Key PointsQuestion
Which individuals are at risk of developing post−COVID-19 condition (PCC)?
Findings
This systematic review and meta-analysis of 41 studies including 860 783 patients found that female sex, older age, higher body mass index, smoking, preexisting comorbidities, and previous hospitalization or ICU admission were risk factors significantly associated with developing PCC, and that SARS-CoV-2 vaccination with 2 doses was associated with lower risk of PCC.
Meanings
The findings of this systematic review and meta-analysis provide a profile of the characteristics associated with increased risk of developing PCC and suggest that vaccination may be protective against PCC.
Importance
Post−COVID-19 condition (PCC) is a complex heterogeneous disorder that has affected the lives of millions of people globally. Identification of potential risk factors to better understand who is at risk of developing PCC is important because it would allow for early and appropriate clinical support.
Objective
To evaluate the demographic characteristics and comorbidities that have been found to be associated with an increased risk of developing PCC.
Data sources
Medline and Embase databases were systematically searched from inception to December 5, 2022.
Study Selection
The meta-analysis included all published studies that investigated the risk factors and/or predictors of PCC in adult (≥18 years) patients.
Data Extraction and Synthesis
Odds ratios (ORs) for each risk factor were pooled from the selected studies. For each potential risk factor, the random-effects model was used to compare the risk of developing PCC between individuals with and without the risk factor. Data analyses were performed from December 5, 2022, to February 10, 2023.
Main Outcomes and Measures
The risk factors for PCC included patient age; sex; body mass index, calculated as weight in kilograms divided by height in meters squared; smoking status; comorbidities, including anxiety and/or depression, asthma, chronic kidney disease, chronic obstructive pulmonary disease, diabetes, immunosuppression, and ischemic heart disease; previous hospitalization or ICU (intensive care unit) admission with COVID-19; and previous vaccination against COVID-19.
Results
The initial search yielded 5334 records of which 255 articles underwent full-text evaluation, which identified 41 articles and a total of 860 783 patients that were included. The findings of the meta-analysis showed that female sex (OR, 1.56; 95% CI, 1.41-1.73), age (OR, 1.21; 95% CI, 1.11-1.33), high BMI (OR, 1.15; 95% CI, 1.08-1.23), and smoking (OR, 1.10; 95% CI, 1.07-1.13) were associated with an increased risk of developing PCC. In addition, the presence of comorbidities and previous hospitalization or ICU admission were found to be associated with high risk of PCC (OR, 2.48; 95% CI, 1.97-3.13 and OR, 2.37; 95% CI, 2.18-2.56, respectively). Patients who had been vaccinated against COVID-19 with 2 doses had a significantly lower risk of developing PCC compared with patients who were not vaccinated (OR, 0.57; 95% CI, 0.43-0.76).
Conclusions and Relevance
This systematic review and meta-analysis demonstrated that certain demographic characteristics (eg, age and sex), comorbidities, and severe COVID-19 were associated with an increased risk of PCC, whereas vaccination had a protective role against developing PCC sequelae. These findings may enable a better understanding of who may develop PCC and provide additional evidence for the benefits of vaccination.
Trial Registration
PROSPERO Identifier: CRD42022381002
Since the first SARS-CoV-2 infections were identified in December 2019, the COVID-19 pandemic has significantly increased morbidity and mortality around the world.1 Previous epidemics of viruses from the coronavirus family, such as SARS-CoV and the Middle East Respiratory Syndrome coronavirus (MERS-CoV), have developed into persistent symptoms in infected individuals, including severe fatigue, decreased quality of life (QOL), and shortness of breath, as well as behavioral and psychological problems.1 These persistent postviral symptoms have been associated with a substantial burden to health care systems.
Similarly, a constellation of various clinical symptoms has been described in a subset of patients who have survived the acute phase of SARS-CoV-2−induced COVID-19.1 This constellation of symptoms has received many labels, including post-acute COVID-19 syndrome, persistent post-COVID-19 syndrome, and Long COVID-19. These terms have been used interchangeably for several years. The UK National Institute for Health and Care Excellence proposed Long COVID to describe the presence of symptoms that persist for 4 or more weeks after acute COVID-19 infection.2 The World Health Organization (WHO) defined post−COVID-19 condition (PCC) as having symptoms usually 3 months from the onset of COVID-19 with a duration of at least 2 months.3 Typical clinical symptoms include dyspnea, fatigue, autonomic dysfunction, headache, and persistent loss of smell and/or taste—although a wide range of symptoms has been described.1,4 Given that individuals with PCC may need long-term clinical support,4 the economic consequences have been estimated to be substantial.5
Not only is it important to recognize which individuals may be at high risk of developing PCC and to offer follow-up care; it is imperative to plan population-level public health measures. Several studies have been published investigating clinical and epidemiologic risk factors and/or predictors of PCC.5-8 However, these studies often had relatively few patients. Furthermore, wide discrepancy exists among published data, yielding uncertainty on the clinical utility of their findings. Therefore, the aim of this study was to search the available literature for published studies that found clinical and epidemiologic risk factors associated with the development of PCC and to pool their results.
This study was exempt from ethics review because it used only previously published data; informed consent was also waived for this reason. The study followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline.9
Search Strategy and Selection Criteria
MEDLINE and Embase databases were systematically searched for studies investigating the risk factors or clinical predictors for PCC in patients diagnosed with COVID-19, from inception to December 5, 2022. Search terms included “long-COVID-19,” “post-COVID-19,” and “chronic COVID-19,” as well as the corresponding MeSH (Medical Subjects Heading) terms. Only peer-reviewed articles were included; preprints were excluded. The full search strategy is available in the eMethods in Supplement 1.
Search results were imported for abstract screening; duplicates and irrelevant studies were removed based on predetermined inclusion and exclusion criteria. All studies that investigated the risk factors or predictors of PCC, as defined by the WHO definition (≥1 symptom for ≥3 months), in a cohort of adult (≥18 years) patients were included. The risk factors evaluated for this meta-analysis were: age; biological sex; body mass index (BMI), calculated as weight in kilograms divided by height in meters squared; smoking status; comorbidities including anxiety and/or depression, asthma, chronic kidney disease (CKD), chronic obstructive pulmonary disease (COPD), diabetes, immunosuppression, and ischemic heart disease (IHD); and COVID-19 vaccination status. Studies were excluded if they investigated persistent COVID-19 symptoms of less than 3 months duration; did not provide data for any of the risk factors listed; or used only univariate regression because we wanted to identify independent risk factor association.
Subsequently, full texts of studies were retrieved and scrutinized against the criteria. The relevant data from the included studies were independently extracted by 2 authors (H.E., V.T.) who were blinded to the authors and institutions involved. Any disagreements were resolved by discussion with the senior author (V.V.). Some cohorts were studied and/or published more than once, for example, neurological PCC and respiratory PCC were evaluated in the same cohort in some studies.10-12 To avoid double-counting patients of these cohorts, we initially meta-analyzed all the PCC symptoms and produced a single odds ratio (OR) for the specific cohort; this OR was then used in the meta-analysis. The Newcastle-Ottawa Scale,13 a 9-point measure assessing the quality of cohort studies and case-control studies or case series, was used to evaluate the observational studies included.
Quantitative synthesis of included studies was performed using RStudio 2022.07.1 + 554 and R, version 4.0.5 (The R Foundation for Statistical Computing). The ORs for each risk factor were pooled with the random-effects model. This was deemed more appropriate than the fixed-effects model because the studies included in this meta-analysis represented samples from different populations. For studies reporting rate ratios, those were converted to ORs using the methods defined in the Cochrane Handbook for Systematic Reviews of Interventions.14 Summary statistics were expressed as ORs and 95% CIs. Prediction intervals were also reported. Statistical heterogeneity was assessed using the I2statistic. Publication bias was assessed qualitatively by visual inspection of inverted funnel plot asymmetry and Egger test was performed to assess small study effects. Statistical tests were 2-tailed, and the statistical significance threshold was P < .05. Data analyses were performed from December 5, 2022, to February 10, 2023.
The search of MEDLINE and Embase databases yielded a total of 5334 records. After removal of duplicates, 3363 were screened at title and abstract level, and 255 studies underwent full-text evaluation. Of those, 41 records with a total of 860 783 patients met the inclusion criteria and were included in the meta-analysis. The PRISMA flowchart for study selection is available in eFigure 1 in Supplement 1. The Table summarizes the population cohorts and the study design characteristics of all the included studies.10-12,15-61 Of the 41 observational studies, 30 were ranked as high quality and 11 moderate quality on the Newcastle-Ottawa Scale (eTable 1 in Supplement 1).
All previously identified risk factors for PCC were evaluated, including patient age, sex, BMI, smoking status, comorbidities (ie, anxiety/depression, asthma, CKD, COPD, diabetes, immunosuppression, IHD), and hospitalization or ICU admission for COVID-19. In addition, the role of vaccination as a risk factor for PCC was evaluated. Funnel plots for all the meta-analyses are shown in eFigure 2 in Supplement 1. Race and ethnicity were not evaluated as this information was not provided consistently across the included studies.
Of the 41 studies, 38 studies including a total of 727 630 patients investigated sex as a risk factor for PCC. Overall, the pooled ORs showed that female sex was significantly associated with PCC (OR, 1.56; 95% CI, 1.41 to 1.73; I2 = 94%; Figure 1). However, the 95% prediction interval (95% PI, 0.94 to 2.61) suggested that this may not be demonstrated in all future studies. To investigate this further, we undertook subgroup analysis separating the studies that included only hospitalized patients from those that included only nonhospitalized and those that included a mixture of hospitalized and nonhospitalized patients (eFigure 3 in Supplement 1). This showed that heterogeneity was lower in studies that included only hospitalized or only nonhospitalized patients compared with those that included patients from both settings (58%, 24%, and 97%, respectively), with the correlation remaining significant and the prediction intervals showing evidence supporting this significance for future studies. Subgroup analysis was also performed by study quality (high vs moderate) per the Newcastle-Ottawa Scale, with no significant between-group differences demonstrated (eFigure 4 in Supplement 1). Meta-regression analysis by study size showed no significance (effect size = 0.0001; 95% CI, −0.0001 to 0.0001; P = .26). Egger test for small study effects was not significant (intercept = 0.36, 95% CI, 0.21 to 0.50; P = .15).
Of the 41 studies, 9 studies including a total of 324 950 patients investigated age as a risk factor for PCC. For the meta-analysis, the risk of PCC among 3 age groups (40-69 years and ≥70 years vs 18-40 years) was analyzed. We found that patients in both of the older groups had a significantly higher risk of PCC when compared with adult patients younger than 40 years, with no significant between-group differences (OR, 1.21; 95% CI, 1.11 to 1.33; I2 = 95%; Figure 2). However, this may not be demonstrated in all future studies (95% PI, 0.84-1.76). Subgroup analysis by study size demonstrated a high rate of heterogeneity in the group of large studies (eFigure 5 in Supplement 1). Meta-regression analysis by study size was significant (effect size = −0.0001; 95% CI, −0.0002 to −0.0001; P = .02) indicating that study size may have influenced the results (eFigure 6 in Supplement 1). Egger test for small study effects was not significant (intercept = 0.20; 95% CI, 0.06 to 0.50; P = .34). Subgroup analysis by study population (not hospitalized patients vs combined settings) showed no significant between-group differences (eFigure 7 in Supplement 1), whereas sensitivity analysis by study quality demonstrated that high quality studies have higher heterogeneity (eFigure 8 in Supplement 1).
Of 41 studies, 16 studies including a total of 701 807 patients investigated obesity (high BMI, defined as ≥30) as a risk factor for PCC. Obesity was found to be significantly associated with PCC (OR, 1.15; 95% CI, 1.08 to 1.23; I2 = 91%; eFigure 9 in Supplement 1). However, this significant correlation may not be shown in all future studies (95% PI, 0.94 to 1.42). Subgroup analysis by study population (hospitalized vs nonhospitalized vs combined) showed that the correlation remained significant in all 3 subgroups; however, the studies of nonhospitalized patients had the lowest heterogeneity (eFigure 10 in Supplement 1). Subgroup analysis by study quality showed that the significant correlation between obesity and PCC was evident only in high quality studies (eFigure 11 in Supplement 1). Egger test was found to be significant (intercept = .06; 95% CI, −0.02 to 0.15; P < .001), suggesting publication bias as shown in the funnel plot (eFigure 2C in Supplement 1). Meta-regression analysis by study size was not significant (effect size = −0.0001; 95% CI −0.00001 to 0.0001; P = .86).
Of the 41 studies, 20 studies including a total of 455 204 patients investigated whether current smokers had higher risk of developing PCC compared with nonsmokers. Overall, the pooled ORs showed that smoking was significantly associated with PCC (OR, 1.10; 95% CI, 1.07 to 1.13; I2 = 0%; Figure 3). Subgroup analysis by study quality showed no significant differences (eFigure 12 in Supplement 1). Egger test suggested no significant publication bias (intercept = .08; 95% CI, 0.05 to 0.11; P = .07), and meta-regression analysis by study size also showed no significance (effect size = −0.0001; 95% CI, −0.0001 to 0.001; P = .14).
Meta-analysis was performed for 34 studies that investigated the presence of comorbidities potentially associated with the risk of PCC syndrome. Specifics for each comorbidity follow.
Anxiety and/or Depression
Four studies including 634 734 patients investigated the risk of PCC in patients with anxiety and/or depression. Pooled analysis of these studies showed a significant association with PCC (OR, 1.19; 95% CI, 1.02 to 1.40; I2 = 96%; eFigure 24 in Supplement 1). Egger test for small study effects was not significant (intercept = 0.22; 95% CI, 0.03 to 0.41; P = .47). Meta-regression analysis for study size was not performed owing to the small number of studies in each group. Subgroup analysis by study quality is shown in eFigure 25 in Supplement 1.
Meta-analysis of 13 studies including 639 397 patients showed that patients with asthma had significantly higher risk of developing PCC (OR, 1.24; 95% CI, 1.15 to 1.35; I2 = 53%; eFigure 13 in Supplement 1). All of these studies were of high quality; therefore, subgroup analysis for this factor was not conducted. Meta-regression analysis for study size showed significance (effect size = −0.0001; 95% CI, −0.0003 to −0.0001; P < .001), which was confirmed by subgroup analysis of studies by their sample size (eFigure 14 in Supplement 1). In this analysis, larger studies demonstrated a significant association between asthma and PCC, whereas smaller studies (<1000 patients) failed to reach significance. Egger test showed no significant publication bias (intercept = 0.23; 95% CI, 0.13 to 0.34; P = .51).
A pooled analysis of 8 studies with a total of 255 791 patients showed that CKD was not a significant risk factor for PCC (OR, 1.12; 95% CI, 0.98 to 1.28; I2 = 22%; eFigure 19 in Supplement 1). Subgroup analysis by study quality is shown in eFigure 20 in Supplement 1. Meta-regression analysis for study size showed no significance (effect size = 0.10; 95% CI, −0.05 to 0.25; P = .20), and Egger test showed no publication bias (intercept = 0.04; 95% CI, −0.21 to 0.29; P = .56).
Chronic Obstructive Pulmonary Disease
Analysis of 10 studies including 257 340 patients showed that COPD was a risk factor associated with persistent symptoms after COVID-19 infection (OR, 1.38; 95% CI, 1.08 to 1.78; I2 = 77%; eFigure 15 in Supplement 1). Nevertheless, this significance may not be shown in all future studies (95% PI, 0.70 to 2.74). Subgroup analyses by study quality is shown in eFigure 16 in Supplement 1. Meta-regression analysis for study size and Egger test were both nonsignificant (effect size = −0.0002; 95% CI, −0.0003 to 0.0001; P = .66; and intercept = 0.23; 95% CI, 0.14 to 0.33; P = .69, respectively).
Meta-analysis of 18 studies including 259 978 patients showed that patients with diabetes (OR, 1.06; 95% CI, 1.03 to 1.09; I2 = 0%) had a significant risk of PCC (eFigure 17 in Supplement 1). Subgroup analysis by study quality is shown in eFigure 18 in Supplement 1. Meta-regression analysis showed that study size did not have a significant effect (effect size = 0.001; 95% CI, −0.0002 to 0.0002; P = .15), and Egger test showed no publication bias (intercept = −0.008; 95% CI, −0.14 to 0.12; P = .34).
Three studies with a total of 967 patients evaluated whether patients with immunosuppression exhibited higher risk of PCC. Meta-analysis of these studies showed a significant association of immunosuppression with PCC (OR, 1.50; 95% CI, 1.05-2.15; I2 = 0%; eFigure 23 in Supplement 1). Egger test did not show significant publication bias. Owing to the small number of studies, subgroup analysis and meta-regression were not performed for these studies.
Five studies including 201 906 patients investigated the association of preexisting IHD. Meta-analysis of these studies showed that patients with IHD had 1.28 times higher risk of developing PCC (OR, 1.28; 95% CI, 1.19 to 1.38; I2 = 0%; eFigure 21 in Supplement 1). Subgroup analysis by study quality is shown in eFigure 22 in Supplement 1. Meta-regression analysis for study size and Egger test for small study effects did not show significance (effect size = −0.0001; 95% CI, −0.0003 to 0.001; P = .49, and intercept = 0.23; 95% CI, 0.14 to 0.33; P = .69, respectively).
Hospitalization and ICU Admission
Meta-analysis of 8 studies with a total of 265 466 patients previously hospitalized for COVID-19 infection was performed. The findings showed that patients who required hospitalization during the acute phase of COVID-19 had significantly higher risk of developing PCC (OR, 2.48; 95% CI, 1.97 to 3.13; I2 = 86%; eFigure 26 in Supplement 1). Subgroup analysis by study quality is shown in eFigure 27 in Supplement 1. Meta-regression analysis for study size and Egger test did not demonstrate statistical significance (effect size = 0.002; 95% CI, −0.001 to 0.0001; P = .73 and intercept = 0.96; 95% CI, 0.47 to 1.43; P = .78, respectively).
Similarly, a meta-analysis of 10 studies with a total of 213 441 patients showed that patients who required ICU admission during the acute phase were at higher risk for PCC (OR, 2.37; 95% CI, 2.18 to 2.56; I2 = 0%; eFigure 28 in Supplement 1). Subgroup analysis by study quality is shown in eFigure 29 in Supplement 1. Meta-regression analysis for study size showed that there was an effect (effect size = 0.0001; 95% CI, 0.0001 to 0.002; P = .01); however, subgroup analysis of studies by sample size did not demonstrate significant between-group differences (eFigure 30 in Supplement 1). Egger test showed no significance (intercept = 0.92; 95% CI, 0.82 to 1.02; P = .06).
Four studies with a total of 249 788 patients evaluated the effect of vaccination status on the risk of developing PCC. Meta-analysis of these showed that individuals who had been vaccinated with 2 doses (in all included studies) had a 40% lower risk of developing PCC (OR, 0.57; 95% CI, 0.43 to 0.76; I2 = 91%; Figure 4). This may not be demonstrated in all future studies (95% PI, 0.15 to 2.22). Subgroup analysis by study quality and meta-regression for study size were not performed because all these studies were of high quality and included more than 1000 patients each. Egger test showed no significant publication bias (intercept = −0.44; 95% CI, −1.38 to 0.48; P = .80).
One original publication48 of 10 longitudinal studies49-57 and another study35 included patients that were self- or clinician-diagnosed with COVID-19 during the acute phase. For this reason, in addition to the aforementioned analyses, we performed sensitivity analyses for all the risk factors excluding these studies (eFigure 31 in Supplement 1). Overall, there were no differences in the outcomes of any risk factor investigated. Additional sensitivity analyses were performed based on the studies that investigated 5 or more risk factors (eFigure 32 in Supplement 1). There were no changes noted in the outcomes of each risk factor. Meta-regression analyses by geographic location were also performed for the risk factors (eTable 2 in Supplement 1); however, given the limited geographic diversity (30 studies from Europe; only 1 from Africa, 6 from the Americas [Brazil, Canada, US], and 5 from Asia), the interpretation of results should be guarded.
This meta-analysis of 41 studies that included a total of 860 783 patients demonstrates that there were certain epidemiologic and clinical risk factors that are associated with a higher risk of developing PCC. In particular, female sex, older age, higher BMI, and smoking were significantly associated with increased risk of persistent symptoms of 3 months or more after the acute phase of COVID-19 infection, ie, PCC. In addition, preexisting comorbidities, including anxiety and/or depression, asthma, COPD, diabetes, IHD, and immunosuppression were also found to be significantly associated with higher risk of PCC. Furthermore, patients who needed hospitalization or ICU care during the acute phase of COVID-19 infection were found to have more than twice the risk of developing PCC compared with those who were not. On the other hand, vaccination (with 2 doses) for COVID-19 was noted to have a protective role against PCC—vaccinated patients had a significantly lower risk of developing the persistent symptoms of PCC.
The aforementioned findings confirm that PCC is a multifactorial and complex clinical syndrome.62 These results strengthen the evidence available regarding the association of female sex with PCC.8,63,64 A previous meta-analysis by Maglietta and colleagues65 including 13 340 patients also highlighted that female sex was significantly associated with the persistent COVID-19 symptoms. A recent large analysis and meta-regression of more than 2 million patients64 confirmed this finding. Many authors have hypothesized mechanistic processes to explain the association between certain risk factors, including female sex, and PCC.1,66-69 For example, it has been suggested that hormones may play a role in perpetuating the hyperinflammatory status of the acute phase of COVID-19 even after recovery.66,67 Also, stronger IgG antibodies production in female individuals in the acute phase has been reported68 and could contribute to perpetuating disease manifestations.68,69
As previous research has suggested,48,65 older age appears to be an independent risk factor for PCC. Subgroup analysis showed that individuals 40 to 69 years old and those 70 years or older are at equally high risk of PCC when compared with younger patients. However, it is important to consider that the prevalence of PCC consists of individuals who have survived the acute phase of COVID-19 infection. Older individuals, possibly with multiple underlying comorbidities, may not survive the acute phase of COVID-19 because they are at higher risk of severe illness.70 As highlighted by Di Toro and colleagues,71 PCC reflects the population of COVID-19 survivors, not the epidemiologic characteristics of COVID-19.
Additionally, the results of our meta-analysis revealed that obesity and smoking were significantly associated with higher risk of developing PCC. These findings concur with recent evidence identifying these characteristics as important risk factors for PCC.48,72,73 Obesity and PCC share a metabolic proinflammatory state that promotes inflammatory processes and their associated signs and symptoms to linger for a prolonged period of time.74 Smoking has been shown to be a significant risk factor for both PCC and severe acute COVID-19 infection.75,76 However, it is unclear whether smoking per se or the associated severe illness predisposes this cohort of patients to higher risk of PCC.
Our meta-analysis revealed that patients who were hospitalized or admitted to the ICU had more than double risk of developing PCC. Severe illness has been found to be a significant risk factor for PCC in previous studies. In a multicenter cohort study that included 246 patients, 74.3% had ongoing physical symptoms 1 year after ICU admission for COVID-19.77 However, it should be noted that ICU survivors may experience postintensive care syndrome after an episode of critical care illness.78,79 Postintensive care syndrome is well-recognized and entails a variety of symptoms that may persist for months or years; therefore, there may be an important overlap with PCC sequelae. Nevertheless, the results of our meta-analysis and those of other studies highlight that patients with previous critical illness represent a high-risk population and their follow-up should reflect intensive plans for prevention, rehabilitation, and treatment of the ongoing debilitating symptoms of PCC.
The results of our study showed that vaccination for COVID-19 has a protective role against PCC, with vaccinated individuals having a significantly lower risk compared with unvaccinated individuals. This finding concurs with those of other studies and the recent report from the UK Office of National Statistics that found a 42% lower risk of PCC after 2 doses of a COVID-19 vaccine.80-82 Importantly, emerging evidence suggests that vaccination reduces the risk of PCC and its sequelae even in individuals with other risk factors, such as older age or high BMI,81 expanding the benefits of vaccination beyond the morbidity and mortality benefits seen during the acute COVID-19 phase.
Individuals with PCC may experience long-lasting adverse effects requiring long-lasting support. It has been reported that 15% of individuals with PCC were absent from work owing to illness.5 Follow-up outpatient services may be needed to manage this condition and to better understand the possible association between symptoms and residual organ impairment. Given that health care systems worldwide have been substantially burdened by the COVID-19 pandemic,83 routine follow-up may not be possible to all those living with PCC.
This review had some limitations. Some of the meta-analyses performed had considerable statistical heterogeneity, which may have affected results. Large meta-epidemiologic studies have shown that studies at high risk of bias tend to overestimate the strength of associations. In addition, all the included studies were observational. Consequently, the results of the performed meta-analyses were based on observational data. Although the observational studies were of moderate or high quality per the Newcastle-Ottawa scale, the scale itself is not without limitations.84 Furthermore, by virtue of being observational, all the studies (even those with a high rating) have an unavoidable risk of bias. Despite this and considering that randomized studies (with the current COVID-19 strains) will not be undertaken, studies providing risk factors following multivariable regression allow us to draw important conclusions. Furthermore, as discussed previously, PCC is a clinically heterogenous condition with a range of manifestations and symptoms. In this analysis, we considered all the various manifestations as a single entity. For this meta-analysis, we relied on the diagnosis identified by the authors of the included studies, accepting that the definition of symptoms included among the different studies might not have been exactly the same. Lastly, the studies spanned across various COVID-19 variants, but were all pooled together independent of variant. It is possible that the various variants, including the effect of vaccination, could alter the absolute value of patients with PCC; however, it is unlikely that the risk factors associated with PCC would change.
The findings of this systematic review and meta-analysis demonstrated that certain demographic characteristics (eg, age and sex) and comorbidities were significantly associated with an increased risk of developing PCC, whereas vaccination had a protective role against developing PCC sequelae. Given these results, a holistic approach and integrated care pathways may enable suitable support for patients who develop PCC and may allow physicians to be better prepared to care for patients at high risk of developing PCC. Moreover, in addition to preventing and diminishing the acute phase of the infection, COVID-19 vaccination may protect against PCC, giving vaccination additional evidence of benefit.
Accepted for Publication: February 19, 2023.
Published Online: March 23, 2023. doi:10.1001/jamainternmed.2023.0750
Corresponding Author: Vassilios S. Vassiliou, MBBS, PhD, Bob Champion Research and Education, Norwich Medical School, University of East Anglia, Rosalind Franklin Road, Norwich NR4 7UQ, UK (v.vassiliou@uea.ac.uk).
Author Contributions: Drs Tsampasian and Elghazaly had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Elghazaly and Tsampasian contributed equally. Drs Ntatsaki and Vassiliou were co-senior authors.
Concept and design: Tsampasian, Ntatsaki, Vassiliou.
Acquisition, analysis, or interpretation of data: Tsampasian, Elghazaly, Chattopadhyay, Debski, Garg, Clark, Ntatsaki, Vassiliou.
Drafting of the manuscript: Tsampasian, Elghazaly.
Critical revision of the manuscript for important intellectual content: Tsampasian, Chattopadhyay, Naing, Debski, Garg, Clark, Ntatsaki, Vassiliou.
Statistical analysis: Tsampasian, Elghazaly, Clark, Vassiliou.
Obtained funding: Vassiliou.
Administrative, technical, or material support: Tsampasian, Debski, Ntatsaki, Vassiliou.
Supervision: Ntatsaki, Vassiliou.
Conflict of Interest Disclosures: Drs Chattopadhyay, Debski, and Tsampasian reported being academic clinical fellows funded by the UK National Institute of Health (NIHR) and Research. Dr Clark reported funding from Brainomix, National Institute of Health Research UK, Stroke Association, and Versus Arthritis outside the submitted work. Dr Ntatsaki reported partial funding from an NIHR Clinical Research Network East of England Greenshoot scheme, outside the submitted work. No other disclosures were reported.
Additional Contributions: We would like to acknowledge the assistance of Haipeng Liu, PhD (Coventry University) with the statistical analysis.
Data Sharing Statement: See Supplement 2.
5.Halpin
SJ, McIvor
C, Whyatt
G,
et al. Postdischarge symptoms and rehabilitation needs in survivors of COVID-19 infection: a cross-sectional evaluation.
J Med Virol. 2021;93(2):1013-1022. doi:
10.1002/jmv.26368
PubMedGoogle ScholarCrossref 11.Fernández-de-Las-Peñas
C, Cancela-Cilleruelo
I, Moro-López-Menchero
P,
et al. Exploring the trajectory curve of long-term musculoskeletal post-COVID pain symptoms in hospitalized COVID-19 survivors: a multicenter study.
Pain. 2023;164(2):413-420. doi:
10.1097/j.pain.0000000000002718
PubMedGoogle ScholarCrossref 12.Zisis
SN, Durieux
JC, Mouchati
C, Perez
JA, McComsey
GA. The protective effect of Coronavirus Disease 2019 (COVID-19) vaccination on postacute sequelae of COVID-19: a multicenter study from a large national health research network.
Open Forum Infect Dis. 2022;9(7):ofac228. doi:
10.1093/ofid/ofac228
PubMedGoogle ScholarCrossref 17.Augustine
R, S
A, Nayeem
A,
et al. Increased complications of COVID-19 in people with cardiovascular disease: Role of the renin-angiotensin-aldosterone system (RAAS) dysregulation.
Chem Biol Interact. 2022;351:109738. doi:
10.1016/j.cbi.2021.109738
PubMedGoogle ScholarCrossref 18.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. 2022;9(9):ofac464. doi:
10.1093/ofid/ofac464
PubMedGoogle ScholarCrossref 22.Chudzik
M, Babicki
M, Kapusta
J,
et al. Long-COVID clinical features and risk factors: a retrospective analysis of Patients from the STOP-COVID registry of the PoLoCOV Study.
Viruses. 2022;14(8):1755. doi:
10.3390/v14081755
PubMedGoogle ScholarCrossref 25.Dias
MB, Medeiros
APV, de Melo
SS,
et al; CO-FRAIL Study Group. The long and winding road of COVID-19 in survivors of hospitalisation: symptoms trajectory and predictors of long COVID.
J Intern Med. 2023;293(2):264-268. doi:
10.1111/joim.13583
PubMedGoogle ScholarCrossref 30.Fernández-de-Las-Peñas
C, Torres-Macho
J, Elvira-Martínez
CM, Molina-Trigueros
LJ, Sebastián-Viana
T, Hernández-Barrera
V. Obesity is associated with a greater number of long-term post-COVID symptoms and poor sleep quality: a multicentre case-control study.
Int J Clin Pract. 2021;75(12):e14917. doi:
10.1111/ijcp.14917
PubMedGoogle ScholarCrossref 32.Fernández-de-Las-Peñas
C, Martín-Guerrero
JD, Pellicer-Valero
ÓJ,
et al. Female sex is a risk factor associated with long-term post-COVID related-symptoms but not with COVID-19 symptoms: the LONG-COVID-EXP-CM multicenter study.
J Clin Med. 2022;11(2):413. doi:
10.3390/jcm11020413
PubMedGoogle ScholarCrossref 33.Fernández-de-Las-Peñas
C, Pellicer-Valero
ÓJ, Navarro-Pardo
E,
et al. Symptoms experienced at the acute phase of SARS-CoV-2 infection as risk factor of long-term post-COVID symptoms: the LONG-COVID-EXP-CM multicenter study.
Int J Infect Dis. 2022;116:241-244. doi:
10.1016/j.ijid.2022.01.007
PubMedGoogle ScholarCrossref 35.Jones
R, Davis
A, Stanley
B,
et al. Risk predictors and symptom features of long COVID within a broad primary care patient population including both tested and untested patients.
Pragmat Obs Res. 2021;12:93-104. doi:
10.2147/POR.S316186
PubMedGoogle ScholarCrossref 36.Kisiel
MA, Janols
H, Nordqvist
T,
et al. Predictors of post-COVID-19 and the impact of persistent symptoms in non-hospitalized patients 12 months after COVID-19, with a focus on work ability.
Ups J Med Sci. 2022;127. doi:
10.48101/ujms.v127.8794
PubMedGoogle ScholarCrossref 37.Kostev
K, Smith
L, Koyanagi
A, Jacob
L. Prevalence of and factors associated with post-coronavirus disease 2019 (COVID-19) condition in the 12 months after the diagnosis of COVID-19 in adults followed in general practices in Germany.
Open Forum Infect Dis. 2022;9(7):ofac333. doi:
10.1093/ofid/ofac333
PubMedGoogle ScholarCrossref 39.Munblit
D, Bobkova
P, Spiridonova
E,
et al; Sechenov StopCOVID Research Team. Incidence and risk factors for persistent symptoms in adults previously hospitalized for COVID-19.
Clin Exp Allergy. 2021;51(9):1107-1120. doi:
10.1111/cea.13997
PubMedGoogle ScholarCrossref 40.Pazukhina
E, Andreeva
M, Spiridonova
E,
et al; Sechenov StopCOVID Research Team. Prevalence and risk factors of post-COVID-19 condition in adults and children at 6 and 12 months after hospital discharge: a prospective, cohort study in Moscow (StopCOVID).
BMC Med. 2022;20(1):244. doi:
10.1186/s12916-022-02448-4
PubMedGoogle ScholarCrossref 50.Timpson
N, Molloy
L, Ring
S,
et al. ALSPAC G0 (Avon Longitudinal Study of Parents and Children): Children of the 90s. University of Bristol. Accessed November 11, 2023.
http://www.bristol.ac.uk/alspac/ 64.Wulf Hanson
S, Abbafati
C, Aerts
JG,
et al; Global Burden of Disease Long COVID Collaborators. Estimated global proportions of individuals with persistent fatigue, cognitive, and respiratory symptom clusters following symptomatic COVID-19 in 2020 and 2021.
JAMA. 2022;328(16):1604-1615. doi:
10.1001/jama.2022.18931
PubMedGoogle ScholarCrossref 66.Bienvenu
LA, Noonan
J, Wang
X, Peter
K. Higher mortality of COVID-19 in males: sex differences in immune response and cardiovascular comorbidities.
Cardiovasc Res. 2020;116(14):2197-2206. doi:
10.1093/cvr/cvaa284
PubMedGoogle ScholarCrossref 68.Zeng
F, Dai
C, Cai
P,
et al. A comparison study of SARS-CoV-2 IgG antibody between male and female COVID-19 patients: a possible reason underlying different outcome between sex.
J Med Virol. 2020;92(10):2050-2054. doi:
10.1002/jmv.25989
PubMedGoogle ScholarCrossref 70.Romero Starke
K, Reissig
D, Petereit-Haack
G, Schmauder
S, Nienhaus
A, Seidler
A. The isolated effect of age on the risk of COVID-19 severe outcomes: a systematic review with meta-analysis.
BMJ Glob Health. 2021;6(12):e006434. doi:
10.1136/bmjgh-2021-006434
PubMedGoogle ScholarCrossref 72.Tenforde
MW, Kim
SS, Lindsell
CJ,
et al; IVY Network Investigators; CDC COVID-19 Response Team. 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
PubMedGoogle ScholarCrossref 73.Evans
RA, Leavy
OC, Richardson
M,
et al; PHOSP-COVID Collaborative Group. Clinical characteristics with inflammation profiling of long COVID and association with 1-year recovery following hospitalisation in the UK: a prospective observational study.
Lancet Respir Med. 2022;10(8):761-775. doi:
10.1016/S2213-2600(22)00127-8
PubMedGoogle ScholarCrossref 81.Antonelli
M, Penfold
RS, Merino
J,
et al. Risk factors and disease profile of post-vaccination SARS-CoV-2 infection in UK users of the COVID Symptom Study app: a prospective, community-based, nested, case-control study.
Lancet Infect Dis. 2022;22(1):43-55. doi:
10.1016/S1473-3099(21)00460-6
PubMedGoogle ScholarCrossref