Clinical Features of COVID-19 Patients in the First Year of Pandemic: A Systematic Review and Meta-Analysis

Background The new coronavirus disease (COVID-19) carries a high risk of infection and has spread rapidly around the world. However, there are limited data about the clinical symptoms globally. The purpose of this systematic review and meta-analysis is to identify the prevalence of the clinical symptoms of patient with COVID-19. Methods A systematic review and meta-analysis were carried out. The following databases were searched: PubMed, CINAHL, MEDLINE, EMBASE, PsycINFO, medRxiv, and Google Scholar, from December 1st, 2019 to January 1st, 2021. Prevalence rates were pooled with meta-analysis using a random-effects model. Heterogeneity was tested using I-squared (I2) statistics. Results A total of 215 studies, involving 132,647 COVID-19 patients, met the inclusion criteria. The pooled prevalence of the four most common symptoms were fever 76.2% (n = 214; 95% CI 73.9–78.5); coughing 60.4% (n = 215; 95% CI 58.6–62.1); fatigue 33.6% (n = 175; 95% CI 31.2–36.1); and dyspnea 26.2% (n = 195; 95% CI 24.1–28.5). Other symptoms from highest to lowest in terms of prevalence include expectorant (22.2%), anorexia (21.6%), myalgias (17.5%), chills (15%), sore throat (14.1%), headache (11.7%), nausea or vomiting (8.7%), rhinorrhea (8.2%), and hemoptysis (3.3%). In subgroup analyses by continent, it was found that four symptoms have a slight prevalence variation—fever, coughing, fatigue, and diarrhea. Conclusion This meta-analysis found the most prevalent symptoms of COVID-19 patients were fever, coughing, fatigue, and dyspnea. This knowledge might be beneficial for the effective treatment and control of the COVID-19 outbreak. Additional studies are required to distinguish between symptoms during and after, in patients with COVID-19.

A large number of studies have been published aiming to identify the clinical features of COVID-19. Several systematic reviews and meta-analyses have been published which have assessed the prevalence of baseline clinical characteristic and associated factors (Fu et al., 2020;Ghayda et al., 2020;Grant et al., 2020;Li et al., 2021;Rodriguez-Morales et al., 2020;Sun et al., 2020;Wan et al., 2020;Yang et al., 2020;Zhu et al., 2020). While such reviews are helpful, they have included relatively few studies. Moreover, a limitation was observed in those reviews in terms of geographical location and the sample characteristics included. Even though there have been a large number of publications, there is a need for an indepth understanding of the clinical symptoms of COVID-19 which can help in tackling this pandemic and preventing future outbreaks of infectious diseases. Therefore, the aim of this study was to conduct a systematic review and metaanalysis to assess the prevalence of the symptoms associated with COVID-19.

Methods
This systematic review and meta-analysis were undertaken according to the PRISMA standards.

Search Strategy
A systematic literature search for the period between December 1st, 2019 and January 1st, 2021, was conducted using the following databases: PubMed, CINAHL, MEDLINE, EMBASE, PsycINFO, medRxiv, and Google Scholar. Search terms used both free text words and medical subject headings, that is, MeSH terms, to search for papers to be included in the review; that is, (

Study Selection
Two investigators (MM; SM) performed the search, scrutinizing all titles and abstracts for eligibility against the inclusion and exclusion criteria. Any disagreements were resolved by discussion with a third investigator (KB). Studies were included in the review based on the following inclusion criteria: (1) diagnosed with  (2) reported prevalence of clinical symptoms; (3) subjects aged 15 or older; (4) all types of settings; (5) cross-sectional or cohort surveys (only the baseline data were extracted); and (6) sample size greater than 40 to avoid selection bias from small studies. The exclusion criteria were the following: (1) protocol papers, and conference abstracts; and (2) case reports and studies with a sample size of less than 40. For any additional information, the study authors were contacted.

Quality Assessment
Upon retrieval of the applicable studies, quality assessment was completed using the Newcastle-Ottawa Scale (NOS; Well et al., 2020). This scale consists of eight items that evaluate nonrandomized studies in terms of three criteria: the selection of the participants, the comparability of study groups, and outcome assessment. The NOS uses a score system with the lowest possible score of 0 and the highest possible score of 9. The total points awarded indicate the overall quality of the study. A study was determined to be of low risk of bias when the score was 7-9, of moderate risk of bias if the score was 5-6, and of high risk of bias if the score was 0-4 (Li & Katikireddi, 2019).

Data Analyses
The mean point of prevalence, the odds ratio (OR) with a 95% Confidence Interval (CI), was calculated as the effect size by using a random-effects model. Heterogeneity was tested using I-squared (I 2 ) statistics. A value of I 2 was considered to be low in terms of heterogeneity with 0-25%, moderate with 25-50%, and high with 50-75% (Higgins et al., 2003). The variation in terms of the prevalence of continent, NOS, and month of conducting the study was assessed by subgroup analyses if there were more than four studies in a subgroup. Meta-regression analyses were performed for moderating continuous variables (mean age, male gender, and comorbidities). A sensitivity analysis was performed by removing one study at a time to evaluate the impact of the pooled prevalence of the remaining studies (Patsopoulos et al., 2008).
Publication bias was estimated using Egger's linear regression test (Egger et al., 1997). A p value of less than 0.05 was considered as indicating statistical significance. Meta-analysis was conducted using the Comprehensive Meta-Analysis software, version 2.2 (Englewood, New Jersey, USA). Forest plots were constructed using a Microsoft Excel spreadsheet constructed by Neyeloff et al. (Neyeloff et al., 2012).

Results
The database search identified 4365 papers. Of these, 4018 papers were excluded during the title and abstract screening process. A further 132 papers were excluded during full text review for the following reasons: 34 papers were not conducted during the COVID-19 period; 9 did not give data about symptoms; 39 were duplicated papers; 50 were commentary, editorial or letter papers. As such, 215 studies were identified as being eligible for meta-analysis ( Figure 1 shows the PRISMA flow chart).
One hundred and fifty-one studies originated from China, 16 from the USA, 6 from Spain, 5 from Iran, 4 each from Italy, Korea, and the UK, 3 each from Belgium, France, the Kingdom of Saudi Arabia and Kuwait, 2 from Jordan, and 1 from each of the following: Bangladesh, Bulgaria, Canada, Germany, Japan, the Netherlands, Pakistan, Poland, Singapore, Somalia, and Turkey (see Supplementary Table 1 for the general characteristics of the studies).

Quality Assessment
The studies were assessed using the NOS checklist. Thirtynine studies were classified as having a low risk of bias, and 166 as moderate risk. The detailed results of the quality assessment of the studies included in this meta-analysis are listed in Supplementary Table 2.

Fever
Fever was estimated to occur in 214 studies. The overall pooled point estimates of prevalence for fever varied between 7.5% and 99.1%. All meta-analyses of prevalence estimates of fever reported by the 214 studies yielded a summary prevalence of 76.2% (84,823/132,436 participants, 95% CI 73.9-78.5) ( Figure 2) (see Table 1). In the subgroup analyses by continent in terms of where the study was conducted, the pooled prevalence of fever was 71.7% in Asia, 68.2% in Europe, and 51.5% in North America. The pooled prevalence     of fever was highest in February (82.5%) followed by January with 81.2% in (see Table 2).

Chills
Chills were estimated in 45 studies. The overall pooled point estimates of prevalence for chills varied between 1.6% and 53.3%. All meta-analyses of the prevalence estimates of chills reported by the 45 studies yielded a summary prevalence of 15% (2728/17,303 participants, 95% CI 11.9-18.6).

Sore Throat
Sore throat was estimated in 80 studies. The overall pooled point estimates of prevalence for sore throat varied between 1.2% and 57.4%. All meta-analyses of prevalence estimates of sore throat reported by the 80 studies yielded a summary prevalence of 14.1% (5389/41,810 participants, 95% CI 11.6-16.9). In the subgroup analyses, the highest prevalence of sore throat was in Europe (24.1%) and North America (21.6%) compared to studies conducted in Asia (12.3%). Pooled prevalence of sore throat was highest in April (25.4%).

Headache
Headache was estimated in 141 studies. The overall pooled point estimates of prevalence for headache varied between 1.5% and 75%. All meta-analyses of prevalence estimates of headache reported by the 141 studies yielded a summary prevalence of 12.1% (

Nausea or Vomiting
Nausea or vomiting was estimated in 95 studies. The overall pooled point estimates of prevalence for nausea or vomiting varied between 1% and 96.3%. All meta-analyses of prevalence estimates of nausea or vomiting reported by the 95 studies yielded a summary prevalence of 8.7% (4093/41,319 participants, 95% CI 7.1-10.5). A subgroups analysis by continent showed that nausea or vomiting prevalence was highest North America (20.1%), compared to Europe (13.3%) and Asia (7.1%).
In meta-regression analyses, patients with cardiovascular, renal, and cerebrovascular disease were significantly associated with the nausea or vomiting prevalence rate (p < 0.001).

Rhinorrhea
Rhinorrhea was estimated in 76 studies. The overall pooled point estimates of prevalence for rhinorrhea varied between 0.2% and 60.1%. All meta-analyses of prevalence estimates of rhinorrhea reported by the 76 studies yielded a summary prevalence of 8.2% (3452/30,150 participants, 95% CI 6.2-10.6). In the subgroup analyses, the prevalence of rhinorrhea was higher in both North America (28.4%) and Europe (25.1%), whereas in Asia it was reported as 5.9%.

Hemoptysis
Hemoptysis was estimated in 41 studies. The overall pooled point estimates of prevalence for hemoptysis varied between 0.1% and 41.4%. All meta-analyses of the prevalence estimates of hemoptysis reported by the 41 studies yielded a summary prevalence of 3.3% (12,578/49,942 participants, 95% CI 1.8-6.2).

Sensitivity Analysis
Sensitivity analysis was conducted with regard to all subgroups by excluding one study each time. This demonstrated that there were no differences in the overall estimation by more or less than 1%.

Discussion
COVID-19 is viewed as a major threat to public health due to the incredible damage it is doing to the medical services and the economies of almost all countries across the global (Wiebers & Feigin, 2020). Given that it is a new infectious disease, assessing the clinical signs and symptoms of COVID-19 is imperative for early detection and for the isolation of infected patients in order to reduce the spread of the disease, and to identify appropriate management strategies. This meta-analysis was undertaken to estimate the aggregate prevalence of clinical signs and symptoms of COVID-19 patients. This meta-analysis found that there was a minor variation in terms of COVID-19 patients between male (53.7%) and female (46.3%). This is consistent with other meta-analyses that have found that there were slightly more male than female patients (Li et al., 2021;Zhu et al., 2020). However, in contrast, a meta-analysis conducted by Peckham et al. (2020) involving 3,111,714 subjects found no difference with regard to the number of males and females with COVID-19. This difference might be linked to sex hormones and X-linked genes that may inactivate immunity with regard to responses (Agrawal et al., 2021). Clearly, the gender disparity with regard to COVID-19 is not fully understood.
The results of the current meta-analyses are even lower when compared with studies which reported symptoms at onset of fever, coughing, fatigue, and dyspnea for the Middle East Respiratory Syndrome (MERS) and Severe Acute Respiratory Syndrome (SARS) epidemics (Assiri et al., 2016;Azhar et al., 2019;Cleri et al., 2010;Yin & Wunderink, 2018;Z. Zhao et al., 2003). Furthermore, in sub-group analyses by continent, differences were found in the reported prevalence of symptoms. Studies conducted in Asia reported the highest prevalence of fever, expectorant, chills, and hemoptysis. In Europe, fatigue, dyspnea, anorexia, myalgias, sore throat, and headache were ranked as the most prevalent, whereas studies conducted in North America reported the highest prevalence of coughing, nausea or vomiting, and rhinorrhea. The prevalence of diarrhea was approximately the same in three continents. This variation of the symptoms between continents may be explained by variations between the patients in terms of social conditions, environment, and geographical location.
The findings of this analysis suggested that four symptoms have slight variations in prevalence between the continentsfever, coughing, fatigue, and diarrhea. However, it is recommended that every country should have its own symptoms list to evaluate patients.
The meta-regression analysis also revealed that male patients were significantly associated with a pooled estimation of anorexia, diarrhea, and nausea or vomiting. In addition, mean age and respiratory disease yielded a higher prevalence of dyspnea. These finding warrant further examination.
The major strength of this meta-analysis is the large sample size of over 132,647 subjects drawn from 215 studies which estimated the symptom onset of patients diagnosed with COVID-19. However, there are several potential limitations to this meta-analysis. First, this review searched medRxiv's preprint studies which, at the time of searching, were not peer reviewed. This might introduce publication bias. Second, there is a possibility that some studies have not been included in this meta-analysis, even though this analysis used different MeSH terms and several databases. In addition, only studies published, unpublished, or translated into English were included in this analysis. Third, around 151 of the studies originated from China, while the other 64 studies originated from 22 countries. Two studies from Spain included 25,615 subjects. Fourth, there were insufficient data available with regard to the demographic and clinical characteristics, so not all information could be eliminated thoroughly. Finally, all findings were derived from retrospective designs, which means that we cannot rule out selection bias. This is a systematic review and meta-analysis reporting pooled estimates with regard to the prevalence of symptoms associated with COVID-19. Thus, the findings can be used to help healthcare professionals and policy makers identify and monitor patients as part of the early screening process of COVID-19 patients. This might help ensure the appropriate utilization of healthcare resources, which in turn might help to reduce the severity of the impact of COVID-19 and be beneficial when it comes to effective management and treatment.

Conclusion
This meta-analysis has been undertaken to estimate the aggregate prevalence of the clinical signs and symptoms of COVID-19 patients. This meta-analysis found that the most prevalent symptoms with regard to COVID-19 patients were fever, coughing, fatigue, and dyspnea. This knowledge might be beneficial for the effective treatment and control of the COVID-19 pandemic. Additional studies are required to distinguish between symptoms during and after in patients with COVID-19.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.

Supplementary Material
Supplementary material for this article is available online.