Haemophagocytic lymphohistiocytosis complicating visceral leishmaniasis in the UK: a case for detailed travel history, a high index of suspicion and timely diagnostics

Abstract

A 4-month-old male infant presented acutely unwell with fever. He was initially treated for sepsis but failed to improve with IV broad spectrum antibiotics. Haemophagocytic lymphohistiocytosis (HLH) was diagnosed due to his fever, pancytopenia, splenomegaly, hypertriglyceridaemia, hypofibrinogenaemia and significant hyperferritinaemia. An array of differentials for HLH including both immunological and infectious causes were considered and excluded. He had travelled to Madrid, and hence visceral leishmaniasis (VL) was suspected, but was not confirmed on the initial bone marrow aspirate (BMA) microscopy or culture. He improved with empirical treatment with dexamethasone and liposomal amphotericin B. VL was later confirmed on BMA PCR. He made a good recovery and remained well at 12 month follow-up. Non-endemic countries need rapid and sensitive VL diagnostics. A thorough travel history and high clinical index of suspicion are necessary to avoid the pitfall of treatment with intense immunosuppression recommended in treatment guidelines for HLH.

Background

Fever in an infant and fever in the returning traveller both require consideration of a wide range of differentials. For the treating clinician, ‘the devil is in the detail’ with rarer diagnoses being easily missed without a concise history and examination. One such diagnosis is haemophagocytic lymphohistiocytosis (HLH) which is a rare but life-threatening condition characterised by inflammation and immune hyperactivation causing uncontrolled cytokine release. The clinical picture initially resembles sepsis, but then progresses in an insidious fashion with multi-system involvement including: pancytopenia, hepatosplenomegaly and hyperferritinaemia.^{1 2} There are two types of HLH: familial HLH (fHLH) and secondary HLH (sHLH). The latter occurs when there are no identified genetic mutations that cause fHLH. sHLH can be associated with infections, including visceral leishmaniasis (VL).²

VL is a differential for fever in a returning traveller which similarly presents as an acutely unwell child with hepatosplenomegaly and is clinically indistinguishable from HLH. It is challenging to recognise VL-driven HLH, as both are rare and initial investigations for Leishmania spp. may be delayed. Due to false negatives, they can also be falsely reassuring. Furthermore, both present with non-specific features mimicking sepsis, thereby obscuring the diagnoses. European travel history is sometimes disregarded in acute presentation of fever and thus VL may not be appropriately considered in the UK.

We present the case of an infant in whom early empirical liposomal amphotericin B treatment was successfully initiated based on the travel history and clinical suspicion. Normal perforin, granule release assay (GRA), signaling-lymphocytic-activation-molecule-associated protein (SAP) and X-linked inhibitor of apoptosis (XIAP) excluded the common causes of fHLH.³ Rapid VL diagnostics were not available. However, the diagnosis was later confirmed by PCR on bone marrow aspirate (BMA) and by serology. The authors suggest, in the context of suspected VL, caution should be taken when considering intensive immunosuppression (as recommended in standard HLH protocols) and empirical liposomal amphotericin B should be considered while awaiting diagnostic confirmation.

Case presentation

A previously well 4-month-old male infant presented with a cough and 2 days of fever. In view of his initial examination and unremarkable blood tests he was diagnosed with a viral upper respiratory tract infection and discharged home. However, he re-presented with a 5-day history of fever and inconsolable crying. He had reduced oral intake, vomiting and a blanching rash.

He was a previously well child who was up to date with his vaccinations. He was born in the UK. While pregnant his mother had travelled to South Africa, Cuba and Madrid. Routine antenatal screening bloods and ultrasounds were unremarkable. She had prolonged rupture of membranes and received intrapartum intravenous antibiotics. Two weeks postpartum, his mother was given oral antibiotics for a lower respiratory tract infection. Her investigations for Leishmania and VL were negative. His parents were both well at the time he presented. The family had travelled to Madrid twice when he was 2 and 3 months of age. There was no known history of Phlebotomus spp. sandfly exposure, however he had visited the local parks.

Cardiorespiratory and neurological examination were unremarkable. Abdominal examination revealed 6 cm hepatomegaly and 5 cm splenomegaly with no associated lymphadenopathy. A blanching erythematous rash was noted. He subsequently developed two pustular papules on his legs.

Investigations

The blood results from admission showed a pancytopenia with haemoglobin of 74 g/L (reference range 110–140 g/L), a WBC of 3.7×10⁹/L (reference range 6.0–15.0×10⁹/L) with neutrophils of 0.7×10⁹/L (reference range 1.5–8.0×10⁹/L) and platelets of 77×10⁹/L (reference range 150–400×10⁹/L). The blood film was reported as showing no blasts, hypochromia and thrombocytopenia. His ESR was 8 mm/hour (reference range <20 mm/hour) and CRP raised at 75 mg/L (reference range <10 mg/L). Urea and electrolytes were normal. A mild transaminitis was observed.

Differential diagnosis

He was treated for sepsis initially, however he failed to make any clinical improvement despite IV antibiotics and his blood cultures were negative. Failure to respond to antibiotics in the context of hepatosplenomegaly led clinicians to think of HLH. An abdominal ultrasound showed hepatosplenomegaly and heterogeneous infiltrates of spleen.

In view of his hepatosplenomegaly and pancytopenia, bloods were sent to investigate for HLH. More than five of HLH diagnostic markers^{4 5} were abnormal, including a hyperferritinaemia of 12 900 µm/L (reference range 23–540 µm/L), hypertriglyceridaemia of 5.38 mmol/L (reference range 0–2.0 mmol/L), a high soluble CD25 (soluble interleukin-2 receptor alpha) of 10 600 U/mL (reference range <2500 U/mL) and a low fibrinogen (0.7 g/L; reference range 1.5–4.0 g/L) (see box 1).

Box 1. Revised Diagnostic Guideline for HLH adapted from Henter ^{4 5} .

1. A molecular diagnosis consistent with HLH is made or

2. Diagnostic criteria for HLH are fulfilled (5 of the eight criteria below):

   • Fever >38.5*

   • Splenomegaly*

   • Cytopenias (affecting 2–3 lineages in the peripheral blood): haemoglobin<90 g/L, platelets<100× 10⁹/L, neutrophils<1.0× 10⁹ /L)*

   • Hypertriglyceridemia* and/or hypofibrinogenemia*: fasting: (triglycerides>3.0 mmol/L, fibrinogen<1.5 g/L)

   • Haemophagocytosis in bone marrow*, spleen, or lymph nodes

   • Low or absent NK-cell activity (according to local laboratory reference)

   • Ferritin>500 g/L*

   • Soluble CD25 >2400 U/mL*

fHLH was excluded on the basis of his immunological work up showing normal lymphocyte subsets (T, B, NK cells) and normal expression of Perforin SAP, XIAP and a normal granulocyte release assay. In addition, there was also no evidence of Griscelli or Chediak-Higashi syndrome on hair microscopy. Next generation sequencing for a panel of 71 genes associated with primary immunodeficiency disorders and HLH disorders was also negative.

As fHLH had been excluded, infectious agents causing sHLH needed to be investigated. Serology was negative for: herpes simplex viruses (HSV) 1+2, human herpes virus 6 (HHV6), hepatitis A, B and C, cytomegalovirus (CMV), Epstein-Barr virus (EBV), measles, varicella zoster virus and HIV 1+2. Extended panel nasopharyngeal aspirates were negative. PCR of the patient’s serum was also negative for: CMV, adenovirus, EBV, HSV1 +2, HHV6, VZV, parvovirus B19 virus and HIV. Bacterial serology was negative for syphilis, Toxoplasma gondii and Mycoplasma spp. An extensive imported fever panel showed no evidence of spotted fever, typhus, Coxiella burnetii and Brucella spp. Microscopy and culture from his cerebrospinal fluid did not show any growth. Acid fast bacilli were not seen and Mycobacteria not isolated from his gastric aspirates.

There was a high clinical suspicion that VL could be driving his sHLH, as it is known to be endemic in Madrid and no alternative causative infectious agent had been found. Microscopy of his first BMA showed no blasts, megakaryocytes granulopoiesis showed all stages of maturation and adequate erythropoiesis. Mild haemophagocytosis was noted. No abnormal infiltrates and no Leishmania donovani amastigotes were noted. Bone marrow immunophenotyping was unremarkable. Bone marrow trephine was not performed. Microscopy of a second BMA did not show haemophagocytosis or Leishmania amastigotes and Leishmania was not cultured.

Results from skin swabs taken from the papules on his legs for bacterial microscopy and culture were negative. The pustular papules were not biopsied and self-resolved.

Treatment

He was initially treated for sepsis, but he failed to improve with intravenous ceftriaxone. Ceftriaxone was stopped when his blood cultures were negative at 48 hours. His tachycardia at presentation was responsive to fluid resuscitation.

VL was suspected from postnatal travel history, notably to Madrid, alongside exclusion of alternative diagnoses. While awaiting diagnostic confirmation, dexamethasone 10 mg/m²/d was commenced on day 5 of admission as per the HLH-94 protocol.^{4 5} Due to a growing concern that this could be VL, in view of the travel history, etoposide was omitted and intravenous liposomal amphotericin B 3 mg/kg/day was empirically started. He made an excellent clinical improvement. He received 5 days of daily liposomal amphotericin B⁶ followed by further single doses on day 14, day 21, day 28, day 35 and day 42. He was afebrile by day 2 of therapy with ambisome and dexamethasone. His ferritin had normalised by day 5. Steroids were weaned (2 weeks in total). Pancytopenia and hepatosplenomegaly steadily resolved and FBC was normal 1 month after starting treatment.

Outcome and follow-up

On day 12 of admission the BMA PCR results detected Leishmania donovani complex DNA. After the patient was discharged the Leishmania K39 antibody serology was reported as positive with a Direct Antigen Test +1:25 600. He remains well, with no recurrence at 12 month follow-up.

Discussion

Patients with HLH present as systematically unwell with fever and hence are typically treated for sepsis. The clinical picture progresses and is characterised by hepatosplenomegaly, pancytopenia and significantly deranged ferritin, fibrinogen, triglycerides and soluble CD25. A diagnosis is made when >5/8 of the HLH criteria are met^{4 5} as in our case (see box 1). Hyperferritinaemia>10 000 is the most sensitive marker.¹

sHLH is diagnosed when there are no identifiable changes in the familial genes causing fHLH. sHLH is often associated with infections. fHLH can also be triggered by infection. Parallel immunological and infectious disease work up are critical to distinguish between fHLH and sHLH. This is important as fHLH requires significant immunosuppression and hematopoietic stem cell transplant when an underlying genetic mutation is identified. Whereas an underlying condition, including a possible infection, needs to be sought and treated in sHLH.

The pathogenesis of sHLH remains unclear. Both the dysregulation of CD8 +T lymphocytes and non-malignant proliferation of macrophages lead to uncontrolled hypercytokinaemia.^{1 7 8} This, in turn, causes migration of these inflammatory cells and organ infiltration causing cell necrosis and multi-organ failure.

The most common protozoal infection associated with sHLH is visceral leishmaniasis. Leishmania spp. are flagellated protozoa that are transmitted by the Phlebotomine sandfly. Leishmania spp. (including Leishmania donovani) cause VL which typically presents with fever, weight loss and hepatosplenomegaly; a collection of signs and symptoms with considerable overlap with HLH.

VL is a public health problem within South East Asia, Brazil, East Africa and the Mediterranean basin. Leishmania spp. have >70 animal reservoirs including a dog reservoir in the Mediterranean basin⁹ and in wild hare in Madrid.² Despite the international importance of VL, cases of VL and sHLH are rare and can be overlooked in non-endemic countries such as the UK. This case highlights the importance of asking parents detailed travel history including non-tropical destinations. The authors also note that seeking advice from colleagues in haematology, immunology, microbiology and infectious disease is key in such cases.

Attempts to diagnose VL in the context of sHLH are hampered by many challenges. First, microscopy of splenic aspirations have a high specificity, but risk haemorrhage and hence are rarely used in children.⁷ Second, haemophagocytosis may be seen on the BMA but is not always present and is a non-specific finding.^{2 7 10} Furthermore, the first microscopy of bone marrow does not reveal Leishmania amastigotes in 36%–64% of cases due to the pauci-microbial nature of this infection.^{1 9–11} Repeating BMA leads to an increased diagnostic yield.^{7 10} Some patients require up to 5 BMAs to reach a diagnosis.¹⁰ Blázquez-Gamero et al.¹ found bone marrow culture was positive for Leishmania in 30% of 24 children with VL. However, this systematic review of children in Spain found Leishmania PCR in serum was positive in 83% children and in bone marrow in 100%.² Of these 24 VL cases in Spain, 41% also had clinically confirmed HLH.²

Serology for Leishmania may also be negative^{1 2 10} in up to 50% of patients, with some seroconverting during the clinical course.⁹ If available, the newer rapid diagnostic anti-rk39 ELISA rapid test has a high sensitivity (97%) and sensitivity (98%).^{10 11} The median time to diagnosis is 1–2 months^{7 8 10} with some diagnoses delayed until 134 days⁹ or postmortem. The rk39 ELISA rapid test has the capacity to make a difference to patient outcome, as sHLH with untreated VL is invariably fatal.¹⁰

Current guidelines for sHLH involve immunosuppression with dexamethasone and etoposide.^{2 4} However, this can aggravate VL and contribute to poor outcome when untreated infections are present.^{8 9 11} Some cases have been treated safely with liposomal amphotericin B alone without dexamethasone.⁸ Liposomal amphotericin B is less toxic and more effective than previous alternatives, for example, sodium stibogluconate^9–11 as it has a preferential uptake in the reticuloendothelial system.¹⁰

VL should be considered in cases of HLH who have visited endemic countries. This infant’s presentation was typical of children with VL-driven HLH.⁹ If clinical suspicion is high, empirical liposomal amphotericin B should be started while awaiting diagnostics. Early consideration of VL allows the potential of rapid testing and to receive liposomal amphotericin B, thus avoiding intense immunosuppression.

Patient’s perspective.

Leishmaniasis is not a well-known condition in this country. Yet despite the disease being common in some of the most popular holiday destinations for Britons, there seems to be very little knowledge of it. We were extremely lucky that our baby was treated at a world-renowned hospital by a very capable team focused on infectious diseases. And they struggled; the first test for leishmaniasis came negative, as dozens of other diseases were investigated. Still, they had a strong suspicion and requested a PCR that finally came back positive. Thanks to that result our child did not receive the chemotherapy already prepared for him as his only diagnosis up until then was haemophagocytic lymphohistiocytosis (HLH). We knew the doctor had in mind similar cases where leishmaniasis was not recognised and treated early resulting in tragic outcomes. The team at Great Ormond Street Hospital  were determined to prevent that from happening again. Knowledge and perseverance saved our boy. We hope the drama we went through can ultimately help medical professionals to test patients with similar symptoms for leishmaniasis, and not stop when HLH is diagnosed. (Written by next of kin)

Learning points.

• Importance of a thorough travel history, including non-tropical destinations such as Spain where visceral leishmaniasis (VL) is endemic.

• A high index of suspicion, and thorough and timely investigations are required to diagnose VL. Non-endemic countries should therefore have access to rapid, sensitive and specific VL diagnostics.

• Importance of rapid screening for primary haemophagocytic lymphohistiocytosis.

• Our patient had clinical improvement with liposomal amphotericin B and high dose dexamethasone.

• Increased awareness of this rare but potentially fatal disease will avoid morbidity and mortality from missed diagnoses.