Association of genetic polymorphisms with COVID-19 infection and outcomes: An updated meta-analysis based on 62 studies

Abstract

Background

The relationship between genetic polymorphisms and coronavirus disease 2019 (COVID-19) remains to be inconsistent. This meta-analysis aimed to provide an updated evaluation of the role of genetic polymorphisms in the infection, severity and mortality of COVID-19 based on all available published studies.

Methods

A systematic search was performed using six databases: PubMed, Embase, Web of Science, Cochrane Library, Chinese National Knowledge Infrastructure (CNKI) and Wanfang. Summary odds ratios (ORs) and corresponding 95 % confidence intervals (CIs) were used to calculate the genotypic comparison. All statistical analyses were conducted in Stata 12.0.

Results

A total of 62 studies with 19600 cases and 28899 controls was included in this meta-analysis. For COVID-19 infection, ACE Ins/Del polymorphism might be related with significantly decreased risk of COVID-19 infection under dominant, homozygote and allelic models. Meanwhile, the IFITM3 rs12252 and TMPRSS2 rs12329760 polymorphisms were significantly associated with the increased risk of COVID-19 infection under one or more models. Regarding COVID-19 severity, ACE2 rs2074192, ACE2 rs2106809, IFITM3 rs12252 and VDR rs1544410 polymorphisms might be related with significantly increased risk of COVID-19 severity in one or more models. Moreover, the analysis of TMPRSS2 rs2070788 indicated that a variant A allele decreased the risk of COVID-19 severity in recessive model. For COVID-19 mortality, the variant C allele of IFITM3 rs12252 polymorphism might be related with significantly increased risk of COVID-19 mortality under all genetic models.

Conclusions

This meta-analysis indicated that he infection, severity or mortality of COVID-19 were related to the above genetic polymorphisms, which might provide an important theoretical basis for understanding the clinical feature of COVID-19 disease.

1. Introduction

Coronavirus disease 2019 (COVID-19), known as severe acute respiratory syndrome coronavirus 2 (SARSCoV-2), is a rapidly infectious disease caused by a novel coronavirus emerging from the end of 2019 [1]. Due to the infection of COVID-19, it has spread worldwide and currently affects more than 200 countries. Although massive public health measures and vaccination initiatives are applied against COVID-19, there is currently no effective treatment for patients who have developed severe lung injury [2]. Worryingly, COVID-19 remains still a major challenge for public health worldwide, resulting in huge social and economic burdens. Hence, it is urgent to uncover the pathogenic mechanism of COVID-19.

It's worth noting that the clinical course and severity of COVID-19 patients exist distinct individual differences, ranging from asymptomatic to severe pneumonia

with multiple organ failure. Some studies indicate that the development and severity of COVID-19 patients have been linked to different clinical risk factors, including old, male and previous comorbidities [3,4]. Increasingly, researchers have focused on the role of host genetic factors in the progression and severity of COVID-19 disease including angiotension converting enzyme (ACE), interleukin 6 (IL6), tumor necrosis factor α (TNFα), interferon-induced transmembrane protein 3 (IFITM3), interferon lambda type 3 (IFNL3), transmembrane serine protease 2 (TMPRSS2), vitamin D receptor (VDR) and so on [5]. However, there is inconsistency between these findings. Furthermore, the same genetic polymorphism might have different effects on the infection, severity or mortality of COVID-19. For example, some studies showed that TMPRSS2 rs12329760 polymorphism was positive related to SARS-CoV-2 infection risk or severity of COVID-19 [6,7]. On the contrary, some articles indicated that TMPRSS2 rs12329760 polymorphism was associated with a reduced risk of COVID-19 disease [8,9].

Although several meta-analyses analyzed the relationship between ACE, TMPRSS2, and IFITM3 polymorphisms with COVID-19 risk [10,11], many new studies including different gene locus variants [12,13] and the same locus variants of ACE, TMPRSS2 and IFITM3 with larger populations [6,14,15] have been published. Therefore, we conducted an updated meta-analysis to analyze the association of genetic polymorphisms with COVID-19 infection and outcomes.

2. Materials and methods

The meta-analysis was carried out according to the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) statement [16].

2.1. Search strategy

Electronic databases including PubMed, Embase, Web of Science, Cochrane Library, Chinese National Knowledge Infrastructure (CNKI) and Wanfang were comprehensively searched with time span from inception to December 20, 2022. The following keywords and MeSH terms were used in each database: polymorphism, genetic, gene, COVID-19, Coronavirus Disease 2019, SARS-CoV-2 Infection. The detailed search results in PubMed were summarized in Supplemental Table 1. The literatures were preliminarily screened by assessing titles and abstracts, and then retrieved in full text to evaluate for eligibility.

2.2. Definition of disease severity

There were a lot of different classifications of COVID-19 severity by individual studies and these differences can influence the reproducing of the results. Hence, this meta-analysis accepted all definitions of COVID-19 severity by individual studies and reclassified all patients into two groups (non-severe and severe) based on WHO COVID‐19 disease severity classification in order to match the World Health Organization guidelines [17]. Specifically, severe disease patients require oxygen support, and 5 % have critical disease with complications such as respiratory failure, ARDS, sepsis and septic shock, thromboembolism, and/or multi-organ failure, including acute kidney injury and cardiac injury. Subgroups analysis were performed to recognize the differences.

2.3. Definition of control groups

For infection, this meta-analysis accepted all definitions of non-infection groups by individual studies including healthy volunteers, PCR negative, and pre-pandemic population controls. Subgroups analysis was performed to recognize these differences.

2.4. Inclusion and exclusion criteria

Eligible articles should meet all the following criteria: (1) studies about the relationship between genetic polymorphisms and COVID-19 directly and indirectly; (2) a case-control or cohort study; (3) literatures providing sufficient data including genotypes and sample size; (4) English or Chinese articles. Exclusive criteria were as follows: (1) reviews, meta-analyses, conference abstracts, letters or commentaries; (2) literatures without sufficient data; (3) only one or two reports of a genotype; (4) data not available.

2.5. Data extraction and quality assessment

Two authors extracted the useful information independently in accordance with a standardized extraction, and any discrepancies were solved by discussion or consulting a third author if needed. The extracted data mainly included as follows:

the first author, publication time, country, ethnicity, genotyping method, genetic polymorphisms, outcomes of COVID-19, genotype counts in the case and the control group, the Hardy-Weinberg equilibrium (HWE) of the controls. The Newcastle-Ottawa scale (NOS) was used to evaluate the quality of the eligible studies [18]. The quality assessment values ranged from 0 to 9 stars. Studies that scored ≥5 were defined as high quality, and 0–4 stars were regarded as low quality respectively.

2.6. Statistical analysis

Data analyzing and processing were performed using STATA version 12.0 (Stata Corp, College Station, TX, USA). Four gene models were evaluated, which were the dominant model (MM + MW vs. WW), recessive model (MM vs. MW + WW), homozygote model (MM vs. WW), and Allelic model (M vs. W) (M: mutation allele, W: wild allele). Genotype frequencies of the control group were analyzed by HWE using the Chi-square test. Summary odds ratios (ORs) and corresponding 95 % confidence intervals (CIs) were calculated using a random effect model. Heterogeneity among studies was tested using Cochran's Q-test and I² statistics, and values of I² ≥ 50 % or P ≤ 0.05 indicated significant heterogeneity. Subgroup analyses based on ethnicity and study quality were performed to investigate the source of heterogeneity. Sensitivity analysis was conducted to assess the effect of individual study on pooled results through deleting a single study each time. Egger's test and Begg's test were adopted to evaluate the publication bias. All tests were two-sided and P < 0.05 was supposed to have a statistically significance.

3. Results

3.1. Characteristics of included studies

Flowchart of the study selection process was shown in Fig. 1. A total of 2337 records (PubMed, 656; Embase, 189; Web of Science, 718; Cochrane Library, 22; CNKI, 311; Wanfang, 441) were preliminarily identified according to our search strategy. Of these, 1692 records remained after removing duplicates. And then, 361 records screened after reading titles and abstracts. Finally, 62 records included in this meta-analysis after a full text screen, which contained 19600 cases and 28899 controls [[6], [7], [8], [9],[12], [13], [14], [15],[19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72]]. Among them, 28 records with 8296 cases and 8822 controls were eligible on the association between COVID-19 infection and genetic polymorphisms including seven SNPs in the four genes (ACE Ins/Del, ACE1 rs1799752, ACE1 rs4646994, ACE2 rs2285666, IFITM3 rs12252, TMPRSS2 rs12329760, VDR rs2228570); 52 records with 9975 cases and 17250 controls were selected to estimate the association between COVID-19 severity and genetic polymorphisms including seventeen SNPs in the seven genes (ACE Ins/Del, ACE1 rs1799752, ACE1 rs4646994, ACE2 rs2074192, ACE2 rs2106809, ACE2 rs2285666, IFITM3 rs12252, IFNL3 rs12979860, IL6 rs1800795, IL6 rs1800797, TMPRSS2 rs12329760, TMPRSS2 rs2070788, TNFα rs1800629, VDR rs1544410, VDR rs2228570, VDR rs731236, VDR rs7975232); 7 records with 1329 cases and 2827 controls were selected to assess the association between COVID-19 mortality and genetic polymorphisms including IFITM3 rs12252 and TMPRSS2 rs12329760. The main characteristics of the 62 articles were shown in Table 1. Of these, 41 articles were performed in Caucasian populations, 13 articles in mixed populations, and 8 articles in Asian populations. For the detection of genotype method, 19 articles were using PCR-AFLP, 15 articles using sequencing, 11 articles using PCR-RFLP, 10 articles using TaqMan SNP genotyping assay, other articles using Allele-specific SNP type assays, KASPar genotyping chemistry with validated assays and so on. For the quality of included studies, 51 articles were regarded as high quality, and 11 articles were considered as low quality. Genotype frequencies of control groups in all studies were calculated by the HWE test (Supplemental Table 2). In addition, control groups of included studies for COVID-19 infection in the meta-analysis were summarized in Supplemental Table 3.

Fig. 1.

Flow chart of study selection.

Table 1.

Main characteristics of included studies in the meta-analysis.

Study [Reference]	Country	Ethnicity	Genotype Method	COVID-19 outcomes and SNP	NOS score	Quality
Abbaszadeh 2022 [19]	Iran	Caucasian	PCR-AFLP	Infection: ACE1 rs1799752; severity: ACE1 rs1799752	5	High
Abdelsattar 2022 [6]	Egypt	Mixed	TaqMan SNP genotyping assay	Infection and severity: ACE2 rs2285666, TMPRSS2 rs12329760	5	High
Abdollahzadeh 2021 [20]	Iran	Caucasian	PCR-RFLP	Severity: VDR rs7975232, rs1544410, rs2228570, rs731236	4	Low
Agwa 2021 [21]	Egypt	Mixed	TaqMan SNP genotyping assay	Severity: IFNL3 rs12979860	6	High
Ahmadi 2022 [15]	Iran	Caucasian	Sequencing	Severity and mortality: IFITM3 rs12252	6	High
Akbari 2022 [14]	Iran	Caucasian	Sequencing	Infection and severity: ACE1 rs1799752	5	High
Aladag 2021 [22]	Turkey	Caucasian	PCR-AFLP	Infection and severity: ACE Ins/Del	5	High
Aladawy 2022 [23]	Egypt	Mixed	TaqMan SNP genotyping assay	Severity: IL6 rs1800795	6	High
Al-Anouti 2021 [24]	UAE	Mixed	Infinium global screening array	Severity: VDR rs7975232, rs1544410, rs2228570, rs731236	5	High
Alghamdi 2021 [25]	Saudi Arabia	Asian	Fluorescence prove assay	Severity and mortality: IFITM3 rs12252	5	High
Ali 2022 [12]	Iraq	Caucasian	PCR-AFLP	Severity: TNFα rs1800629	5	High
Alimoradi 2022 [26]	Iran	Caucasian	PCR-RFLP	Infection and severity: ACE2 rs2285666	5	High
Amodio 2020 [27]	Italy	Caucasian	TaqMan SNP genotyping assay	Severity: IFNL3 rs12979860	8	High
Andrade 2022 [35]	Brazil	Caucasian	TaqMan SNP genotyping assay	Severity: TMPRSS2 rs2070788, rs12329760; Mortality: TMPRSS2 rs12329760	8	High
Angulo-Aguado 2022 [28]	Colombia	Mixed	Sequencing	Infection: ACE1 rs4646994, ACE2 rs2285666	8	High
Annunziata 2021 [29]	Italy	Caucasian	PCR-AFLP	Infection: ACE Ins/Del	5	High
Apaydin 2021 [30]	Turkey	Caucasian	PCR-RFLP	Severity: VDR rs7975232, rs1544410, rs2228570, rs731236	5	High
Balzanelli 2022 [31]	Italy	Caucasian	PCR-AFLP	Infection: ACE1 rs1799752, VDR rs2228570	5	High
Baştuğ 2021 [32]	Turkey	Caucasian	PCR-AFLP	Severity: ACE1 rs1799752	7	High
Cafiero 2021 [34]	Italy	Caucasian	Commercial kits	Severity: ACE1 rs1799752; ACE2 rs2074192, rs2106809	4	Low
Çelik 2021 [44]	Turkey	Caucasian	PCR-RFLP	Severity: ACE Ins/Del; ACE2 rs2106809, rs2285666	5	High
Falahi 2022 [36]	Iran	Caucasian	PCR-RFLP	Severity: IL6 rs1800795, rs1800797	5	High
Faridzadeh 2022 [37]	Iran	Caucasian	PCR-RFLP	Infection and severity: ACE1 rs1799752	6	High
Gómez 2020 [39]	Spain	Caucasian	PCR-AFLP	Infection and severity: ACE1 rs4646994, ACE2 rs2285666	5	High
Gómez 2021 [38]	Spain	Caucasian	Sequencing	Infection and severity: IFITM3 rs12252	5	High
Gong 2022 [40]	China	Asian	PCR-AFLP	Infection and severity: ACE Ins/Del	4	Low
Gunal 2021 [41]	Turkey	Caucasian	PCR-AFLP	Severity: ACE Ins/Del	6	High
Hubacek 2021 [42]	Czech Republic	Caucasian	PCR-AFLP	Infection and severity: ACE1 rs4646994	5	High
Jafarpoor 2022 [13]	Iran	Caucasian	PCR-AFLP	Infection: VDR rs2228570	4	Low
Jevnikar 2022 [43]	Slovenia	Caucasian	KASPar genotyping chemistry with validated assays	Infection: ACE1 rs4646994, ACE2 rs2285666	5	High
Kerget 2021 [45]	Turkey	Caucasian	Allele-specific SNP type assays	Severity: IL6 rs1800795, rs1800797	4	Low
Khalilzadeh 2022 [46]	Iran	Caucasian	PCR-RFLP	Severity: ACE2 rs2285666	5	High
Kotur 2021 [47]	Serbia	Caucasian	TaqMan SNP genotyping assay	Severity: VDR rs2228570	5	High
Lapić 2022 [48]	Croatia	Caucasian	Commercial multilocus genotyping assays	Severity: ACE Ins/Del	5	High
Mahdi 2022 [49]	Iraq	Caucasian	Sequencing	Severity: TMPRSS2 rs2070788, rs12329760	5	High
Martínez-Gómez 2022 [50]	Mexico	Mixed	TaqMan SNP genotyping assay	Severity: ACE Ins/Del; ACE2 rs2074192, rs2285666	4	Low
Mir 2021 [51]	Saudi Arabia	Caucasian	PCR-AFLP	Infection and severity: ACE1 rs4646994	5	High
Möhlendick 2021 [52]	Germany	Caucasian	Sequencing	Infection and severity: ACE1 rs1799752, ACE2 rs2285666	4	Low
Molina 2022 [62]	Spain	Caucasian	Sequencing	Severity: ACE Ins/Del; ACE2 rs2074192, rs2106809, rs2285666	5	High
Najafi 2022 [53]	Iran	Caucasian	Sequencing	Severity: ACE1 rs4646994, ACE2 rs2285666	4	Low
Nhung 2022 [54]	Vietnam	Asian	Sequencing	Infection and severity: TMPRSS2 rs12329760	7	High
Pan 2021 [55]	China	Asian	Sequencing	Infection and severity: IFITM3 rs12252	5	High
Papadopoulou 2021 [56]	Greece	Caucasian	PCR-AFLP	Infection: ACE1 rs1799752	5	High
Peralta 2021 [57]	Cuba	Mixed	PCR-RFLP	Severity: VDR rs731236	5	High
Posadas-Sánchez 2022 [9]	Mexico	Mixed	TaqMan SNP genotyping assay	Infection: TMPRSS2 rs12329760	5	High
Rahimi 2021 [58]	Iran	Caucasian	PCR-RFLP	Severity: IFNL3 rs12979860	6	High
Reviono 2022 [59]	Indonesia	Asian	PCR-AFLP	Severity: TNFα rs1800629	7	High
Rokni 2022 [60]	Iran	Caucasian	PCR-AFLP	Infection, severity and mortality: TMPRSS2 rs12329760	8	High
Saad 2021 [61]	Lebanon	Caucasian	PCR-AFLP	Infection and severity: ACE1 rs1799752	5	High
Saleh 2020 [63]	Egypt	Mixed	TaqMan SNP genotyping assay	Severity: TNFα rs1800629	4	Low
Schönfelder 2021 (1) [64]	Germany	Caucasian	Sequencing	Infection and severity: IFITM3 rs12252	7	High
Schönfelder 2021 (2) [65]	Germany	Caucasian	PCR-RFLP	Infection and severity: TMPRSS2 rs12329760; Severity: TMPRSS2 rs2070788	8	High
Sekiya 2022 [8]	Japan	Asian	Sequencing	Infection and severity: TMPRSS2 rs12329760	5	High
Shirazi 2022 [33]	Iran	Caucasian	Sequencing	Severity and mortality: TMPRSS2 rs12329760	6	High
Sienko 2022 [66]	Poland	Caucasian	PCR-AFLP	Severity: ACE2 rs2074192, rs2285666	4	Low
Sotomayor-Lugo 2022 [67]	Cuba	Mixed	PCR-AFLP	Severity: TNFα rs1800629	5	High
Verma 2021 [68]	India	Mixed	PCR-AFLP	Severity: ACE1 rs4646994	7	High
Verma 2022 [69]	India	Mixed	PCR-RFLP	Severity: IL6 rs1800795, rs1800797	4	Low
Vitello 2022 [7]	Italy	Caucasian	Sequencing	Infection: TMPRSS2 rs12329760	5	High
Wulandari 2021 [70]	Indonesia	Asian	TaqMan SNP genotyping assay	Severity and mortality: TMPRSS2 rs12329760	6	High
Zeidan 2022 [71]	Egypt	Mixed	Allelic discrimination RT-PCR	Infection and severity: VDR rs2228570	5	High
Zhang 2020 [72]	China	Asian	Sequencing	Severity and mortality: IFITM3 rs12252	6	High

Annotation: ACE: angiotensin-converting enzyme type; AFLP: amplified fragment length polymorphism; COVID-19: coronavirus disease 2019; IFITM3: interferon-induced transmembrane protein 3; IFNL3: interferon lambda type 3; IL6: interleukins-6; NOS: Newcastle-Ottawa scale; RFLP: restriction fragment length polymorphism; RT-PCR: real-time reverse transcription polymerase chain reaction; SNP: single nucleotide polymorphism; TMPRSS2: transmembrane serine protease type 2; TNFα: tumor necrosis factor-α; UAE, United Arab Emirates; VDR: vitamin D receptor.

3.2. Meta-analysis results

The meta-analysis results of the association between genetic polymorphisms with the infection, severity and mortality of COVID-19 were presented in Table 2. Subgroup analysis was performed according to ethnicity and study quality if the included studies were greater than or equal to five (Supplemental Table 4).

Table 2.

Meta-analysis of the association of genetic polymorphisms with COVID-19 infection and outcomes.

Study group	Study (n)	Dominant model			Recessive model			Homozygote model			Allelic model
		OR (95%CI)	P	I2 (%)	OR (95%CI)	P	I2 (%)	OR (95%CI)	P	I2 (%)	OR (95%CI)	P	I2 (%)
Infection vs. Non-infection
ACE Ins/Del	3	0.37 (0.24–0.59)	<0.001	50.2	0.28 (0.06–1.36)	0.114	91.1	0.20 (0.08–0.49)	0.001	68.5	0.42 (0.25–0.69)	0.001	79.7
ACE1 rs1799752	7	0.93 (0.67–1.29)	0.674	64.9	0.95 (0.61–1.47)	0.821	65.5	0.92 (0.65–1.29)	0.626	33.4	0.95 (0.83–1.10)	0.516	18.7
ACE1 rs4646994	5	1.01 (0.77–1.33)	0.917	48.7	0.80 (0.50–1.29)	0.365	73.7	0.83 (0.47–1.46)	0.519	75.9	0.93 (0.71–1.21)	0.589	74.6
ACE2 rs2285666	6	0.80 (0.42–1.54)	0.504	87.6	0.78 (0.26–2.32)	0.652	89.1	0.72 (0.23–2.23)	0.570	89.2	0.82 (0.40–1.65)	0.573	93.8
IFITM3 rs12252	3	1.45 (0.97–2.16)	0.070	28.9	1.87 (0.76–4.61)	0.173	11.2	2.07 (0.65–6.55)	0.216	20.0	1.49 (1.02–2.18)	0.040	46.0
TMPRSS2 rs12329760	7	1.18 (0.93–1.49)	0.179	50.8	1.45 (1.13–1.87)	0.004	0	1.46 (1.01–2.12)	0.042	31.7	1.19 (0.98–1.44)	0.076	55.8
VDR rs2228570	3	1.01 (0.22–4.71)	0.985	93.8	0.90 (0.63–1.28)	0.560	0	0.65 (0.23–1.79)	0.401	68.9	1.16 (0.57–2.37)	0.678	88.8
Severe vs. Non-severe
ACE Ins/Del	7	0.83 (0.62–1.10)	0.195	18.9	0.85 (0.66–1.09)	0.195	0	0.80 (0.59–1.08)	0.140	0	0.87 (0.74–1.02)	0.083	0
ACE1 rs1799752	7	0.76 (0.43–1.36)	0.359	77.2	0.70 (0.40–1.24)	0.220	62.9	0.59 (0.26–1.34)	0.206	76.4	0.78 (0.51–1.18)	0.237	82.0
ACE1 rs4646994	5	1.27 (0.55–2.94)	0.576	87.7	1.01 (0.50–2.04)	0.979	76.8	1.16 (0.41–3.32)	0.776	84.8	1.14 (0.66–1.96)	0.638	88.8
ACE2 rs2074192	4	1.61 (0.69–3.75)	0.267	87.6	3.38 (1.47–7.78)	0.004	82.3	3.91 (1.19–12.79)	0.024	88.7	2.06 (1.05–4.04)	0.035	90.9
ACE2 rs2106809	3	1.10 (0.55–2.20)	0.793	49.9	1.92 (1.09–3.41)	0.025	0	1.82 (1.02–3.25)	0.041	0	1.30 (0.73–2.31)	0.381	57.0
ACE2 rs2285666	10	0.82 (0.49–1.38)	0.447	87.0	0.87 (0.42–1.80)	0.700	85.9	0.83 (0.37–1.85)	0.651	87.2	0.84 (0.50–1.42)	0.518	92.8
IFITM3 rs12252	6	1.46 (0.67–3.17)	0.342	91.0	3.06 (1.35–6.94)	0.008	59.9	2.28 (0.71–7.39)	0.168	74.8	1.71 (0.81–3.61)	0.159	93.5
IFNL3 rs12979860	3	3.77 (0.29–48.22)	0.307	98.2	4.14 (0.34–49.77)	0.263	95.3	7.71 (0.20–300.04)	0.274	97.5	2.69 (0.41–17.67)	0.304	98.5
IL6 rs1800795	4	0.76 (0.30–1.92)	0.558	81.2	0.53 (0.10–2.86)	0.462	34.2	0.51 (0.08–3.06)	0.460	37.9	0.88 (0.46–1.65)	0.682	75.5
IL6 rs1800797	3	0.88 (0.58–1.35)	0.565	0.0	1.13 (0.72–1.77)	0.604	0	0.83 (0.37–1.88)	0.654	0	0.99 (0.77–1.29)	0.957	0
TMPRSS2 rs12329760	10	1.17 (0.47–2.91)	0.744	96.7	0.97 (0.43–2.03)	0.940	87.4	1.14 (0.39–3.31)	0.808	92.8	1.05 (0.52–2.13)	0.883	96.9
TMPRSS2 rs2070788	3	0.54 (0.18–1.59)	0.264	77.1	0.60 (0.39–0.93)	0.023	0	0.52 (0.22–1.32)	0.138	48.1	0.68 (0.42–1.09)	0.109	59.0
TNFα rs1800629	4	4.74 (0.25–88.59)	0.298	95.9	2.20 (0.20–24.14)	0.520	81.1	8.07 (0.02–3395.26)	0.498	93.5	2.05 (0.35–11.92)	0.423	96.8
VDR rs1544410	3	1.37 (0.89–2.11)	0.149	57.1	1.26 (0.87–1.82)	0.226	0	1.56 (1.03–2.36)	0.038	0	1.18 (0.99–1.41)	0.066	0
VDR rs2228570	5	1.02 (0.78–1.34)	0.871	0	1.03 (0.69–1.53)	0.885	22.1	1.11 (0.65–1.91)	0.707	23.7	1.00 (0.80–1.26)	0.976	27.8
VDR rs731236	4	0.96 (0.76–1.21)	0.734	0	1.28 (0.86–1.91)	0.220	0	1.23 (0.82–1.86)	0.320	0	1.03 (0.86–1.23)	0.777	0
VDR rs7975232	3	1.03 (0.77–1.39)	0.832	0	1.10 (0.77–1.58)	0.601	28.4	1.02 (0.69–1.52)	0.902	0	1.05 (0.88–1.25)	0.569	0
Deceased vs. Survived
IFITM3 rs12252	3	2.75 (1.26–6.00)	0.011	77.2	6.46 (4.60–9.08)	$<$0.001	0	7.86 (4.36–14.19)	$<$0.001	5.1	2.84 (1.17–6.88)	0.021	86.3
TMPRSS2 rs12329760	4	0.54 (0.12–2.36)	0.415	94.9	0.44 (0.12–1.55)	0.201	83.3	0.43 (0.07–2.76)	0.373	90.0	0.53 (0.16–1.75)	0.296	96.0

Annotation: ACE: angiotensin-converting enzyme type; COVID-19: coronavirus disease 2019; IFITM3: interferon-induced transmembrane protein 3; IFNL3: interferon lambda type 3; IL6: interleukins-6; TMPRSS2: transmembrane serine protease type 2; TNFα: tumor necrosis factor-α; VDR: vitamin D receptor.

3.2.1. Genetic polymorphisms with COVID-19 infection

Our meta-analysis showed that the variant I allele of ACE Ins/Del polymorphism might be related with significantly decreased risk of COVID-19 infection under dominant (II + ID vs. DD, OR = 0.37, 95 % CI: 0.24–0.59, P < 0.001, Fig. 2A), homozygote (II vs. DD, OR = 0.20, 95 % CI: 0.08–0.49, P = 0.001, Fig. 2B) and allelic models (I vs. D, OR = 0.42, 95 % CI: 0.25–0.69, P = 0.001, Fig. 2C). The IFITM3 rs12252 polymorphism was significantly associated with the risk of COVID-19 infection under allelic model (C vs. T, OR = 1.49, 95 % CI: 1.02–2.18, P = 0.040, Fig. 2D). And our results revealed a significant association between TMPRSS2 rs12329760 polymorphism and an increased risk of COVID-19 infection in the recessive (AA vs. AG + GG, OR = 1.45, 95 % CI: 1.13–1.87, P = 0.004, Fig. 2E) and homozygote models (AA vs. GG, OR = 1.46, 95 % CI: 1.01–2.12, P = 0.042, Fig. 2F). However, there were no significant differences between COVID-19 infection and other genetic polymorphisms including ACE1 rs1799752, ACE1 rs4646994, ACE2 rs2285666 and VDR rs2228570 under four genetic models (all P > 0.05). Meanwhile, subgroup analysis by ethnicity revealed that ACE2 rs2285666 polymorphism was associated with a decreased risk of COVID-19 infection in Caucasian population under recessive (TT vs. TC + CC, OR = 0.56, 95 % CI: 0.40–0.79, P = 0.001), homozygote (TT vs. CC, OR = 0.52, 95 % CI: 0.30–0.89, P = 0.018) and allelic models (T vs. C, OR = 0.59, 95 % CI: 0.36–0.97, P = 0.036). Subgroup analysis of TMPRSS2 rs12329760 polymorphism and COVID-19 infection by ethnicity also showed statistically significant results in Caucasian population under four genetic models (dominant model: AA + AG vs. GG, OR = 1.54, 95 % CI: 1.16–2.06, P = 0.003; recessive model: AA vs. AG + GG, OR = 1.80, 95 % CI: 1.27–2.56, P = 0.001; homozygote model: AA vs. GG, OR = 2.20, 95 % CI: 1.47–3.31, P < 0.001; allelic model: A vs. G, OR = 1.46, 95 % CI: 1.21–1.75, P < 0.001). In addition, subgroup analysis of ACE1 rs4646994 polymorphism and COVID-19 infection by control groups indicated statistically significant results in healthy volunteers (recessive model: AA vs. AG + GG, OR = 0.59, 95 % CI: 0.41–0.84, P = 0.003; homozygote model: AA vs. GG, OR = 0.57, 95 % CI: 0.39–0.84, P = 0.004; allelic model: A vs. G, OR = 0.79, 95 % CI: 0.64–0.97, P = 0.027) and other groups (dominant model: AA + AG vs. GG, OR = 1.32, 95 % CI: 1.04–1.67, P = 0.021; recessive model: AA vs. AG + GG, OR = 1.30, 95 % CI: 1.03–1.64, P = 0.026; homozygote model: AA vs. GG, OR = 1.49, 95 % CI: 1.12–1.99, P = 0.006; allelic model: A vs. G, OR = 1.22, 95 % CI: 1.06–1.41, P = 0.006).

Fig. 2.

Forest plot for the association between COVID-19 infection and significant genetic polymorphisms including ACE Ins/Del polymorphism in dominant (A), homozygote (B) and allelic (C) models; IFITM3 rs12252 polymorphism in allelic model (D); TMPRSS2 rs12329760 polymorphism in recessive (E) and homozygote models (F).

3.2.2. Genetic polymorphisms with COVID-19 severity

This meta-analysis showed the variant A allele of ACE2 rs2074192 polymorphism might be related with significantly increased risk of COVID-19 severity in recessive (AA vs. AG + GG, OR = 3.38, 95 % CI: 1.47–7.78, P = 0.004, Supplemental Fig. 1A), homozygote (AA vs. GG, OR = 3.91, 95 % CI: 1.19–12.79, P = 0.024, Supplemental Fig. 1B) and allelic models (A vs. G, OR = 2.06, 95 % CI: 1.05–4.04, P = 0.035, Supplemental Fig. 1C). The ACE2 rs2106809 polymorphism increased the risk of COVID-19 severity in recessive (CC vs. CT + TT, OR = 1.92, 95 % CI: 1.09–3.41, P = 0.025, Supplemental Fig. 1D), homozygote models (CC vs. TT, OR = 1.82, 95 % CI: 1.02–3.25, P = 0.041, Supplemental Fig. 1E). And the IFITM3 rs12252 polymorphism was significantly associated with the risk of COVID-19 severity in recessive model (CC vs. CT + TT, OR = 3.06, 95 % CI: 1.35–6.94, P = 0.008, Supplemental Fig. 1F). The analysis of TMPRSS2 rs2070788 indicated that a variant A allele decreased the risk of COVID-19 severity in recessive model (AA vs. AG + GG, OR = 0.60, 95 % CI: 0.39–0.93, P = 0.023, Supplemental Fig. 1G). The VDR rs1544410 polymorphism increased the risk of COVID-19 severity in homozygote model (TT vs. CC, OR = 1.56, 95 % CI: 1.03–2.36, P = 0.038, Supplemental Fig. 1H). However, there were no obvious differences between COVID-19 severity and other genetic polymorphisms including ACE Ins/Del, ACE1 rs1799752, ACE1 rs4646994, ACE2 rs2285666, IFNL3 rs12979860, IL6 rs1800795, IL6 rs1800797, TMPRSS2 rs12329760, TNFα rs1800629, VDR rs2228570, VDR rs731236 and VDR rs7975232 under four genetic models (all P > 0.05). Similarly, subgroup analysis by ethnicity revealed that IFITM3 rs12252 polymorphism was associated with an increased risk of COVID-19 severity in Caucasian population under recessive (CC vs. CT + TT, OR = 4.85, 95 % CI: 1.44–16.34, P = 0.011), and homozygote models (CC vs. TT, OR = 5.37, 95 % CI: 1.29–22.34, P = 0.021).

3.2.3. Genetic polymorphisms with COVID-19 mortality

Our results indicated that the variant C allele of IFITM3 rs12252 polymorphism might be related with significantly increased risk of COVID-19 mortality under all genetic models (dominant model: CC + CT vs. TT, OR = 2.75, 95 % CI: 1.26–6.00, P = 0.011, Figure Supplemental Fig. 2A; recessive mode: CC vs. CT + TT, OR = 6.46, 95 % CI: 4.60–9.08, P = 0.001, Supplemental Fig. 2B; homozygote model: CC vs. TT, OR = 7.86, 95 % CI: 4.36–14.19, P = 0.001, Supplemental Fig. 2C; allelic model: C vs. T, OR = 2.84, 95 % CI: 1.17–6.88, P = 0.021, Supplemental Fig. 2D). But we did not find any significant association between TMPRSS2 rs12329760 polymorphism and COVID-19 mortality under four genetic models (all P > 0.05).

3.3. Heterogeneity and sensitivity analyses

To detect the sources of obvious heterogeneity for four groups of significant genetic polymorphisms in this meta-analysis, Galbraith graph was performed. In detail, for the association of COVID-19 infection and ACE Ins/Del polymorphism under allelic model, the article studied by Gong et al. might be the main cause of heterogeneity (Fig. 3A). After removing this study, the heterogeneity of ACE Ins/Del polymorphism in allelic model decreased from 79.7 % to 0 %. As shown in Fig. 3B, the articles studied by Molina et al. and Martínez-Gómez et al. might be the main causes of heterogeneity for the relationship between COVID-19 severity and ACE2 rs2074192 polymorphism under allelic model. For the association of COVID-19 severity and IFITM3 rs12252 polymorphism under recessive model, the article studied by Ahmadi et al. might be the main cause of heterogeneity (Fig. 3C). After removing this study, the heterogeneity of IFITM3 rs12252 polymorphism in recessive model decreased from 59.9 % to 23.1 %. Meanwhile, the article studied by Ahmadi et al. might also be the main cause of heterogeneity for the relationship between COVID-19 mortality and IFITM3 rs12252 polymorphism under allelic model (Fig. 3D). Among the above four groups of significant genetic polymorphisms which were conducted to assess the sources of heterogeneity, sensitivity analyses were also performed to evaluate the stability of the results via sequential elimination of each study. As shown in Fig. 4A-D, there were no obvious changes of combined OR before and after the removal of each study in the above four groups, suggesting that our findings were stable and robust.

Fig. 3.

Heterogeneity analysis for this meta-analysis including COVID-19 infection and ACE Ins/Del polymorphism in allelic model (A), COVID-19 severity and ACE2 rs2074192 polymorphism under allelic model (B), COVID-19 severity and IFITM3 rs12252 polymorphism under recessive model (C), COVID-19 mortality and IFITM3 rs12252 polymorphism under allelic model (D).

Fig. 4.

Sensitivity analysis for this meta-analysis including COVID-19 infection and ACE Ins/Del polymorphism in allelic model (A), COVID-19 severity and ACE2 rs2074192 polymorphism under allelic model (B), COVID-19 severity and IFITM3 rs12252 polymorphism under recessive model (C), COVID-19 mortality and IFITM3 rs12252 polymorphism under allelic model (D).

3.4. Publication bias

We performed the Begg's test and Egger's test to evaluate the publication bias across the included studies. As shown in Table 3, we observed no publication bias for the infection and mortality of COVID-19 with the significant genetic polymorphisms in this meta-analysis under all comparison models (all P > 0.05). For COVID-19 severity, there is no obvious evidence of publication bias by Begg's test under all four genetic models (all P > 0.05). The results of Egger's test for ACE2 rs2106809 under allelic model (t = −97.80, P = 0.007) and IFITM3 rs12252 under homozygote model (t = −3.21, P = 0.033) in the population of COVID-19 severity were represented by P-value <0.05 for the Egger's test, suggesting that further larger and higher-quality studies were still needed to demonstrate the finding. Publication bias of all genetic polymorphisms in this meta-analysis were available in Supplemental Table 5.

Table 3.

Publication bias of significant genetic polymorphisms in this meta-analysis.

Group and genetic model	Begg's test		Egger's test
	z value	P value	t value	P value
Infection vs. Non-infection
ACE Ins/Del
Dominant model	0	1.000	−0.59	0.662
Recessive model	0	1.000	−0.87	0.545
Homozygote model	0	1.000	−0.83	0.560
Allelic model	0	1.000	−1.37	0.401
IFITM3 rs12252
Dominant model	0	1.000	−1.36	0.403
Recessive model	0	1.000	4.65	0.135
Homozygote model	0	1.000	4.62	0.136
Allelic model	0	1.000	0.89	0.535
TMPRSS2 rs12329760
Dominant model	1.20	0.230	1.20	0.285
Recessive model	0.30	0.764	−0.24	0.816
Homozygote model	0.60	0.548	−0.06	0.953
Allelic model	0.90	0.368	0.75	0.486
Severe vs. Non-severe
ACE2 rs2074192
Dominant model	1.02	0.308	1.30	0.323
Recessive model	0.34	0.734	1.92	0.195
Homozygote model	1.02	0.308	1.74	0.224
Allelic model	1.02	0.308	1.25	0.337
ACE2 rs2106809
Dominant model	1.04	0.296	−4.37	0.143
Recessive model	0	1.000	−1.07	0.480
Homozygote model	1.04	0.296	−1.86	0.313
Allelic model	1.04	0.296	−97.80	0.007
IFITM3 rs12252
Dominant model	0	1.000	−1.26	0.275
Recessive model	0.75	0.452	−2.69	0.055
Homozygote model	0	1.000	−3.21	0.033
Allelic model	0.75	0.452	−1.46	0.218
TMPRSS2 rs2070788
Dominant model	0	1.000	−1.18	0.447
Recessive model	0	1.000	−0.47	0.719
Homozygote model	0	1.000	−0.90	0.532
Allelic model	0	1.000	−1.16	0.452
VDR rs1544410
Dominant model	1.04	0.296	3.91	0.159
Recessive model	0	1.000	0.18	0.884
Homozygote model	0	1.000	2.46	0.245
Allelic model	0	1.000	1.74	0.332
Deceased vs. Survived
IFITM3 rs12252
Dominant model	0	1.000	−0.78	0.579
Recessive model	1.04	0.296	−2.16	0.276
Homozygote model	0	1.000	−3.25	0.190
Allelic model	0	1.000	−0.66	0.630

Annotation: ACE: angiotensin-converting enzyme type; IFITM3: interferon-induced transmembrane protein 3; TMPRSS2: transmembrane serine protease type 2; VDR: vitamin D receptor.

4. Discussion

There is mounting evidence that genetic variants may be related with the risk of specific infections, which can predict unfavorable disease outcomes, and provide more effective therapeutic interventions [73]. In recent years, a growing body of researches have revealed that COVID-19 disease is closely related to genetic variants in renin-angiotensin-aldosterone system (ACE1, ACE2) [26], cytokines (IL6, TNFα) [45], type I interferon-related genes (IFITM3, IFNL3) [65], the ABO blood group system and the human leukocyte antigen (HLA) [74]. However, the associations between genetic polymorphisms with infection and outcomes of COVID-19 remain partially controversial. With the publication of new high-quality and well-designed studies, an updated meta-analysis is required to confirm the roles of genetic polymorphisms in COVID-19 disease. Although several meta-analyses analyzed the relationship between specific gene polymorphisms with a certain risk of COVID-19 [11,75], the present data aimed to provide a more comprehensive estimate on three aspects of researches including the relationship between all genetic polymorphisms and the infection, severity and mortality of COVID-19, rather than one of them.

In this meta-analysis, a total of 62 studies with 19600 cases and 28899 controls was included. Firstly, 28 records with 8296 cases and 8822 controls were included to assess the association between COVID-19 infection and genetic polymorphisms. The analysis of ACE Ins/Del polymorphism might significantly decrease the risk of COVID-19 infection under dominant, homozygote and allelic models. Meanwhile, the IFITM3 rs12252 and TMPRSS2 rs12329760 polymorphisms significantly increased the risk of COVID-19 infection under one or more models. Secondly, the analysis of ACE2 rs2074192, ACE2 rs2106809 and VDR rs1544410 polymorphisms might be related with significantly increased risk of COVID-19 severity in one or more models. Moreover, IFITMS rs12252 and TMPRSS2 rs2070788 polymorphisms showed significant associations with COVID-19 severity in recessive model, but not in other models, which might be related to heterogeneity of the included literatures. Finally, the variant C allele of IFITM3 rs12252 polymorphism might be associated with obvious increased risk of COVID-19 mortality under all genetic models. In addition, most of the included studies were detected to be of good quality with an acceptable risk of bias.

The above genes (ACE, TMPRSS2, IFITM3, VDR) that we analyzed play important roles in the development of COVID-19. SARS-CoV-2 belonging to coronavirus is highly selective binding the ACE2-expressing cells, which is widely distributed in blood vessels, tissues and organs including lungs, heart, kidney and eye, contributing to the widespreading of SARS-CoV-2 in humans [76]. There is evidence that TMPRSS2 interacts with the spike protein of SARS-CoV-2 to allow the fusion with host-membrane and endocytosis [77]. IFITM3 encoded the interferon-induced transmembrane proteins play a critical role in the antiviral defense in the adaptive and innate immune response. The variants of IFITM3 gene are known risk factor for severe viral diseases including SARS-CoV [78]. Moreover, genetic variations in the VDR gene can influence the activity, stability and expression levels of VDR products. And the deficiency and insufficiency of Vitamin D (VD) further lead to many pathogenic outcomes, including respiratory infections [79].

Our findings had some limitations which should be considered. Some genetic polymorphisms just reported in one or two articles are not included in this meta-analysis. Another limitation is the occurrence of heterogeneity between some studies, which may be related to the ethnicity and study quality. Meanwhile, there are very few results in the absence of publication bias. In addition, it has been reported that certain comorbidities are known risk factors for COVID-19 severity and mortality [80]. Our study did not address the possible impact of comorbidities on the relationship between genetic polymorphisms and COVID-19 severity and mortality, due to the small amounts of studies. Given these limitations, our results should be further assessed by the larger, better-designed studies.

In conclusion, this study provides an overview for the role of genetic polymorphisms in COVID-19 disease. Polymorphisms in ACE Ins/Del, ACE2 rs2074192, ACE2 rs2106809, IFITM3 rs12252, TMPRSS2 rs2070788, TMPRSS2 rs12329760, and VDR rs1544410 influence the infection, severity or mortality of COVID-19. Our findings may provide more extensive and indepth strategies to reveal the pathogenesis of COVID-19.

Data availability statement

Data will be made available on request.

CRediT authorship contribution statement

Hongyue Ren: Writing – original draft, Software, Funding acquisition, Formal analysis, Data curation. Yanyan Lin: Methodology, Data curation. Lifeng Huang: Methodology, Formal analysis. Wenxin Xu: Software, Methodology. Deqing Luo: Writing – review & editing, Methodology, Funding acquisition, Formal analysis. Chunbin Zhang: Writing – review & editing, Methodology, Funding acquisition.

Declaration of competing interest

The author (s) declare that they have no competing interests.

Appendix A. Supplementary data

The following are the Supplementary data to this article.

figs1.

figs2.