Covid-19 and low molecular weight heparin therapy: retrospective study of 257 patients.

Objective To evaluate the role of low molecular weight heparin in COVID-19 treatment. Design Retrospective cohort study Setting Patients with COVID-19 pneumonia consecutively admitted to Castel San Giovanni COVID-Hospital from February 29, to April 7, 2020. Main outcome measure Hospital mortality and safety in patients treated with low molecular weight heparin. Results Of the 257 patients enrolled, 49 (19.1%) died during the hospitalization. Hospital mortality was signicantly lower in patients treated with therapeutic dose of low molecular weight heparin (T-LMWH) (17/126, 13.5%), compared with patients treated with prophylactic dose (P-LMWH) (32/131, 24.4%; χ²=4.98, p = 0.02). Crude and adjusted odds ratios of mortality for patients treated with T-LMWH were OR=0.483, 95% CI 0.252-0.923 and OR=0.374, 95% CI 0.177-0.792. In a stratied analysis by ventila- tion type, the only subgroup of patients who beneted from therapeutic doses of heparin were those receiving non-invasive mechanical ventilation (OR=0.099, 95% CI 0.028-0.354, p<0.001). No fatal bleed-ings were observed. Conclusion Treatment with therapeutic doses of T-LMWH is safe and seems to reduce mortality in COVID-19 patients with pneumonia, especially among those who need non-invasive mechanical ventila- tion. We look forward to prospective studies to conrm this observation and evaluate the appropriate dose of LMWH in the treatment of COVID-19 patients.


Introduction
The coronavirus SARS-CoV-2 infected thousands of people in Wuhan in December 2019 and spread rapidly worldwide. Since February 29, 2020, Castel San Giovanni Hospital in Italy was entirely dedi-cated to the treatment of coronavirus disease 2019  patients. Pneumonia is the main clin-ical feature of COVID-19, however we observed several patients dying with very high D-dimer se-rum levels, suggesting a hypercoagulability state, as recently reported [1][2][3].
The relationship between in ammation and thrombosis is well recognized and the anticoagulant and anti-in ammatory activity of heparin may exert a bene cial effect on COVID-19 disease [4][5]. No speci c pharmacological protocols with proved e cacy were, and actually are, available for treat-ment of Coronavirus disease [6]. We therefore hypothesized that treatment with Low Molecular Weight Heparin (LMWH) at a therapeutic dose could bene t COVID-19 patients and we decided, with clinical equipoise, to treat patients with therapeutic doses of LMWH taking advantage of its an-ticoagulant, anti-in ammatory, and in vitro antiviral properties [7][8][9][10].
The aim of this study is to evaluate the hospital mortality of COVID-19 patients treated with LMWH, administered at prophylactic doses (P-LMWH) and therapeutic doses (T-LMWH).

Methods
Adult patients, consecutively admitted to the Castel San Giovanni Hospital, from February 29, to April 7, 2020 with con rmed COVID-19 pneumonia were included in the study. The diagnosis was consistent with the World Health Organization interim guidelines [11] and was con rmed by RNA detection of the SARS-CoV-2 from oropharyngeal and nasopharyngeal swab sample. In all patients Computed Tomogra-phy scan diagnosis of pulmonary pneumonia was performed.
We excluded patients on hemodialysis and patients with a hospital stay ≤ 7 days to exclude either mild forms of disease with an early discharge or more advanced forms of disease with early death or early transfer to another hospital for whom heparin treatment would not be appropriate [12][13].
From 24 February 2020 to 15 March 2020, our patients received P-LMWH corresponding to 4000 IU once a day, as suggested by International guidelines for bedridden patients [14]. On 16 March 2020, dur-ing a medical staff brie ng, we decided to treat our patients with T-LMWH corresponding to 70-100/kg IU twice a day. The approval of the Healthcare Directorate of Piacenza Local Health Authority was ob-tained. Patients with a HAS BLED score ≥ 3 were excluded from T-LMWH treatment [15].

Data Collection
Data were extracted from electronic medical records and included age, gender, clinical characteristics, inhospital pharmacological treatments and type of high-ow oxygen needed. The comorbidities recorded were arterial hypertension, diabetes mellitus, cardiovascular diseases (de ned as history of myocardial infarction, ischemic stroke, and peripheral atheromasia), atrial brillation, chronic obstructive pulmonary disease, chronic kidney disease (Cockroft-Gault e-GFR< 60 ml/min/1.73m2) and the presence of active or previous cancer. Bleeding events also recorded and classi ed according to International Society of Thrombosis and Hemostasis (ISTH) de nition of bleeding [16]. Patients were followed up until discharge, death or April 20, 2020, whichever came rst.

End points
The primary endpoint of the study was hospital mortality and the secondary endpoint was safety.

Statistical analysis
The cohorts of patients treated with P-LMWH and T-LMWH were compared on continuous variables using the t-test and on categorical variables using the χ²-test. Crude odds ratios of hospital mortality T-LMWH were obtained using univariate logistic regression analysis. In order to control for the different proportion of patients receiving concomitant treatments in the two groups, adjusted odds ratios were estimated using multivariable logistic regression. All statistical analyses were performed using IBM SPSS, version 25 [16].

Results
During the enrolment period, 351 patients were admitted to Castel San Giovanni COVID Hospital. Nine-tyfour patients were excluded from the analysis for the following reasons: 30 were discharged, 43 died and 17 were transferred to other care units (7 to Intensive Care Units outside our province) within 7 days.
Three were not COVID-19 con rmed patients and 1 patient was in hemodialysis. The characteristics of the 257 patients included in the analyses are shown in Table 1.
Patients treated with P-LMWH and T-LMWH had similar demographic and clinical characteris-tics (Table  1). However, patients in the T-LMWH group received more frequently corticosteroids, hy-droxychloroquine and tocilizumab.
When analyses were strati ed by type of ventilation, patients with non-invasive mechanical ventilation (Vent-2) were those who mostly bene ted from higher doses of heparin, since in these patients the mortality was lower than in those treated with P-LMWH (OR=0.099, 95% CI 0.028-0.354, p<0.001). In the other two types of ventilation, no signi cant differences in mortality rates were observed (Vent1: OR=0.853, 95% CI 0.309-2.355, p=0.759; Vent 3: OR=0.792, 95% CI 0.220-2.852, p=0.721. Multivariable linear regression was conducted to determine whether T-LMWH was more effective than P-LMWH in reducing hospital mortality after controlling for concomitant administration of corticosteroids, hydroxychloroquine and tocilizumab. The results indicate a signi cant 62.6% reduction in the mortality risk among those treated with T-LMWH (OR=0.374, 95% CI 0.177-0.792, p=0.01) ( Table 2).

Safety of T-LMWH
Two major and 1 minor bleeding episodes were recorded in group T-LMWH patients (3/126; 2.4 %): two patients had psoas muscle hematoma that required two packed red blood cells transfusion each and one patient had gluteus muscle hematoma that recovered spontaneously. No patient needed invasive treatment.

Discussion
In this retrospective study, T-LMWH appeared to be the only treatment associated with a signi cant reduction of mortality, after controlling for the concomitant in-hospital treatment.
Very few data exist on the relationship between therapeutic doses of LMWH and the reduction of mortality rate in COVID-19 patients. In a recent retrospective study, Tang et al. treated COVID-19 patients with therapeutic doses of heparin, reporting a low mortality rate at 28 days' follow-up only in a subgroup of patients with SIC score ≥ 4 and D-dimer > 6-fold the upper limit [13]. During the rst two weeks of our involvement in this emergency setting, D-dimer was not systematically recorded in all patients and no analysis could be carried out comparing this predictive parameter apart from few patients. In our study, patients who mostly bene ted from therapeutic doses of heparin were those included in the Vent-2 group, who needed non-invasive mechanical ventilation with helmet CPAP (p<0.001).
We can speculate that in this speci c clinical setting, the hypercoagulability state, which worsens the respiratory clinical picture, can bene t from the administration of therapeutic doses of heparin.
Conversely, in the mechanical ventilated patients' group (Vent-3), a prognostic advantage from heparin at therapeutic doses was not found, probably because in Intensive Care Unit (ICU) patients the disease is too advanced to bene t from the anti-in ammatory and anticoagulant activity of heparin. Our ndings are consistent with a recent retrospective analysis of 150 COVID-19 ICU patients, in which a signi cantly high thromboembolic event rate was found also in those treated with anticoagulant therapy [18]. Probably in such an advanced state of the disease even a therapeutic dose of heparin may not improve the outcome [19].
SARS-Cov-2 infection induces diffuse endothelial in ammatory status due to direct viral infection of the endothelial cells in different organs of the human body [20]. Endothelial damage is the main determinant of microvascular dysfunction that led to vasoconstriction and subsequent organ ischemia, in ammation, tissue edema and a pro-coagulant state [21]. Hypercoagulability state has been demonstrated in other viral infections [22,23]. During the epidemic SARS-CoV-2, vascular endothelial damage, in both small and mid-sized pulmonary vessels, was observed and resulted in pulmonary infarction [24,25]. In COVID-19 patients, the radiological features observed on CT angiography images suggest venous and arterial throm-bosis [26] and also in autopsy studies, diffuse thrombosis of the peripheral small vessels has been found [27,28]. Therefore, the protective effect of T-LMWH is probably due to the prevention/treatment of such thrombotic events, which is in keeping with the bene t of anticoagulation in a particular subset of pa-tients with severe pulmonary involvement.
Apart from its anticoagulant properties, LMWH has an anti-in ammatory effect that reduces the uncontrolled activation of the cytokine cascade by inhibiting the release of IL-6 [29]. High serum IL-6 levels are observed in the advanced stages of COVID-19, signi cantly and directly correlate with the severity of the disease [30,31,32].
Lastly, no fatal bleeding occurred in the study population and only 2 out of 126 patients (1.6 %) had bleeding requiring transfusions, which indicates that this treatment is safe [33].
Limitations in current retrospective observational study include the potential selection bias and residual confounding associated with a small sample size. It is also possible that the improving learning curve may have produced a non-pharmacological bene t in the management of patients over time. However our ob-servation maybe useful to stimulate prospective studies to evaluate which heparin dose could be optimal in the treatment of Covid-19 patients.

Conclusion
In this retrospective study T-LMWH treatment reduces in-hospital mortality in COVID-19 patients and seems to be safe and well-tolerated. Future randomized clinical trials are still needed to con rm this observation.

Declarations
Funding No speci c funding has supported this study.
Ethics approval The Healthcare Directorate of Piacenza Local Health Authority approved the therapeutic protocol.
Consent to participate The Healthcare Directorate of Piacenza Local Health Authority granted a waiver of informed consent.
Consent for publication No consent for publication was requested, because data are presented in aggregate form and do not include sensitive data.
Availability of data and material Data are available upon request from the rst author. Tables   Table 1. Demographic, clinical characteristics and treatment of the study population and comparisons between patients treated with P-LMWH (N=131)and T-LMWH (N=126). P-LMWH = Low Molecular Weight Heparin at prophylactic doses 4000 IU once a day; T-LMWH = Low Molecular Weight Heparin at therapeutic doses 70-100 IU twice a day. Vent 1: nasal cannula with oxygen ow < 6 l/min or non-rebreather mask with oxygen ow from 10 to 15 l/min; Vent 2: non-invasive mechanical ventilation; Vent 3: invasive mechanical ventilation. ^Chi-square test. °t-test.