More than 50 long-term effects of COVID-19: a systematic review and meta-analysis

COVID-19 can involve persistence, sequelae, and other medical complications that last weeks to months after initial recovery. This systematic review and meta-analysis aims to identify studies assessing the long-term effects of COVID-19. LitCOVID and Embase were searched to identify articles with original data published before the 1st of January 2021, with a minimum of 100 patients. For effects reported in two or more studies, meta-analyses using a random-effects model were performed using the MetaXL software to estimate the pooled prevalence with 95% CI. PRISMA guidelines were followed. A total of 18,251 publications were identified, of which 15 met the inclusion criteria. The prevalence of 55 long-term effects was estimated, 21 meta-analyses were performed, and 47,910 patients were included (age 17–87 years). The included studies defined long-COVID as ranging from 14 to 110 days post-viral infection. It was estimated that 80% of the infected patients with SARS-CoV-2 developed one or more long-term symptoms. The five most common symptoms were fatigue (58%), headache (44%), attention disorder (27%), hair loss (25%), and dyspnea (24%). Multi-disciplinary teams are crucial to developing preventive measures, rehabilitation techniques, and clinical management strategies with whole-patient perspectives designed to address long COVID-19 care.


INTRODUCTION
The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was detected in China in December 2019.Since then, more than 90 million people worldwide have been infected after a year, and over 2 million people have died from the coronavirus disease 2019 (COVID-19) 1 .Although unprecedented efforts from the scientific and medical community have been directed to sequence, diagnose, treat, and prevent COVID-19, individuals' lasting effects after the acute phase of the disease are yet to be revealed.
To date, there is no established term to coin the slow and persistent condition in individuals with lasting sequelae of COVID-19.Different authors have used the terms "Long-COVID-19", "Long Haulers", "Post-acute COVID-19", "Persistent COVID-19 Symptoms", "Post COVID-19 manifestations", "Long-term COVID-19 effects", "Post COVID-19 syndrome", among others.In the absence of an agreed definition, we convened for this review to refer to "Long-term effects of COVID-19".Symptoms, signs, or abnormal clinical parameters persisting two or more weeks after COVID-19 onset that do not return to a healthy baseline can potentially be considered long-term effects of the disease 2 .Although such alteration is mostly reported in severe and critical disease survivors, the lasting effects also occur in individuals with a mild infection who did not require hospitalization 3 .It has not yet been established if sex, gender, age, ethnicity, underlying health conditions, viral dose, or progression of COVID-19 significantly affect the risk of developing long-term effects of COVID-19 4 .
Since first reported, there has been a vast amount of social media patient groups, polls, comments, and scientific articles aiming to describe the chronicity of COVID-19.In parallel, hundreds of scientific publications, including cohorts studying specific effects of the disease and lists of case reports, have been described 5 .However, a broad overview of all the possible longstanding effects of COVID-19 is still needed.The aim of our study was to perform a systematic review and metaanalysis of peer-reviewed studies to estimate the incidence of all the symptoms, signs, or abnormal laboratory parameters extending beyond the acute phase of COVID-19 reported to date.

Search strategy and selection criteria
The search's objective was to identify peer-reviewed human studies in English that reported symptoms, signs, or laboratory parameters of patients at a post-COVID-19 stage (assessed two weeks or more after initial symptoms) in cohorts of COVID-19 patients.Only studies with a minimum of 100 patients were included.
The databases used to identify the studies were LitCOVID 6 (PubMed and Medline) and Embase.Studies were included if they were published before January 1 st 2021.The search terms used for both searches were: COVID long* OR haulers OR post OR chronic OR term OR complications OR recurrent OR lingering OR convalescent OR convalescence.Given that LitCOVID includes all articles from MedLine, in the search in Embase we excluded the articles from MedLine and those not related to COVID-19.All types of studies, including randomized controlled trials, cohorts, and cross-sectional studies, were analyzed only when the cases (numerator) were part of a COVID-19 cohort (denominator).Titles, abstracts, and full texts of articles were independently screened by two authors (SLL and TWO).The complete article was reviewed in case of difference of opinion on the inclusion based on title or abstract.Disagreement on the inclusion of a full-text article was discussed with all the authors.The exclusion criteria were: (1) not written in English; (2) have less than 100 patients included in the study.The systematic review followed the Preferred reporting Items for Systematic Reviewers and Meta-analysis (PRISMA) guidelines 7,8 .

Data extraction
Data were extracted by 4 authors (AC, PR, RS, SV) and QCed by two authors (TWO, CP).The descriptive variables extracted were country, setting, follow-up time, the severity of COVID-19, sample size, mean age and percentage of gender, outcomes, and names used to describe the Longterm effects of COVID-19.

Outcomes
All the diseases, disorders, symptoms, signs, and laboratory parameters reported total numbers or percentages were included.When two-time points were reported in the study, the outcomes assessed after the most extended follow-up were used.

Data-analysis
For effects reported only in a single study, the prevalence was estimated by dividing the number of patients with each symptom by the total number of COVID-19 patients in the sample multiplied by 100 to estimate the percentage.For effects reported in two or more studies, meta-analyses using a random-effects model were performed using the MetaXL software to estimate the pooled prevalence, which uses a double arcsine transformation 9 .Prevalence with 95% Confidence Intervals (CI) was presented.
Given the heterogeneity expected, a random-effects model was used.FHeterogeneity was assessed using the I 2 statistics.Values of 25%, 50%, and 75% for I 2 represented low, medium, and high heterogeneity.Sensitivity analyses were performed to assess the contribution of each study.Each study's quality was assessed and described using the MetaXL Guidelines, which is specific to evaluate the quality of studies assessing incidence and prevalence.A description of what was considered is found in supplement Table 1.

RESULTS
The title and abstract of 18,251 publications were screened.Of these, 82 full publications were reviewed.Nineteen studies were excluded because they involved less than 100 persons.A total of 15 studies were selected to be analyzed (Table 1, Figure 1).Most of the studies assessed specific symptoms included in a previously applied questionnaire.The process of study selection is presented in Figure 1.There were 9 studies from UK/ Europe, 3 from the US, 1 each from Australia, China, Egypt, and Mexico.The patient follow-up time ranged from 14 days to 110 days.Six out of the 11 studies included only patients hospitalized for COVID-19.The rest of the studies mixed mild, moderate, and severe COVID-19 patients.The number of patient cohorts that were followed up in the studies ranged from 102 to 44,799.Adults ranging from 17-87 years of age were included.There were no studies with overlapping samples.A couple of studies reported that fatigue was more common in female, and one study reported that postactivity polypnoea and alopecia were more common in female 3,10 .The rest of the studies did not stratify their results by age or sex.
We identified a total of 55 long-term effects associated with COVID-19 in the literature reviewed (Table 3).Most of the effects correspond to clinical symptoms such as fatigue, headache, joint pain, anosmia, ageusia, etc. Diseases such as stroke and diabetes mellitus were also present.Measurable parameters included 6 elevated laboratory parameters, i.e., interleukin-6 (IL-6), procalcitonin, serum ferritin, C-reactive protein (CRP), N-terminal (NT)-pro hormone BNP (NT-proBNP), and D-dimer.Abnormal chest XRay/computed tomography (CT) was also identified (Figure 2).Table 2 presents the prevalence of all the effects that were reported.It was possible to perform 21 meta-analyses, for the rest the prevalence was estimated using 1 cohort.The meta-analysis of the studies (n=7) that included an estimate for one symptom or more reported that 80% (95% CI 65-92) of the patients with COVID-19 have long-term symptoms.
One study was excluded because it did not provide a denominator, and therefore it was not possible to estimate the prevalence 11 .In such a study, the authors performed a survey in a Facebook group of patients who previously had COVID-19 and compared the symptoms of those hospitalized with mild to moderate symptoms.They concluded that both groups had symptoms after three months of having COVID-19.Symptoms that were not mentioned in any of the articles we studied include sudden loss of body weight, ear pain, eye problems, sneezing, cold nose, burning feeling in the trachea, dizziness, heart palpitations, pain/ burning feeling in the lungs, pain between the shoulder blades, Sicca syndrome, vertigo, body aches, and confusion 2,12 .

DISCUSSION
The recovery from COVID-19 should be more developed than checking for hospital discharge or testing negative for SARS-CoV-2 or positive for antibodies 13 .This systematic review and metaanalysis shows that 80% (95% CI 65-92) of individuals with a confirmed COVID-19 diagnosis continue to have at least one overall effect beyond two weeks following acute infection.In total, 55 effects, including symptoms, signs, and laboratory parameters, were identified, with fatigue, anosmia, lung dysfunction, abnormal chest XRay/CT, and neurological disorders being the most common (Table 1, Figure 2).Most of the symptoms were similar to the symptomatology developed during the acute phase of COVID-19.However, there is a possibility that there are other effects that have not yet been identified.In the following paragraphs, we will discuss the most common symptoms to illustrate how complex each one can be.However, further studies are needed to understand each symptom separately and in conjunction with the other symptoms.The five most common effects were fatigue (58%), headache (44%), attention disorder (27%), hair loss (25%), and dyspnea (24%).
Fatigue (58%) is the most common symptom of long and acute COVID-19 14 .It is present even after 100 days of the first symptom of acute COVID-19 3,14 .There are syndromes such as acute respiratory distress syndrome (ARDS), in which it has been observed that after a year, more than two-thirds of patients reported clinically significant fatigue symptoms 15 .The symptoms observed in post-COVID-19 patients, resemble in part the chronic fatigue syndrome (CFS), which includes the presence of severe incapacitating fatigue, pain, neurocognitive disability, compromised sleep, symptoms suggestive of autonomic dysfunction, and worsening of global symptoms following minor increases in physical and/or cognitive activity [16][17][18][19][20] .Currently, myalgic encephalomyelitis (ME) or CFS is a complex and controversial clinical condition without established causative factors, and 90% of ME/CFS has not been diagnosed 21 .Possible causes of CFS include viruses, immune dysfunction, endocrine-metabolic dysfunction, and neuropsychiatric factors.The infectious agents related to CFS have been Epstein-Barr virus, cytomegalovirus, enterovirus and herpesvirus 22 .It is tempting to speculate that SARS-CoV-2 can be added to the viral agents' list causing ME/CFS.Several neuropsychiatric symptoms have been reported, headache (44%), attention disorder (27%), and anosmia (21%).There are other symptoms reported, which were not included in the publications, including brain fog and neuropathy 23,24 .The etiology of neuropsychiatric symptoms in COVID-19 patients is complex and multifactorial.They could be related to the direct effect of the infection, cerebrovascular disease (including hypercoagulation) 25 , physiological compromise (hypoxia), side effects of medications, and social aspects of having a potentially fatal illness 26 .Adults have a double risk of being newly diagnosed with a psychiatric disorder after the COVID-19 diagnosis 26 , and the most common psychiatric conditions presented were anxiety disorders, insomnia, and dementia.Sleep disturbances might contribute to the presentation of psychiatric disorders. 27Prompt diagnosis and intervention of any neuropsychiatric care is recommended for all patients recovering from COVID-19.An increase in mental health attention models in hospitals and communities is needed during and after the COVID-19 pandemic.Hair loss after COVID-19 could be considered as telogen effluvium, defined by diffuse hair loss after an important systemic stressor or infection, and it is caused because of premature follicular transitions from active growth phase (anagen) to resting phase (telogen).It is a self-limiting condition that lasts approximately 3 months, but it could cause emotional distress 28 .Dyspnea and cough were found in 24% and 19% of patients, respectively (Table 2, Figure 2).In addition, abnormalities in CT lung scans persisted in 35% of patients even after 60-100 days from the initial presentation.In a follow-up study conducted in China among non-critical cases of hospitalized patients with COVID-19, radiographic changes persisted in nearly two-thirds of patients 90 days after discharge 29 .Although most of the available studies do not include baseline pulmonary dysfunction or radiographic abnormalities, findings indicate improvement or resolution of abnormal CT findings.Previous data from recovered patients with other viral pneumonia 30,31 , also found residual radiographic changes.Abnormalities in pulmonary function, such as decreased diffusion capacity for carbon monoxide, were present among 10% of patients in this meta-analysis.Although these findings are not as high as compared to other available studies of survivors with COVID-19 or SARS, where the estimate of lung dysfunction is 53% and 28% respectively 32,33 , the reasons behind these differences could be distinct follow-up periods, definitions of pulmonary dysfunction, or characteristics of the patient population.Nevertheless, residual radiographic findings or lung function abnormalities require additional investigation on their clinical relevance and long-term consequences.
The immune-mediated tissue damage in COVID-19 involves cellular and humoral responses, but the immunity to SARS-CoV-2 and the protection to reinfection or a final viral 29,34 clearance is unknown.Also, the reason why some patients experience long-term symptoms after COVID-19 is uncertain.This could be partially explained by host-controlled factors that influence the outcome of the viral infection, including genetic susceptibility, age of the host when infected, dose and route of infection, induction of anti-inflammatory cells and proteins, presence of concurrent infections, past exposure to cross-reactive agents, etc.Whether SARS-CoV-2 can cause substantial tissue damage leading to a chronic form of the disease such as the chronic lesions in convalescence observed in other viruses such as HIV, hepatitis C virus (HCV), hepatitis B virus (HBV), and some herpesviruses is still unknown.
The results assessed in the present study are in line with the current scientific knowledge on other coronaviruses, such as those producing SARS and MERS, both sharing clinical characteristics with COVID-19, including post symptoms.Studies on SARS survivors have shown lung abnormalities months after infection.After a one-year follow-up, a study showed that 28% of the survivors presented decreased lung function and pulmonary fibrosis signs 33,35,36 .In addition, MERS survivors showed pulmonary fibrosis (33%) 37 .Regarding psychiatric symptoms, a study reported high levels of depression, anxiety, and post-traumatic stress disorder (PTSD) 26 in the long term in patients previously infected with other coronaviruses.
To assure that future healthcare providers, researchers, and educators recognize the effects of longterm COVID19 that are sex-and age-specific related, it is of high importance to classify the groups according to such variables to make better decisions about prevention, diagnosis and disease management.
Limitations of this systematic review and meta-analyses include the small sample size for some outcomes, which makes it difficult to generalize these results to the general population.The variation in the definition of some outcomes and markers and the possibility of bias.For example, several studies that used a self-reported questionnaire could result in reporting bias.In addition, the studies were very heterogeneous, mainly due to the follow-up time references and the mixture of patients who had moderate and severe COVID-19.All of the studies assessed had performed their internal pre-definition of symptoms, and therefore there is the possibility that important outcomes were not reported.Another limitation is that, given that COVID-19 is a new disease, it is not possible to determine how long these effects will last.In order to decrease heterogeneity and have a better understanding of the long-term effects of COVID-19, there is a need for studies to stratify by age, previous comorbidities, the severity of COVID-19 (including asymptomatic), as well as the duration of each symptom.To determine whether these long-term effects either complicate previous diseases or are a continuation of COVID-19, there is a need for prospective cohort studies.The baseline characteristics should be well established.
There is a need to standardize biological measures such as peripheral blood markers of genetic, inflammatory, immune, and metabolic function to compare studies.Besides studying pre-defined symptoms and markers, an open question should be included.Proper documentation in medical charts by health care providers and the flexibility and collaboration from the patients to report their symptoms are of equal importance.

CONCLUSIONS
More evidence and research from multi-disciplinary teams are crucial to understanding the causes, mechanisms, and risks to develop preventive measures, rehabilitation techniques, and clinical management strategies with whole-patient perspectives designed to address the after-COVID-19 care.From the clinical point of view, physicians should be aware of the symptoms, signs, and biomarkers present in patients previously affected by COVID-19 to promptly assess, identify and halt long COVID-19 progression, minimize the risk of chronic effects and help reestablish pre-COVID-19 health.Management of all these effects requires further understanding to design individualized, dynamic cross-sectoral interventions in Post-COVID clinics with multiple specialties, including graded exercise, physical therapy, continuous checkups, and cognitive behavioral therapy when required 38,39 .

Figure 1 .
Figure 1.Study selection.Preferred items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.Out of 15,917 identified studies and after application of the inclusion and exclusion criteria, 15 studies were included in the quantitative synthesis.