Rationale for Testing Anticoagulants Against COVID-19


We have seen and heard about the classic symptoms of COVID-19 at UCSF Medical Center, where I work as a cardiologist. Patients keep coming in with pulmonary distress, pneumonia, and ultimately, Acute Respiratory Distress Syndrome (ARDS) – the life-and-death situation that requires ventilators.

However, I’m beginning to learn about other symptoms that some doctors are noticing. There are numerous reports of other complications, especially in advanced disease.

One of the most interesting involves disruption of the blood’s coagulation system. New anecdotal reports have described clotting in test tubes and lines, derangements of clinical clotting assays, pulmonary embolilarge clots in the heart, as well as microvascular thrombosis.

Elevation in D-Dimer, (a biomarker of coagulation system activation) has been associated with dramatically increased risk of death from COVID-19. This has led some to speculate that empiric treatment with anticoagulants might improve outcomes in these critically ill patients. Indeed, there was this recent publication of a retrospective analysis of anticoagulation with heparin or low molecular weight heparin showing an association with improved outcomes in COVID-19 patients in China.

These are early observations, but they are coming from enough peers from different institutions that it makes me wonder if there could be a pattern. Clots, after all, are something we know how to treat, especially in an acute care setting.  

Now let’s back up a little and think about what might be going on in the biology.

Hemostasis is defined as the arrest of bleeding by the physiological properties of vasoconstriction and coagulation (clotting). Thrombosis is defined as the formation of a thrombus: an aggregation of blood factors, primarily platelets and fibrin with entrapment of cellular elements. A thrombus can be big trouble. They frequently cause vascular obstruction at the point of its formation, which can lead to heart attacks, strokes, and death.

There is a long history of association between systemic inflammatory states and infections and hemostasis and thrombosis. We’ve learned to look for certain well-established biomarkers like C-reactive protein, as early warning signals of systemic inflammation that can raise the risk of serious cardiovascular events. And there is a theory that the coagulation system itself evolved as an important component of the innate immune response. The system is highly regulated in both a positive and negative manner and must be tuned to respond rapidly to injury so as to prevent blood loss. But it must also be specific. Excessive clotting, when the patient isn’t suffering from an acute bleeding episode,  can lead to rapid destruction of tissues and organs and death. This is accomplished via complex interactions between pro-coagulant molecules and anti-coagulant inhibitors as well as a finely-tuned system of molecules that dissolve clots (fibrinolysis). These all interface with cellular elements such platelets, endothelial cells and inflammatory cells.  This delicate balance has been exploited by predators such as snakes. Their venom attacks important regulators of this highly regulated system to cause bleeding or clotting in their prey.

The relationship between inflammation and the coagulation system has been well-studied by Chuck Esmon at the University of Oklahoma Health Sciences Center and many others. Early work focused on the impact of inflammatory cytokines on increasing levels of pro-coagulant factors such as tissue factor and reducing anti-coagulant inhibitors such as activated protein C. Protein C levels are strongly impacted by overwhelming infectious conditions such as sepsis. Changes in protein C levels or function have been associated with adverse outcomes in patients with sepsis.

This led Esmon and colleagues to speculate that treatment with recombinant activated protein C might improve outcomes in sepsis. Early work in animals eventually led to the development of a clinical drug, Eli Lilly’s drotrecogin alfa (Xigris), which was eventually tested in a large randomized controlled trial, the PROWESS Trial, which showed reduced mortality but increased bleeding in patients with sepsis. Subsequent studies confirmed these results. The drug, lacking a compelling risk-benefit ratio, was pulled off the market in 2011.

Still, the learnings from those studies and others can be informative today. There is a well-explored link between the contact (intrinsic) pathway and inflammation. The contact pathway is activated by exposure to negatively charged molecules or polyphasphates and activation can lead to activation of the coagulation cascade via cleavage of Factor XI by Factor XIIa. Factor XIIa can also activate inflammatory pathways via release of bradykinin, a potent inflammatory mediator that causes blood vessels to dilate.  While most of the work to date has focused on the effect of bacteria in activating the contact pathway, there is evidence that hantavirus can also activate the pathway and increase levels of bradykinin as well. Preclinical studies have supported the development of factor XIIa inhibitors for the treatment of sepsis.

It may be worth asking whether the new coronavirus might be acting along similar pathways as the hantavirus, triggering the release of inflammatory mediators that affect the coagulation cascade.

There have been many other attempts to modulate the coagulation system in severely ill patients with inflammation or systemic infections. The most notable of these was the APEX trial which compared extended use betrixaban (a novel oral factor Xa inhibitor marketed by Portola Pharmaceuticals as Bevyxxa) versus generic enoxaparin (a low molecular weight heparin anticoagulant) in medically ill patients. The results of this study, published in 2016, showed a 24% reduction in the primary efficacy outcome of asymptomatic proximal deep-vein thrombosis and symptomatic venous thromboembolism in subjects treated with betrixaban versus enoxaparin in the whole study.  There was a statistical trend toward benefit in a subset of patients with elevated levels of D-dimer. A recent analysis of the trial using D-dimers measured in a central lab showed significant improvement in patients with elevated D-dimer with a 31% risk reduction in the primary endpoint in those with D-Dimer ≥ 2 × upper limit of normal. 

None of these clinical studies has explored the potential role of anticoagulants in viral infections. To date, there haven’t been large, randomized studies asking questions of this nature.

However, there is abundant evidence that severe viral infections also impact the coagulation system and can lead to increased morbidity and mortality. This was first described during the 2003 SARS outbreak (now called SARS-CoV-1). Comprehensive pathology studies demonstrated thrombus formation and fibrin deposition in small vessels and organs (so-called microvascular thrombosis). These studies suggest activation of the coagulation system, or perhaps diminished levels or activity of inhibitors. This was followed by a study showing the SARS-CoV-1 infection led to increases in levels of plasminogen activator inhibitor 1 (PAI1) in patients with SARS-CoV-1 compared to normal controls and patients with non-SARS pneumonias. PAI-1 is a serine protease inhibitor that inhibits the activity of multiple proteins in the hemostasis and thrombosis system, most notably tissue-type plasminogen activator (tPA). The traditional recombinant clot-buster, tPA activates the fibrinolytic system. Inhibiting tPA leads to less fibrin breakdown and increased clot burden. There was also an observed increase in the levels of vitronectin, which has been shown to be a co-factor for PAI-1 in inhibiting activated protein C (APC). While protein C levels or APC activity were not measured, this result suggested that SARS-CoV-1 infection could lead to a decrease in APC function and a functional pro-thrombotic state.

This brings us back to the clinical observations in patients with COVID-19. It is early and we have a lot to learn, but the clinical anecdotes suggest that SARS-CoV-2 may also impact the coagulation system as does SARS-CoV-1. And we certainly know that in advanced disease, there is marked elevation in inflammatory cytokines. What is most interesting is that there is now good evidence that the coagulation system is activated in severe medical illness including non-infectious illnesses. There is further evidence coagulation is activated in severe infections.  Moreover, treatment with anticoagulant medications has improved outcomes in both severe medical illness and systemic infections. And there is good evidence that both SARS-CoV-1 and SARS-CoV-2 activate the coagulation cascade especially in the critically ill, leading to significant morbidity and mortality.

Given the nature of the illness and the prevalence of severe infections, now seems like the ideal time to perform rigorous clinical studies to explore the potential benefits/risks of drugs that modulate coagulation system activity in COVID-19. Questions as to which patients (perhaps those with modest or severe increases in D-Dimer) and which targets and which drugs remain to be answered. Such studies could enroll severely ill patients quickly at multiple centers around the world, and could conceivably yield statistically valid results on safety and efficacy in a matter of weeks. Based on what we know already about the underlying biology and the clinical need, there is clearly a very strong rationale to begin these studies as soon as possible.

Update (4/14/20):

Since publishing this just a week ago, there have been numerous developments in this rapidly moving field. I will not recount them exhaustively but will pick some highlights. First, I had previously missed this review of changes in coagulation markers in patients with COVID-19. Again we see dramatic changes especially with D-Dimer. Several recent studies put the prevalence of thrombotic complications in severe COVID-19 infections at 25 to 31% again with changes in coagulation markers (most notably D-Dimer). There was an autopsy series from New Orleans published in pre-print form showing prominent micro- and macro-vascular thrombosis with an interesting finding of the presence of megakaryocytes in the lungs. There was a small case report of 3 patients with severe COVID-19 infections and antiphospholipid antibodies and clinical thrombosis (all intra-cerebral).

Lastly, the International Society of Thrombosis and Hemostasis (ISTH) published guidance suggesting anticoagulating all hospitalized patients with COVID-19. And of course there continue to be anecdotes of doctors experimenting with many therapies for sick COVID patients with thrombosis including this story of the use of tPA. I will try to continue to update regularly.

Ethan Weiss, MD is an Associate Professor of Medicine at UCSF, is a Preventive Cardiologist, and is a Core Investigator at the Cardiovascular Research Institute.

This article originally appeared on the Timmerman Report here.