Dis. Sep 2022;9(9):ofac464. doi: 10.1093/ofid/ofac464. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Mizrahi B, Sudry T, Flaks-Manov N, et al. Long covid outcomes at one year after mild SARS-CoV-2 infection: nationwide cohort study. BMJ. Jan 11 2023;380 doi: 10.1136/bmj-2022-072529. [DOI] [PMC free article] [PubMed] [Google Scholar]