COUNTERPOINT: Should Therapeutic Heparin Be Administered to Acutely Ill Hospitalized Patients With COVID-19? No.

SARS-CoV-2 causes a systemic illness that is unique from other respiratory viruses. Chief among the differences compared with other viruses is the propensity to activate clotting cascade within afflicted patients. Specifically, initial reports identified VTE in as many as 35% of patients with COVID-19. , Lung specimens from autopsies of patients with COVID-19 demonstrate widespread thrombosis with capillary microthrombi nine times more prevalent compared with autopsy specimens from patients with severe influenza A. Anticoagulation potentially could decrease thrombotic events. Further, heparins (both unfractionated and low-molecular-weight) have some antiviral properties by binding to receptor binding domain spike protein of SARS-CoV-2. This, in turn, may prevent organ injury from SARS-CoV-2.

For critically ill patients with COVID-19, the INSPIRATION trial and a multiplatform randomized control trial (RCT) failed to demonstrate a net clinical benefit for heparin-based intermediate-dose thromboprophylaxis or therapeutic anticoagulation, respectively. , These study findings are in accordance with a previously published CHEST guideline and expert panel report that recommended standard-dose VTE prophylaxis in critically ill populations. This leads many clinicians to ask: “Should patients with COVID-19 who are hospitalized in a ward (not ICU) setting receive therapeutic anticoagulation?”

Four open-label multicenter RCTs attempt to provide an answer to this question.8, 9, 10, 11 The study designs and characteristics of these trials are described in Table 1 . Only two of these trials achieved their primary efficacy end points. The multiplatform trial (Randomized Embedded Multifactoral Adaptive Platform Trial for Community-Acquired Pneumonia [REMAP-CAP], Acute Inpatient Anti-Thrombotic Study [ACTIV-4a, and Anti-Thrombotic Therapy To Ameliorate Complications of COVID-19 [ATTACC]) investigated a novel primary end point of organ support free days (OSFD). It reached primary efficacy end point with OR of 1.27 (95% credible interval, 1.03 to 1.58) in favor of heparin-based anticoagulation. The HEP-COVID trial diverted attention to the combined primary end point of thrombosis events (arterial and/or venous) and/or death. The HEP-COVID trial demonstrated benefit from therapeutic heparin-based anticoagulation with relative risk of 0.46 (95% CI, 0.27-0.81) in patients who were not critically ill with elevated D-dimers. However, the between-group difference in combined end points was driven primarily by venous thrombotic events, not by mortality rate. In contrast to these “positive” trials, the Randomized Clinical Trial To Evaluate a Routine Full Anticoagulation Strategy in Patients with Coronavirus (COVID-19): COALIZAO ACTION trial investigated full-dose rivaroxaban vs standard-dose thromboprophylaxis and failed to meet its primary efficacy end point (time to death, duration of hospitalization, or duration of oxygen use) with a win ratio of 0.86 (95% CI, 0.59-1.22). The Asessing Point-of-Care Influenza and Other Respiratory Virus Diagnostics (RAPID) trial not only compared heparin-based full-dose anticoagulation with standard-dose thromboprophylaxis but also failed to achieve the primary composite end point of death, invasive mechanical ventilation, or admission to ICU with an OR of 0.69 (95% CI, 0.43-1.10). To summarize these results, a recently conducted meta-analysis of RCTs with 3,305 patients who were not critically ill showed that, compared with standard-dose thromboprophylaxis, therapeutic anticoagulation was associated with significantly lower rates of VTE (risk ratio, 0.53; 95% CI, 0.34-0.83), significantly higher rates of any bleeding (risk ratio, 3.92; 95% CI, 1.92-8.00), major bleeding (risk ratio, 1.86; 95% CI, 1.04-3.33), and minor bleeding (risk ratio, 5.23; 95% CI, 1.54-17.77), and no significant reduction in all-cause death (risk ratio, 0.80; 95% CI, 0.40-1.61).

Table 1

Antithrombotic Trials for Hospitalized Patients With COVID-19 Who Are Not Critically Ill

Trial	Elevated D-Dimer as Inclusion Criterion	Intervention and Duration	Primary Efficacy Outcome	Safety Outcome (ISTH)
ACTION8 (N = 615; 576 not critically ill)	Yes (above the local ULN)	Therapeutic rivaroxaban for 30 days	Win ratio for death, duration of hospitalization, or oxygen use for first 30 days	Major or clinically relevant nonmajor bleeding (ISTH criteria)
Multiplatform randomized controlled trial9 (N = 2,219 not critically ill)	No	Therapeutic LMWH for 14 days or recovery	Survival to hospital discharge and days free of organ support through first 21 days	Major bleeding (ISTH criteria)
RAPID10 (N = 465)	Yes (above the local ULN)	Therapeutic heparin until 28 days or discharge or death	Composite of death or noninvasive mechanical ventilation or ICU admission.	Major bleeding (ISTH criteria)
HEP-COVID11 (N = 253; 170 not critically ill)	Yes (> 4 times the ULN) or sepsis-induced coagulopathy ≥ 4	Therapeutic LMWH until hospital discharge	VTE or arterial thromboembolism or death in first 30 days	Major bleeding (ISTH criteria)

ISTH = International Society of Thrombosis and Haemostasis; LMWH = low-molecular-weight heparin; ULN = upper limit of normal.

There were several general limitations among these RCTs. Therapeutic anticoagulation regimens varied not only in terms of agents used (rivaroxaban, unfractionated, or low-molecular-weight) but also in total duration (2 weeks, 4 weeks, or until hospital discharge/death).8, 9, 10, 11 All the trials were open-label in design, which means that clinicians might have provided other aspects of background care differently in the control vs active treatment groups. These trial designs are also prone to confirmation bias, as clinicians might have been more inclined to order diagnostic tests in search for VTE or provide organ support for patients in the control group. The definition of critically ill vs not critically ill was not consistent among the trials.8, 9, 10, 11 Given the nature of the interventions, all RCTs had extensive exclusion criterion (60% to 98% of the screened patients were excluded), which makes it very hard to apply therapeutic benefit to a general non-ICU population without risking harm from full-dose anticoagulation. 8, 9, 10, 11 Notably, despite very high rates of thrombosis in COVID-19, none of the RCTs beyond the HEP-COVID trial included VTE as a primary end point (Table 1).

One of the largest RCTs that exclusively looked at noncritically ill patients is the multiplatform trial (ATTACC, ACTIV-4a, and REMAP-CAP). Although we recognize the efforts to conduct such a large-scale study during the pandemic, we have several reservations about the trial. Authors postulate that potential benefits of therapeutic anticoagulation in patients who are not critically ill act early during the disease, but the REMAP-CAP trial included patients up to 14 days. Approximately 20% of patients who were in the therapeutic anticoagulation arm were not actually receiving therapeutic anticoagulation. There was an exhaustive list of exclusion criterion based on bleeding risk profile (even varied among three trial platforms: ATTACC, ACTIV-4a, and REMAP-CAP) further diminishing generalizability (ie, how well the outcomes can be expected to apply to other settings) to general ward populations. Only 62% and 36% of the patients in the multiplatform trials received steroids and remdesivir. One must consider whether the antiinflammatory effects of heparins might be less beneficial when this now current standard of care is provided. Primary outcome was OSFDs, evaluated on an ordinal scale that combined in-hospital death and the number of days free of cardiovascular or respiratory organ support up to day 21 among patients who survived to hospital discharge. Because most of the patients in the two treatment groups survived until hospital discharge without receipt of critical care-level organ support, the median value for OSFDs was 22 in both groups, and the study had to report the proportion of patients in each treatment group who survived until hospital discharge without receipt of organ support. In addition, organ support is a subjective, potentially problematic outcome. Each step on the scale is not necessarily of equivalent clinical significance. For example, moving from the clinical state of “low-flow oxygen” to “high-flow supplemental oxygen” is less important than moving to “requiring mechanical ventilation.”

Similarly, the HEP-COVID trial has its own shortcomings. In the HEP-COVID trial, the absolute risk of VTE, arterial thrombosis, or death was very high (36%) in the control group. Complication rates (mainly bleeding) were much lower than in routine practice, which generates consequent doubts about external validity. Hence, the current results might not be applied to patients who are thought to be at lower risk for VTE or higher risk for bleeding.

In summary, after review of the literature: (1) Therapeutic anticoagulation significantly decreases rates of VTE during COVID, although its effects might be exaggerated because of confirmation bias in these open-labeled studies. (2) Although the trials enrolled COVID-19 at a very low risk for bleeding and were underpowered for safety, therapeutic anticoagulation significantly increases rates of major bleeding. (3) Current standards of care, including steroids/antiviral therapies, were not used consistently in some studies, which raises concerns for external validity.

As a whole, the best evidence to date suggests that therapeutic anticoagulation likely benefits some patients who are moderately ill with COVID-19, but the concerns listed earlier urge us to avoid blindly following the guidelines and instead to consider carefully the risks and benefits for each individual patient.