Plasma exchange as an effective salvage therapy in AZD1222 vaccine-induced thrombotic thrombocytopenia: a case report

The AZD1222 vaccine (Astrazeneca, Cambridge, UK) is effective and safe against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)¹. In real-life experience, cases of vaccine-induced thrombocytopenia and thrombosis (VITT) have been described, with associated high morbidity and mortality rates. Thrombocytopenia and thrombosis in unusual venous sites are the common clinical presentation, but arterial thrombosis has also been described^2,3. Some treatment recommendations have been published⁴. Plasma exchange (PE) has recently been reported to be effective in three patients⁵. We discuss the case of a patient successfully treated with PE after a severe VITT with atypical arterial involvement.

CASE REPORT

A 60-year-old woman with no important medical history arrived at our Emergency Department from a secondary centre with suspicion of a posterior circulation stroke and moderate thrombocytopenia (87×10⁹/L). She had received the first dose of AZD1222 vaccine seven days before. On her arrival, brain computed tomography (CT) angiography showed occlusion of the left vertebral artery, of the M2 segment of the right middle cerebral artery, and mural thrombosis in both carotid bulbs with no significant stenosis. Whole body CT with contrast showed an extensive patched mural thrombosis of the aorta, from the aortic arch to the lower abdominal aorta, and acute infarctions in the spleen and left kidney (Figure 1). Elevated serum I-troponin was observed. Although the electrocardiogram was normal, a thoracic echocardiogram showed hypokinesia of the heart inferolateral face, compatible with a probable acute myocardial infarction. No left to right shunt was found. There was no associated haemorrhage.

Figure 1.

Brain angiography, body computed tomography scan with contrast and brain nuclear magnetic resonance imaging

Different thrombo-ischemic events presented: (a) aortic patched mural thrombosis, (b) right middle cerebral artery M2 occlusion, (c) spleen and left kidney infarctions (left) and left vertebral artery occlusion (right), (d) hyperintense lesion on fluid-attenuated inversión recovery in the left posterolateral medulla (above), with restricted diffusion (below), compatible with an acute ischaemic lesion.

Platelet count decreased to 46×10⁹/L, with no associated anaemia or leukopenia; blood smear confirmed thrombocytopenia. Internationalised normal ratio, activated partial thromboplastin time, and fibrinogen levels were all within the normal range, but D-Dimer (DD) levels were increased (2,584 ng/mL). No alterations in hepatic or renal function were found. The patient was not eligible for recanalisation treatment because more than 24 hours had passed since clinical onset and she had a National Institutes of Health Stroke Scale (NIHSS) score of only 4. The patient was admitted to the intensive care unit for close monitoring, and empirical treatment with methylprednisolone at 1 mg/kg/24 h and intravenous immunoglobulin (IVIG) at 1 g/kg/2 days was then started. A positive result (241 IU/mL, normal <12 IU/mL) in an ELISA test (AESKULISA-HiTII, AESKU-DIAGNOSTICS, Wendelsheim, Germany) was obtained 18 h after admission. These data suggested probable VITT according to the Expert Haematology Panel⁶. Endovenous argatroban was started 48 h after onset of the initial symptoms.

After 48 h of treatment, platelet count decreased to 28×10⁹/L and no neurological improvement was observed. Due to apparent refractoriness, PE was initiated: a single platelet transfusion was performed after central vascular catheter (CVC) canalisation. The PE scheme included one blood volume replacement per day with 50% albumin and 50% fresh frozen plasma; IVIG was maintained at 0.5 g/kg/24 h after each PE session. Platelet count gradually started to increase and DD levels progressively decreased (Figure 2). Antibodies decreased six times the basal level at diagnosis. Given the previous results, PE sessions were interrupted after five sessions once platelet count had remained consistently >50×10⁹/L over 48 h. DD showed an important increase on day 9 due to a left deep venous thrombosis related to the CVC; the CVC was removed, and anticoagulation was changed to dabigatran 150 mg/12 h after 16 days.

Figure 2.

Analytical evolution of platelet count and D-Dimer under the different treatments

DVT: deep venous thrombosis; PLEX: plasma exchange; IVIG: intravenous immunoglobulin.

No new thrombotic complications developed, and antibody levels were negative 20 days after the diagnosis. Although the patient’s haematological parameters improved, she suffers neurological deficits consistent with a left Wallenberg syndrome; brain magnetic resonance imaging showed an ischaemic lesion in the left medulla (Figure 1).

DISCUSSION

Vaccine-induced thrombotic thrombocytopenia is a severe complication that must be suspected in patients presenting an acute neurological deficit and thrombocytopenia 1–3 weeks after the first dose of AZD1222 vaccine. Although the majority of the reported cases affected venous sites, arterial thrombosis has been also reported^7,8.

An Expert Haematology Panel has proposed diagnostic criteria for VITT based on the presence of typical clinical onset, thrombosis, thrombocytopenia, elevated DD, and anti-PF4 antibody positivity on ELISA^6,9. According to this, our patient was considered as probable VITT. Interestingly, isolated anti-PF4 positivity must be considered as unlikely to be due to VITT due to the presence in healthy subjects after vaccination with the AZD1222 vaccine⁶. Functional platelet tests can confirm VITT suspicion⁴, but these were not available at our centre.

Although there is no strong evidence to suggest the best treatment, it has been recommended to treat VITT with IVIG, glucocorticoids, and non-heparin anticoagulation, and to avoid platelet transfusion^4,10. In our patient, glucocorticoids and IGIV were the first treatment administered, as recommended⁴. Argatroban was chosen among other non-heparin treatments because of our previous experience in heparin-induced thrombocytopenia, its short life-time, and the possibility of close patient monitoring¹¹.

The VITT mortality rate reaches 40%, so new therapeutical approaches must be explored. The maximum utility of PE in these cases is the rapid elimination of the autoantibody. In a recently published experience⁵, three patients with refractory VITT underwent PE achieving consistent responses, although in one case rituximab was added. As in our case, PE was maintained once platelets were at safe range; two patients received five PE sessions, as did our patient. In our patient, PE was performed combining 50% of frozen plasma and 50% of albumin 5% as replacement fluid¹². This novel approach was designed to diminish the risk of allergic reactions to plasma and because a fresh frozen plasma supply is not vital in VITT. Although platelet transfusions are not recommended, we considered a single transfusion before the placement of a CVC for the PE therapy, according to our background experience in thrombocytopenic thrombotic purpura.

Vaccine-induced thrombotic thrombocytopenia is a rare and severe adverse reaction to vaccination with the AZD1222 vaccine with a few confirmed cases reported that can involve any vascular site. PE therapy seems to be effective in refractory cases, and this should undergo further studies as a possible treatment for this pathology.