Heparin-Induced Thrombocytopenia in COVID-19.

COVID-19 (coronavirus disease-2019) infection is a highly prothrombotic state, resulting from a dysregulation of the coagulation cascade. Therefore, thromboprophylaxis is strongly recommended in these patients, with some experts even advocating for therapeutic dosing to prevent thromboembolic events. Heparin-induced thrombocytopenia (HIT) is a well-known complication of heparin therapy. In this article, we report a case of HIT in a patient with COVID-19. A 63-year-old male presented with 1 week of dry cough and diarrhea. He had a positive nasopharyngeal COVID-19 reverse-transcriptase-polymerase chain reaction. On admission, the platelet count and liver function tests were within normal limits. During his hospitalization, he developed a right femoral deep venous thrombosis and was started on therapeutic anticoagulation. Due to worsening respiratory failure, he was intubated and mechanically ventilated. Between days 11 and 12 of hospitalization, platelet count dropped from 304 000 to 96 000 cells/µL. He had a high pretest probability for HIT with a 4T score of 6 and a positive anti-PF4/heparin antibody. Heparin drip was discontinued and was switched to argatroban. The serotonin release assay eventually returned positive, which confirmed the diagnosis of HIT. We also discuss potential overdiagnosis of HIT in COVID-19 through 4 cases with false-positive HIT antibodies.

Introduction

An outbreak of novel coronavirus (2019-nCoV) that began in Wuhan, China, is now a major public health concern worldwide. After the first reported case in the state of Washington, the number of cases in the United States has rapidly increased to over 2 000 000 with more than 100 000 deaths.[1] In addition to the typical presentation of respiratory symptoms, hematologic complications have been a significant concern in these patients.

Thrombocytopenia has been reported in up to 55% of patients with COVID-19.[2] A meta-analysis of 9 studies has shown that platelet counts were significantly lower in those with more severe COVID-19 infection, and thrombocytopenia predicts mortality.[2] Heparin-induced thrombocytopenia (HIT) is a differential for thrombocytopenia, especially in the intensive care setting. In one review, the authors suggested that clinicians routinely evaluate all COVID-19 patients on heparin for indices of HIT by performing the 4T score.[3] We report a case of HIT in a patient with critical COVID-19 infection. Also, we briefly describe 4 other cases that were initially managed as HIT, until proven later not to be true HIT, based on the functional assay.

Case Description

A 63-year-old male without comorbidities presented with 1 week of dry cough, myalgias, and diarrhea. On examination, he was afebrile but appeared to be in distress, saturating at 89% on room air. Chest X-ray showed bilateral interstitial opacities. Initial laboratory tests revealed an elevated C-reactive protein of 24.49 mg/dL and a high interleukin-6 level of 58 pg/mL. Fibrinogen and D-dimer were also elevated to 708 mg/dL and 1.44 fibrinogen equivalent units, respectively. The platelet count and liver function tests were within normal limits. He was diagnosed with COVID-19 by nasopharyngeal reverse-transcriptase–polymerase chain reaction. Low-molecular-weight heparin was started for thromboprophylaxis. On day 2 of admission, he developed progressive worsening of dyspnea and was transferred to the intensive care unit (ICU).

On day 7, his right lower extremity was noted to be swollen. Ultrasound duplex revealed a nonocclusive thrombus within the right common femoral vein. He was started on therapeutic anticoagulation for deep vein thrombosis with enoxaparin, which was switched to unfractionated heparin (UFH) later, due to worsening renal function. On the same day, due to worsening respiratory failure, he was intubated and mechanically ventilated.

Between days 11 and 12 of hospitalization, platelet count dropped from 304 000 to 96 000 cells/µL. He had a high pretest probability for HIT with a 4T score of 6. HIT antibody testing (anti-PF4/heparin antibody, by enzyme-linked immunosorbent assay) was sent, which returned positive with an optical density (OD) of 1.243 units. Heparin drip was discontinued and was switched to argatroban. Unfortunately, he died a day later, following a cardiac arrest. Serotonin release assay (SRA) eventually returned positive, with 49% serotonin release at low UFH dose (0.1 IU/mL) with a reduction to 0% release at high UFH dose (100 IU/mL). This confirmed the diagnosis of HIT.

Other Patients With COVID and Possible HIT

We encountered 4 additional patients admitted to our ICU with a diagnosis of acute respiratory distress syndrome secondary to COVID-19 infection with thrombocytopenia, and an intermediate to high pretest probability for HIT. Due to a positive platelet factor 4 (PF4) antibody, they were started on an argatroban, pending send out results of OD levels and functional assay. However, the SRA was negative. Interestingly, their clinical characteristics (reported in Table 1) were similar to those of our patients with true HIT. Two of these patients had a confirmed deep vein thrombosis, and one with suspected pulmonary embolism. Also, 2 patients met International Society on Thrombosis and Haemostasis criteria for disseminated intravascular coagulation (DIC). All the patients were in hypotensive shock, requiring vasopressor support.

Table 1.

Characteristics of Patients Who Tested Positive for PF4 Antibody.

Characteristics	Patient 1[a]	Patient 2	Patient 3	Patient 4	Patient 5
Age in years	63	53	63	70	46
Sex	Male	Male	Male	Female	Female
Indication for anticoagulation	DVT	Acute coronary syndrome and atrial fibrillation	DVT	DVT	Suspected PE
Confirmed thrombosis via imaging	Yes	No	Yes	Yes	No
Days from initiation of heparin to drop in platelets	11	7	6	8	2
4T score	6	5	7	7	4
PF4 antibody (optical density)	Positive (1.243)	Positive (0.707)	Positive (0.767)	Positive (0.042)	Positive (0.307)
Serotonin release assay	Positive	Negative	Negative	Negative	Negative
Presence of DIC by ISTH criteria	No	No	No	Yes	Yes

Abbreviations: DVT, deep vein thrombosis; PE, pulmonary embolism; PF4, platelet factor 4; DIC, disseminated intravascular coagulation; ISTH, International Society on Thrombosis and Haemostasis; HIT, heparin-induced thrombocytopenia.

Patient 1 had true HIT. Patients 2 to 5 had false-positive HIT antibody.

Discussion

HIT is an immune-mediated condition, characterized by platelet drop of 50% or greater, typically within 5 to 10 days of exposure to heparin, leading to hypercoagulability and presence of platelet-activating IgG antibodies.[4-6] These antibodies target complexes of PF4 and heparin.[6] The IgG-PF4-heparin immune complexes cross-link Fcy (γ) receptors on platelets and monocytes and activate them. Activation of platelets and monocytes increases thrombin generation, leading to thromboembolic complications.[4]

HIT affects 1 in 5000 hospitalized patients and paradoxically results in a prothrombotic state and usually does not induce bleeding.[4] Thromboembolic phenomena are seen in half of the patients with confirmed HIT.[4] The mortality rate is around 5% and 10%.[6] In our patient with HIT, the diagnosis of venous thromboembolism preceded thrombocytopenia, which is not uncommon. In a series of 408 patients of HIT with thrombosis, 33.5% of patients developed thrombosis before the decrease in platelet counts.[7]

American Society of Hematology guidelines recommends the use of 4T score to estimate the pretest probability of HIT.[6] In those with intermediate or high pretest probability, testing for the presence of HIT antibodies should be done, by using immunoassays or platelet activation assays.[4,5]

The first step in the management of highly suspected or confirmed HIT is to stop heparin (including heparin flushes), and to initiate a non-heparin anticoagulant, to prevent thrombotic events, in the setting of ongoing massive thrombin generation.[4] Available options for anticoagulation include argatroban, bivalirudin, danaparoid, fondaparinux, or a direct oral anticoagulant.[6] Argatroban is a reversible inhibitor of thrombin[5] with a short half-life and is not renally cleared. It is commonly used in the United States for HIT.[4] In 2 multicenter trials of patients with HIT, argatroban resulted in reduced composite endpoint of death, amputation, and thrombosis, when compared with historic controls.[5] A major concern with argatroban is potential under-anticoagulation in patients with elevated partial thromboplastin time, secondary to additional coagulopathies (hepatic dysfunction, prior anticoagulation with warfarin, consumptive coagulopathy, and more recently, COVID-19).[4] Partial thromboplastin time confounding may also occur in the presence of a nonspecific inhibitor (such as lupus anticoagulant).[8] This would increase the risk of thrombosis and limb loss. Data supporting the efficacy of argatroban is limited to patients with suspected or confirmed HIT.[8,9]

A major problem of HIT is its overdiagnosis. Only around 50% of EIA+ sera have platelet-activating properties and, therefore, do not have true HIT.[10] Antiphospholipid syndrome has been reported in COVID-19 patients,[11] and these patients often test false positive with HIT serology.[9] In the ICU setting, HIT explains about 1 out of 100 cases of thrombocytopenia.[8] Critically ill patients have other plausible non-HIT mechanisms to explain their thrombocytopenia, including septicemia, consumptive coagulopathy in non-HIT-related pulmonary embolism, or catastrophic antiphospholipid syndrome.[8]

Many laboratories report EIA results qualitatively as positive or negative.[10] Interpretation of this immunoassay is improved with quantitative reporting in the form of OD levels.[1,10,12] Higher OD levels correspond to a higher likelihood of true HIT.[10] Weak-positive EIA (OD 0.4 to 1 unit) points strongly against the diagnosis of HIT (≤5% have a strong-positive SRA). Patients with EIA OD ≥2 have a ~90% chance of strong-positive SRA.[10,12] Accuracy of diagnosing HIT is significantly improved by combining an immunoassay with functional tests, such as platelet activation assays (eg, SRA).[4] A negative functional assay essentially rules out HIT.[4] Often, SRA is a “send-out test” and results are unavailable to assist with initial decision making.[13]

Liu et al[14] suggested that anti-heparin-PF4 antibodies are induced in critical COVID-19 patients, resulting in HIT. However, OD analysis and SRA were not performed in these patients due to limited resources. Our 4 additional patients with COVID-19 were positive for anti-heparin PF4 antibodies. However, they had a negative confirmatory test. Perhaps, their intermediate to high pretest probabilities was partly related to complications from COVID-19 itself (venous thromboembolism, sepsis, multiorgan dysfunction, and DIC), which led to testing for HIT.

A misdiagnosis of HIT could potentially lead to exposure to alternate anticoagulants with risks of major hemorrhage, nonavailability of reversal agents, a potential increase in thrombosis from discontinuation of heparin, and increased medical expenditure when compared with heparin.[13]

Conclusion

COVID-19 infection is a highly prothrombotic state, resulting from a dysregulation of the coagulation cascade. COVID-19 and HIT are both prothrombotic and can lead to thrombocytopenia. Both entities may be complicated by DIC. Given a considerable overlap, there is potential for devastating consequences, if a diagnosis of HIT is missed. As most patients with COVID-19 receive prophylactic heparin, and those with severe disease are likely to develop venous thromboembolism and thrombocytopenia, unrelated to HIT, there could be an even greater potential for overdiagnosis of HIT.