Haemophagocytic lymphohistiocytosis associated with immune checkpoint inhibitors: a descriptive case study and literature review

Abstract

Immune checkpoint inhibitors (ICI), such as anti-PD-1 and anti-CTLA-4 antibodies, have proven effectiveness in treating many cancers; thus, their field of action is rapidly expanding.1 These immunotherapies enhance T cell activity, which can lead to immune-related adverse events (irAEs) in 70–90% of patients, according to the ICI used.2 Haemophagocytic lymphohistiocytosis (HLH) is a rare and severe condition characterised by T cell hyperactivation that can be measured by soluble IL-2r (CD25).3 It may be of primary or secondary origin, due mainly to infections, cancers or autoimmune disorders. Patients with HLH usually present with recurrent high fever and pancytopenia, and blood tests show hyperferritinaemia, hypertriglyceridaemia and hypofibrinogenaemia. Features of haemophagocytosis can be seen on bone marrow smear examination. HLH is generally life-threatening, with a mortality rate of up to 66%.4 Diagnostic criteria are well-established for children (HLH-2004),5 and a probability score (HScore) was recently proposed for adults.6, 7 In this study, we investigated all cases of ICI-associated HLH reported in three large pharmacovigilance databases in France. We aimed to extend the clinical description of ICI-associated HLH, its outcome and management. In this descriptive observational study, we included all adult (over the age of 18) patients experiencing HLH associated with ICI treatment, identified from three pharmacovigilance databases in the Parisian region in France. We then performed a literature review of ICI-associated HLH. Materials and methods are detailed in supplementary data. We identified 5 patients who experienced HLH on ICI treatment in the databases. The description of these cases is available in supplementary data. The literature review identified another 15 cases of HLH occurring during immunotherapy. The main characteristics of these cases are summarised in Table I. Most of the 20 cases were men (n = 13), and the mean age of the patients was 57+/− 17 years. Immunotherapy was prescribed for metastatic melanoma (n = 10), lung cancer (n = 4) or other cancers (n = 6). The immunotherapy prescribed consisted of antibodies against PD-1/PD-L1 (n = 13), CTLA4 (n = 2) or both (n = 5). The mean time between the first infusion and the onset of HLH was 18 weeks (range: 5 days to 1 year). One patient was treated concomitantly by radiotherapy, one with targeted therapy (dabrafenib, trametinib) and one with eribulin. Eight patients developed other immune-related toxicities, such as hepatitis (n = 4), meningitis (n = 2), hypophysitis (n = 2), colitis (n = 1), pneumonitis (n = 1), auto-immune haemolytic anaemia (n = 1), thyroiditis (n = 1) Stevens-Johnson syndrome (n = 1) and flare of known RA (n = 1). A complete (n = 3), or partial (n = 5) tumour response, stable disease (n = 4) or progression (n = 6) was observed. Satzger et al. Eur J Cancer. 2018 Nivolumab 1 mg/kg IV Q3W + ipilimumab 3 mg/kg IV Q3W × 4 cycles 5 months (1 month after cessation) Pneumonitis Stevens-Johnson syndrome Nivolumab NA for 9 months then stopped 3 months then ipilimumab NA × 1 cycle (progression) Pembrolizumab 2 mg/kg Q3W × 7 cycles Then dabrafenib (300 mg/day) and trametinib (2 mg/day) due to progression Dabrafenib Trametinib Pembrolizumab 200 mg Q3W × 5 cycles + eribulin mesylate 1·4 mg/m2 IV D1 and D8 Q3W Our study (also in Michot et al., Ann Oncol 2017), case #3 Pembrolizumab (adjuvant treatment) (3 cycles) then nivolumab 1 mg/kg Q3W + ipilimumab 3 kg/mgQ3W × 2 cycles (progression) The clinical and biological characteristics of the 20 patients are shown in Table II. 15 patients fulfilled the HLH-2004 criteria for HLH, and the mean of the available HScore was 243 ± 48 (mean probability: 94% ± 10%). The clinical symptoms were fever (18/20 patients), splenomegaly (12/20), hepatomegaly (6/20) and skin rash (5/20). Blood tests revealed bicytopenia (n = 9) or pancytopenia (n = 9) in most cases. The mean of the available haemoglobin concentrations was 8·5 g/dL and mean platelet count was 84 x 109/l. In patients with leukopenia, the mean leukocyte count was 2 085/mm3. Hyperferritinaemia (ferritin concentration> 500 µg/l) was observed in 20/20 patients (mean: 27 000 µg/l); hypofibrinogenaemia (<1·5 g/l) was observed in 8/12 patients; and an increase in LDH concentration to values >200 IU/l was observed in 8/8 patients (mean: 1 564 IU/l). A biopsy or blood smear was performed in 16 patients, revealing features of haemophagocytosis in 14 (88%) patients. Another possible cause of HLH was identified in 11 patients (including infection in three, progression of the cancer in six). All the patients received steroids. Six patients were treated with etoposide, two with Ig IV, two with cytokine blockade (tocilizumab in one case and anakinra in the other) and two received immunosuppressive therapy (mycophenolate mofetil, tacrolimus and cyclosporin). HLH was controlled by the treatment in 15 patients. Three patients died from HLH. Three patients were rechallenged with immunotherapy, with no recurrence of HLH. HLH outcome Cancer response Pancytopenia, hyperferritinaemia, hypofibrinogenaemia, ↗sCD25, haemophagocytosis (liver) Prednisone 2 mg/kg/d MMF 360 × 2 then 720 mg × 2/d Resolution of HLH Complete response for cancer MTP 1·5 mg/kg/d × 3 for 4 d Then prednisone 1 mg/kg Resolution of HLH Complete response for cancer Pancytopenia, hyperferritinaemia, hypofibrinogenaemia, ↗ sCD25, ↘NK cell function, haemophagocytosis (BM) Dexamethasone Etoposide [HLH-2004] Resolution of HLH Complete response for cancer Stabilisation of HLH Partial response for cancer Dead after 6 weeks (fungal pneumopathy) Cancer response: NA Systemic corticosteroid treatment Antibiotic treatment Resolution of HLH Cancer response: NA High-dose steroids, IV Ig, and anakinra Steroid pulse therapy then prednisolone Resolution of HLH Partial response for cancer High-dose steroids, daily plasmapheresis, then cyclosporin A, then etoposide 250 mg IV x 2, then tacrolimus monotherapy Resolution of HLH Complete response for cancer Resolution of HLH Partial response for cancer Resolution of HLH Partial response for cancer Resolution of HLH Partial response for cancer Bicytopenia, hyperferritineamia, hypofibrinogenaemia, ↗ sCD25, no BM examination Resolution of HLH Complete response for cancer Prednisolone 60 mg/d Second cycle of pembrolizumab Corticosteroids 500 mg IV then PO Etoposide 150 mg/m2 IV Then IV Ig 1 g/kg, then tocilizumab Partial control of HLH Cancer progression Yes Nivolumab 1 mg/kg Q2W + targeted therapy No (after 3 infusions) Prednisone 1 mg/kg/d Etoposide 150 mg/m2 IV Resolution of HLH Cancer progression Yes Nivolumab 3 mg/kg + Triptorelin + Bicalutamide No (after 1 infusion) Resolution of HLH Cancer response: NA We report here the first comprehensive clinical description of the patients who have experienced HLH associated with ICI. The rationale behind the use of ICI in cancer treatment is the restoration of cytotoxic T cell activity against tumour cells. In this context, ICI treatment may induce a deleterious hyperactivation of CD8 T cells, causing systemic irAEs. We confirm that HLH is a rare complication of ICI, as only one case was identified among the 745 patients included in one of the three pharmacovigilance databases, the French Reisamic prospective registry enrolling patients treated with anti-PD-1 or anti-PD-L1 at a single center between 2014 and 2019. As HLH is a rare disease, it may therefore be difficult to diagnose, particularly in patients presenting several concomitant irAEs, hence the importance of systematically checking for other toxicities. Diagnosis is based on a set of clinical and biological evidence, and could be helped with HScore.6 A quick decline of cytopenia with fever and high ferritin level should lead doctors to discuss HLH, perform bone marrow cytology to look for haemophagocytosis pictures (which could be lacking) and start treatment. Treatment with immune checkpoint inhibitors is an emerging cause of HLH, but it should be borne in mind that cancer patients may have other associated causes of HLH, such as an infectious disease or progression of cancer itself. An exhaustive infectious workup should be considered, as well as a cancer reevaluation, in order to focus on the most appropriate treatment. What is more, ICI-induced HLH might be associated with underlying genetic risk factors, as suggested by one case in a patient with a polymorphism in perforin-1 (PRF1) gene, a gene associated with HLH.8 Interestingly, a recent study showed that Nivolumab was effective for EBV-induced HLH by restoring the unbalanced anti-EBV program of the immune system, suggesting that anti-PD1 could play a dual role as both a potential trigger for HLH in malignancy as well as a potential treatment for EBV-HLH.9 Unfortunately, EBV status was often unknown in the case reports. HLH treatment is based on corticosteroids and etoposide in most cases of secondary HLH, together with aetiological treatment.5,7 For the treatment of HLH induced by ICI, the expert advice from the HLH Society recommends first-line treatment with corticosteroids and anti-IL6R antibodies, such as tocilizumab, with possible second-line treatment with etoposide (75–100 mg/m2) after 48 h in case of an insufficient response.7 In our study, only one patient with refractory-HLH was treated with tocilizumab, which was effective. Three patients were rechallenged with immune checkpoint inhibitors, with no recurrence of HLH, but with short durations of follow-up (less than 2 months). One limitation of this study is its retrospective nature, with a lack of information about corticosteroid dosage and tapering, or the time to HLH resolution, for example. Haemophagocytic lymphohistiocytosis is a rare but severe, life-threatening complication of checkpoint inhibitor therapy. Early diagnosis is important for HLH management and can be guided by the HScore. Tocilizumab is an interesting therapeutic option in steroid-refractory HLH associated with ICI. A.D. performed the research. A.D., J.M., O.L. designed the study. All authors contributed to the interpretation of study results. All authors critically revised the paper and approved the final version. We thank Dr Hiroe Kayatani, Dr Nausicaa Malissen, Dr Masood Sadaat and Dr Philipp Moog for providing additional information about their published cases. J.M.M. declares: principal/sub-Investigator of clinical trials: Abbvie, Aduro, Agios, Amgen, Argen-x, Astex, AstraZeneca, Aveo pharmaceuticals, Bayer, Beigene, Blueprint, BMS, Boeringer Ingelheim, Celgene, Chugai, Clovis, Daiichi Sankyo, Debiopharm, Eisai, Eos, Exelixis, Forma, Gamamabs, Genentech, Gortec, GSK, H3 biomedecine, Incyte, Innate Pharma, Janssen, Kura Oncology, Kyowa, Lilly, Loxo, Lysarc, Lytix Biopharma, Medimmune, Menarini, Merus, MSD, Nanobiotix, Nektar Therapeutics, Novartis, Octimet, Oncoethix, Oncopeptides AB, Orion, Pfizer, Pharmamar, Pierre Fabre, Roche, Sanofi, Servier, Sierra Oncology, Taiho, Takeda, Tesaro, Xencor; personal fees (monies paid for services rendered, such as honoraria, royalties or fees for consulting, lectures, speakers bureaus, expert testimony, employment, ad-boards, etc.): Celgene, Bristol-Myers Squibb, AstraZeneca, Janssen; non-financial supports (drugs, equipment supplied by the entity, travel paid by the entity, writing assistance, administrative support, etc.): AstraZeneca, Roche, Novartis, Gilead, Celgene, Bristol-Myers Squibb. C.L. declares: honoraria (Roche/BMS/Novartis/Amgen/MSD/Pierre-Fabre/Pfizer/Incyte), consulting or advisory board (BMS/MSD/Novartis/Amgen/Roche), speaker’s bureau (Roche/BMS/Novartis/Amgen), research funding (Roche/BMS), travel, accommodations, expenses (BMS/MSD), other: Avantis (Board). O.L. declares: paid expert testimony and consultancy fees (BMS France, MSD, Astra Zeneca), consultancy fees (Incyte), expert testimony (Janssen). B.B. declares: board and travel accommodation (BMS, MSD, Pierre Fabre). J.D. declares: travel accommodations (Pierre Fabre, Roche). None for A.D., L.F. or A.S. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts