Background

In December 2019, the Coronavirus disease 2019 (COVID-19) outbreak led to a pandemic [1]. As of November 1, 2022, over 627 million confirmed cases have resulted in more than 6.5 million fatalities worldwide [2]. A significant number of venous thromboembolism (VTE) events have been observed in COVID-19 patients, likely due to endothelium damage, immobility, weakness, and prolonged inflammation [3]. VTE, defined as presenting pulmonary embolism (PE) or deep vein thrombosis (DVT), is a common medical concern associated with potentially fatal complications [4].

Due to different study designs, the prevalence of VTE in hospitalized COVID-19 patients is variable. A previous meta-analysis review, including nearly 2000 COVID-19 patients, reported that the weighted mean prevalence of VTE among Intensive care unit (ICU) and non-ICU patients was 31.3% [5]. In other studies, the VTE pooled prevalence was 17%, with a fourfold higher VTE rate in ICU patients [3, 6]. Due to the reduction in mortality rate and high incidence of VTE in COVID-19 patients, current guidelines recommend using in-hospital thromboprophylaxis for all hospitalized patients, especially critically ill patients [7]. However, even after the disease’s acute phase, patients can still experience VTE after hospital dismissal. In the recent systematic reviews, post-discharge VTE pooled prevalence was reported to be around 1.16–1.8%, suggesting a higher rate than other medically ill patients [8, 9]. 80% of VTE cases occur 30–45 days after hospital discharge [10]. Hence, the appropriate early thromboprophylaxis for COVID-19 discharged patients is essential.

The question to be discussed is the necessity, duration, and selection of the ideal anticoagulant (AC) in post-discharge COVID-19 patients. Several reviews and studies provided evidence regarding the possible benefits of post-discharge AC therapy; for instance, The MICHELLE randomized controlled trial (RCT) studied the necessity and duration of extended thromboprophylaxis using oral ACs [11]. However, as the American Society of Hematology guideline states, studies with a high level of evidence have spoken little about this issue, and the need for systematic review studies to summarize data and provide high-level evidence is required [12]. Furthermore, there are other ongoing RCTs underway, including Post-hospital Thromboprophylaxis RCT (NCT04650087), Hero-19 (NCT04542408), and XACT (NCT04640181), from which no findings have yet been published.

Eventually, still there remains the possibility of COVID-19 pandemic recurrence in the recent future, the spread of new variants, and even similar pandemics [13]. As a result, the question regarding post-discharge thromboprophylaxis in COVID-19 patients remains highly relevant. This practical systematic review seeks to provide a recommendation for physicians based on guidelines, reviews, RCTs, and other current evidence-based data, regarding extended thromboprophylaxis in hospitalized COVID-19 patients without VTE diagnosis at discharge time.

Main text

Protocol and registration

This systematic review was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and is registered on PROSPERO (Registration Number: CRD42022365107) [14].

Eligibility criteria

We included peer-reviewed observational studies, RCT studies, and reviews, especially guidelines and position papers reporting the necessity, type, and duration of VTE thromboprophylaxis in post-discharge COVID-19 patients. We excluded conference papers, conference abstracts, erratums, retracted papers, correspondence papers, book titles, and meta-analyses. We also excluded studies carried out on animal or cellular models. Studies had to be available in English.

Search strategy

We searched PubMed, EMBASE, Web of Science, Scopus, Cochrane, and clinicaltrials.gov from December 1, 2019, to October 6, 2022. We also screened all the review's reference lists by hand-searching. To find relevant literature for the systematic search, we used the search query provided in Additional file 1: Appendix A.1.

Study selection

We initially screened titles and abstracts of studies for duplication and relevance. The full text of all potentially relevant studies was then independently studied by two authors (R. A. and B. D.) to determine the final study selection. Resolution of disagreement was resolved by consensus and the third author's final decision (M. KA.).

Data extraction

The following data were extracted by two authors (R. A. and B. D.) from eligible articles: Study characteristics (study titles, authors, year of publication, publication study type, and study site country), population characteristics (number of patients, gender, and age), percentage of patients in the ICU setting, post-discharge thromboprophylaxis name, dosage, and recommended duration of the used AC, risk assessment tool, post-discharge events (thromboembolic events and major bleedings), and duration of follow-ups.

Risk of bias assessment

Two authors (R. A. and B. D.) assessed the risk of bias and quality of individually selected studies using the Newcastle–Ottawa Quality Assessment Scale (NOS) for cohort studies [15], adapted NOS for cross-sectional studies [16], and the Jadad scale [17] for the RCT studies (Additional file 1: Appendix A.2). NOS and adopted NOS assess the risk of bias within domains, including the study groups' selection, comparability, and the ascertainment of the outcome of interest. The quality of studies was graded using the star system with a maximum possible score of 9 for NOS and 10 for adopted NOS. The Agency for Healthcare Research and Quality (AHRQ) standard was used to convert the NOS (good, fair, and poor) [18]. Thresholds for converting the Adopted NOS (very good, good, satisfactory, and unsatisfactory) were based on a study by Herzog et al. [16]. The Jadad scale evaluated the randomization, blinding, and description of withdrawals with a maximum score of 5. Based on a study by Falagas et al. [19], an RCT with a score of 2 and above was considered a good quality study.

Data analysis

We used a qualitative analysis and presented the findings with a descriptive approach, odds ratio (OR) with a 95% confidence interval (CI), and risk ratio (RR) with a 95% CI based on the included studies and the summative nature of this systematic review. A meta-analysis and statistical calculations were not performed because the studies' design and reporting differed.

Results

Search results

Our initial search in the selected databases yielded 4897 titles, including 37 studies that met the eligibility criteria. The detailed search process is depicted in Fig. 1. 18 out of 37 studies, including eight retrospective cohorts, seven prospective cohorts, two cross-sectional studies, and only one RCT (Table 1). Studies were conducted worldwide, including nine from The United States, two from Russia, and the other six from Norway, Brazil, Spain, Belgium, Singapore, Iran, and England. 19 out of 37 studies were guidelines and reviews, including 14 guidelines, four position papers, and one state-of-the-art review (Table 2). Six Guidelines and reviews were International; the others were from The United States, England, Brazil, Italy, Algeria, Scotland, and Germany.

Fig. 1
figure 1

PRISMA flowchart of the literature search and selection of studies that reported about post-discharge thromboprophylaxis

Full size image
Table 1 Characteristics of included studies reporting on post-discharge thromboprophylaxis
Full size table
Table 2 Characteristics of guidelines and reviews reporting on post-discharge thromboprophylaxis
Full size table

Risk of bias assessment

The systematic review included 18 studies. Four cohorts were of good quality, and the other 11 were studies with lower quality scores, mainly due to comparability issues. One of the cross-sectional studies was of good quality, and the other was of satisfactory quality. The RCT was of good quality. The majority of included studies (12/18) were of low quality, and the others (6/18) were of high quality (Additional file 1: Appendix A.2).

Characteristics of patients and included studies

The major characteristics of included studies are summarized in Table 1. This systematic review included 18 studies with a total of 52,927 patients. Thirteen studies reported the mean age ranging from 40 to 68.6 years. The follow-up period of the included studies was different, ranging from 30 to 393 days after hospital discharge (median of 49.5 days). All the studies reported the rate of thromboembolic events in their follow-up duration, with a total of 1182 VTE events ranging from 0 to 8.19% (median of 0.7%). All but three studies reported that the PE ratio is equal to or greater than DVT. Eight studies reported the rate of post-discharge bleeding ranging from 0 to 3.7%. (median of 0%). Only one study did not use ACs after hospital discharge for any patients, and the others used AC for a total of 10,088 patients (19.25% of all) [20]. Ten studies reported the rate of ICU patients ranging from 7.8 to 54%, where the highest rate of ICU patients was in the MICHELLE study with the highest ratio of post-discharge VTE events [11]. Also, the major characteristics of included guidelines and reviews are summarized in Table 2.

Post-discharge thromboprophylaxis: necessity, evaluation, and AC selection

Based on the results of the included guidelines and studies, there are controversial views on post-discharge thromboprophylaxis. All the guidelines but one [21], and most studies (11/18), were in favor of this matter, but after an individual risk assessment; indicated in post-discharge COVID-19 patients with high VTE risk, low bleeding risk, and no known contraindications (Tables 1 and 2). Li et al. reported a reduced risk post-discharge VTE in patients who received the therapeutic AC at discharge (OR: 0.18 and 95% CI 0.04–0.75); however, the association of post-discharge prophylactic AC with post-discharge VTE was insignificant [22]. As the only published good-quality RCT, the MICHELLE trial study investigated the VTE and bleeding outcomes in the rivaroxaban and control group at day 35. Post-discharge thromboprophylaxis reduced the risk of VTE events by 67% in the rivaroxaban group (95% CI 0.12–0.90), and no major bleeding occurred [11]. Nevertheless, three cohorts and one cross-sectional study implied no role for post-discharge thromboprophylaxis [20, 23,24,25]. Additionally, three cohort studies did not provide a definite opinion on this matter [26,27,28].

Studies used risk assessment models (RAMs), clinical evaluation, and laboratory data to identify COVID-19 patients with high post-discharge thrombotic risk. Almost half of the guidelines (9/19) used RAMs which all mentioned The International Medical Prevention Registry on Venous Thromboembolism (IMPROVE). IMPROVE is a RAM consisting of seven variables, including the previous episode of VTE (3 points), known thrombophilia (2 points), current paralysis or paresis of lower-limb extremity (2 points), Current cancer (2 points), ICU/CCU stay (1 point), immobilization (1 point), and age > 60 years (1 point), categorizing COVID-19 patients into low (0–1 score), moderate (2–3 score), and high VTE risk (≥ 4 scores) [29]. Two high-quality studies, including Courtney et al. and Ramacciotti et al., reported a significant association between a higher IMPROVE VTE risk score and receiving extended thromboprophylaxis [11, 30]. Accordingly, in the MICHELLE trial study, with increasing the modified IMPROVE VTE risk score from 2–3 to ≥ 4, the RR increased by 27% in a way that patients with IMPROVE VTE score ≥ 4 or 2–3 with a D-dimer > 500 ng/mL were suitable for receiving extended thromboprophylaxis [11]. The IMPROVE-DD RAM with eight variables, including the D-dimer (2 points), has a similar cut-off as IMPROVE for high-risk VTE patients, [31]. In a prospective cohort CORE-19 registry, Giannis et al. demonstrated that the IMPROVE-DD RAM score ≥ 4 was significantly associated with an increased risk of VTE, arterial thromboembolism, and mortality in post-discharge COVID-19 patients (OR: 3.64 with 95% CI 2.91–4.55) [32]. Padua Prediction Score (PPS) (4/37) and the Caprini model (2/37) were used less in the included studies. Moreover, most included studies (6/18) and guidelines (10/19) used clinical evaluation as an important factor for assessing the VTE risk. 6/18 studies and 5/19 guidelines mentioned lab data, especially D-dimer (Tables 1 and 2).

Direct oral anticoagulants (DOACs) and low molecular weight heparins (LMWHs) have been used more than other AC classes in the reviewed studies, with 9/18 included studies and 8/19 guidelines suggesting LMWH; 8/18 included studies, and 9/19 guidelines suggesting DOACs. Unfractionated heparin (UFH) and vitamin K antagonist both with 3/18 included studies but none of the guidelines mentioned any of these two AC classes. Cohort studies have reported a post-discharge thromboprophylaxis of 7–28 days for LMWHs and 30–35 days for DOACs, while in guideline studies, the range is between 14 and 45 days for both AC classes (Tables 1 and 2).

Discussion

COVID-19 disease seems to be associated with a higher risk of VTE incidence, especially in more severe cases [8]. This practical systematic review aimed to determine the need to receive thromboprophylaxis and whether post-discharge thromboprophylaxis improves outcomes, including decreasing VTE events accompanying low bleeding risks. Then, identifying high-risk VTE patients and post-discharge thromboprophylaxis management, including the type of drug, dosage, and medication duration, will be discussed. Figure 2 provides a pragmatic approach for managing post-discharge thromboprophylaxis in COVID-19 patients without VTE diagnosis at discharge time based on the available evidence.

Fig. 2
figure 2

Suggested algorithm for post-discharge thromboprophylaxis in COVID-19 patients. COVID-19 = coronavirus disease of 2019; VTE = venous thromboembolism; IMPROVE-DD = International Medical Prevention Registry on Venous Thromboembolism and D-dimer; ICU = intensive care unit; CCU = cardiac care unit; AC = anticoagulant; DOAC = direct oral anticoagulants; P.O = per os; LMWH = low molecular weight heparin; S.C = subcutaneous

Full size image

Should COVID-19 patients receive post-discharge VTE thromboprophylaxis?

While several observational cohort studies, RCTs, and guidelines studied thromboprophylaxis during and after hospitalization, the role of post-discharge VTE thromboprophylaxis remains controversial [12, 22, 33, 34]. Most guidelines recommend against routinely continuing VTE prophylaxis after hospital discharge [34, 35]. Still, they suggest an individual risk assessment and using ACs after discharge in patients with high thrombotic risk, low bleeding risk, and no contraindications (Table 2).

Likewise, most of the included studies in this systematic review agreed with post-discharge VTE thromboprophylaxis if the patient's risk assessment indicated a high-risk situation for VTE. In between, three good-quality cohort studies reported a significant association between post-discharge VTE risk reduction and extended thromboprophylaxis [22, 30, 36]. As in the study by Li et al., this risk reduction was stated to be 82%; although, in the Courtney et al. study, the chance of bleeding increased significantly with post-discharge AC [22, 30]. The MICHELLE trial provided valuable information about post-discharge VTE thromboprophylaxis. The results showed that AC therapy in high-risk patients after discharge reduces the VTE events, and the risk of bleeding will remain unchanged [11].

Three cohort studies and one cross-sectional study suggested against using extended thromboprophylaxis due to their results that only the Eswaran et al. study was of Good quality and the others were of studies with lower quality scores [20, 23, 25, 37]. It is worth saying that these four studies had the lowest average age among the included studies. Tan et al. included patients with few comorbidities and the IMPROVE VTE score of 0 or 1 in 91.3% of all patients [20]. Stawiarski et al. evaluated patients with low D-dimer levels and moderate COVID-19 disease [23]. These three poor-quality studies had few ICU-admitted patients, which has been proven important in increasing the risk of VTE after discharge [20, 23, 37]. The Eswaran et al. study found no correlation even after adjusting for possible confounders such as age and ICU admission [25]. This matter can be attributed to the lack of accurate follow-up and AC thromboprophylaxis in high-risk patients, which may have led to a low incidence of VTE.

Recommendation

Due to the inflammatory state and the chance of post-discharge recurrence of VTE in COVID-19 patients, we suggest that the physicians decide on extended thromboprophylaxis based on individual assessment of VTE and bleeding risk.

Which COVID-19 patients should receive post-discharge thromboprophylaxis? Tools, lab data, and clinical evaluation

Predicting VTE risk, identifying hospitalized patients with COVID-19 at high VTE risk, and discriminating who may benefit from post-discharge thromboprophylaxis with a low risk of major bleeding remains a critical clinical issue [38]. Several tools and models, including the Caprini model, the IMPROVE VTE RAM, the modified IMPROVE RAM, the IMPROVE-DD RAM, the PPS, and the Wells model have been used in COVID-19 patients to assess the need for thromboprophylaxis. IMPROVE RAM was the most applied RAM among the studies to assess the VTE risk in post-discharge COVID-19 patients, and the other RAMs were less used by studies or recommended by guidelines. In a study by Goldin et al. in 9407 patients, the IMPROVE VTE RAM without D-dimer demonstrated a sensitivity of 83.9% and specificity of 29.2% [31]. MICHELLE RCT used modified IMPROVE RAM assigned to COVID-19 patients with IMPROVE score of ≥ 4 or 2–3 with an elevated D-dimer (> 2 times the upper limit of normal or as stated in MICHELE RCT with a D-dimer > 500 ng/mL) for patients with increased risk of VTE [11, 39]. For this reason, IMPROVE-DD eliminates the need for separate grouping using a D-dimer and increases validity scores to a sensitivity of 97.1% and specificity of 21.5% simultaneously [31]. Furthermore, various guidelines have also suggested the IMPROVE RAM, which is either the IMPROVE-VTE RAM with D-dimer or IMPROVE-DD itself (Table 2).

Tsaplin et al. [40] used the original Caprini score (2005 version) and eight modified versions to predict VTE frequency. Among the four modifications used to predict the risk of symptomatic VTE 6 months after discharge, all the versions demonstrated high sensitivity and specificity, especially Caprini with D-dimer and Caprini with COVID-19 risk scores with a sensitivity of 75% and a specificity of 81%. However, the original Caprini score correlates significantly with the VTE risk with the cut-off score of seven [40]. More studies are needed to evaluate the modified versions of the Caprini score. A retrospective cohort study also validated Caprini and IMPROVE RAM as a practical RAM independent of each other [39].

Not all VTE risk assessments are based on models and scores but on the patient's lab data and clinical evaluations. Lab data including D-dimers > two times upper the normal limit (threshold adjusted according to age) [11, 23, 30, 41,42,43,44], and pre-discharge C-reactive protein (CRP) level > 10mg/dl [22, 42] are important factors having significant association with increasing the risk of VTE [33]. In this regard, Li et al. reported a 3.76-fold (95% CI 1.86–7.57) and 3.02-fold (95% CI 1.45–6.29) higher risk of VTE with patient's peak D-dimer levels greater than 3μg/mL and pre-discharge CRP levels greater than 10mg/dL, respectively [22].

Clinical evaluations have long been essential, with easy access to assess the thrombosis risk. Prolonged immobilization [41, 43,44,45,46], advanced age (> 70–75 years) [43, 44, 47, 48], previous history of VTE [22, 43,44,45,46, 48, 49], active cancer [30, 41, 43,44,45,46, 48, 49], known thrombophilia [44, 45, 48, 49], and chronic heart or respiratory failure [21, 23, 47, 48] are the most important factors increasing the VTE risk that will be examined during the clinical evaluation. Some clinical risk factors are not included in IMPROVE-DD RAM. However, they are mentioned in the included studies, including obesity, use of estrogen, family history of VTE, comorbid chronic inflammatory or autoimmune condition, chronic kidney disease (CKD), recent major surgery (e.g., orthopedic procedure), and atrial fibrillation (Tables 1 and 2). Pregnancy is a controversial indication; two included studies reported pregnancy as an indication [10, 30], while the ISTH guideline [50], due to the risk of bleeding, has reported it as a contraindication, demonstrating greater consideration during the risk of bias assessment.

Recommendation

Clinical evaluation and laboratory data are practical factors in AC thromboprophylaxis. The most important clinical risk factors are prolonged immobilization, advanced age, previous history of VTE, active cancer, known thrombophilia, and chronic heart or respiratory failure. In this regard, IMPROVE-DD RAM is designed based on most of the mentioned risk factors and has shown good efficiency in assessing high-risk VTE events in COVID-19 patients without VTE diagnosis at discharge time.

Post-discharge VTE AC thromboprophylaxis in patients with COVID-19: which and how?

The choice of medications, dosing, and duration of thromboprophylaxis should be based on high-quality, evidence-based data and guideline recommendations. Recommended drug medication to prevent thrombosis can be placed in four popular classes of ACs, including LMWHs, DOACs, UFH, and vitamin K antagonists. Several studies recommended DOACs as a post-discharge thromboprophylaxis agent. Three high-quality studies, including the MICHELLE trial, recommended rivaroxaban 10mg daily for 30–35 days. Alternatively, apixaban 2.5mg BID and dabigatran 110mg BID can be used as the choices of DOACs [11, 25, 36]. Also, ISTH, the anticoagulation forum, the VAS, and the health system anticoagulation task force guidelines favored rivaroxaban 10mg daily for 30–42 days [35, 50,51,52]. The VAS guideline also recommended betrixaban 80mg daily for 40 days [52].

Several cohort and guideline studies recommended LMWHs, especially enoxaparin. In this regard, Quiros Ambel et al. provided a protocol in which patients in the absence of hemorrhagic risk and high risk of thrombosis should receive weight or albumin/creatinine ratio (ACR) adjusted LMWH (enoxaparin or bemiparine) for 4–6 weeks [27]. Patients weighted ≤ 50 kg or elderly patients with ACR < 30 ml/min should receive 2500 IU sc/day of bemiparine or 20mg sc/day of enoxaparin, patients weighted 51–80 kg should receive 40mg sc/day of enoxaparin or 3500IU sc/day of bemiparine. Finally, patients who weighed 81-100 kg and > 100 kg were suggested to receive 60mg sc/day of enoxaparin and 80mg sc/day of enoxaparin, respectively. In the same direction, Engelen et al. suggested enoxaparin 0.5 mg/kg daily for 14 days, and Giannis et al. used any dose of enoxaparin < 80 mg daily [42, 47]. In addition, the health system anticoagulation task force guideline recommends enoxaparin 40mg Qday subcutaneously for 6 weeks as an alternative over DOACs [51]. Generally, apart from Li et al., all other included studies emphasize the preference for prophylactic dosage over therapeutic dosage [22]. Regarding the selection of the recommended duration for extended prophylaxis, the included studies have suggested a shorter duration than the guidelines [33, 42, 47]. However, the majority of the guidelines have suggested 40–45 days [41, 43, 51, 52]. Finally, due to limitations, such as INR checks for warfarin and the need for injection for UFH and fondaparinux, the two classes of drugs, LMWH and DOACs, seem to be more acceptable.

Recommendation

If a COVID-19 patient needs extended thromboprophylaxis, we suggest oral AC medications such as DOACs, especially rivaroxaban 10mg daily for 30–35 days, and subcutaneous AC drugs such as the LMWH family, especially weight-adjusted enoxaparin, for 40–45 days. Depending on the specialist's evaluation and the persistence of VTE risk factors, an individual risk assessment should be repeated, and, if necessary, the length of thromboprophylaxis should be continued.

Role of lifestyle modification

The immune system and hemostasis have a close relationship, with each system protecting the host and preventing the spread of foreign diseases [53]. In patients with COVID-19, immunothrombosis has been hypothesized as a pathogenic mechanism in which endothelial dysfunction, hypercoagulability, and activation of innate immune cells contribute to the observed prothrombotic condition [54]. In addition, several environmental factors can affect a person's immune system. In order to have a healthy lifestyle and thus a better immunity system, we can refer to [E(e)SEEDi], which includes five fundamental items: "External and internal environment—Sleep—Emotion—Exercise—Diet" Interventions, also known as magic polypill [55].

Modifications such as communication with loved ones, washing hands, 7–9 h of sleep at night, control of obstructive sleep apnea, decreasing anxiety and depression, maintaining a healthy weight by exercise, anti-inflammatory/antioxidant diet, quitting smoking and reducing alcohol consumption are beneficial E(e)SEEDi for every COVID-19 patients [55].

Cardiovascular events, including VTE, are closely related to a person's lifestyle, and E(e)SEED imbalance can reduce the body's immunity and, as a result, increase the risk of cardiovascular events. In this regard, in addition to pharmacological treatment in post-discharge VTE prophylaxis, every physician should consider lifestyle modification to manage such patients thoroughly [55].

Limitation

The limitations of this study include the use of only one published RCT and other related clinical trial studies are ongoing and have not yet been published. For this reason, most of the data presented in this practical systematic review are from cohort studies and guidelines. Due to the rapid rate of newly published articles on patients with COVID-19 about post-discharge thromboprophylaxis, relevant studies may have been published since the end of our search date.

Conclusions

COVID-19 disease is associated with a hypercoagulable state that has increased VTE risk. Since COVID-19 coagulopathy persists after the acute phase of the disease, extended thromboprophylaxis remains controversial. Based on this systematic review, which included studies and guidelines, after a risk/benefit assessment, post-discharge AC therapy can be reasonable in high-risk patients. Clinical characteristics and laboratory data accompanying RAMs, particularly IMPROVE-DD, can help predict VTE risk. After distinguishing patients who need post-discharge AC therapy, DOACs for 30–35 days and LMWHs for 40–45 days can be the drug of choice. Further studies, particularly the results of the ongoing RCTs, are required to choose better the type of AC, dosage, and duration of prophylaxis. In addition, lifestyle modification is also an aspect to consider when deciding to use AC for post-discharge COVID-19 patients.