Clinical Determinants of Severe COVID-19 Disease – A Systematic Review and Meta-Analysis

Abstract

Background:

A systematic review and meta-analysis of available studies was performed to investigate the clinical characteristics that can predict COVID-19 disease severity.

Materials and Methods:

Databases including PubMed, Embase, and Web of Science were searched from December 31, 2019, to May 24, 2020. Random-effects meta-analysis was used for summarizing the Pooled odds ratio (pOR) of individual clinical characteristics to describe their association with severe COVID-19 disease.

Results:

A total of 3895 articles were identified, and finally, 22 studies comprising 4380 patients were included. Severe disease was more common in males than females (pOR: 1.36, 95% confidence interval [CI]: 1.08–1.70). Clinical features that were associated with significantly higher odds of severe disease were abdominal pain (pOR: 6.58, 95% CI: 1.56–27.67), breathlessness (pOR: 3.94, 95% CI: 2.55–6.07), and hemoptysis (pOR: 3.35, 95% CI: 1.05–10.74). pOR was highest for chronic obstructive pulmonary disease (pOR: 2.92, 95% CI: 1.70–5.02), followed by obesity (pOR: 2.84, 95% CI: 1.19–6.77), malignancy (pOR: 2.38, 95% CI: 1.25–4.52), diabetes (pOR: 2.29, 95% CI: 1.56–3.39), hypertension (pOR: 1.72, 95% CI: 1.23–2.42), cardiovascular disease (pOR: 1.61, 95% CI: 1.31–1.98) and chronic kidney disease (pOR: 1.46, 95% CI: 1.06–2.02), for predicting severe COVID-19.

Conclusion:

Our analysis describes the association of specific symptoms and comorbidities with severe COVID-19 disease. Knowledge of these clinical determinants will assist the clinicians in the risk-stratification of these patients for better triage and clinical management.

INTRODUCTION

The novel coronavirus, named as severe acute respiratory syndrome coronavirus 2, was identified in Wuhan, China, in December 2019. The disease caused by the virus, COVID-19, has created havoc all over the world and has been declared pandemic by the World Health Organization (WHO). As of March 21, 2020, 11,183 patients have succumbed to this disease and with the rapid spread of the disease, these numbers might run into millions.[1]

The clinical spectrum of COVID-19 disease is wide, ranging from nonsevere (asymptomatic infection and mild respiratory tract infection) to severe disease (severe pneumonia and critical illness, including multiorgan dysfunction).[2] In a case series of 44,672 confirmed COVID-19 patients, 14% had severe, and 5% had critical disease.[2] However, most of the patients present with fever, dry cough, myalgia and have a favorable prognosis.[2] Older patients and those with comorbidities progress to severe disease and have worse outcomes.[3]

With overwhelmed health-care systems and no proven treatment, it is important to identify the patients who could have a high likelihood of progression to severe disease. This will help the concerned physicians to allocate the resources judiciously. The goal of this investigation was to identify the clinical determinants which are associated with severe COVID-19 disease.

MATERIALS AND METHODS

Data sources and searches

This systematic review was performed according to the Preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA). Databases including PubMed, Embase, and Web of Science were searched from December 31, 2019, to May 24, 2020. There were no restrictions in terms of country, publication language or publication date. Reference lists of all relevant articles and “related citation” search tool of PubMed were checked for any additional publications. The detailed search criteria used are available in Supplement.

Selection criteria

Study selection was performed by two independent investigators (A. S. and P. A.). We included studies that focused on individual symptoms and comorbidities of laboratory-confirmed COVID-19 patients and reported the data according to disease severity or ICU admission. Case reports, duplicate publications, reviews, editorials, letters, and family-based studies, studies with insufficient data on symptoms/comorbidities on admission in either severe or non-severe disease groups, and studies reporting exclusively on pediatric (<18 years age) or pregnant populations were excluded. Discrepancies between the reviewers were resolved in the presence of a third reviewer (J. N.).

Data abstraction and quality assessment

Data collected included: study characteristics – authors, publication date, study design, country, sample size; patient's characteristics – median age with interquartile range, sex (% men); criteria for severe disease; total number of severe and non-severe patients; and clinical characteristics (clinical features and comorbidities) at admission – overall prevalence and prevalence among severe and non-severe patients. One reviewer extracted the data (A. S.) and second reviewer (S. S) verified the data independently. The methodological quality of the study was assessed with the Appraisal tool for Cross-Sectional Studies (AXIS) tool.[4] Two authors (S. S, A. S.) performed the quality assessment separately, and disagreements were resolved by consensus in the presence of a third reviewer (P. A.). In the AXIS tool, for every correct answer, score of one was assigned to each of the twenty questions.

Quantitative data synthesis

Patient numbers were extracted across all the included studies for each group (severe and non-severe) according to the individual symptoms and comorbidities. The odds ratio (OR, 95% confidence intervals [CIs]) of individual clinical characteristics was used to describe their association with severe COVID-19 disease. These ORs were further pooled using random-effects meta-analysis. To assess the heterogeneity among studies, inconsistency statistics (I²) were calculated. I² >50% was considered as significant heterogeneity. Publication bias was visually analyzed from Funnel plots and Egger's regression was also performed. P value for Egger's regression coefficient < 0.05 was considered as significant publication bias. All data were collected in Microsoft Excel Spreadsheet (MS Office – 2018). Random-effects analysis, generation of forest plot, assessment of heterogeneity, and publication bias were performed with the METAN platform for STATA (version-14.2); StataCorp, College Station, TX.

As the study design was a systematic review and meta-analysis, Institute Ethics Committee approval was not sought.

RESULTS

Search results and study characteristics

The literature search flow diagram is summarized in PRISMA format [Figure 1]. Using our search criteria (available in supplement), we identified 3895 studies, of which 3645 were from PubMed, 50 were from EMBASE, and 200 were from Web of Science. A total of 209 records were screened after the removal of duplicates. A total of 87 full-text articles were assessed for eligibility and 65 articles were excluded due to various reasons, as shown in Figure 1. Finally, 22 studies were included in this meta-analysis.

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram

Characteristics of included studies

A total of 22 studies, consisting of 4380 patients, were selected for this meta-analysis [Table 1]. Studies were published recently between January 24, 2020 and May 24, 2020. Individual study population size ranged between 12 and 1494 patients. Fifty-six percent of the study population were males. Median age of the patients in severe disease cohort varied from 45.2 to 67 years, whereas median age in non-severe disease cohort varied from 37 to 68.5 years. Individual symptoms studied were cough,[5,6,7,9,10,12,16,18,19,20] expectoration,[5,6,7,9,10,14,18,19,21] fever,[5,6,9,10,11,13,14,15,16,17,18,19,20,21] breathlessness,[5,6,9,10,11,13,14,15,16,17,18,19,20,21] hemoptysis,[5,6] sore throat,[5,7,9,10,15,16,18,21] fatigue,[5,6,9,10,11,13,14,16,17,18] myalgia,[6,7,9,10,12,16,18,19,21] headache,[5,6,7,8,9,10,12,16,18,21] nausea/vomiting,[5,9,11,12,16,18,21] diarrhea,[5,7,9,11,12,15,16,17,18,21] abdominal pain,[9,11] anorexia,[9,11] and anosmia.[16,18] The various comorbidities studied were chronic obstructive pulmonary disease (COPD)[5,6,7,9,11,12,16,17,18,19,21,22,23,24,25,26] diabetes[5,6,7,9,11,12,13,14,16,17,18,19,21,22,23,24,25,26] obesity,[16,18,22] hypertension,[5,6,7,9,11,12,13,14,16,17,18,19,21,22,23,25,26] cardiovascular disease (CVD), [5,6,7,9,11,12,13,14,16,17,18,19,21,22,24,25] cerebrovascular accidents,[5,9,11,16,18,19,21,24] chronic kidney disease (CKD),[5,9,11,12,16,18,21,24,25,26] chronic liver disease,[6,9,11,19,21,24] malignancy,[5,6,9,16,17,19,21,23,26] and immunocompromised state.[5,18,24] Majority of the studies (13) were from China,[5,6,7,8,9,10,11,12,13,14,22,23,26] however, three studies were from the United States,[16,18,4] two from Italy[17,21] and one each from Singapore,[15] Norway,[20] South Korea[19] and Israel.[25] Each study was retrospective observational in design. The number of clinical characteristics including comorbidities reported in each study, varied from 3 in one study[20] to 21 in another study.[5] Patients with severe disease were older compared to those with non-severe disease (59.8 years vs. 50.8 years, P = 0.008). According to the WHO-China joint mission,[2] severe disease was defined as tachypnea (≥30 breaths/min) or oxygen saturation ≤ 93% at rest, or ratio of arterial oxygen saturation and fraction of inspired oxygen < 300 mmHg, and critical disease was defined as respiratory failure requiring mechanical ventilation, shock, or other organ failure that requires intensive care. Severe/critical disease were considered “Severe” in most of the studies.[5,7,8,10,12,16,23] Intensive care unit (ICU) admission was considered as “Severe/critical disease” in six studies.[18,19,20,21,24,25] Results of quality assessment of the included studies are summarized as AXIS scores in Table 1. Overall quality of studies was good, with thirteen out of twenty-two studies having scores above average (score ≥ 15).

Table 1.

Characteristics of the included studies

Author	Publication date	Country	Sample size			Age (median, IQR)		Male (%)		Clinical characteristics included		Definition of severity	Quality of study (score out of 20)**
			Total	Severe	Nonsevere	Severe	Nonsevere	Severe	Nonsevere	Numbers	Characteristics*
Huang C	January 24, 2020	China	41	13	28	49 (41-61)	49 (41-57.5)	85	68	13	a, b, c, d, e, g, i, n, o, p, q, t, v	Requiring ICU care	16
Wang D	February 07, 2020	China	138	36	102	66 (57-78)	51 (37-62)	61.1	52	20	a, b, c, d, f, g - t, v	Requiring ICU care	16
Liu Y	February 09, 2020	China	12	6	6	64	43.5	50	83.3	11	a, c, h, i, j, k, n, o, p, q, s	Severe and critical disease	12
Zhang J	February 18, 2020	China	140	58	82	64 (25-87)	51.5 (26-78)	56.9	46.3	16	a, c, d, g, j, k, l, m, n - t	Severe and critical disease	15
Xu Y	February 21, 2020	China	50	13	37	NA	NA	54	59.5	8	a, b, c, d, f, g, h, i	Severe and critical disease	14
Tian S	February 27, 2020	China	262	46	216	61.4 (1-94)	44.5 (1-93)	56.5	46.8	4	a, c, d, i	Severe and critical disease	16
Guan W	February 28, 2020	China	1099	173	926	52 (40-65)	45 (34-57)	57.8	58.2	21	a - s, v, w	Severe and critical disease	16
Liu W	February 28, 2020	China	78	11	67	66 (51-70)	37 (32-41)	63.6	47.8	4	n, o, p, v	Clinical deterioration to severe or critical disease or death	16
Li K	February 29, 2020	China	83	25	58	53.7 (12.3)	41.9 (10.6)	60	50	12	a, b, c, d, f, h, i, k, n, o, p, q	Severe and critical disease	16
Yudong P	March 02, 2020	China	112	16	96	57.5 (54-63)	62 (55-67.5)	56.25	45.83	7	a, c, d, g, o, p, q	Critical disease	15
Young B	March 03, 2020	Singapore	18	6	12	56 (47-73)	37 (31-56)	33	58	5	a, c, d, f, k	Requiring supplemental oxygen	12
Wu C	March 13, 2020	China	201	84	117	58.5 (50-69)	48 (40-54)	71.4	58.1	8	a, b, c, d, g, o, p, q	Acute respiratory distress syndrome	17
Gao Y	March 17, 2020	China	43	15	28	45.2 (7.68)	43 (14)	60	60.7	5	n, o, p, q, u	Not clear	14
Chow N (CDC US)	March 31, 2020	US	1494	457	1037	NA	NA	NA	NA	7	n, o, q, r, s, t, w	Requiring ICU care	15
Ihle-hansen H	April 10, 2020	Norway	42	9	33	71.8	66.8	NA	NA	3	a, c, d	Requiring ICU care	17
Colaneri M	April 23, 2020	Italy	44	17	27	NA	NA	76.5	55.6	10	a, c, d, g, k, n-q, v	Requirement for highflow oxygen	17
Hong K	April 24, 2020	South Korea	98	13	85	NA	NA	46.2	37.6	12	a-d, h, n-r, t, v	Requiring ICU care	16
Aggarwal S	April 29, 2020	US	16	8	8	67 (38-70)	68.5 (41-95)	63	88	18	a, b, d, f-k, x, n-s, u, v	Critical disease	14
Zhao X	April 29, 2020	China	91	30	61	NA	NA	46.7	57.4	6	n-q, s, v	Not clear	15
Lagi F	April 30, 2020	Italy	84	16	68	67 (58-71)	62 (50-72)	87.5	60.3	17	a-d, f, h-k, n-t, v	Requiring ICU care	18
Itelman E	May 01, 2020	Israel	162	26	136	NA	NA	80.8	51.9	5	n-q, s	Requiring ICU care	17
Ferguson J	May 14, 2020	US	72	21	51	57.6 (42.2-70.1)	61.7 (46.6-72.9)	61.9	49	19	a-d, f-k, x, n-s, u, w	Requiring ICU care	19

*Clinical characteristics: Clinical symptoms - a: Cough, b: Expectoration, c: Fever, d: Dyspnea, e: Hemoptysis, f-: Sore throat, g: Fatigue, h: Myalgia, i: Headache, j: Nausea or vomiting, k: Diarrhea, l-abdominal pain, m: Anorexia; n: Chronic obstructive pulmonary disease, o: Diabetes, p: Hypertension, q: Cardiovascular diseases, r: Cerebrovascular accidents, s: Chronic kidney disease, t: Chronic liver disease, u: Obesity, v: Malignancy, w: Immunodeficiency, x: Anosmia. ^Severe disease (any of the following conditions): I, respiratory distress, RR ≥30 breaths/min; II, oxygen saturation ≤93% at rest; III, PaO2/FiO2 ≤300 mmHg (1 mmHg=0.133 kPa); AND critical disease (any of the following conditions): I, respiratory failure and a requirement for mechanical ventilation; II, shock; III, concomitant failure of other organs and requirement for ICU monitoring and treatment[2], **Scores for each study in AXIS tool. RR: Respiratory rate, PaO2/FiO2: Partial pressure of oxygen/fraction of inspired oxygen, ICU: Intensive care unit, AXIS: Appraisal tool for Cross-Sectional Studies, IQR: Interquartile range, NA: Not available

Quantitative data synthesis results

ORs of severe disease were pooled for each of the individual symptoms and comorbidities. Forest plots of pOR and funnel plots for each of the clinical determinants are depicted in Supplementary Figure S1 ^{(220.9KB, tif)} -S50 ^{(220KB, tif)} , Table 2 and Figure 2 summarizes the pOR for each clinical determinant (clinical feature at admission and comorbidities). Severe disease was more common in males than females (pOR: 1.36, 95% CI: 1.08–1.70). Clinical features associated with significantly higher odds of disease severity were abdominal pain (pOR: 6.58, 95% CI: 1.56–27.67) and breathlessness (pOR: 3.94, 95% CI: 2.55–6.07). Fever (pOR: 1.48, 95% CI: 1.19–1.85) and hemoptysis (pOR: 3.35, 95% CI: 1.05–10.74) were also associated with severe disease, although their lower confidence levels were approaching near one. Patients with comorbidities were also at higher odds of presenting with severe COVID-19 disease. pOR was highest for COPD (pOR: 2.92, 95% CI: 1.70–5.02), followed by obesity (pOR: 2.84, 95% CI: 1.19–6.77), malignancy (pOR: 2.38, 95% CI: 1.25–4.52), diabetes (pOR: 2.29, 95% CI: 1.56–3.39), hypertension (pOR: 1.72, 95% CI: 1.23–2.42), CVD (pOR: 1.61, 95% CI: 1.31–1.98) and CKD (pOR: 1.46, 95% CI: 1.06–2.02). With the exception of the studies considered for breathlessness, nausea/vomiting, anorexia, and diabetes, none of the studies included in the meta-analysis for comorbidities had statistical heterogeneity (I² < 50%). Funnel plot analyses [Supplementary Figures: S1 ^{(220.9KB, tif)} -S50 ^{(220KB, tif)} ] and Egger's regression [Table 2] demonstrated the evidence of publication bias in the meta-analysis of studies focussing on fever, COPD and CVD.

Table 2.

Summary of meta-analyses for each of the clinical symptoms and comorbidities that are associated with severe COVID-19 infection

Clinical characteristics	Odds ratio	Lower CL	Upper CL	Number of studies	Total patients included in meta-analysis	Prevalence of characteristics in severe disease (n/N)	Prevalence of characteristics in mild disease (n’/N’)	I2 (%)	Publication bias (Egger’s P value)
Demographic characteristics
Male gender	1.36	1.08	1.70	20	2844	-	-	13.4	0.16
Clinical characteristics
Cough	1.24	0.98	1.56	17	2512	392/560	1242/1952	5.0	0.26
Expectoration	1.15	0.73	1.82	9	1866	132/394	454/1472	47.8	0.65
Fever	1.48	1.19	1.85	17	2512	369/560	1055/1952	0.0	0.03
Dyspnea	3.94	2.55	6.07	16	2500	251/554	339/1946	56.1	0.12
Hemoptysis	3.35	1.05	10.74	2	1140	5/186	7/954	0.0	NA
Sore throat	1.39	0.77	2.49	8	1560	47/298	168/1262	29.6	0.79
Fatigue	1.22	0.83	1.81	10	1913	196/439	607/1474	41.6	0.23
Myalgia	1.25	0.87	1.79	9	1652	73/311	249/1341	8.0	0.34
Headache	1.15	0.80	1.64	9	1857	44/357	174/1500	0.0	0.89
Nausea/vomiting	0.68	0.30	1.51	7	1561	29/318	104/1243	55.6	0.31
Diarrhea	1.43	0.93	2.21	10	1706	36/366	88/1340	0.0	0.74
Abdominal pain	6.58	1.56	27.67	2	278	9/94	2/184	0.0	NA
Anorexia	2.54	0.74	8.70	2	278	32/94	40/184	72.3	NA
Anosmia	0.61	0.11	3.48	2	88	2/29	5/59	0.0	NA
Comorbid illness
Chronic obstructive pulmonary disease	2.92	1.70	5.02	16	3695	124/925	177/2770	23.3	<0.01
Diabetes	2.29	1.56	3.39	18	4008	258/1025	413/2983	50.5	0.08
Obesity	2.84	1.19	6.77	3	131	18/44	18/87	0.0	0.43
Hypertension	1.72	1.23	2.42	17	2514	182/568	412/1946	41.6	0.41
Cardiovascular disease	1.61	1.31	1.98	16	3839	199/984	372/2855	0.0	0.01
Cerebrovascular accidents	1.68	0.73	3.84	8	3141	21/782	44/2359	32.1	0.45
Chronic kidney disease	1.46	1.06	2.02	10	3308	70/831	112/2477	0.0	0.16
Chronic liver disease	1.55	0.75	3.18	6	1995	12/593	20/1402	0.0	0.90
Malignancy	2.38	1.25	4.52	9	1689	19/317	29/1372	0.0	0.81
Immunocompromised state	1.46	0.98	2.17	3	2665	42/651	70/2014	0.0	0.33

CL: Confidence limits, n: Number of patients with the clinical determinant among patients with severe disease, N: Total number of patients with severe disease, n’: Number of patients with the clinical determinant among patients with mild disease, N’: Total number of patients with severe disease, I²: Heterogeneity statistics, Egger’s P<0.05: Publication bias present

Figure 2.

Summary of pooled odds ratio for each of the presenting clinical features and comorbidities. OR – pooled odds ratio, LCL – lower confidence limit of OR, UCL – upper confidence limit of OR, COPD – chronic obstructive pulmonary disease, CVD – cardiovascular diseases, CVA – cerebrovascular accidents, CLD – chronic liver disease, CKD – chronic kidney disease

DISCUSSION

COVID-19 is a rapidly progressing pandemic affecting millions of people worldwide. With the surge of cases, it is expected to overwhelm health-care systems, thereby making it important for physicians to identify clinical characteristics that could point toward progression-to-severe illness. In our meta-analysis of 4380 patients, we found that patients presenting with complaints of breathlessness, hemoptysis and/or abdominal pain, and comorbidities had significantly higher odds of having severe disease.

Multiple studies have shown that patients with breathlessness on arrival had a higher likelihood development of acute respiratory distress syndrome and ICU requirements.[7,14,9] In studies conducted by Guan et al. and Huang et al., the incidence of hemoptysis was higher among patients with severe disease as compared to that of non-severe disease, although its proportion was lower in both the study groups.[5,6] In a study by Zhang et al., few COVID-19 patients presented with atypical abdominal pain and were initially admitted to the surgical ward but subsequently required ICU. These patients were presumed to infect others during their hospital stay, and the newly infected patients also had abdominal pain at presentation. Hence, some authors have suggested the gastrointestinal tract as an alternative route for viral transmission.[27] Hence, it is necessary to not miss abdominal pain as a rare but important predictor of severe disease. Therefore, any patient presenting with SARI with suspicion of COVID-19 and complaints of breathlessness, hemoptysis and/or abdominal pain should be admitted and evaluated further before deciding further course of treatment. These symptoms, along with fever and cough, might act as warning signs of severe disease.

In most of the included studies, the patients in the severe group had a higher median age when compared to the non-severe group, which was consistent with previous reports.[14,23] Our meta-analysis showed that patients with COPD had the highest risk of the development of severe disease, followed by obesity, malignancy, diabetes, hypertension, CVD, and CKD. A previous meta-analysis of eight studies had shown CVD, respiratory illness, and hypertension as significant predictors of severe disease.[28] The study differs in terms of the inclusion of a greater number of studies and comorbidities. A weaker immune system might explain the higher likelihood of the development of severe disease among older patients with comorbidities.

There are certain limitations of this meta-analysis. The studies included are retrospective in nature with considerable heterogeneity. Further, 13 out of 22 of the studies are from a single country. The criteria of severe disease were also not similar across all the included studies, thereby limiting the strength of our observations. We have also not included the studies exclusively reporting predictors of mortality in COVID-19 patients. Finally, it is possible that newer studies might have been published between the completion of this literature review and its publication.

CONCLUSION

Our analysis describes the presence of a significant association of the severe disease with the male gender and specific presenting symptoms such as breathlessness, abdominal pain, hemoptysis, fever, and cough. The presence of comorbidities, namely, COPD, CKD, diabetes, CVD and hypertension were also significant risk factors for severe disease, which is in line with previous studies. Knowledge of these clinical determinants will assist the clinicians in the risk-stratification of the patients for better triage and clinical management.

What is already known on the subject

• Patients with COVID-19 presents with a wide spectrum of clinical manifestations, i.e., asymptomatic, mild upper respiratory tract symptoms, severe disease, and critical disease

• It is difficult to predict the disease progression early in the course of illness

• Multiple laboratory parameters, comorbid illness, and advanced age have been shown to predict the disease prognosis.

Study's main messages

• This updated meta-analysis consisted of 22 studies comprising 4380 patients

• Severe disease was more common in males than females

• Clinical features that were associated with significantly higher odds of severe disease were abdominal pain, breathlessness, and hemoptysis

• pOR was highest for chronic obstructive pulmonary disease, followed by obesity, malignancy, diabetes, hypertension, CVD , and CKD, for predicting severe COVID-19

• Knowledge of these clinical determinants will help the clinician to triage and manage the patients carefully, and appropriately allocate the resources in this resource-constraining pandemic.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Figure S1

Forest plot of odds ratio for cough as a predictor of disease severity

Figure S2

Funnel plot of odds ratio for cough as a predictor of disease severity

Figure S3

Forest plot of odds ratio for expectoration as a predictor of disease severity

Figure S4

Funnel plot of odds ratio for expectoration as a predictor of disease severity

Figure S5

Forest plot of odds ratio for fever as a predictor of disease severity

Figure S6

Funnel plot of odds ratio for fever as a predictor of disease severity

Figure S7

Forest plot of odds ratio for dyspnea as a predictor of disease severity

Figure S8

Funnel plot of odds ratio for dyspnea as a predictor of disease severity

Figure S9

Forest plot of odds ratio for hemoptysis as a predictor of disease severity

Figure S10

Funnel plot of odds ratio for hemoptysis as a predictor of disease severity

Figure S11

Forest plot of odds ratio for sore throat as a predictor of disease severity

Figure S12

Funnel plot of odds ratio for sore throat as a predictor of disease severity

Figure S13

Forest plot of odds ratio for fatigue as a predictor of disease severity

Figure S14

Funnel plot of odds ratio for fatigue as a predictor of disease severity

Figure S15

Forest plot of odds ratio for myalgia as a predictor of disease severity

Figure S16

Funnel plot of odds ratio for myalgia as a predictor of disease severity

Figure S17

Forest plot of odds ratio for headache as a predictor of disease severity

Figure S18

Funnel plot of odds ratio for headache as a predictor of disease severity

Figure S19

Forest plot of odds ratio for nausea / vomiting as a predictor of disease severity

Figure S20

Funnel plot of odds ratio for nausea / vomiting as a predictor of disease severity

Figure S21

Forest plot of odds ratio for diarrhea as a predictor of disease severity

Figure S22

Funnel plot of odds ratio for diarrhea as a predictor of disease severity

Figure S23

Forest plot of odds ratio for abdominal pain as a predictor of disease severity

Figure S24

Funnel plot of odds ratio for abdominal pain as a predictor of disease severity

Figure S25

Forest plot of odds ratio for anorexia as a predictor of disease severity

Figure S26

Funnel plot of odds ratio for anorexia as a predictor of disease severity

Figure S27

Forest plot of odds ratio for anosmia as a predictor of disease severity

Figure S28

Funnel plot of odds ratio for anorexia as a predictor of disease severity

Figure S29

Forest plot of odds ratio for COPD as a predictor of disease severity

Figure S30

Funnel plot of odds ratio for COPD as a predictor of disease severity

Figure S31

Forest plot of odds ratio for diabetes as a predictor of disease severity

Figure S32

Funnel plot of odds ratio for diabetes as a predictor of disease severity

Figure S33

Forest plot of odds ratio for obesity as a predictor of disease severity

Figure S34

Funnel plot of odds ratio for obesity as a predictor of disease severity

Figure S35

Forest plot of odds ratio for hypertension as a predictor of disease severity

Figure S36

Funnel plot of odds ratio for hypertension as a predictor of disease severity

Figure S37

Forest plot of odds ratio for cardiovascular diseases as a predictor of disease severity

Figure S38

Funnel plot of odds ratio for cardiovascular diseases as a predictor of disease severity

Figure S39

Forest plot of odds ratio for cerebrovascular accidents as a predictor of disease severity

Figure S40

Funnel plot of odds ratio for cerebrovascular accidents as a predictor of disease severity

Figure S41

Forest plot of odds ratio for CKD as a predictor of disease severity

Figure S42

Funnel plot of odds ratio for CKD as a predictor of disease severity

Figure S43

Forest plot of odds ratio for chronic liver disease as a predictor of disease severity

Figure S44

Funnel plot of odds ratio for chronic liver disease as a predictor of disease severity

Figure S45

Forest plot of odds ratio for malignancy as a predictor of disease severity

Figure S46

Funnel plot of odds ratio for chronic liver disease as a predictor of disease severity

Figure S47

Forest plot of odds ratio for immunocompromised state as a predictor of disease severity

Figure S48

Funnel plot of odds ratio for immunocompromised state as a predictor of disease severity

Figure S49

Forest plot of odds ratio of gender for disease severity

Figure S50

Funnel plot of odds ratio of gender for disease severity