One of the most controversial issues regarding management of COVID-19 is the appropriate dosing of anticoagulation.  It’s universally agreed upon that patients with COVID-19 should receive some dose of anticoagulation (at least prophylaxis), barring contraindications.  However, the optimal dose is unclear.  Numerous reports indicate that patients with COVID-19 have high rates of venous thromboembolic disease, even despite prophylactic anticoagulation.^​1–3​  Likewise, autopsy series commonly feature micro- and macro-thrombotic events.^​4​

Randomized controlled trials will be needed to answer this question definitively.  However, with ~25,000 patients infected daily with COVID-19 in the United States alone, this is a pressing clinical question which won’t await the publication of multicenter RCTs.  The following new papers might help guide our management for now:   

Paranjpe I et al.  Association of treatment dose anticoagulation with in-hospital survival among hospitalized patients with COVID-19

This is a retrospective analysis of 2,733 patients with confirmed COVID-19 admitted at five New York City hospitals within the Mount Sinai system between 3/14 and 4/11.^​5​  In-hospital mortality was related to the use of therapeutic anticoagulation (including oral, subcutaneous, or intravenous forms).  The study may be broken into three parts: 

Analysis #1:  Survival among all admitted patients

Among all 2,733 patients, 786 (28%) received systemic anticoagulation during their hospital course.  Patients treated with anticoagulation were generally sicker:

• Anticoagulated patients were more likely to require invasive mechanical ventilation (29.8% vs. 8.1%, p< 0.001).
• Anticoagulated patients had significantly increased baseline prothrombin time, activated partial thromboplastin time, lactate dehydrogenase, ferritin, C-reactive protein, and D-dimer values – all predictors of poor outcome.

The Kaplan-Meier survival curve is shown below.  Despite being sicker, patients treated with anticoagulation tended to survive slightly longer. 

Analysis #2:  Survival among mechanically ventilated patients

Mechanically ventilated patients represent a more homogeneous population of patients, all of whom are critically ill.  Among 395 intubated patients, anticoagulation didn’t appear to correlate with laboratory markers of worse prognosis. 

The Kaplan-Meier survival curves for mechanically ventilated patients is shown below.  Patients who received anticoagulation tended to survive longer: 

Using a multivariate Cox proportional hazards model, longer duration of anticoagulation treatment was associated with a reduced risk of mortality (adjusted hazard ratio of 0.86 per day, 95% confidence interval 0.82-0.89, p<0.001). 

Analysis #3:  Risk of systemic anticoagulation

Hospital records were evaluated for major bleeding events, which were defined as one of the following:

• Hemoglobin <7 g/dL and any red blood cell transfusion.
• Or
• At least two units of red blood cell transfusion within 48 hours.
• Or
• A diagnostic code for major bleeding (including intracranial hemorrhage, hematemesis, melena, peptic ulcer with hemorrhage, colon, rectal, or anal hemorrhage, hematuria, ocular hemorrhage, and acute hemorrhagic gastritis). 

Bleeding event rates were as follows:

• Patients not treated with anticoagulation:  38/1987 = 1.9% of patients
• Patients treated with full anticoagulation:  24/786 = 3% of patients

This difference is actually not statistically significant (p=0.2). 

Bleeding rates were considerably higher among patients who were intubated (30/395 = 7.5%) compared to patients who weren’t intubated (32/2378 = 1.4%).  Since anticoagulation correlated with greater overall illness severity, it’s possible that the higher rate of bleeding observed in anticoagulated patients could partially reflect that these patients were more often intubated and critically ill. 

A limitation of this analysis is that it doesn’t differentiate between less important bleeding events versus life-threatening bleeding events (e.g., intracranial hemorrhage).  Most of these bleeding events were likely easily manageable, but some may have caused morbidity or mortality. 

Overall, this data suggests that the excess risk of bleeding attributable to anticoagulation in patients with COVID-19 is probably low (e.g. ~1%). 

Weaknesses of this study:  Evaluation of efficacy

Press surrounding this study has generally focused on correlations with mortality.  However, this may be the less compelling aspect of the study. 

As a retrospective study, it cannot prove causality.  Numerous sources of bias exist.  Perhaps most notable is immortal time bias – patients who live longer may survive long enough to be diagnosed with DVT/PE and be treated with anticoagulation (whereas other patients may die rapidly, before being able to be prescribed anticoagulation).  The median time from admission to anticoagulation initiation was 2 days, with an inter-quartile range of 0-5 days.  Since most deaths occurred beyond this 5-day window, this might suggest that immortal time bias wasn’t a major player here.  Nonetheless, this remains an issue (particularly when analyzing “duration of anticoagulation” as a variable within a Cox proportional hazards model). 

Strengths of this study:  Evaluation of safety

This study provides the best available evidence regarding the safety of therapeutic anticoagulation in COVID-19 patients.  Among 786 patients treated with therapeutic anticoagulation, the rate of bleeding (3%) was only slightly higher than that of patients not anticoagulated (2%).  Furthermore, the definition of bleeding was rather broad, so most of these events probably didn’t cause significant morbidity.

Of course, this is a retrospective analysis which is also subject to selection bias.  Specifically, patients were anticoagulated if they were clinically judged to be adequate candidates for anticoagulation (without excessive risk of bleeding).  Thus, this hemorrhage rate wouldn’t apply if every single patient with COVID-19 were anticoagulated, regardless of risk factors.  However, since clinical judgement is always a part of our anticoagulation decisions, this bias may be a reasonable reflection of our clinical reality.  

Atallah B et al.  Anticoagulation for COVID-19 (correspondence in the European Heart Journal)

As these authors note, “in view of the current limited evidence, and the urgency of the topic, the decision to anticoagulate or not is contentious.”^​6​  In the absence of any prospective RCT evidence, guidance from various sources will vary.  Likewise, practitioners may or may not be early adopters of therapeutic anticoagulation, based on their philosophies about critical care (discussed previously here). 

These authors outline an approach to individualizing anticoagulation decisions for patients with COVID-19 (shown above).  In the absence of firm evidence, it’s impossible to know what the perfect algorithm is.  However, this algorithm has many laudable features.  Several options are provided regarding anticoagulation intensity, allowing this to be matched to patient risk.  Numerous patient-specific factors are considered, including coagulation parameters and the patient’s overall clinical risk for thrombosis.  This could provide a useful scaffold for considering anticoagulant dose in a patient with COVID-19.  Of course, additional patient-specific features should be considered as well (e.g., other coagulation abnormalities, other anti-coagulating medications the patient may be on, and various risk factors for hemorrhage). 

One weakness of this algorithm is that pulmonary embolism cannot be excluded by point-of-care ultrasonography (this would require CT angiography, which is often difficult to obtain in COVID-positive patients) .  Therefore, a negative POCUS exam doesn’t necessarily exclude venous thromboembolic disease and cannot necessarily be used to de-escalate anticoagulant therapy. 

• Paranjpe et al. is a multi-center retrospective description of outcomes in 2,733 patients with COVID-19, of whom 786 (28%) received systemic anticoagulation. 
• Systemic anticoagulation appeared to correlate with improved survival within the subset of patients who were intubated.  However, this should be interpreted with extreme caution due to numerous potential sources of bias.
• The risk of systemic anticoagulation was low, with a hemorrhage rate that was only slightly higher than that of patients who didn’t receive anticoagulation (3% vs. 2%). 
• Further evidence is needed regarding the precise risks vs. benefits of therapeutic anticoagulation in COVID-19 (ideally in the form of a multi-center RCT).  Pending the availability of such evidence, this study suggests that therapeutic anticoagulation is a reasonable intervention in many patients.
• Currently, decisions regarding anticoagulant dosing should be individualized (e.g., accounting for coagulation labs, other anticoagulant medications such as antiplatelet agents, clinical risk factors for thrombosis or bleeding, the index of suspicion for DVT/PE, and the availability of testing for DVT/PE). 

related

• IBCC section on thrombosis in COVID-19
• COVID-19: Thrombosis and hemoglobin (RebelEM)
• D-dimer cutoffs to predict thrombosis in COVID-19 (PulmCrit)
• COVID-19: Dimers, clots, and DIC (ERCast with Rob Orman, MD and Thomas Deloughery, MD)

Image credit: Photo by Joline Torres on Unsplash

• Author
• Recent Posts

Social Me

Josh Farkas

Josh is the creator of PulmCrit.org. He is an associate professor of Pulmonary and Critical Care Medicine at the University of Vermont.

Social Me

Latest posts by Josh Farkas (see all)

• PulmCrit Wee: Michelin Chest Syndrome - March 15, 2025
• PulmCrit: ADAPT and SCREEN trials are full of sound and fury, signifying little - December 13, 2024
• PulmCrit: How to quickly create a useful professional account in BlueSky - November 28, 2024