Haemophagocytic lymphohistiocytosis in pregnancy and the postpartum period: A retrospective case series analysis

Introduction

Haemophagocytic lymphohistiocytosis (HLH) is characterised by the coexistence of immune dysregulation and unchecked inflammation, leading to haemophagocytic activity¹. It has seldom been reported in pregnancy. Clinically HLH manifests as relapsing fevers with features of multi-organ failure². The mortality rate from HLH is estimated at 50% even with optimal treatment³. Therefore, prompt investigation and diagnosis are essential for meaningful attempts at curative therapy.

Case presentation

This retrospective case series analysis uses all national maternal death reports reported to MBRRACE-UK attributed to HLH between 2012 and 2021 [n=5], and case notes from two patients diagnosed with HLH during pregnancy. The objective is to describe the presentation, diagnosis, management, and maternal and fetal outcomes of these cases.

Investigations

Treatment

After the diagnosis of HLH had been made, corticosteroids and ciclosporin were the most frequently used treatments. Dexamethasone was administered intravenously per the HLH-2004 protocol⁴, with an initial dose of 10 milligram(mg)/metre(m)² and subsequently tapered. Prior to definitive diagnosis, corticosteroids were administered orally or intravenously and included hydrocortisone, dexamethasone, prednisolone and methylprednisolone. The cytotoxic agent, ciclosporin, was administered orally at a dose of 6 mg/kg daily in two divided doses and titrated to achieve a target serum trough level of 200 microgram/litre(L).

Women also received intravenous (IV) antibiotics, antivirals and antifungals, either in response to proven or clinically suspected infection, or as prophylaxis due to immunocompromise. Antibiotic medications included: co-amoxiclav, teicoplanin, gentamicin, meropenam, piperacillin/tazobactam, metronidazole, vancomycin and amikacin. Prophylactic co-trimazole cover against Pneumocystis jiroveci was prescribed for two women. Antivirals included: aciclovir, tenofovir disoproxil, ganciclovir and lamivudine. Antifungals utilised were fluconozole and caspofungin. Choice and dose of antimicrobials varied according to clinical judgement and local policy. Good response to treatment of underlying infection was achieved in one case.

IV immunoglobulins were prescribed for three women (0.5 milligrams/kilogram IV once every 4 weeks), and one woman was treated with Anakinra (100 milligrams daily by subcutaneous injection). Due to the severity of the maternal condition and failed attempts at other therapies, early delivery by caesarean section was undertaken in two women (between 24–32 weeks of gestation) and termination of pregnancy in one case. Delivery or termination permitted aggressive maternal treatment with agents that are considered unsafe during pregnancy including etoposide and other cytotoxic agents. Etoposide was administered IV at a dose of 150mg/m². Allogenic haemopoietic stem cell transplantion was undertaken in one case post-delivery.

Outcome and follow up

Admission to level three critical care units was necessitated for six women, with the duration of stay ranging from 1 day to 231 days. These women required multi-organ support including invasive ventilation, haemofiltration and inotropic medications. Sadly, five women progressed to multi-organ failure resulting in maternal death; two women survived. The recorded causes of death were multi-organ failure [n=2], acute respiratory distress syndrome (ARDS) [n=1] and sepsis [n=1]; all of which were associated with concomitant HLH.

There were two livebirths at term, two preterm livebirths, one woman had a termination of pregnancy and one a spontaneous miscarriage. One surviving infant, whose mother had received Anakinra, an interleukin inhibitor, during pregnancy, had bone marrow suppression with anaemia at birth and was later diagnosed with sensorineural deafness. Given the timing of the development of the fetal anaemia, it is possible that the bone marrow suppression may have been a result of the Anakinra, although it is impossible to rule out other causes, such as HLH itself, given the lack of published data in pregnancy.

Discussion

HLH is an immune disorder which results in an ‘inflammatory storm’. Known triggers include infection, malignancy, rheumatological and autoimmune conditions. It can also occur spontaneously and it has been suggested that it can be precipitated by pregnancy itself⁵. The women in this series uniformly presented with pyrexia after a prodromal phase of variable duration. All fulfilled the diagnostic criteria for HLH and six of the women were admitted to level 3 critical care with multi-organ failure. Successful treatment is reliant upon early diagnosis and suppression of the excessive inflammatory reaction.

Due to the rarity of the condition, particularly during pregnancy, diagnosis is often delayed⁶. With diagnostic delays meaning the disease may be in advanced stages at diagnosis. In addition to the need to consider the impact of treatment on the fetus, there can be other dilemmas in the management of these cases leading to prolonged delays in initiating effective treatment⁷. Early multi-disciplinary team (MDT) involvement can facilitate prompt formulation of appropriate management plans. Additionally, in view of the high mortality, clinicians should consider HLH early when reviewing pregnant patients with unexplained pyrexia and multi-organ dysfunction, with a low threshold for additional investigations.

Reaching a correct diagnosis is an obstetric challenge. The blood results of HLH are nonspecific⁸. Hyperferritinaemia with a ferritin above 500ng/ml in adults is a non-specific diagnostic feature, which can be found in many disease conditions, but a ferritin level >10,000ng/L is significantly higher than would be expected in inflammatory syndromes or sepsis⁸. Pronounced hyperferritinaemia in the setting of systemic signs and symptoms along with a negative infectious and rheumatological workup should raise suspicion of HLH⁹. Where clinical suspicion is high, treatment should be commenced, as delaying treatment while awaiting results of investigations to confirm the diagnosis may be detrimental⁶.

The women presented in this paper were extensively investigated due to diagnostic uncertainty, and in the majority of instances it was bone marrow biopsy which confirmed the diagnosis of HLH. Histological examination of lymphoid tissue, such as bone marrow, lymph nodes, liver, or spleen, may show haemophagocytosis, which describes the engulfment of hematopoietic cells by activated macrophages acting outside of usual immune system regulations¹⁰. However, despite the fact that haemophagocytosis is prominently featured in the name of this disease, early on in the natural history of the disease, biopsies may be normal or demonstrate nonspecific findings such as increased or decreased unilineage or multilineage hematopoiesis; haemophagocytosis may not be present until later in the disease process. As in one of our reported cases, repeated bone marrow biopsies may be necessary to demonstrate the characteristic findings and repeated samples should be considered where clinical suspicion is high, yet the diagnosis remains uncertain. Early involvement of haematology expertise should be sought, as prompt diagnosis is crucial for meaningful attempts at curative therapy. These data do not allow any inferences to be made about why some women died when others survived. However, from the general literature, an early diagnosis followed by prompt treatment is of benefit to patients.

Other investigations more specific to HLH are detectable earlier in the disease process. Mayama et al. suggest that measuring soluble CD25 levels is the most specific diagnostic criteria for HLH, due to the direct correlation with increased T-cell and natural killer (NK) cell activity¹¹. Soluble CD25 levels were not measured in three of the cases presented in this paper, and we suggest that this investigation should be considered in any case where HLH is suspected.

Adult-onset Still’s disease (AOSD) is an autoimmune condition that can predispose patients to developing HLH¹². Therefore, the opinion of an expert rheumatologist is valuable when assessing patients suspected of having HLH. In this series of cases, one patient had a diagnosis of AOSD prior to becoming pregnant, and another was diagnosed with AOSD after being treated for HLH. One woman experienced multiple episodes of sepsis following a mid-trimester miscarriage. Despite aggressive treatment with antibiotics and immunosuppressive therapy she died.

A small number of published case reports have highlighted the association between AOSD and HLH in pregnancy. Dunn et al.¹³ reported a case of a woman in her early 40’s with history of Still’s disease, who was admitted with high grade fever at 19 weeks’ gestation and subsequently diagnosed with HLH. This patient responded well to steroids and the disease remained in remission for the rest of the pregnancy. Peters et al.¹⁴ reported a case report of new-onset adult-onset Still’s disease in a woman in her mid-20’s at 35 weeks’ gestation, which presented as macrophage-activation syndrome. Her condition improved with immunosuppressive therapy and the fetus was delivered prior to treatment. There are also two cases reported of neonatal HLH associated with maternal AOSD^15,16.

Treatment of HLH in pregnancy poses a number of challenges. There are several important considerations including fetal viability, maternal condition, treatment toxicity and identification of any triggers towards which specific treatments can be directed. There is no universally agreed treatment for HLH in pregnancy, controlling and removing the underlying cause is a priority. Infections should be treated initially with broad spectrum antibiotics, narrowing to bacteria specific when sensitivities are available. Similarly, rheumatological and autoimmune conditions and lymphoid malignancies should have disease specific treatment. Supportive treatment with blood products is also often necessary. Treatment options directed at HLH include IV and oral steroids, IV immunoglobulins (IVIG), cytotoxic drugs, Etoposide, Anakinra (a recombinant IL-1 receptor antagonist) and highly active antiretroviral therapy (HAART)^3,17,18. HLH has also been treated with bone marrow transplant (BMT)³.

Although Etoposide is both embryotoxic and teratogenic, in catastrophic cases it has been suggested that the benefit likely outweighs the potential harm to the fetus and should be strongly considered⁶, and favourable outcomes after etoposide treatment during pregnancy have been reported^7,19–22. In the cases presented in this paper, early delivery of viable fetuses to allow maternal treatment with Etoposide was favoured over the risks of treatment during pregnancy.

One case received an allogenic BMT following delivery of her infant, but despite an initial promising response continued to suffer multiple septic episodes, recurrent gastrointestinal bleeds, acute kidney failure and profound electrolyte disturbances. This patient eventually died from ARDS.

There are limited studies showing long term effects of Anakinra on the fetus if used during pregnancy, although there are reports of successful use of Anakinra with favourable fetal and maternal outcomes^23,24. Anakinra was used in one of our patients with a favourable maternal response, however the fetus had bone marrow suppression at birth and long term follow up of the baby revealed sensorineural deafness. It is thought this was due to the use of Anakinra, but other causes are possible, including HLH itself.

In summary, it is important to consider HLH as a potential diagnosis when reviewing pregnant patients with unexplained pyrexia and multi-organ dysfunction to avoid delays in diagnosis and treatment initiation. Early MDT involvement can facilitate prompt diagnosis and identification of underlying causes. Treatment should be directed against the underlying cause where one is identified, and the risks of treatment toxicity need to carefully weighed against the benefits. Early delivery may permit aggressive maternal treatment in catastrophic cases.

Conclusions

Haemophagocyctic lymphohisitiocytosis should be considered in the differential diagnosis in any pregnant woman who presents with a relapsing fever and incipient or overt multi-organ failure. Although identification of haemophagocytosis on a bone marrow biopsy is diagnostic, this may only appear later in the course of the disease, therefore repeated biopsy may be necessary. Any underlying causes in secondary HLH must be treated promptly and effectively. Management of women with HLH requires a multi-disciplinary approach, with involvement of obstetrics, rheumatology, haematology and critical care. Long term follow-up is required to ensure the continued health of the mother and to detect any adverse consequences of the disease, or treatment, in the children.