Immune thrombocytopenia following vaccination during the COVID-19 pandemic

To date, there have been over 3.2 million doses of ChAdOx1 nCoV-19 (ChAd) COVID-19 vaccine (AstraZeneca) and 1 million doses of BNT162b2 (BNT) COVID-19 vaccine (Pfizer-BioNTech) administered in Australia. Among the numerous safety signals that have been raised, we present our case series of immune thrombocytopenia (ITP) after COVID-19 vaccination.^1-4 ITP following vaccination has been previously described in other settings and after mRNA-based COVID-19 vaccines. ^5-8 A Scottish National Registry study examined general practice data and identified a small increased incidence of ITP diagnoses between days 0-27 after vaccination with ChAD.⁹ We present the clinical characteristics and treatment outcomes of patients diagnosed with ITP following COVID vaccinations (ChAd or BNT) in Australia.

After obtaining independent ethics committee approval, we contacted hemostasis hematologists across Australia to participate in our comprehensive survey of clinical presentations of vaccine-associated ITP as defined by the temporal relationship of ITP within 42 days following COVID-19 vaccination, without an otherwise apparent alternative cause or thrombosis. Patients with thrombosis or elevated D-dimer levels were investigated and excluded for vaccine-induced immune-mediated thrombotic thrombocytopenia according to international guidelines.¹⁰ Response was defined as per international consensus guidelines as a platelet count ≥30x10⁹/L, 2-fold increase over baseline and absence of bleeding. A complete response was defined as a platelet count ≥100x10⁹/L and absence of bleeding.¹¹

A total of 14 patients were diagnosed with ITP following vaccination. Twelve of these cases followed administration of the ChAd vaccine. Ten cases were de novo ITP, presented in Table 1. Four cases were relapses in patients with previously stable chronic ITP, presented in Table 2. None of the 14 patients had concurrent thrombosis. Among the 12 cases of ITP following administration of the ChAd vaccine, an enzyme-linked immunosorbent assay for platelet factor 4 (PF4) was performed in six and all of these tested negative.

Table 1.Demographics and clinical features of patients with newly diagnosed immune thrombocytopenia after COVID-19 vaccination.

Table 2.Demographics and clinical features of patients with relapsed chronic immune thrombocytopenia after COVID-19 vaccination.

The median age of the patients was 75 years (range, 22-94), the median time to presentation after vaccination was 10 days (range, 2-31), and the platelet count at presentation was 7x10⁹/L (range, 0-22x10⁹/L). World Health Organization bleeding scores were mild: ten patients had grade 0 or 1, two patients had grade 2, and one patient each had grades 3 and 4.¹²

Ten cases had no prior history of ITP and all received treatment upfront: seven received prednisone, and three high-dose dexamethasone pulses. Eight patients also received between 1-2 g/kg intravenous immunoglobulins (IVIg) as part of first-line therapy. The median time to response was 3.5 days (range, 1-18). Ten evaluable patients achieved a complete response by a median of 9 days (range, 3-47). Day 30 data were available for nine of these ten patients without a prior history of ITP, as one left Australia: the median platelet count was 151x10⁹/L (range, 8-259x10⁹/L); eight were still on corticosteroids (median prednisone equivalent 20 mg daily), one was on eltrombopag (commenced as second-line treatment) and another was receiving mycophenolate mofetil that had been commenced in first-line treatment in combination with prednisone.

One 80-year-old female presented with life-threatening bleeding (influenza vaccination 1 day prior and ChAd 21 days prior to presentation) and after no initial response to escalating prednisone doses and IVIg, eltrombopag was commenced on day 15. Platelets began to respond by day 18, and the platelet count rose to 157x10⁹/L by day 30 after presentation while only on prednisone.

One 82-year-old male presented with a platelet count of 3x10⁹/L, and widespread bruising 9 days after his first ChAd vaccination. He was treated with high-dose dexamethasone and platelets responded, reaching 97x10⁹/L by day 16 (Figure 1A). He received influenza vaccination the following day, but his ITP relapsed by day 32. He responded promptly to a second pulse of high-dose dexamethasone with a platelet ount of 65x10⁹/L by day 36. He had never previously developed ITP despite numerous influenza vaccinations in the past.

One 83-year-old female presented with a platelet count of 10x10⁹/L, facial petechiae, and upper chest ecchymoses 23 days after her first ChAd vaccination (Figure 1B). She responded promptly to a dexamethasone pulse 20 mg daily for 4 days and IVIg infusion 0.4 g/kg for 3 days. She relapsed on day 19 with platelets 23x10⁹/L and new lower limb bruising, and was treated with another pulse of dexamethasone and IVIg 0.4 g/kg for 2 days.

In total, there were four patients with chronic ITP who relapsed following COVID-19 vaccination. Three patients receiving ChAd had stable chronic ITP, and were off ITP-directed therapies at the time of COVID-19 vaccination. They were treated with standard first-line therapies and all responded within 3 days.

IVIg monotherapy alone was successful in one 72-yearold female with chronic ITP who presented with a platelet count of 11x10⁹/L but responded by day 3, achieving a complete response on day 5; her day 30 platelet count was 215x10⁹/L (Figure 1C), and she had no need for steroids at any time despite having had refractory ITP requiring splenectomy in 1994. Her most recent prior platelet count was 255x10⁹/L less than 3 weeks before vaccination. Her most recent prior ITP treatment had been rituximab monotherapy in 2011.

A second chronic ITP patient, a 77-year-old male who received influenza vaccination prior to ChAd vaccination, presented with a platelet count of 2x10⁹/L, achieved a response and complete response by days 3 and 8 respectively, had a day 30 platelet count of 144x10⁹/L, and was on a weaning schedule of prednisone at day 30 after initially being treated with prednisone/IVIg upfront.

The third patient with chronic ITP, a 73-year-old male with a pre-vaccination platelet count of 120x10⁹/L, was thrombocytopenic (platelet count, 5x10⁹/L) 31 days after ChAd vaccination. He was started on prednisone monotherapy and achieved a response within 2 days, a complete response by day 4, and a platelet count of 234x10⁹/L by day 30 while on prednisone 10 mg daily.

The fourth chronic ITP patient in this analysis was a 94-year-old female who received her first dose of BNT 9 days prior to presentation. She had previously enjoyed a stable platelet response on romiplostim for her chronic ITP with a recent platelet count of 86x10⁹/L, falling to 12x10⁹/L without any bleeding; her platelet count returned to baseline within 5 days of presentation. She proceeded to receive her second dose of BNT 21 days after the first, relapsing again on day 15 with a platelet count of 14x10⁹/L before returning to her stable baseline within a further 7 days.

Figure 1.Clinical course of three separate cases of immune thrombocytopenia following COVID-19 vaccination. (A) An 82-year-old male with newly diagnosed immune thomrbocytopenia (ITP) was treated with dexamethasone an initially responded, received influenza vaccination, relapsed, and responded again to another pulse of dexamethasone and weaning prednisone taper. (B) An 83-year-old female with newly diagnosed ITP was treated with two pulses of dexamethasone/IVIg. (C) A 73-year-old female had a relapse of chronic ITP after receiving ChAd vaccination, received IVIg 2 g/kg over 2 days as monotherapy. ChAd: ChAdOx1 nCoV-19 (AstraZeneca); IVIg: intravenous immunoglobulin.

Our case series of vaccine-associated ITP comprises more cases of ITP following administration of the ChAd vaccine than after the BNT vaccine (12 from 3.2 million ChAd vaccinations vs. 2 from 1 million BNT), although there may be an ascertainment bias due to greater scrutiny of patients following ChAd vaccination, as suggested in a recent Scottish study even though this paper also concluded that there was an increased rate of ITP diagnoses of 1.13 per 100,000 doses.⁹ In contrast, a Scandinavian epidemiological study was unable to identify an increased rate of ITP diagnoses although rates of “unspecified thrombocytopenia” and bleeding events were increased significantly.¹³ Our study was not designed to address the questions of frequency or causality. Our designation of these cases as “vaccine-associated” ITP as opposed to co-incident ITP is based on the clinical diagnosis of ITP as one of exclusion. As vaccine association cannot be excluded, we cannot conclude that these patients have primary ITP, conceding that future outcomes may eventually justify revision of our diagnosis, which is common in ITP.¹⁴

Two of 14 cases are confounded at presentation by the recent administration of influenza vaccination, and another patient received influenza vaccination shortly after initial recovery from ITP before relapsing. However, these limitations reflect an unavoidable real-world dilemma as public health imperatives to protect populations at risk during a pandemic will likely outweigh the considerably smaller numerical risk of uncertain outcomes and vaccination side effects when immunization programs overlap.

Most cases responded rapidly to first-line therapy although the majority remained on corticosteroids for at least 30 days (median prednisone equivalent dose 13.75 mg daily for all cases, 20 mg daily for those with newly diagnosed ITP). Patients whose chronic ITP relapsed after vaccination responded rapidly to first-line therapies, consistent with other observations,⁸ and reassuringly for those with underlying ITP who are at present hesitant to receive COVID-19 vaccination. So far, in three patients, a single pulse of high-dose dexamethasone was insufficient to maintain remission in this cohort, but repeat courses have been successful and well tolerated. Additional strategies used successfully include eltrombopag and mycophenolate mofetil. Further data will be needed to understand the durability of these responses.

We anticipate that there may be cases along a spectrum of clinical presentations between vaccine-induced immune-mediated thrombotic thrombocytopenia and vaccine-associated ITP, as have already been noted elsewhere. ¹⁵ In our cohort, overlapping characteristics have not yet been identified, and all six patients with samples tested were negative for anti-PF4 antibodies.

Both local and international registries are currently collecting data that will be useful for investigating treatment strategies and clinical outcomes for patients developing ITP following COVID-19 vaccination.

Footnotes

• Received June 15, 2021
• Accepted August 2, 2021

Correspondence

PHILIP YOUNG-ILL CHOI - phil.choi@act.gov.au

Disclosures: no conflicts of interest to disclose.

Contributions: RB and PY-LC designed the study, analysed the data, and wrote the manuscript; DH, HAT, CWT, AE, VMYC, BHC, JC and DP reviewed and edited the manuscript.