NICE guidance on the management of vaccine-induced immune thrombocytopenia and thrombosis (VITT) has recently been published.1 This includes recommendations on which laboratory assays should be performed at diagnosis and during recovery, as well as recommendations on anticoagulation and other therapeutic strategies.

In our centre we conducted a study involving six patients (Table 1) with clinically-confirmed VITT (three men, three women, age 22–53 years, median 38 years). All patients presented with thrombosis and thrombocytopenia 8–16 days (median 11 days) after vaccination with ChAdOx1 nCoV-19 (Oxford–AstraZeneca). Patient 1, who was on long-term rivaroxaban for congenital heart disease, fulfilled the criteria for probable VITT rather than definite VITT2 due to a D-dimer at presentation of 2120 ng/ml fibrin equivalent units [FEU], rather than >4000 ng/ml FEU, but all other features were present. All other patients fulfilled the criteria for definite VITT.2 Patients were followed up for a median of six months (range 142–204 days).

TABLE 1. Patient details at presentation Patient Age Sex Days post vaccine Site of thrombosis PLT DD Anti-PF4 HIPA PIPA 1 22 F 10

CVST Portal vein thrombosis

Central line associated DVT extending into IVC

37 2 120 1.902 Negative Positive 2 34 M 10

Bilateral PE

Right Leg DVT

36 >80 000 1.181 Negative Negative 3 53 F 12 Popliteal artery thrombus 74 4 110 1.851 Negative Positive 4 46 M 16 Multiple PE 38 >80 000 1.462 Negative Positive 5 34 F 8

CVST Bilateral adrenal infarction

Femoral vein DVT

88 >80 000 1.860 Positive Positive 6 41 M 11

CVA Internal carotid artery thrombus

Popliteal artery thrombus

Internal jugular vein thrombus

99 20 570 1.694 Negative Positive Abbreviations: Anti-PF4, optical density from Lifecodes PF4 IgG enzyme-linked immunosorbent assay; CVA, cerebrovascular accident; CVST, cerebral venous sinus thrombosis; DD, D-Dimer (ng/ml FEU); DVT, deep vein thrombosis; HIPA, heparin-induced platelet aggregation assay; IVC, inferior vena cava; PE, pulmonary emboli; PIPA, PF4 induced platelet aggregation assay; PLT, platelet count (x109/l).

For the study, all patients were tested by enzyme-linked immunosorbent assay (ELISA) using the Lifecodes PF4 IgG (Immucor, Solihull, UK), Zymutest HIA IgG and Zymutest HIA IgGAM (Sysmex UK, Milton Keynes, UK) assays, at diagnosis and during the follow-up period; samples were tested alongside routine blood testing as an in-patient, and then weekly or fortnightly after discharge.

All patients were also tested by the in-house anti-PF4/heparin IgG ELISA of the Greifswald laboratory3 and by the PF4-induced platelet aggregation (PIPA) assay4 at diagnosis and on the first sample post discharge that was ELISA-negative by any method. Samples received on or around 1 July 2021 were also tested by these two methods.

Results are shown in Figure 1.

Sequential plots for patients 1–6. Platelet count plotted on left axis. Anti-PF4 enzyme-linked immunosorbent assay (ELISA) optical density (OD) plotted on right axis; dotted red line represents ELISA cut-off of OD 0.50. neg, Negative; PIPA, PF4 induced platelet aggregation assay; PLT, platelet count; POS, Positive

On admission, all patients were positive [optical density (OD) > 0.50] by all four ELISA methods and by PIPA, except the sample from patient 2 (with Definite VITT2) that was positive by all ELISA methods but negative by PIPA. Patients 1–4 were also positive by Asserachrom HPIA IgG, Asserachrom HPIA (Stago UK Ltd, Theale, UK) and AESKULISA HiT II (AEKSU UK, London, UK) assays as previously described5; for operational reasons the first two samples from patient 6 (on admission and two days later) were referred to another laboratory for testing and were both negative using the Asserachrom HPIA assay.

All patients were still ELISA-positive by the Lifecodes PF4 IgG on the last sample tested (median 181 days, range 178–204), which was recently described by Schönborn et al.6 in the in-house assay of the Greifswald laboratory. However, five patients in our study had two consecutive negative results by the Zymutest IgG and/or Zymutest IgGAM by day 100 post vaccination (median 83 days, range 13–100); patient 3 had one negative result by both of these assays before day 100.

PIPA assays were negative in three patients when performed on the first sample that was ELISA-negative by either of the Zymutest assays but were still positive in the other three patients with a negative ELISA, on day 20 (patient 6), 56 (patient 5) and 96 (patient 3). Therefore, it appears that the Zymutest assays become negative before the PIPA, whereas the Lifecodes PF4 IgG and in-house assay of the Greifswald laboratory remain positive after the PIPA has become negative.

NICE guidance suggests that patients should continue anticoagulation for at least three months after discharge and until anti-PF4 antibodies are no longer detected,1 but our data, along with those of Schönborn et al.6 suggests that the presence or absence of these antibodies may be assay-dependent.

The platelet-activating antibodies in the PIPA have disappeared by 12 weeks post vaccination in most patients,6 although we had at least one patient out of six who had a positive PIPA assay after 14 weeks. This suggests that a platelet activation assay may be a better indicator than an ELISA in deciding when to cease anticoagulation.

If a suitable platelet activation assay cannot be performed and anticoagulation is discontinued, either because of a negative ELISA (which could be a false negative depending on the assay) or in spite of a positive ELISA (which could be a false positive depending on the assay), we would suggest to continue monitoring D-dimer and platelet count to identify patients with an ongoing or recurring prothrombotic state so that consideration can be given to restarting anticoagulation if D-dimer increases.7

SP designed the study and wrote the paper; LS and SP performed laboratory assays; SC, MB, JB, TB, AT, LB and PM contributed clinical data and interpretation. All authors critically reviewed the paper and contributed to the final manuscript.