Updated: September 5, 2020
- In ICU patients, cumulative incidences range from 9% to 70% in patients on varying levels of prophylactic anticoagulation, and whether patients were screened with compression ultrasonography or imaged for change in clinical status (Klok et al, Thrombosis Res 2020.; Middeldorp et al, Preprint 2020.; Klok et al., Thrombosis Res April 30 2020; Llitjos et al, J Thromb Haemostat 2020; Nahum et al. JAMA Net Open 2020.; Moll et al. CHEST 2020). One study suggests COVID-19 patients at increased risk for thrombosis and bleeding (Xu et al, Pulmonology 2020 [preprint, under review]).
- We found that in 102 COVID-19 positive ICU patients, there were 9 radiographically-confirmed DVT or PE, based on imaging obtained for a change in clinical status; all patients received standard dose prophylactic anticoagulation (enoxaparin 40 mg daily or unfractionated heparin 5000 IU three times daily). No events occured in 108 wards patients (Moll et al., CHEST 2020). Similar findings were reported in Indianapolis (Maatman et al. Crit Care Med 2020).
- Higher D-dimer and FDP levels track with multi-organ dysfunction syndrome and poorer prognosis (Wang et al, JAMA, 2020; Zhou et al, Lancet, 2020).
- The mechanism for VTE are unknown and likely multifactorial:
- Systemic inflammatory response as seen in sepsis
- Stasis/critical illness
- Possibly direct endothelial damage from viral injury/ACE2 binding
- An autopsy series of 10 patients from New Orleans reported thrombotic and microangiopathic pathology (and diffuse alveolar damage) (Fox et al. Lancet Resp Med, 2020). Our discussions with pathology colleagues indicate more cellular debris than microthrombi.
- There is a theory from the SARS epidemic that SARS-CoV1 Spike protein can be cleaved by FXa and FIIa. Cleavage of the Spike protein activates it which promotes infectivity (Du, Biochem Biophys Res Com, 2007). By extension, it is hypothesized that anticoagulation might inhibit SARS-CoV-2 replication, however this remains unproven.
- There is a small case series suggesting dipyridamole may be useful, though anticoagulation and antiplatelet agents require further investigation prior to being used therapeutically (Liu et al, medRxiv, 2020 preprint; Lin et al, Emerging Microbes & Infections, 2020).
- There is evidence that A blood type is a risk factor for COVID-19 respiratory failure, and O may be protective. This was based on a genome-wide association study (GWAS) of 835 patients and 1255 control participants from Italy and 775 patients and 950 control participants from Spain. Respiratory failure was defined as a patient requiring supplemental oxygen or mechanical ventilation (Ellinghaus et al. NEJM 2020).
- Rationale: For ward patients, VTE rates at our institution were similar to general ward patients amongst those receiving standard prophylactic anticoagulation (0 out of 108). There are likely a subset of ward patients who would benefit from therapeutic anticoagulation, but the best way to identify these patients remains to be determined.
ICU patients are clearly at elevated risk for VTE events even on standard prophylaxis (Klok et al, Thrombosis Res 2020.; Middeldorp et al, Preprint 2020.; Klok et al., Thrombosis Res April 30 2020; Llitjos et al, J Thromb Haemostat 2020; Nahum et al. JAMA Net Open 2020.; Moll et al. CHEST 2020); for ICU patients experts are divided (Bikdeli et al. JACC 2020) as to whether intermediate or full dose anticoagulation provides the optimal balance of benefits of anticoagulation with risk of bleeding. One study reported improved survival with therapeutic anticoagulation, but was severely limited in the ability to control for confounding (Paranjpe et al. JACC 2020). A propensity score-matched analysis suggested there was no survival benefit of therapeutic anticoagulation (Tremblay et al., Blood 2020). Meanwhile, a small study of 49 ICU patients in Belgium suggests intermediate dose prophylaxis is associated with an odds ratio of 0.13 for PE (Taccone et al. Crit Care Med 2020); clearly a larger study is needed. Given the unclear evidence, we implemented standard prophylactic dosing in ward (i.e. floor) patients and intermediate dosing in ICU patients. We are currently in the process of evaluating the efficacy and safety of this approach, and several randomized trials are underway. To enroll BWH patients in the COVID-PACT trial, email Mat Lopes (MLOPES3@partners.org) and Christian Ruff (firstname.lastname@example.org). To become a participating center, please email COVID-PACT@bwh.harvard.edu.
- For floor patients:
- Given BWH data do not support significantly increased VTE risk in this population (1%, see above), prophylaxis for SPU (Covid-19) patients remains the same:
VTE Dosing Weight Adjustment
CrCl ≥ 30mL/min
CrCl < 30mL/min
Enoxaparin 40mg Daily
UFH 5000 units Q8H
Obese (≥120kg or BMI ≥ 35)
Enoxaparin 40mg BID or 0.5mg/kg Daily
(max dose 100mg Daily)
UFH 7500 units Q8H
Low Body Weight (< 50kg*)
Enoxaparin 30mg Daily
UFH 5,000 units BID or TID
*LBW does not have a universal definition for LMWH dosing, we define it differently in the non-ICU (<50kg) and ICU (<60kg) populations to help achieve our targeted anticoagulant effect, though this remains an active area of research
- For ICU patients and post-ICU patients:
- Given the elevated VTE risk relative to baseline (1.5-2 fold, see above), our recommendation for prophylaxis doses for ICU patients and post-ICU patients is now higher (see table below)
- Covid-19 confirmed and PUI patients requiring ICU level of care during ICU course and after transfer to the floor
- Platelets > 25
- If Platelets <25,000 or bleeding, hold prophylaxis and start TEDs and SCDs
- Other exclusion criteria remain the same (e.g., including but not limited to certain neurosurgery patients, active hemorrhage etc.).
VTE Dosing Weight Adjustment
CrCl ≥ 30mL/min
CrCl < 30mL/min
Enoxaparin 40mg BID
UFH 7,500 units Q8H
Obese (≥120kg or BMI ≥ 35)
Enoxaparin 0.5mg/kg BID*
(max dose 100mg BID)
UFH 10,000units Q8H
Low Body Weight (< 60kg)**
Enoxaparin 30mg BID*
UFH 7,500 units Q8H
*Consider anti-Xa monitoring to adjust regimens if additional risk of bleeding or thrombosis. However, routine anti-Xa monitoring is not required for the average patient. Goal peak anti-Xa for VTE ppx is between 0.2 and < 0.5 (peak should be taken 4-6 hours after 3-4 injections)
**LBW does not have a universal definition for LMWH dosing, we define it differently in the non-ICU (<50kg) and ICU (<60kg) populations to help achieve our targeted anticoagulant effect, though this remains an active area of research
- Recommendations for therapeutic anticoagulation of patients with known DVT or PE remain the same as prior.
- While some institutions are considering full dose anticoagulation in severe COVID disease without known VTE, our interpretation of the data is that the risks outweigh the benefits at this time, unless documented DVT or PE. Preliminary data from Wuhan suggest that prophylactic LMWH or UFH may be of benefit in those patients with severe COVID-19 and D-dimer levels > 6 times the upper limit of normal (Tang et al, JTH, Mar 27, 2020)
- A propensity score-matched cohort study of 3,772 participants compared COVID-19 patients receiving anticoagulation/anti-thrombotic therapy prior to diagnosis to patients without prior anticoagulation/anti-thrombotic therapy; no statistically significant difference in survival or time to mechanical ventilation was observed (Tremblay et al. Blood, 2020.)
- If the patient is on direct oral anticoagulants (DOACs) or Warfarin for Afib or VTE, switch to full dose anticoagulation (LMWH or UFH, as indicated based on renal function or clinical scenario).
Speculative use of therapeutic anticoagulation or tissue plasminogen activator (TPA)
- While therapeutic anticoagulation has been used empirically in some severe COVID-19 patients in Wuhan given the possible microthrombi in pulmonary vasculature (see “Pathophysiology” above), our interpretation of the data is that the risks outweigh the benefits at this time, unless documented DVT or PE (Hardaway et al, Am Surg 2001).
- Similarly, TPA has been proposed as a possible therapeutic. We recommend against TPA for ARDS.
- If high clinical suspicion of PE, consider TPA for salvage/lysis as per usual indications
- Limited data: 16 of 183 hospitalized patients in Wuhan had DIC (Tang et al, J Thromb Haemost, 2020).
- Laboratory changes in coagulation parameters and FDP track with multi-organ dysfunction (Zhou et al, Lancet, 2020).
- Median time to onset of DIC was 4 days into hospital admission (Tang et al, J Thromb Haemost, 2020).
- Identify and treat underlying condition
- ISTH DIC score (MDcalc online calculator): If score < 5, no DIC; recalculate in 1-2 days
- Elevated PT/PTT and D-dimer correlate with worse prognosis: trend PT/INR, PTT, D-dimer, fibrinogen every 3 days until discharge or death
- If not bleeding, supportive care:
- If fibrinogen < 150: FFP, cryoprecipitate or fibrinogen concentrate (RiaSTAP or Fibryga)
- RiaSTAP and Fibryga are less volume, but dose must be discussed with HAT/pharmacy
- Transfuse platelets if < 30K
- Consider holding anticoagulation if the patient requires blood products for supportive care, though clinician should weigh risks and benefits.
- If bleeding, give blood products:
- For elevated PT/PTT and bleeding, use FFP or 4F-PCC (KCentra is less volume, but must discuss dose with HAT/pharmacy)
- Hold anticoagulation for active bleeding.
- Start systemic anticoagulation only if:
- Overt thromboembolism or organ failure due to clot (i.e., purpura fulminans)
- There has been no mortality benefit of therapeutic anticoagulation in DIC (Levi et al, Blood, 2018).
- DIC is associated with worse survival in COVID-19 patients. Out of 183 COVID-19 patients in Wuhan, 71% of non-survivors had DIC (ISTH score ≥ 5; MDcalc online calculator) compared to 0.6% of survivors (Tang et al, J Thromb Haemost, 2020).
- Many patients with COVID-19 have either normal WBC or leukopenia
- Leukocytosis (>10,000/µL) in 13% and leukopenia (<4000/µL) in 15.5% (Goyal et al, N Engl J Med, 2020)
- Lymphocytopenia, or lymphopenia, typically defined as an absolute lymphocyte count < 1000/µL, is the most common abnormality on the CBC in COVID-19 and is found in over 80% of hospitalized patients (Guan et al, N Engl J Med, 2020, Huang et al, Lancet, 2020)
- In 393 adult patients hospitalized with COVID-19 in New York City, 90 percent had a lymphocyte count <1500/µL (Goyal et al, N Engl J Med, 2020)
- ICU patients with COVID-19 had a median lymphocyte count of 800 cells/mm (Wang et al, JAMA, 2020)
- Neutrophil counts in the ICU were on average 4.6 x109/L compared to 2.7 x109/L in non-ICU patients (Wang et al, JAMA, 2020)
- The use of steroids may skew these results
- In COVID-19, the degree of lymphopenia correlates with disease severity and survival (Tan et al, Signal Transduct Target Ther, 2020; Yang et al Lancet Respir Med, 2020; Ruan et al, Intensive Care Med, 2020).
- Patients who died from COVID19 are reported to have had significantly lower lymphocyte counts than survivors, and a lymphocyte percentage <10% on the WBC differential is strongly associated with decreased survival. (Ruan et al, Intensive Care Med, 2020)
- Compared to survivors, non-survivors have persistent lymphopenia (Zheng et al, Cellular and Molecular Immunology, 2020, Figure 1)
- Numerous possible explanations for lymphopenia in COVID-19 have been proposed.
- Invasion/ destruction of lymphocytes via ACE2 receptor
- ACE2 receptor is expressed on lymphocytes, mainly those within oral mucosa, digestive system, and the lungs
- Acidemia, nutrition, bone marrow suppression
- In critically ill patients, all of these factors may suppress proliferation of lymphocytes
- Cytokine Storm
- Cytokines (such as TNFα, IL-6) may induce apoptosis of lymphocytes and has been previously documented in critical patients with SARS (Lam, The Clinical Biochemist Review, 2004). However, this explanation is likely inadequate, as in Cytokine Release Syndrome with CAR-T therapy there are elevated cytokine levels but not consistent documented lymphopenia
- Lymphatic organ damage (thymus, spleen)
- This possibility still requires pathological evidence and remains speculative (Tan et al, Signal Transduct Target Ther, 2020)
- Host Endothelial function
- With age and chronic disease, there is more leukocyte adhesion and extravasation (Bermejo-Martin, Journal of Infection, 2020)
- Sequestration of lymphocytes
- Cytokine release leads to movement of the lymphocytes to the site of infection, the lung tissue, which could contribute to peripheral lymphopenia (Rahimmanesh, Preprint, 2020)
- Cytotoxic T lymphocytes decline sharply in COVID-19, which could be due to movement of these cells to the lower respiratory tract (Rahimmanesh, Preprint, 2020)
- Lymphocyte Subgroups
- Increased CD8+ T cells tended to be an independent predictor for COVID-19 severity and treatment efficacy, which is likely due to their known function of helping to mediate clearance of viruses. (Wang, Journal of Infectious Disease, 2020)
- No current treatment regimen management changes based on lymphopenia
- There is no evidence for giving pneumocystis jiroveci prophylaxis given the transient nature of lymphopenia with COVID-19
- Potential future options may include drugs that improve endothelial dysfunction such as adrecizumab (Bermejo-Martin, Journal of Infection, 2020)
- Thrombocytopenia has been described in hospitalized patients with COVID-19 with variable reported incidence:
- Initial reports in small cohorts: 5% of patients presented with a platelet count of less than 100,000/µL (Huang et al, The Lancet, 2020); 72.5% of patients developed thrombocytopenia during admission (Chang et al, JAMA, 2020)
- More recently: 20.7% to 36.2% of patients in larger cohorts are thrombocytopenic (Yang et al, J Thromb Haemost, 2020; Guan et al, NEJM, 2020)
- Nadir platelet count of 0-50,000/µL is associated with a 92.1% rate of in-hospital mortality, compared to mortality rates of 62.1% for platelet count 50,000-100,000 /µL, 17.5% for platelet count 100,000-150,000/µL, and 4.7% for platelet count greater than 150,000/µL
- Relative risk of death increased with decreasing platelet count, with RR 13.68 (95% CI 9.89-18.92) for 0-50,000/µL group
- Multiple proposed mechanisms of thrombocytopenia (Xu et al, Ann Hematol, 2020; Amgalan & Othman, J Thromb Haemost, 2020)
- Decreased primary platelet production
- Cytokine storm leading to suppression of bone marrow progenitor cells
- Direct infection of hematopoietic and bone marrow stromal cells
- Increased platelet destruction
- Increase of autoantibodies and immune complexes leading to platelet clearance by the immune system
- Decreased circulating platelets
- Lung injury, leading to:
- Increased platelet consumption via platelet aggregation and wrapping into microthrombi
- Decreased platelet production secondary to damaged pulmonary capillary beds and thus decreased megakaryocyte fragmentation
- Increased platelet consumption
- Activation of coagulation pathway via increased thrombomodulin levels
- Increased fibrinolysis
- Disseminated intravascular coagulation (DIC) may also contribute as a related or independent process (see section on DIC)
- Click here for a flow chart summarizing possible mechanisms of thrombocytopenia
Diagnosis and workup
- Consider other potential contributing etiologies of thrombocytopenia
- Medication, additional infection(s), liver disease, splenomegaly, heparin-induced thrombocytopenia (HIT), thrombotic microangiopathy (TTP, HUS, DIC), alcohol, malignancy, pregnancy, rheumatologic/autoimmune, bone marrow disorders
- Initial workup
- Labs: PT, aPTT, fibrinogen, LDH, LFTs, B12, folate
- Peripheral blood smear
- Useful to exclude platelet clumping (pseudothrombocytopenia) and to evaluate for other contributing causes
- Pretest probability of HIT can be calculated by 4Ts score (MDCalc 4Ts calculator)
- Laboratory testing for HIT should typically only be sent in patients with at least intermediate probability of HIT (4 or more points on 4Ts score), although need to consider clinical context.
- If send PF4, use a non-heparin anticoagulant (e.g. bivalirudin or other direct thrombin inhibitor per institutional protocols) while awaiting PF4 results. Serotonin release assays may be necessary to confirm positive PF4 results.
- If concern for DIC, refer to DIC protocol section
- If not bleeding, transfuse platelets if < 10,000/µL
- If bleeding, transfuse platelets according to clinical situation