Systemic Anticoagulation is Associated With Decreased Mortality in COVID-19 Patients: A Propensity Score-Matched Cohort Study

Background: Accumulating evidence has revealed that coagulopathy and widespread thrombosis in the lung are common in patients with Coronavirus Disease 2019 (COVID-19). This raises questions about the ecacy and safety of systemic anticoagulation (AC) in COVID-19 patients. Method: This single-center, retrospective, cohort study unselectively reviewed 2272 patients with COVID-19 admitted to the Tongji Hospital between Jan 25 and Mar 23, 2020. Propensity score-matching between patients adjusted for potential covariates was carried out with the patients divided into two groups depending on whether or not they had received AC treatment (AC group, ³7 days of treatment; non-AC group, no treatment). This yielded 164 patients in each group. Result: In-hospital mortality of the AC group was signicantly lower than that of the non-AC group (14.0% vs. 28.7%, P =0.001). Treatment with AC was associated with a signicantly lower probability of in-hospital death (adjusted HR=0.273, 95% CI, 0.154 to 0.484, P<0.001). The incidence of major bleeding and thrombocytopenia in the two groups was not signicantly different. Subgroup analysis showed the following factors were associated with a signicantly lower in-hospital mortality in patients who had received AC treatment; severe cases (13.2% vs. 24.6%, P=0.018), critical cases (20.0% vs 82.4%, P=0.003), patients with a D-dimer level ≥ 0.5 μg/mL (14.8% vs. 33.8, P<0.001), and moderate (16.7% vs. 60.0%, P=0.003) or severe acute respiratory distress syndrome (ARDS) cases at admission (33.3% vs. 86.7%, P=0.004). During the hospital stay, critical cases (38.3% vs. 76.7%, P<0.001) and severe ARDS cases (36.5% vs. 76.3%, P<0.001) who received AC treatment had signicantly lower in-hospital mortality. Conclusions: AC treatment decreases the risk of in-hospital mortality, especially in critically ill patients, with no additional signicant, major bleeding events or thrombocytopenia being observed. Trials registration - ChiCTR2000039855


Introduction
Coronavirus disease 2019 (COVID-19) has developed into a pandemic disease and affected nearly every country in the world. There is no comprehensive and strong clinical evidence to support the e cacy of any drugs that speci cally target the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1].
Previous research has found that coagulopathy is very common in COVID-19 patients, and includes thrombosis and coagulation abnormalities and dysfunction such as an elevated D-dimer level and prolonged prothrombin time (PT), respectively [2]. Autopsy histopathologic analysis has identi ed widespread thrombosis and microangiopathy in small vessels and capillaries of the lung [3; 4; 5], that are different from the pathologies observed in respiratory failure caused by other diseases [3; 6; 7; 8]. Some scholars have therefore proposed anticoagulation (AC) treatment as an integral part of systemic therapy in the early stage of COVID-19 [9]. Retrospective research has suggested that AC may decrease mortality in COVID-19 patients [9; 10], although these conclusions are not completely reliable nor applicable to all COVID-19 patients due to limitations in methodology such as no prospective control or matching cohort, large heterogeneity in anticoagulant therapy, and a lack of subgroup analysis. As a consequence, the recommendations for empiric systemic AC treatment currently differ between COVID-19 management guidelines [11; 12; 13], with some recommending to use anticoagulant drugs preventively for patients with no contraindications to AC and a signi cantly increased D-dimer level, while others recommend that all hospitalized adults with COVID-19 should receive pharmacologic thromboprophylaxis with low molecular weight heparin (LMWH) rather than unfractionated heparin (UFH).
We conducted a propensity score-matched cohort study using a comprehensive database of COVID-19 patients to investigate whether AC treatment was protective and safe for COVID-19 patients. Innovative analyses were carried out in which the propensity score (PS) matching was performed by balancing demographic variates, disease severity, and major treatment between patients with or without AC treatment. A second aim of the study was to identify the patients that bene ted most from AC treatment using subgroup analysis that involved stratifying the data according to the severity of the acute respiratory distress syndrome (ARDS) [14], COVID-19 clinical classi cation [13], and D-dimer levels.

Ethics and registration
This retrospective cohort study was approved by the ethics committee of Tongji Medical College, Huazhong University of Science and Technology (No. 2020-S220). The clinical trial was registered and veri ed by the Chinese Clinical Trial Registry (ChiCTR2000039855).

Patient population and study design
This single-center retrospective cohort study was conducted in two exclusive branches of COVID-19 patients treated at Tongji Hospital, an academic hospital a liated to Tongji Medical College, Huazhong University of Science and Technology in Wuhan, China. All patients with con rmed COVID-19 admitted consecutively to these two institutions between Jan 25 to Mar 23, 2020 were enrolled retrospectively in the study. Approval was obtained from the ethics committee at our institution that the patients did not have to provide informed consent for inclusion in the study. Patients who received systemic therapeutic dose AC treatment for at least 7 days during hospitalization were assigned to the AC group, while patients who did not receive any AC treatment were assigned to the non-AC group. The medications administered and clinical outcomes were followed-up to June 4, 2020, when these two branches for exclusive COVID-19 treatment were closed. All COVID-19 patients were diagnosed according to the World Health Organization interim guidelines [15] and the Diagnosis and Treatment Protocol for COVID-19 Patients (Trial Version 8) [13]. The exclusion criteria for the study were younger than 18 years, pregnant, length of stay <24 hours, insu cient medical information, a history of severe comorbidities requiring surgical operation including, but not limited to, multiple trauma, a severe infection that required debridement, amputation or laparotomy, and patients who were classi ed again as COVID positive after RNA for SARS-CoV-2 was detected following their discharge from hospital.
To minimize bias caused by the nonrandom allocation of potential confounding covariates between the AC and non-AC groups, we adopted PS matching [16] methods. PS was calculated using a logistic regression model, adjusted for the following covariates: level of oxygen therapy, clinical classi cation, high-sensitivity C reactive protein (hs-CRP) and D-dimer levels, CURB-65 score for the severity of pneumonia [17] at hospital admission, and the highest level of oxygen therapy during hospitalization. The match ratio was set at 1 to 1 and the maximum allowable distance (caliper) at 0.1 [18].
2.3 Outcomes, de nitions and data collection AC treatment was de ned as receiving either therapeutic-dose UFH (intravenous), LMWH (subcutaneous injection) or direct-acting oral anticoagulants (DOACs) (mainly Rivaroxaban and Argatroban) for at least 7 days.
The primary outcome of this study was in-hospital mortality. The safety endpoints included bleeding events and thrombocytopenia. Bleeding events included major bleeding, gastrointestinal bleeding, hemoptysis, hematuria, and bleeding in other parts. Major bleeding was de ned according to the International Society on Thrombosis and Haemostasis statement [19] as those that resulted in death, were life-threatening, caused chronic sequelae or consume major healthcare resources. All other bleeding episodes were classi ed as minor bleeding [20]. Thrombocytopenia was de ned as a platelet count <100×10 9 /L [21].
All the characteristics and clinical information of the patients were obtained from electronic medical and nursing record systems. This data included age, gender, current smoking history, comorbidities, laboratory results at admission, CURB-65 score and qSOFA score at admission, ARDS classi cation and COVID-19 clinical classi cation at admission and during the hospital stay, antiviral therapies and other treatments during hospitalization, the level of oxygen therapy at admission, and the most intense level of oxygen therapy during hospitalization.

Statistical analysis
Quantitative variables were expressed as medians and interquartile ranges (IQR) and compared using the Mann-Whitney U test. Categorical variables were compared using the Pearson χ2 test, continuity correction, or Fisher's exact test, as appropriate. A Kaplan-Meier curve was used to analyse survival during hospitalization, with the data strati ed according to AC treatment. Univariate Cox proportional hazards regression was used to determine the risk factors for in-hospital mortality. Factors with a P value <0.05 were then included in a multivariate Cox proportional hazard regression model. SPSS version 22.0 software (IBM Corp., Armonk, New York, U.S.) was used for the statistical analyses and PS matching. The Kaplan-Meier survival plot was constructed using GraphPad Prism version 4.0 software (GraphPad Software Inc., La Jolla, CA, USA). All tests were two-tailed, with a P-value <0.05 considered statistically signi cant.

Clinical characteristics of the patients at presentation
2469 con rmed COVID-19 patients were admitted to Tongji Hospital between Jan 25 and Mar 23, 2020. Exclusion of 197 patients who did met the study entry criteria left 2272 patients to be consecutively and unselectively identi ed as candidates for PS matching. In this cohort, 78 patients who received AC treatment for < 7 days were excluded before PS matching. Finally, PS matching yielded 164 patients in the AC group and 164 in the non-AC group after adjusting for covariates between the two groups ( Figure  1). Compared to the non-AC group, patients in the AC group were older (69 [60-78] yr vs 67 [56-73] yr, P=0.017) and had more comorbidities at admission (76.2% vs 61.6%, P=0.004), higher white blood cell counts (7.05 ×10 9 /L vs 6.52 ×10 9 /L P=0.044), higher neutrophil counts (5.56×109/L vs 4.84×109/L, P=0.035), and higher D-dimer levels (2.26 μg/mL vs 1.40 μg/mL, P=0.003). At hospital admission, no signi cant difference in oxygen therapy, COVID-19 clinical classi cation, and ARDS classi cation was observed between the two groups. As shown in Table 1, a higher proportion patients in the AC group were receiving intravenous immunoglobulin (IVIG) (50.0% vs 31.1%, P<0.001), corticosteroid (64.0% vs 51.8%, P=0.025), and convalescent plasma (8.5% vs 1.8%, P=0.006) than patients in the non-AC group. The median duration of hospitalization in all the patients was 28 days (IQR, 15-41days), while the median duration of AC treatment was 15 days (IQR, 10-22 days).
Variables represented the poorest value of the rst day at admission. AC=anticoagulation; IQR=interquartile range; ARDS=acute respiratory distress syndrome; qSOFA=quick sequential organ failure assessment; hs-CRP=high sensitive C reacting protein; IVIG=intravenous immunoglobulin; ECMO= extracorporeal membrane oxygenation

Subgroup strati cation
In-hospital mortality between the AC treatment and the non-AC treatment groups was compared in individuals strati ed according to ARDS classi cation, COVID-19 clinical classi cation, and D-dimer levels at both hospital admission and during hospitalization ( Table 3). characterized by thrombocytopenia, mildly prolonged prothrombin time, and elevated serum D-dimer levels [21]. Recent research clearly indicates that coagulopathy is not only common in COVID-19 patients, but is also associated with increased mortality [9]. The potential mechanism for the development of coagulopathy in COVID-19 patients may be related to endothelial cell dysfunction [26] and hypoxiainduced thrombosis [27] following a SARS-CoV-2 infection. Because the endothelium plays an important role in regulating hemostasis, brinolysis, and vessel wall permeability, endothelial dysfunction in pulmonary microvessels may act as a trigger for immunothrombosis, resulting in coagulopathy.
Histological analysis of pulmonary vessels in COVID-19 patients shows more widespread thrombosis with microangiopathy compared to that observed in patients with in uenza [3]. Based on this preliminary evidence, AC treatment may be bene cial for COVID-19 patients by inhibiting thrombin generation and thereby reducing mortality. The International Society on Thrombosis and Hemostasis suggests that a prophylactic dose of LMWH should be considered in all patients who do not have any contraindications.
Moreover, the Chinese Diagnosis and Treatment Protocol for COVID-19 Patients (Version 8.0) also suggested using AC treatment in selected patients. However, these recommendations require additional clinical evidence to determine the association between AC treatment and outcome of COVID-19 patients, as well as clarify the indications, contradictions and optimal duration, dose, and time to use AC. We conducted this propensity score-matched cohort study using a comprehensive source of COVID-19 patients and showed that receiving a therapeutic-dose of AC for 7 days or longer was associated with a decrease in-hospital mortality in these patients, with no increase in the incidence of major bleeding events. A subgroup analysis was also carried out to identify patients who might obtain greater bene t from AC treatment. At hospital admission, patients who were diagnosed as severe or critical cases or those who had either moderate or severe ARDS or a D-dimer level ≥ 0.5µg/mL were more likely to bene t from AC therapy. During hospitalization, patients who developed into a critical case or severe ARDS were more likely to bene t from AC therapy (Table 3).
To date, several research works have investigated systemic AC therapy in COVID-19 patients. A retrospective cohort study of 2,773 COVID-19 patients from the Mount Sinai Health System [28] suggested that systemic treatment-dose AC may provide COVID-19 patients with potential survival bene ts by adjusting mechanical ventilation despite an increase in the risk of bleeding events. However, this singlecenter study did not match the AC and non-AC groups according to classi cation of disease severity. It is also important to note that the duration of hospitalization (median 5 days, IQR 3-8 days) and the course of AC treatment (median 3 days, IQR 2-7 days) were relatively short. The authors also suggested that a longer duration of AC treatment was associated with reduced mortality risk (adjusted HR of 0.86 per day). Within the current consensus on anticoagulant therapy for venous thromboembolism it is generally considered that patients with con rmed deep vein thrombosis or pulmonary embolism need LMWH treatment for at least ve days followed by dabigatran or edoxaban [29]. Two studies have reported that AC treatment for 7 days or longer is associated with decreased mortality in COVID-19 patients [9; 30]. In our study, we also de ned systemic AC treatment as receiving UFH, LMWH and/or DOACs for 7 days or longer. A retrospective study on 449 patients in Wuhan with severe COVID-19 [9] examined the relationship between heparin treatment for 7 days or longer and 28-day mortality. Although no difference was observed between the AC and the non-AC groups patients who meet the sepsis-induced coagulopathy (SIC) criteria or had a markedly elevated D-dimer level appeared to bene t from AC. Similar to the study from the Mount Sinai Health System, this study also did not match patients with or without AC treatment according to the classi cation of disease severity, which may have led to unbalanced variation between the two groups. In addition, the target population in this study was con ned to patients with severe COVID-19 disease, which may have resulted in selection bias and did not provide evidence to support the use of AC treatment in patients with mild or moderate COVID-19 disease. A multicenter, retrospective study in Italy [31] and a single-center, retrospective study from Wuhan [30] also investigated the relationship between AC therapy and in-hospital mortality in COVID-19 patients. Both studies found that heparin administration was associated with lower in-hospital mortality, especially in patients with severe COVID-19 disease. Propensity score-weighting analysis was used in both these studies to control differences at baseline between the AC and non-AC users. However, an imbalance of several variates still existed between the study groups, including the disease severity classi cation at admission and other treatment during hospitalization which are known to be associated with the prognosis of patients with COVID-19. In our study, we used propensity score matching to create comparable study groups with balanced variates that were potential confounders associated with prognosis. However, previous research did not match the patients before grouping that would have achieved a balance between variables. We conducted the propensity score matching using covariates that included the level of oxygen therapy at admission, the most intense level of oxygen therapy during hospitalization, clinical classi cation at admission, hs-CRP and D-dimer levels at admission, and CURB-65 score at admission. As a consequence, the clinical classi cation of COVID-19 severity, ARDS classi cation, and oxygen therapy at admission were balanced between patients in the AC and non-AC groups. These balanced variables therefore re ected the severity of illness in COVID-19 patients at admission. One of the advantages of the current study is the strategy to use propensity score matching before grouping, which would be expected to minimize bias caused by inconsistency in the severity of illness between groups. A propensity scorematched cohort study based on hospitalized and ambulatory COVID-19 patients in New York City examined the impact on clinical outcomes of therapeutic AC administered before COVID-19 infection [32].
The covariates used for matching in that study included age, sex, race, Charlson Comorbidity Index, and obesity. However, no variates related to the severity of illness were included in the matching model. The results of the study suggested AC treatment was not protective for morbidity and mortality of COVID-19 patents.
In our retrospective cohort, we observed that therapeutic AC treatment for 7 days or longer was associated with lower in-hospital mortality of COVID- 19 [13; 37] indicate the severity of hypoxia and are used frequently by clinicians to evaluate and triage patients and to decide major treatments (e.g., levels of oxygen therapy). In our study, we analyzed the ARDS classi cation and COVID-19 clinical severity classi cation at admission as well as during hospitalization. As a result, our ndings provided more information for clinicians to determine whether to use AC treatment, not only at admission, but also when patients have progressed to severe disease levels during hospitalization.
In the analysis of safety endpoints, we observed that although the incidence of bleeding events, including hemoptysis and hematuria, were higher in the AC group compared to the non-AC group, there was no signi cant difference in the rate of major bleeding events, gastrointestinal bleeding, bleeding in other parts, or thrombocytopenia. In brief, the above ndings suggested that the use of therapeutic AC treatment for 7 days or longer in hospitalized COVID-19 patients was associated with increased hemoptysis and hematuria but not with other bleeding events, especially major bleeding. These key observations are consistent with those reported in recent studies [28; 30].
This study had several limitations. First, it was a single-center retrospective design, and large-scale, multicenter, randomized, controlled trials are urgently needed to fully assess the e cacy of AC in patients with COVID-19. Second, as we pooled UFH, LMWH and DOACs together as AC treatments in the study, further research is required in order to determine which type of anticoagulant has the best therapeutic effect and safety pro le. In addition, our study did not include patients who had receive AC treatment for < 7 days but included patients with AC treatment for ≥7 days in the AC group. Whether AC treatment < 7 days is also associated with similar outcomes and also the impact of the duration of AC treatment on the outcomes needs further investigation.
In summary, COVID-19 patients who received therapeutic AC treatment for 7 days or longer had a signi cantly lower in-hospital mortality with no signi cant increase in major bleeding events. Moreover, patients had greater survival bene t from AC treatment if they were diagnosed at admission as severe or critical cases or had moderate or severe ARDS or a D-dimer level ≥ 0.5 µg/mL or alternatively became critical clinical cases or developed severe ARDS during hospitalization. Our study provides new insights into the role of AC treatment in hospitalized COVID-19 patients, although further high-quality randomized, control trials are urgently needed to validate these ndings.