Heparin-Induced Thrombocytopenia in COVID-19: A Systematic Review

Abstract

Background:

It has been more than a year since the whole world is struggling with COVID-19 pandemic and may experience resurgences in the near future. Along with severe pneumonia, this disease is notorious for extensive thromboembolic manifestations. That is why experts advocated aggressive anticoagulation as a part of the therapy since the beginning. However, from May 2020 onwards, cases of heparin-induced thrombocytopenia (HIT) are being reported. HIT in itself is an autoimmune entity leading to life-threatening thrombosis in the setting of thrombocytopenia. Continuation of heparin can have disastrous consequences in case of unrecognized HIT. Hence, timely recognition of HIT is of utmost value to modify the anticoagulation strategy and salvaging lives. We performed a systemic review trying to find all reported cases of HIT in COVID-19.

Methods:

It involved extensive search of the databases including PubMed, Google Scholar, Scopus, and Embase in an attempt to find all reported literature in the last 1 year (November 1, 2019–December 25, 2020) using keywords in various combinations. Literature search resulted in a total of 27 articles and 12 articles were finally selected based on the study design and their relevance pertaining to the intervention done and the outcome of interest.

Results:

A total of 35 patients were included (mean age 56.7 ± 12.8 years, male-to-female ratio = 2:1). The most frequent comorbidity was hypertension. Fifty-seven percent of cases were with low-molecular weight heparin and the rest with unfractionated heparin. Confirmatory functional assay was done in 85.7% of cases (67% by serotonin-release assay [SRA] and 33% by heparin-induced platelet aggregation [HIPA]). All cases tested with HIPA were positive, while with SRA, only 30% were positive. The most common alternate anticoagulation used was argatroban infusion. The new arterial thrombotic event was seen in only 5.7% of cases as repeat myocardial infarction, stroke, and splenic infarction, while clinically significant bleeding was seen in 17.1% of cases. Fifty percent of bleeding episodes were seen where conventional doses of argatroban were used, while no mortality was seen with low-dose argatroban infusion. However, only 45.7% of patients were discharged, 31.4% of patients died, while the outcome was pending for 23% of patients.

Conclusion:

Severe endotheliitis and immune dysregulation giving rise to HIT antibodies and antiphospholipid antibodies have been demonstrated in COVID-19 and modifying our therapy becomes indispensable when it is pathogenic with potentially fatal consequences. In the light of interim results of REMAP-CAP study in severe COVID-19 cases where heparin does not improve the outcome, the present anticoagulation strategy needs re-evaluation. Unrecognized HIT can be catastrophic and close clinical monitoring is required for patients on heparin therapy.

INTRODUCTION

The whole world is going through the crisis of COVID-19 pandemic. With the ongoing pandemic and effective vaccination still in primitive phases, the treatment options largely remain supportive and symptomatic. Demonstration of disseminated thromboembolism in this disease has led to the inclusion of anticoagulation as an integral part of the COVID-19 treatment protocol.[1,2] Low-molecular weight heparin (LMWH) has been advocated for anticoagulation[3] and also for its inhibitory potential on the release of pro-inflammatory cytokines (interleukin-6).[4] Despite using therapeutic anticoagulation, there were reports of thromboembolism in the form of myocardial infarction, pulmonary embolism, stroke, deep-vein thrombosis, limb gangrene, etc.[5,6,7] Meanwhile, the literature on the incidence of heparin resistance in COVID-19 was published, and the coagulation studies showed very high D-dimers, fibrinogen, Factor VIII, and antithrombin in these patients.[8] This led the medical world to research on the coagulation profile in the COVID-19 patient and introspecting the use of heparin as anticoagulant therapy in COVID-19. Interestingly, there were other reports of thrombocytopenia and disseminated thromboembolism where a similar coagulation profile was seen.[9] Even the recently published thromboelastogram study in COVID-19 demonstrated high fibrinogen, D-dimer, Factor VIII, and von Willebrand factor (vWF) levels.[10] Progressive thrombocytopenia leading to fatal outcomes in severe COVID-19 patients leading to the fatal outcome was demonstrated by Liu et al.[11] The pattern closely resembled that of heparin-induced thrombocytopenia (HIT), which was eventually confirmed by the demonstration of antibodies against heparin-platelet factor 4 antigen complex (HPF4 antibody). This revelation led to the debate that whether this immune complex is a consequence of heparin use or is this an effect of COVID-19 pathophysiology, whose role is yet to be unveiled? This article aims to critically review the literature available till date and analyze the origin and role of HIT in COVID-19 pathophysiology and treatment. The need of this modification is further emphasized in light of the recent reports of REMAP-CAP trial results, which stated that therapeutic anticoagulation with heparin does not improve clinical outcome in COVID-19 patients requiring organ support unless having standard indication such as deep-venous thrombosis (DVT) or pulmonary thromboembolism (PTE).[12] It discusses the probability of contribution of HIT in the ongoing widespread thromboembolic phenomenon in COVID-19 despite aggressive anticoagulation. Detection of HIT antibodies puts the treating clinician into a dilemma because of the occurrence of “spontaneous HIT phenomenon” even in “heparin-naïve population.” Such cases underscore the need of rigorous literature search to modify our current anticoagulation strategy keeping in mind the devastating complications of heparin continuation in HIT cases. This article also tries to justify the need for strategic modification of the present COVID-19 therapy in the light of available literature and emerging HIT cases in COVID-19.

DATA RETRIEVAL

The systematic review was done as per the PRISMA statement. We performed a systematic review of English language reports of HIT in COVID-19, searching the databases such as PubMed, Google Scholar, Scopus, and Embase in an attempt to find all reported literature in last 1 year (November 1, 2019–December 25, 2020). The search strategy included terms Heparin-induced thrombocytopenia, HIT, HIT antibodies, Heparin-PF4 antibody complex, H-PF4 antibody, antibodies against the heparin-PF4 complex and COVID 19, SARS-CoV-2 disease, Coronavirus in various combinations. Selected studies enrolled where reverse transcription–polymerase chain reaction-positive COVID cases treated with heparin showed the presence of HIT antibodies. All the articles concerning the management of suspected or confirmed HIT in COVID-19 were eligible. From these studies, data on the incidence of HIT by clinical probability, enzyme immunoassays, and platelet activation assays in COVID-19, alternative anticoagulant used, major thrombotic or bleeding episodes, and mortality were extracted. Quality of evidence was not assessed due to the general low quality of evidence in the form of case reports and case series. All articles were selected for relevance based on title and independently reviewed by two authors for relevance and inclusion. Additional reports were sought for in the reference list of all the relevant studies. All data were independently extracted by both the authors.

RESULTS

The literature search resulted in 25 articles. Manual searching of references of relevant literature identified two additional studies. Of them, one was excluded due to duplication. An additional seven articles were excluded due to nature of the article (review, expert opinion, study protocol, letter to the editor-in response to a published article). Two cases of HIT managed on argatroban have been reported in a retrospective single-center study of 31 adult cases of COVID-19, but it lacked the outcome of interest to be involved in the systemic review.[13] Another retrospective analysis was excluded from this review because it mentioned the presence of HIT antibodies in COVID-19, but there was no mention of any intervention and HIT antibodies were found in “heparin naïve” patients as well.[11] A retrospective cohort study by Lozano and Franco was excluded because it neither used enzyme-linked immunosorbent assay (ELISA) nor serotonin-release assay (SRA) for demonstration of HIT. It completely relied on 4Ts score for the clinical probability of HIT.[14] Finally, 12 articles were included, 5 case reports, 3 case series, and 4 retrospective cohort studies. Flowchart of the literature search is depicted in Figure 1.

Figure 1.

Study selection (search strategy)

A total of 35 patients were identified with a mean age was 56.7 ± 12.8 years and male-to-female ratio was 2:1.The most frequent pre-existing comorbidity was hypertension, followed by diabetes mellitus. About 43% of cases were seen with unfractionated heparin, while incidence with LMWH was 57%. HIT antibodies were detected by ELISA (70%), HemosIL chemiluminescent immunoassay (20%), 3% by qualitative assay (exact method not mentioned), and 3% case was prediagnosed. The confirmatory functional assay was done in 85.7% of cases (67% by SRA and 33% by heparin-induced platelet aggregation [HIPA]). Results with HIPA were 100% positive, while with SRA only 30% were positive. The most common alternate anticoagulation used was argatroban infusion. About 7.5% of cases required renal replacement therapy (RRT), while extracorporeal life support was required in 28.6% of cases. The new arterial thrombotic event was seen in only 5.7% of cases as repeat myocardial infarction, stroke, and splenic infarction, while clinically significant bleeding was seen in 17.1% of cases. Fifty percent of bleeding episodes were seen where conventional doses of argatroban were used.[15] Outcomes were favorable in all the cases where low-dose argatroban was used.[16,17] Regarding outcomes, 45.7% of patients were discharged, 31.4% of patients died, while it was pending for 23% of patients. Details are provided in Table 1.

Table 1.

Description of included studies

	Author/year	Study design	Number of HIT patients	Mean patient age (years)	Male: female	Comorbidities	Diagnosis of HIT	Heparin type UFH/LMWH	Mean duration of heparin exposure before thrombocytopenia
1	Riker et al., May 2020	Case series	3	65.6±11.1	3:0	HTN/BPH HTN/CAD/COPD/CLD AF/IBS/OSA	ELISA	UFH	10
2	Lingamaneni et al., June 2020	Case series	5	59±9.5	3:2	1 - none Rest 4 - NA	ELISA	LMWH	11
3	Aly et al., July 2020	Case report	1	66	Female	DM/HTN/CAD/ESRD	Prediagnosed	NA	NA
4	Gubitosa et al., September 2020	Case report	1	65	Male	DM/HTN/CKD	Qualitative (exact method not mentioned)	LMWH	3
5	Patell et al.	Retrospective cohort	5	NA	NA	NA	HemosIL	UFH	11 days
6	Tran et al.	Case report	1	62	Male	DM	ELISA	Initially LMWH then UFH	4
7	Ogawa et al.	Case report	1	37	Male	None	HemosIL	UFH	8
8	Huang et al.	Case report	1	44	Male	Thalassemia	ELISA	UFH	12 h
9	May et al., July 2020	Case series	7	56.4±16.3	3:4	NA	ELISA	3UFH/4LMWH	NA
10	Daviet et al., September 2020	Retrosp-ective cohort	7	57	6:1	NA	HemosIL	LMWH	10 days
11	Phan et al.	Case report	1	43	Male	None	HemosIL	LMWH f/b UFH	5-10 days
12	Bidar et al.	Case reports	2	62	Female	HTN/asthma	ELISA	UFH	10 days
				38	Female	None	ELISA	UFH	16 days
	Total		35	56.7±12.8 years	20/10/5 NA			15/19/1 prediagnosed
	Functional assay report positive/negative	Alternative anticoagulant used A/B/D/F/DOACs*	Thrombocytopenia recovery	Fibrinogen	Bleeding	Thrombotic event	New thrombotic event	MV/ECMO/RRT	Outcome (died/discharged/pending)
1	SRA 1/2	Bivalirudin Fondaparinux f/by bivalirudin Argatroban (2 ug kg−1 min−1 f/b apixaban	No recovery NA 5 days	NA	1 GI bleeding (on fondaparinux)	PTE/upper extremity DVT/skin necrosis	None	3/0/0	2/1/0
2	SRA 1/4	Argatroban	NA	NA	None	3 DVT	None	1-MV Rest 4NA	1/2/2
3	Prediagnosed	Argatroban f/b apixaban	NA	NA	NA	PTE	None	0/0/1	0/1/0
4	SRA 0/1	Apixaban f/b Fondaparinux f/b argatroban	NA	Initially Low <100 mg/dl, then normal	None	Initially popliteal artery thrombus then bleeding	None	1/0/0	1/0/0
5	SRA 3/1/1 not performed	4-argatroban 1-bivalirudin	NA	NA	3	NA	1 (ischemic stroke + splenic infarction)	NA	3/1/1
6	HIPA (+)	Bivalirudin 0.19 mg.kg−1.h−1	3 days	661 mg/dl	None	PTE	None	1/0/0	0/1/0
7	Not done	Argatroban (0.2 μg.kg−1.min−1) f/b warfarin	NA	NA	During removal of ECMO	PTE/thrombosis of ECMO exchange membrane	None	1/1/0	0/1/0
8	Not done	Not given (died within 23 h)	Expired within 23 h	NA	Upper GI bleeding	MI	Repeat MI	1/1/0	1/0/0
9	SRA 1/6	NA	Not mentioned	NA	None	1 PTE 1 Stroke 1 DVT	NA	NA/2/2	3/1/3
10	HIPA (all +)	Danaparoid/argatroban	Few days	NA	None	5/0	None	6/3/NA	0/5/2
11	Not done	Rivaroxaban for 10 days then argatroban then rivaroxaban	4 days	NA	None	Frequent oxygenator thrombosis	None	1/1/0	0/1/0
12	HIPA (+)	Argatroba 0.25 ug kg−1 min−1	Few days	820	None	No additionl thromboembolism	None	2/2/0	0/2/0
	HIPA (+)	Argatroban 0.25 ug kg−1 min−1	Few days	660	None	Membrane oxygenator thrombosis	None
	Total				6		2		11/16/8

HIT=Heparin induced thrombocytopenia, UFH=Unfractionated heparin, LMWH=Low-molecular weight heparin, MV=Mechanical ventilation, ECMO=Extracorporeal membrane oxygenation, RRT=Renal replacement therapy, HTN=Hypertension, BPH=Benign prostatic hyperplasia, CAD=Coronary artery disease, COPD=Chronic obstructive pulmonary disease, CLD=Chronic liver disease, AF=Atrial fibrillation, IBS=Irritable bowel syndrome, OSA=Obstructive sleep apnea, DM=Diabetes mellitus, ESRD=End stage renal disease, CKD=Chronic kidney disease, ELISA=Enzyme-linked immunosorbent assay, SRA=Serotonin-release assay, HIPA=Heparin-induced platelet aggregation assay, A/B/D/F/*DOACs=Argatroban/Bivalirudin/Danaparoid/Fondaparinux/Direct oral anticoagulants, DVT=Deep-vein thrombosis, PTE=Pulmonary thromboembolism, GI=Gastrointestinal, MI=Myocardial infarction, f/b=Followed by, NA=Not applicable

DISCUSSION

The presence of HIT antibodies has been shown in patients with sepsis, orthopedic and cardiac surgeries, on RRT or extracorporeal membrane oxygenation (ECMO).[18,19,20] The study by Liu et al. showed a significant presence of HIT antibodies along with progressive thrombocytopenia in cases of COVID-19 nonsurvivors. The presence of HIT antibodies was independent of prior heparin exposure.[11] Beun et al. reported heparin resistance in COVID-19 patients, but they did not evaluate for the presence of HIT antibodies.[8] Hence, it is worth introspecting the most appropriate anticoagulation strategy in COVID-19, keeping in mind the background knowledge of the presence of “spontaneous HIT,” “heparin-resistance,” and the widespread disseminated thromboembolic phenomenon in COVID-19, along with the disastrous consequences of heparin continuation in HIT.

This review suggests that in all cases presenting with thrombocytopenia in COVID-19, HIT should be kept as a differential even though the patient has not been exposed to heparin previously. All cases of COVID-19 on heparin therapy should be followed closely with 4Ts test and evaluated for HIT if the score is high/intermediate except in ECMO patients where this test has lower performance.[21] Alternate anticoagulation should be used until the results are available. Functional assays are recommended in all the positive cases for HIT by ELISA. However, it is advisable to continue alternate anticoagulation if there is thrombocytopenia recovery on stopping heparin due to the possibility of “SRA negative HIT.”[22,23] In the case series by Riker et al., out of three cases, one case received fondaparinux that was later changed to bivalirudin. That case had progressive upper extremity thrombosis and could not survive.[24] In the same case series, one case, despite being ELISA positive and SRA negative, was managed with argatroban followed by oral apixaban. It showed remarkable platelet count recovery after stopping heparin and uneventful discharge from the hospital. Having several laboratory methodologies for detection of HIT, and all tests having several limitations due to sensitivity, specificity, technical complexity, or resource limitation, it is always advisable to have a good clinical acumen and closely monitor the clinical response on stopping heparin and switching to alternate anticoagulation.[25] The case report by Gubitosa et al. is worth mentioning where HIT antibodies were positive and SRA was negative. They mentioned the initial 4Ts score as 2, but thrombocytopenia, presence of arterial thrombus, and positive HIT antibodies all together put the patient into intermediate-to-high-risk category for HIT.[26] The use of fondaparinux in suspected cases of HIT could be avoided due to the possibility of cross-reactivity with HIT antibodies which can exacerbate or rarely cause HIT,[27] although there are no large randomized trials to confirm it.

May et al. also reported seven cases of HIT (ELISA + ve) where only one case was SRA positive. They recommended it to be false ELISA + ve. However, 50% of cases had intermediate-to-high 4Ts score. In addition, they did not mention the alternate anticoagulant used before SRA results were available, and it is worth mentioning that only one out of seven patients was discharged.[28] Similarly, Lingamaneni et al. reported five cases where 60% had high risk and 40% had intermediate pretest probability of HIT. Only one out of five was SRA positive. Sixty percent of patients had DVT, 20% had acute coronary syndrome and atrial fibrillation, and 20% was suspected of PTE. The SRA-positive case eventually expired and the rest all cases were in severe shock. They also did not comment on platelet recovery trend on switching to alternate anticoagulation.[29] Cases have been reported where initial tests for HIT antibodies and SRA were falsely negative and thrombocytopenia recovered on switching to alternate anticoagulation. However, the patient developed recurrent HIT on restarting heparin due to initial negative results. His subsequent tests (ELISA and SRA) were found to be positive.[30] All these evidences indicate a very significant role of clinical suspicion and repeated scoring by 4Ts test in all patients on heparin.

Argatroban was the most commonly used direct thrombin inhibitor (DTI) which was later switched to warfarin or direct oral anticoagulants (DOACs). In two cases due to nonavailability of DTI, DOACs were initiated as alternative anticoagulation, followed by argatroban (when available) in one and fondaparinux in the other. The case managed with rivaroxaban and argatroban survived to discharge, while one treated with apixaban followed by fondaparinux showed progressive arterial thrombosis and had fatal outcome.[26,31] Availability and frequent coagulation profile monitoring make DTI infusions resource intensive. There is accumulating evidence for the use of DOACs in HIT, though most of them are used after bridging with DTI. Further studies are required to know the risk–benefit ratio of directly switching to DOACs in HIT without any bridging anticoagulation.[32,33] Besides using rivaroxaban, Phan et al. also used the immunoadsorption column to filter out antibodies.[31]

Fibrinogen levels were elevated in most of the cases (where values were mentioned). High fibrinogen levels might be the body's compensation for ongoing thrombocytopenia due to immune dysregulation in COVID-19 witnessed as the presence of antiphospholipid antibodies or HIT antibodies.[34] There have been cases reported with high fibrinogen, vWF, antiphospholipid and anticardiolipin antibodies in few patients at admission who developed HIT antibodies in due course of time.[17,35,36] Severe endothelial inflammation has been shown in severe COVID-19 infection and extensive tissue trauma and inflammation sensitizing the immune system to produce antibodies might be one explanation to the “spontaneous HIT phenomenon.”[37]

The strength of our study is the comprehensive literature search. However, there can be limitations due to language restrictions and evidence mostly in the form of case reports and case series. The retrospective nature of many case series may lead to selection bias. Overburdened health-care resources due to ongoing pandemic might be the reason for under-reporting of co-existence of HIT antibodies in COVID-19.

CONCLUSION

The presence of HIT antibodies is frequently being seen in COVID-19. A high index of suspicion should be kept for patients showing worsening of coagulation profile in the form of rising D-dimers despite LMWH therapy. Alternate anticoagulation with DOACs can be tried in such circumstances as therapy with argatroban and other DTIs are resource intensive. Many reports of HIT in COVID-19 were labeled false positive based on negative SRA and few cases diagnosed as HIT showed progressive worsening despite switching to fondaparinux. Keeping into consideration the existence of serotonin negative HIT which eventually turns positive on modified functional assays, future case reports should keep a consolidated approach before refuting the diagnosis of HIT. Furthermore, fondaparinux should be preferably avoided in all the cases with the presumptive diagnosis of HIT. However, the evidence available is controversial for it.[27] In the light of results of the REMAP-CAP study in severe COVID-19 cases where heparin does not improve the outcome, the present anticoagulation strategy needs re-evaluation. Unrecognized HIT can be catastrophic and close clinical monitoring is required for patients on heparin therapy.

Key Points

• Close attention should be paid to platelet counts in COVID-19 patients on anticoagulation therapy

• All patients of COVID-19 on heparin therapy should be screened with 4Ts score

• ELISA for HIT antibodies in intermediate/high pretest probability cases should be performed

• Presumptive diagnosis of HIT should be considered in cases with high pretest probability and highly positive immunoassay report and managed accordingly

• Alternative anticoagulation (other than heparin) can be considered in cases of COVID-19 with rising D-dimer levels and widespread thromboembolic phenomenon on heparin therapy where ELISA for HIT antibodies testing not feasible, as in resource-limited settings.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.