Risk of thromboembolic events in non-hospitalized COVID-19 patients: A systematic review

Abstract

The risk of thromboembolism in non-hospitalized COVID-19 patients remains uncertain and was assessed in this review to better weigh benefits vs. risks of prophylactic anticoagulation in this population. A search was performed through three databases: Medline, Embase, and Cochrane Library until 2022. Self-controlled case series, case-control and cohort studies were included, and findings summarized narratively. Meta-analyses for risk of thromboembolism including deep vein thrombosis (DVT), pulmonary embolism (PE), and myocardial infarction (MI) between COVID-19 and non-COVID-19 non-hospitalized patients were conducted. Frequency, incidence rate ratio (IRR), and risk ratio (RR) of stroke were used to assess risk in non-hospitalized COVID-19 patients considering the lack of studies to conduct a meta-analysis. Ten studies met inclusion criteria characterized by adult non-hospitalized COVID-19 patients. Risk of bias was relatively low. Risk of DVT (RR: 1.98 with 95% CI: 1.03–3.83) and PE (OR: 6.72 with 95% CI: 4.81–9.39 and RR: 4.44 with 95% CI: 1.98–9.99) increased in non-hospitalized COVID-19 patients compared to controls. Risk of MI (OR: 1.91 with 95% CI: 0.89–4.09) is possibly increased in non-hospitalized COVID-19 patients with moderate certainty when compared to controls. A trend in favor of stroke was documented in the first week following infection. Our meta-analyses support the increase in risk of DVT and PE, and likely increase of MI, in non-hospitalized COVID-19 patients. The risk of stroke appears significant in the first week following infection but drops to insignificance two weeks later. More studies are needed to establish evidence-based recommendations for prophylactic anticoagulation therapy in non-hospitalized COVID-19 patients.

1. Introduction

On March 2020, a devastating COVID-19 pandemic was declared in response to the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) outbreak (World Health Organization, 2020). With the occurrence of rapid mutations, COVID-19 raged on with the appearance of the Delta variant followed by the Omicron variant in 2021 and the BA.2 strain in 2022. In two years, COVID-19 has claimed more than six million lives and affected billions more, as reported by the World Health Organization (WHO).

Although a compromised respiratory system has been considered a fundamental characteristic delineating COVID-19, investigations that reveal an association with insults in the vascular system continue to accumulate. The incidence of thrombotic events ranged from 15.2% to 79% in patients with severe COVID-19 and the vascular complications correlated with disease severity and fatality (Chen et al., 2020; Cui et al., 2020; Klok et al., 2020a; Malas et al., 2020; Nahum et al., 2020; Ren et al., 2020; Voicu et al., 2020; Zhang et al., 2020). Concurrently, autopsy studies confirmed the prothrombotic state and indicated the presence of multiorgan damage (Menter et al., 2020; Wichmann et al., 2020a, 2020b). Coagulopathy in patients with COVID-19 has been demonstrated as a hallmark predisposing to development of both venous and arterial thromboembolism (Malas et al., 2020; Tang et al., 2020). Patients with severe disease manifestations had greater levels of d-dimer, fibrinogen degradation products, a longer prothrombin time, and a longer thrombin time (Wang et al., 2020). Numerous mechanisms have been postulated to cause the thrombotic phenotype. The pathogenesis is likely provoked by a dysregulated immune response following the introduction of the viral pathogen commonly emerging under the description “immunothrombosis”. Endothelial cell damage, dysregulated activation of the coagulation cascade, and cytokine storm contribute to the pro-thrombotic state (Portier et al., 2021).

A spotlight has been shed on thrombotic events particularly occurring in severe cases of hospitalized COVID-19 patients, which spurred the action for a prophylactic anticoagulation prescription. However, prescribing prophylactic anticoagulation to asymptomatic or non-hospitalized COVID-19 patients remains a controversial issue as evidence is not yet established. This systematic review aimed to study the risk of developing thromboembolic events in non-hospitalized COVID-19 patients compared to non-COVID-19 patients to better assess the benefit versus risk of prophylactic anticoagulation in this population.

2. Methods

2.1. Study Eligibility

Observational studies including self-controlled case series, case-control and cohort studies reporting the risk of thromboembolic events in non-hospitalized COVID-19 patients were included in this review. Non-hospitalized setting included outpatients and emergency department patients. Thromboembolic events including deep venous thrombosis (DVT), pulmonary embolism (PE), stroke and myocardial infarction (MI) were considered. Unpublished studies were excluded.

2.2. Participant selection criteria

Adults more than 18 years of age, non-hospitalized with and without COVID-19 infection were included. Studies including children, hospitalized patients, patients exposed to respiratory tract infections were excluded.

2.3. Types of outcome measures

Primary outcomes included DVT, PE, stroke, and MI. Studies reporting one or more of the primary outcomes were included.

2.4. Search strategy and data sources

The search involved three different databases: Medline, Embase and Cochrane. It was not restricted to any language. MeSH and keywords for the population, exposure, and outcomes were applied. The search strategy is detailed in Supplement 1 of this review. Abstracts, unpublished studies, and reference lists were also reviewed. No organizations or individuals working in the field were contacted. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines were followed (Moher et al., 2009). This study was registered at the International Prospective Register of Systematic Reviews (PROSPERO) (No. CRD42021242101).

2.5. Data extraction and management

Two independent authors reviewed the titles and abstracts of articles. All articles meeting the inclusion criteria were included, and the full article was reviewed. Eligibility was assessed independently using a standard form and based on the information provided by the article. Disagreements between reviewers were resolved by consensus. Two authors independently extracted and compared data using standard sheets. The sheets included study characteristics, participant characteristics (age, sex, and comorbidities) and the outcomes of interest. The study details are listed in Table 1a, Table 1b, Table 2 .

Table 1a.

Characteristics of included studies.

Study	Country	Journal	Study Design	Follow Up	N Followed Up	Exclusion Criteria
Aktaa et al. (PMID:33711754)	United Kingdom	Thrombosis Research	Cohort study	Feb 1, 2018–July 31, 2020	84,728 COVID +	Re-hospitalized patients during the study period
					1849 COVID -
Alquezar-Arbe et al. (PMID:35065863)	Spain	The Journal of Emergency Medicine	Retrospective case-control study	March 1- April 30, 2020	74,814 COVID +	N/A
					1,388,879 COVID -
Chevinsky et al. (PMID:33909072)	United States of America	Clinical Infectious Diseases	Matched cohort study	March–June 2020	44,489 Outpatients COVID + ≥18	Pregnancy, congenital malformations, external causes of morbidity, factors influencing contacts with health services
					44,489 Outpatients COVID - ≥18
Ho et al. (PMID:34607634)	United States of America	Mayo Clinic Proceedings	Self-controlled case series	March 1, 2018–October 5, 2020	1449 COVID +	Absence of thromboembolic events
Jimenez et al. (PMID:33904528)	Spain	European Journal of Emergency Medicine	Retrospective case-control study	March 1- April 30, 2020	74,814 COVID+	N/A
					1,388,879 COVID -
Katsoularis et al.1(PMID:35387772)	Sweden	BMJ	Self-controlled case series (SCCS) and matched cohort study	Feb 1- May 25, 2021	1,057,174 COVID +	N/A
					4,076,342 COVID -
Lund et al. (PMID:33984263)	Denmark	The Lancet	Cohort study	Feb 27- May 31, 2020	8983 COVID +	Less than 1 year of residency in Denmark, inconclusive test results, died in the 2 weeks after their test
					80,894 COVID -
Miro et al. (PMID:34164664)	Spain	European Heart Journal	Retrospective case-control study	March 1- April 30, 2020	74,814 COVID +	N/A
					1,388,879 COVID -
Piazza et al. (PMID:33121712)	United States of America	Journal of the American College of Cardiology	Cohort study	March 13- April 3, 2020	715 Outpatients COVID +	Outcomes prior to PCR diagnosis
Rass et al. (PMID:33682276)	Austria	European Journal of Neurology	Prospective study	April–September 2020	32 Outpatients COVID +	N/A

¹Data used in the meta-analysis belongs to the cohort part of the study.

Table 1b.

Characteristics of included studies.

Study	Age	Male	Female	Smoking	Hypertension	Diabetes	Previous Cardiovascular Events
Aktaa et al. (PMID:33711754)	0–49 12.8% COVID +	N/A	40,594 (47.9%) COVID + 859 (46.5%) COVID -	N/A	N/A	N/A	N/A
	10.2% COVID -
	50–59 12.7% COVID +
	11% COVID -
	60–69 17.7% COVID +
	16.6% COVID -
	70–79 25.4% COVID +
	24.9% COVID -
	80 + 31.4% COVID +
	37.2% COVID -
Alquezar-Arbe et al. (PMID:35065863)	Mean ± SD 74 (13) COVID + and ACS	N/A	33 (30.00%) COVID + and ACS	11 (10.00%) COVID + and ACS	86 (78.18%) COVID + and ACS	33 (30%) COVID + and ACS	47 (42.73%) COVID + and ACS
	63 (18) COVID + non-ACS		156 (47.27%) COVID + non-ACS	22 (6.67%) COVID + non-ACS	150 (45.45%) COVID + non-ACS	57 (17.27%) COVID + non-ACS	25 (7.58%) COVID + non-ACS
	67 (14) COVID - and ACS		98 (29.70%) COVID - and ACS	80 (24.61%) COVID - and ACS	212 (24.24%) COVID - and ACS	108 (32.73%) COVID - and ACS	92 (27.88%) COVID - and ACS
Chevinsky et al. (PMID:33909072)	18–39 35.7% COVID +	17,220 (38.7%) COVID + 17,413 (39.1%) COVID -	27,225 (61.2%) COVID + 27,048 (60.8%) COVID -	N/A	N/A	N/A	N/A
	36.6% COVID -
	40–49 18.1% COVID +
	16.4% COVID -
	50–64 25.9% COVID +
	25.9% COVID -
	65–74 10.3% COVID +
	12.1% COVID -
	75–84 6.0% COVID +
	6.8% COVID -
	≥85 4% COVID +
	2.8% COVID -
Ho et al. (PMID:34607634)	Median (IQR) 77 (65–85)	625 (49.96%)	626 (50.04%)	N/A	N/A	N/A	N/A
Jimenez et al. (PMID:33904528)	Mean ± SD 68 (59–78) COVID + with DVT	N/A	44 (60.7%) COVID + with DVT	8 (7.1%) COVID + with DVT	66 (58.9%) COVID + with DVT	19 (17%) COVID + with DVT	7 (6.3%) COVID + with DVT
	65 (52–77) COVID + without DVT		160 (47.6%) COVID + without DVT	22 (6.5%) COVID + without DVT	152 (45.2%) COVID + without DVT	60 (17.9%) COVID + without DVT	26 (7.7%) COVID + without DVT
Katsoularis et al. (PMID:35387772)	Mean 40.2 COVID +	517,434 (48.9%) COVID +	539,740 (51.1%) COVID +	N/A	N/A	N/A	N/A
	40.2 COVID -	2,000,973 (49.1%) COVID -	2,075,369 (50.9%) COVID -
Lund et al. (PMID:33984263)	Median (IQR) 43 (30–56) COVID +	3512 (39·1%) COVID +	5471 (60·9%) COVID +	246 (2.7%) COVID +	2127 (23.7%) COVID +	526 (5.9%) COVID +	837 (9.3%) COVID +
	43 (29–56) COVID -	29,263 (36·2%) COVID -	51,631 (63·8%) COVID -	4548 (5.6%) COVID -	21,808 (27.0%) COVID -	4910 (6.1%) COVID -	9120 (11.3%) COVID -
Miro et al. (PMID:34164664)	Mean ± SD 66 (55–78) COVID + with PE	215 (58.4%) COVID + with PE	N/A	24 (6.5) COVID + with PE	181 (49.2) COVID + with PE	66 (17.9) COVID + with PE	23 (6.3) COVID + with PE
	66 (52–77) COVID + without PE	199 (54.1) COVID + without PE		25 (6.8) COVID + without PE	168 (45.7) COVID + without PE	69 (18.8) COVID + without PE	29 (7.9) COVID + without PE
	71 (58–80) COVID - with PE	180 (48.9) COVID - with PE		60 (16.3) COVID - with PE	196 (53.3) COVID - with PE	62 (16.8) COVID - with PE	18 (4.9) COVID - with PE
Piazza et al. (PMID:33121712)	Mean ± SD 44.8 ± 16.2 Outpatients COVID +	N/A	433 (60.6%) Outpatients COVID +	Current 29 (4.1%)	175 (24.5%)	67 (9.4%)	22 (3.1%)
				Former 104 (14.6%)
				Never 511 (71.5%)
Rass et al. (PMID:33682276)	Median (IQR) 45 (35–55) Outpatients COVID +	N/A	22 (69%) Outpatients COVID +	Current smoker 1 (3%) Ex-smoker 8 (25%)	2 (6%) Outpatients COVID +	1 (3%) Outpatients COVID +	2 (6%) Outpatients COVID +

Table 2.

– Details of included study outcomes.

Study	DVT	DVT Diagnostic Criteria	Incidence	PE	PE Diagnostic Criteria	Incidence	Stroke	Stroke Diagnostic Criteria	Incidence	MI	MI Diagnostic Criteria	Incidence	Follow Up Duration
Aktaa et al. (PMID:33711754)	Yes	ICD-10	1943 per 100,000 COVID +	Yes	ICD-10	886 per 100,000 COVID +	No			No			4 months
			1597 per 100,000 COVID -			322 per 100,000 COVID -
Alquezar-Arbe et al. (PMID:35065863)	No			No			No			Yes	Medical records	92.7 per 100,000/year COVID + 69.8 per 100,000/year COVID -	N/A
Chevinsky et al. (PMID:33909072)	No			Yes	ICD-10 CM	N/A	No			Yes	ICD-10 CM	N/A	4 months
Ho et al. (PMID:34607634)	Yes	ICD 10	3.77 (1.36–10.46) 7 days	Yes	ICD 10	3.92 (1.83–8.40) 7 days	Yes	ICD 10	4.22 (2.50–7.12) 7 days	Yes	ICD 10	3.38 (1.50–7.65) 7 days	56 days
			2.75 (1.18–6.40) 8–28 days			3.54 (2.03–6.16) 8–28 days			0.51 (0.19–1.37) 8–28 days			1.85 (0.86–3.96) 8–28 days
			2.35 (1.01–5.46) 28–56 days			2.99 (1.74–5.15) 28–56 days			0.60 (0.27–1.36) 28–56 days			1.15 (0.47–2.81) 28–56 days
Jimenez et al. (PMID:33904528)	Yes	Complete compression Doppler ultrasound	98.38 per 100,000/year COVID +	No			No			No			N/A
			33.25 per 100,000/year COVID -
Katsoularis et al. (PMID:35387772)	Yes	ICD-9 and ICD-10	N/A	Yes	ICD-9 and ICD-10	N/A	No			No			30 days
Lund et al. (PMID:33984263)	No			No			Yes	ICD-10 I82	12/8748 (0.1) COVID +	No			2 weeks- 6 months
									117/78411 (0·2) COVID -
Miro et al. (PMID:34164664)	No			Yes	CTPA	310.1 per 100,000/year COVID +	No			No			N/A
						28.5 per 100,000/year COVID -
Piazza et al. (PMID:33121712)	No			No			Yes	Imaging/Confirmed by Neurologist	0/715	Yes	Imaging/Elevated Cardiac Biomarkers	0/715	30 days
Rass et al. (PMID:33682276)	No			No			Yes	Clinical diagnosis	0/32	No			3 months

2.6. Risk of bias assessment

Two reviewers independently assessed the risk of bias in included studies using the Newcastle-Ottawa Scale (NOS). The risk of bias was assessed as a score reaching a maximum of 9, the lower the score, the higher the bias. Domains assessed for risk of bias for cohort studies included the representativeness of the exposed cohort (non-hospitalized COVID-19 patients), selection of the non-exposed cohort (non-hospitalized non-COVID-19 patients), ascertainment of exposure, demonstration that outcome of interest was not present at start of the study, comparability of cohorts based on the design or analysis, assessment of outcome, duration, and adequacy of follow up. Whereas assessment for risk of bias for case control studies included adequate definition and representativeness of the cases, selection and definition of the controls, comparability of cases and controls based on the design and analysis, ascertainment of exposure and non-response rate. Discrepancies among results were resolved upon discussion. The assessment was recorded and provided in Table 3, Table 4 of this review.

Table 3.

The Newcastle-Ottawa Scale (NOS) risk of bias assessment for cohort studies.

Cohort Studies	Selection				Comparability	Outcome			Total score
	Representativeness of the exposed cohort	Selection of the non-exposed cohort	Ascertainment of exposure	Demonstration that outcome of interest was not present at start of study	Comparability of cohorts based on the design or analysis	Assessment of outcome	Duration of follow up	Adequacy of follow up
Aktaa et al. (PMID:33711754)	1	1	1	1	2	1	1	0	8
Chevinsky et al. (PMID:33909072)	1	1	1	1	2	1	1	0	8
Ho et al.a(PMID:34607634)	1	N/A	1	1	N/A	1	1	0	N/A
Katsoularis et al. (PMID:35387772)	1	1	1	1	2	1	1	0	8
Lund et al. (PMID:33984263)	1	1	1	1	2	1	1	1	9
Piazza et al. (PMID:33121712)	1	1	1	1	2	1	0	1	8
Rass et al. (PMID:33682276)	1	1	1	1	1	1	1	0	7

^aSelf-controlled case series, assessed using the Newcastle-Ottawa Scale for cohort studies while accounting for non-applicable domains.

Table 4.

The Newcastle-Ottawa Scale (NOS) risk of bias assessment for case control studies.

Case-Control Study	Selection				Comparability	Outcome			Total score
	Adequate case definition	Representativeness of the cases	Selection of controls	Definition of controls	Comparability of cases and controls based on the design or analysis	Ascertainment of exposure	Same method of ascertainment of cases and controls	Non-response rate
Alquezar-Arbe et al. (PMID:35065863)	1	1	0	1	2	1	1	0	7
Jimenez et al. (PMID:33904528)	1	1	0	1	2	1	1	0	7
Miro et al. (PMID:34164664)	1	1	0	1	2	1	1	0	7

2.7. Data analysis

The thromboembolic events in included studies were assessed using the incidence rate ratio (IRR), risk ratio (RR), odds ratio (OR), and/or frequency. For each outcome except for stroke, the pooled adjusted risk ratio (RR) or adjusted odds ratio (OR) with their 95% confidence interval (CI) were calculated/displayed in a forest plot. Pooled RR and OR have been calculated using inverse variance method in combination with random effects analytic model. Heterogeneity was assessed using Chi² and I² tests.

2.8. Certainty of the evidence

The certainty of evidence of the primary outcomes including DVT, PE, and MI was assessed using the GRADE approach. It was rated as high, moderate, low or very low. The certainty of evidence was downgraded if risk of bias, imprecision, inconsistency, indirectness, and publication bias were noted. The overall assessment was reported according to the GRADE criteria in Table 5 .

Table 5.

GRADE assessment for certainty of evidence.

Question: What is the risk of thromboembolic events in non-hospitalized COVID-19 patients?
Outcomes	Certainty assessment							Effect	Certainty
	№ of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Relative (95% CI)
DVT	2	observational studies	not serious	seriousa	not serious	not seriousb	none	RR 1.98 (1.03–3.83)	⨁⨁⨁◯ Moderate
PE	2	observational studies	not serious	seriousc	not serious	not serious	none	OR 6.72 (4.81–9.39)	⨁⨁⨁◯ Moderate
PE	2	observational studies	not serious	seriousd	not serious	not serious	none	RR 4.44 (1.98–9.99)	⨁⨁⨁◯ Moderate
MI	2	observational studies	not serious	seriouse	not serious	not seriousf	none	OR 1.91 (0.89–4.09)	⨁⨁⨁◯ Moderate

CI: confidence interval; OR: odds ratio; RR: risk ratio.

Explanations.

^aI2 = 97%. We did not rate down by two levels because the conclusion about an increased risk among COVID-19 patients is consistent across studies; the conclusion about the precise increase in risk remains inconsistent.

^bWe did not rate down because the confidence intervals of the included studies were judged as precise for the judgment of increased risk and the observed imprecision in the pooled effect estimate is a reflection of the inconsistency for which we have already rated down the certainty.

^cI2 = 60%.

^dI2 = 98%. We did not rate down by two levels because the conclusion about an increased risk among COVID-19 patients is consistent across studies; the conclusion about the precise increase in risk remains inconsistent.

^eI2 = 92%. We did not rate down by two levels because the conclusion about an increased risk among COVID-19 patients is consistent across studies; the conclusion about the precise increase in risk remains inconsistent.

^fWe did not rate down because the confidence intervals of the included studies were judged as precise for the judgment of increased risk and the observed imprecision in the pooled effect estimate is a reflection of the inconsistency for which we have already rated down the certainty.

3. Results

3.1. Search results

8678 potentially relevant studies were identified, 8589 were excluded by title and abstract screening. 89 articles were retrieved for full text screening of which 10 studies met the inclusion criteria (Fig. 1 ). Reason for exclusion was provided in Supplement 2 of this review.

Fig. 1.

Flowchart for study eligibility.

3.2. Included studies

3.2.1. General characteristics

Ten studies were included and published between 2020 and 2022. Of which, 6 were cohort studies, 3 were case-controls and 1 was a self-controlled case series. The risk of PE was reported in 5 studies, whereas the risk of DVT, stroke and MI were reported in 4 studies each Table 1a, Table 1b, Table 2

3.2.2. Participants

A total of 1,272,384 COVID-19 and 5,592,453 non-COVID-19 individuals participated in the included studies. They were all adults with an age range of 40–70. Both men and women were somewhat equally represented, although participating women were slightly higher. Several studies reported on multiple comorbidities including cardiovascular diseases, hypertension, and diabetes mellitus. 6 studies included cardiovascular accidents, hypertension as well as smoking status, and 5 studies included diabetes mellitus. Participant status is detailed in Table 1a, Table 1b.

3.2.3. Risk of bias

The quality of the included studies was moderate to high with a score ranging from 7 to 9. Risk of Bias assessment is detailed in Table 3, Table 4

3.3. Outcomes

3.3.1. Deep venous thrombosis

Four studies looked at the risk of DVT among COVID-19 and non-COVID-19 non-hospitalized patients (Aktaa et al., 2021; Jimenez et al., 2021; Lars Christian Lund et al., 2021). Ho et al. reported the IRR of DVT among non-hospitalized COVID-19 patients as 3.77, 95% CI 1.36–10.46 in the first week following infection and 2.75, 95% CI 1.18–6.40 in the second and third week (Ho et al., 2021). Jimenez et al. reported that the incidence of DVT among non-hospitalized COVID-19 patients is 98.38 per 100,000 compared to 33.25 per 100,000/year in non-COVID-19 with an adjusted OR: 2.8 with 95% CI: 2.62–3.08 (Jimenez et al., 2021). Aktaa et al. reported an incidence of 288 per 100,000 COVID-19 patients compared to 237 per 100,000 non-COVID-19 patients over 4 months follow up with an adjusted RR: 1.2 with 95% CI: 1.18–1.22 (Aktaa et al., 2021). Katsoularis et al. reported an adjusted RR: 2.8 with 95% CI: 2.26–3.47 over a one month follow up (Katsoularis et al., 2022). Our meta-analysis of studies reporting on pooled RR using random effect model showed an increased risk of DVT among non-hospitalized COVID-19 patients (RR: 1.98 with 95% CI: 1.03–3.83 I²: 97%) (Fig. 2 ).

Fig. 2.

Risk of DVT in COVID-19 vs non-COVID-19 patients.

3.3.2. Pulmonary embolism

Five studies looked at the risk of PE among COVID-19 and non-COVID-19 non-hospitalized patients. Ho et al. reported the IRR of PE among non-hospitalized COVID-19 patients as 3.92, 95% CI: 1.83–8.40 in the first week following infection and 3.54, 95% CI: 2.03–6.16 in the second and third week (Ho et al., 2021). Miro et al. reported the incidence of PE among non-hospitalized COVID-19 patients as 310.1 per 100,000 compared to 28.5 per 100,000/year in non-COVID-19 patients with an adjusted OR: 7.53 with 95% CI: 7.17–7.9 (Miro et al., 2021). Chevinskey et al. reported an adjusted OR: 2.8 with 95% CI: 2.62–3.08 over a one month follow up (Chevinsky et al., 2021). Aktaa et al. study reported an incidence of 886 per 100,000 COVID-19 patients compared to 322 per 100,000 non-COVID-19 patients over 4 months follow up with an adjusted RR: 2.96 with 95% CI: 2.91–3 (Aktaa et al., 2021). Katsoularis et al. reported an adjusted RR: 6.77 with 95% CI: 5.43–8.45 over a one month follow up (Katsoularis et al., 2022). The meta-analysis of the adjusted OR and RR studies using random effects model showed an increased risk of PE among non-hospitalized COVID-19 patients with OR: 6.72 with 95% CI: 4.81–9.39, I²: 60% and RR: 4.44 with 95% CI: 1.98–9.99, I²: 98% respectively (Fig. 3, Fig. 4 ).

Fig. 3.

Risk of PE in COVID-19 vs non-COVID-19 patients.

Fig. 4.

Risk of PE in COVID-19 vs non-COVID-19 patients.

3.3.3. Stroke

Four studies looked at the risk of stroke among COVID-19 and non-COVID-19 non-hospitalized patients. Ho et al. reported the IRR of stroke among non-hospitalized COVID-19 patients as 4.22, with 95% CI: 2.50–7.12 in the first week following infection and 0.51, with 95% CI: 0.19–1.37 in the second and third week of follow up (Ho et al., 2021). Lund et al. reported an adjusted RR: 1 with 95% CI:0.55–1.82 two weeks to six months following infection (Lars Christian Lund et al., 2021). Piazza et al. reported a frequency of 14/715 (2%) over 30 days follow up (Piazza et al., 2020). Rass et al. reported a frequency of 0/30 over 3 months follow up (Rass et al., 2021). No sufficient studies were available to conduct a meta-analysis for stroke findings.

3.3.4. Myocardial infarction

Four studies looked at the risk of MI among COVID-19 and non-COVID-19 non-hospitalized patients. Ho et al. reported the IRR of MI among non-hospitalized COVID-19 patients as 5.16, with 95% CI: 3.04–8.73 in the first week following infection and 1.51 with 95% CI: 0.77–2.95 in the second and third week (Ho et al., 2021). Alquezar-Albe et al. reported the incidence of MI among non-hospitalized COVID-19 patients as 92.7 per 100,000/year compared to 69.8 per 100,000/year in non-COVID-19 patients with an adjusted OR: 1.33 with 95% CI: 1.23–1.44 (al., 2021). Chevinskey et al. reported an adjusted OR: 2.9 with 95% CI: 1.9–4.6 over one month follow up (Chevinsky et al., 2021). Piazza et al. reported a frequency of MI 8/714 (1.1%) over 30 days follow up (Piazza et al., 2020). The meta-analysis of the studies reporting on adjusted OR using random effects model did not show significant higher odds of MI among non-hospitalized COVID-19 patients with OR: 1.91 with 95% CI: 0.89–4.09, and I²: 92% (Fig. 5 ).

Fig. 5.

Risk of MI in COVID-19 vs non-COVID-19 patients.

4. Discussion

A hypercoagulable state has been generally associated with hospitalized and severely ill COVID-19 patients, drawing attention to anticoagulation therapies in this population. The risk of hypercoagulability, however, remains largely unestablished in non-hospitalized COVID-19 patients. Our analysis shows that COVID-19 increases the risk of DVT in the studied population with moderate certainty. Moreover, our analysis shows that non-hospitalized COVID-19 patients are likely to have a substantial risk of PE compared to non-COVID-19 controls even at asymptomatic or mild infection levels. With moderate certainty, COVID-19 probably increases the risk of MI in non-hospitalized patients. Although the risk of MI appears to be significant as early as the first week following COVID-19 infection, meta-analysis of adjusted OR reported non-significance. Similarly to MI, the risk of stroke appears to be significant as early as the first week following COVID-19 infection, but drops to insignificance as early as two weeks following the infection.

The findings of this systematic review are in line with numerous studies reporting the high risk of thromboembolic events in hospitalized COVID-19 patients (Arachchillage and Laffan, 2020; Klok et al., 2020b). This could be possibly explained by the predominant inflammatory pro-thrombotic state of the acute infection phase that could predispose the patient to a high risk of thromboembolic events during both infection and post-recovery phases (Ho et al., 2021). The co-occurrence of hypoxia, vascular inflammation, and endothelial dysfunction are all potential contributing factors (Mahajan and Chandra, 2020; Pellicori et al., 2021). Viral alveolar endothelial cells infiltration through the ACE2 receptor triggers pro-inflammatory cytokines release including IL-1, IL-6, and IL-8. Simultaneously, endothelial damage leads to the increased expression of tissue factor, von Willebrand factor, platelet and coagulation cascade resulting in clot formation. Thrombin activation from the coagulation cascade contributes to the pro-inflammatory state through its effect on platelets and on extracellular traps (NET) formation in neutrophils. NET have been shown to cause organ damage and widespread thrombosis (Abou-Ismail et al., 2020).

4.1. Significance and perspective

Risk of COVID-19 induced thromboembolism is comparable to sepsis-induced thombosis, hence the emerging recommendation for prophylactic anticoagulation in the management of critically ill hospitalized COVID-19 patients (Hajra et al., 2020; Pellicori et al., 2021). While a solid recommendation regarding prophylactic anticoagulation in severely ill, hospitalized COVID-19 patients is present, a definitive one in the case of non-hospitalized COVID-19 patients is not yet established due to insufficient evidence regarding the risk of thromboembolism in this population.

The purpose of this systematic review was to explore the literature and estimate the size of this risk. All analyzed studies reported an increased risk of DVT, PE, stroke, and MI during the acute phase (first week) of non-hospitalized COVID-19 patients, but an increased risk of DVT and PE only after recovery. Our meta-analyses conducted on DVT, PE, and MI studies support the documented probable increase in the risk of DVT, PE, and MI in non-hospitalized COVID-19 patients. More studies are needed to reach a definitive conclusion and provide a solid recommendation regarding the use of prophylactic anticoagulation therapies in non-hospitalized COVID-19 patients.

4.2. Limitations

The findings highlighted in this review are limited by the lack of susbtantial number of studies and the high level of heterogeneity between published studies. Despite some concerns around using the case-control study design in retrospective database studies (Schuemie et al., 2019), the latter was included due to the limited number of studies tackling the non-hospitalized COVID-19 population. The marked heterogeneity could be attributed to the difference in sample size between pooled studies and different follow up period. Additional studies highlighting the risk of thromboembolic events in the context of non-hospitalized COVID-19 patients are warranted to draw strong conclusions.

Funding

This work was supported by grants from the Centre National de la Recherche Scientifique (CNRS) [grant number 103941/103944], Lebanon to FAZ, and Agence Nationale de la Recherche (ANR) [grant - ANICOV-HF] France to FAZ and MM.

Declaration of competing interest

The authors have no competing interests to declare.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.