Background

Over 200 million cases of COVID-19 infections have been confirmed worldwide. Multiple platforms were used to develop vaccine candidates, including messenger RNA (mRNA) vaccines encoding the SARS-CoV-2 spike glycoprotein. The mRNA vaccines, tozinameran and elasomeran, were approved by the US Food and Drug Administration (FDA) under emergency use authorisations; currently only tozinameran has full FDA approval. Both are highly efficacious in preventing serious outcomes, including hospitalisation, severe illness and death. Current knowledge about the safety of COVID-19 vaccines relies on data from phase 1–3 randomised controlled trials and vaccine safety surveillance systems.1 2 Due to the highly contagious nature of the disease and the pandemic straining healthcare systems, the benefits of mass vaccination efforts are unquestionable.3 Nonetheless it is important to recognise that similar to post-COVID-19 infectious complications, rare life-threatening complications post vaccination have also been reported including thrombocytopaenia, haemolytic anaemia, vasculitis and myocarditis.4–9

Haemophagocytic lymphohistiocytosis (HLH) is a rare, severe, uncontrolled hyperinflammatory reaction where activated lymphocytes and histiocytes infiltrate all organs and secrete large amounts of cytokines, leading to tissue damage and organ failure. Characteristic features include prolonged fever, splenomegaly, pancytopenia and haemophagocytosis. Biochemical markers include hyperferritinaemia, hypertriglyceridaemia and hypofibrinogenaemia. Due to its rarity and variable clinical presentation, an accurate diagnosis may be delayed, leading to substantial morbidity and mortality. Although HLH has been described post-Covid-19 infection as well as postadenoviral vector vaccination,10–14 it has never been reported post-mRNA vaccination. We report two cases of HLH admitted to our hospital from June to August 2021 after administration of mRNA vaccines. We also searched the vaccine adverse event reporting system (VAERS) (last release 09/10/21) for ‘HLH’ and found 50 reports of suspected HLH postvaccination.15 We cannot currently define a causal relationship between vaccination and the development of HLH, additional surveillance is needed to understand whether a relationship exists, if any, between them.

Case presentation

Patient 1 is a 60-year-old man with a history of Barrett’s esophagus who developed altered mental status and slurred speech 6 days after the first dose of tozinameran (BNT162b2 Pfizer-BioNTech vaccine). He was initially diagnosed with transient ischaemic attack, but progressively deteriorated over the next month developing fevers, drenching night sweats, loss of appetite, unintentional weight loss, delirium and became non-ambulatory.

Patient 2 is a 32-year-old woman with no prior medical history who developed high fevers 4 weeks after the second dose of tozinameran. She was briefly hospitalised at an outside hospital with new anaemia (haemoglobin 80 g/L), hyperferritinaemia (35 000 ng/mL) and elevated transaminases. Rheumatological, oncological and infectious workups were negative, and fevers worsened despite 8 weeks of prednisone 50 mg daily. By the ninth week of worsening symptoms as an outpatient, fevers reached 40°C and became debilitating. She was re-admitted and was transferred to our institution for further management. By this time her ferritin had increased to 68 212 ng/mL (normal range 15–150 ng/mL) and neopterin was 9.5 ng/mL (normal range <2.5 ng/mL).

Investigations

Acute COVID-19 infection was ruled out in both patients based on negative PCR tests of nasopharyngeal specimens. Negative nucleocapsid and positive anti-spike antibodies confirmed no prior COVID-19 infection.

Patient 1 met 6/9 HLH criteria including a bone marrow biopsy that showed haemophagocytosis (table 1, figure 1A). Patient 2 met 7/9 HLH criteria including haemophagocytosis on bone marrow biopsy (table 1, figure 1A). The diagnosis of HLH was made in both patients based on these criteria. Table 1 details the extensive infectious, malignant and rheumatological workups that were obtained in both patients to rule out alternate aetiologies for HLH.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

HLH post-COVID-19 vaccine. (A) Depicts bone marrow aspirate smears obtained by means of bone marrow biopsy in patient 1 and patient 2, which show haemophagocytosis (black arrow) in both patients. (B) Depicts trends in white cell count (WCC), absolute neutrophil count (ANC), haemoglobin (Hgb) and platelet (Plt) for patient 1 and patient 2. (C) Depicts trends in ferritin and soluble IL-2 receptor (sIL2R) on a logarithmic scale with dotted line indicating upper limit of normal for ferritin (400 ng/mL) and sILR (858 pg/mL). HLH, haemophagocytic lymphohistiocytosis.

Differential diagnosis

The differential diagnosis was very broad in both patients. Table 1 lists the complete work up for each patient, which aimed to rule out not only an aetiology for their fevers but also an underlying aetiology that could result in secondary HLH. Given their high fevers, infectious aetiologies were high on the differential for both patients. In addition to blood cultures and ruling out SARS CoV-2 infection, evaluation for Aspergillus, Babesia, Bartonella, Brucella, Leptospira, Tuberculosis and Francisella tularensis in the serum were performed. Both patients also had lumbar punctures to evaluate for bacterial, fungal and viral aetiologies; studies included evaluation for Cryptococcus, Epstein-Barr virus, cytomegalovirus and West Nile virus.

When infectious aetiologies were negative, both patients also underwent an extensive rheumatologic workup which included evaluation for systemic lupus erythematosus, mixed connective tissue disorder and vasculitis.

CT, including positron emission tomography CT, was also performed to evaluate for oncological explanations for their symptoms. Although lymphadenopathy was seen in both patients, biopsies of suspicious lymph nodes were non-diagnostic for malignancy. Additionally, both patients had bone marrow biopsies as part of the workup for fever of unknown origin, both to evaluate for HLH as well as to rule out a haematological malignancy which subsequently revealed evidence of haemophagocytosis. A final diagnosis of HLH was even more suggestive following drastic clinical response to HLH-directed treatment as well as further serological testing (elevated CXCL9, neopterin).

Treatment

Patient 1 was initially treated with 5 days of prednisone 1 mg/kg prior to the diagnosis of HLH but his symptoms were refractory to this treatment, and he continued to deteriorate. He was subsequently started on HLH-directed therapy with etoposide and dexamethasone.

Patient 2 had initial improvement in her symptoms and pancytopenia after starting etoposide and dexamethasone but after her fourth etoposide dose she developed neutropenic fevers and haemodynamic instability without identifiable infectious source. Due to concern for disease refractoriness, she was started on emapalumab-lzsg.

Outcome and follow-up

Patient 1 had dramatic improvement in his speech, ambulation and pancytopenia within 48 hours of starting etoposide and dexamethasone (figure 1B,C). He was discharged 13 days after his first etoposide dose and he initially continued to improve on outpatient follow-up visits, however, symptoms (night sweats, malaise) relapse and his HLH biomarkers worsen (worsening thrombocytopaenia, rising triglycerides, elevated ferritin) anytime he is tapered off steroids. He has been started on cyclosporine without meaningful improvement, so we have restarted steroids.

Patient 2 similarly had rapid improvement in her symptoms and laboratory markers after starting second line HLH-directed therapy. She was discharged home after 5 weeks and continues to improve on outpatient follow-up visits. She completed eight treatments of emapalumab-lzsg and on discontinuation, laboratory markers worsened, so we plan for a longer course of emapalumab-lzsg.

For patient 1, we considered intrathecal methotrexate as treatment for the neurological manifestations of HLH. Although intrathecal therapy has been used in thrombocytopenic patients, there is no consensus on a safe cut-off value of platelet counts. The patient had profound thrombocytopaenia (<20 000/µL), which was unresponsive to platelet transfusions during his hospital course. Given the risks of haemorrhagic complications, we were unable to offer this therapeutic intervention during the hospitalisation. Additionally, his neurological symptoms improved rapidly with systemic first line treatment for HLH with etoposide and steroids.

For patient 2, we discussed pursuing an allogeneic haematopoietic stem cell transplant as she initially had worsening laboratory HLH markers after emapalumab was being tapered. We are currently awaiting results of genetic testing to rule out hereditary HLH mutations that may have predisposed her to developing HLH before we subject her to transplant considering that she is being weaned off the emapalumab. Her biomarkers have also improved with prolonged emapalumab so we have currently deferred pursuit of transplant.

Elevated D-dimer can also be seen in HLH as many patients will have hepatitis and abnormal coagulation parameters. Although D-dimer is not included in the diagnostic criteria for HLH, it may be of interest to trend this lab to evaluate for response to HLH directed therapy.