Haemolytic paroxysmal nocturnal haemoglobinuria in patients with myeloid neoplasms: A rare association with specific therapeutic implications

Abstract

Paroxysmal nocturnal haemoglobinuria (PNH) is an acquired non-malignant disease of haematopoietic stem cells, associating haemolysis, bone marrow failure and thrombosis.1 PNH results from a somatic mutation in the phosphatidylinositol glycan anchor biosynthesis class A (PIGA) gene, which encodes an enzyme required to anchor protective molecules against complement to the cell membrane.2 The growth advantage of the PIGA-mutant clone is the result of an immune privilege in the context of immune-mediated bone marrow suppression,3 which is supported by the identification of glycosylphosphatidylinositol-specific T cells.4 Data from whole exome sequencing has revealed additional somatic mutations in ~50% of patients with PNH.5, 6 These mutations, described in the elderly and in myeloid neoplasms (MN), may precede or follow the genetic event responsible for PNH.5 The presence of small PNH clones (<10%) in patients with myelodysplastic syndrome (MDS)7, 8 is well known, as is the classical PNH evolution to MN (11.6% at 10 years).9 Conversely, concurrent or secondary diagnosis of haemolytic PNH in patients with MN is extremely rare1 and is confined to case reports in patients with myeloproliferative neoplasms (MPN).10-12 The purpose of the present work was to describe the disease characteristics and clinical course of patients with a concurrent diagnosis of PNH and MN, without evidence of pre-existing PNH. We retrospectively reviewed files of all patients referred to the French Reference Centre for Aplastic Anaemia and sent two waves of emails to >100 specialised physicians in France. Inclusion criteria were a concurrent or secondary diagnosis of PNH in patients with MN. To mitigate the risk of selecting patients with a clonal evolution of PNH, we excluded all patients with a previous history of haemolytic anaemia, haemoglobinuria or aplastic anaemia. Overall, we identified 49 patients with this suspected association. Figure S1 recapitulates the selection process that led to the choice of 20 patients meeting criteria for concurrent or secondary PNH diagnosis with MN. The study was conducted in accordance with the Declaration of Helsinki. The study was approved by the Institutional Review Board of the National Aplastic Anaemia Centre and the Local Ethics Committee and in accordance with French law, after patients had provided a non-opposition statement. Table 1 summarises the patients’ characteristics. The median (range) age at the diagnosis of MN was 65 (47–83) years. In all, 10 patients had MDS (50%), nine had MPN or MDS-MPN (45%) and one had acute myeloid leukaemia. For patients with MDS, the median Revised International Prognostic Scoring System (IPSS-R) score was 3. At MN diagnosis, before eculizumab initiation, all patients presented an elevated lactate dehydrogenase (LDH) level (median, 3.3 × upper limit of normal [ULN]; interquartile range [IQR] 2.1–4.7) and haptoglobin was undetectable in 15 of them. The median (IQR) reticulocyte count was 95 (23–122) ×109/L. Cytogenetics are described in Table 1. Among the eight patients with MPN, the Janus kinase 2 (JAK2) V617F mutation was found in five patients and the calreticulin (CALR) mutation in two. Figure 1 represents each patient's clinical history. For 11 patients, diagnoses of MN and PNH were concurrent: all had been referred for haemolytic anaemia with cytopenia, while three of them also presented evidence of thrombosis (patients numbers 2, 4 and 20). Four additional patients were diagnosed with PNH in the year following the diagnosis of MN, all with evidence of haemolysis and, for three of them, an associated episode of intra-abdominal thrombosis (two myelofibrosis and one MDS). For the remaining five patients, the median (range) delay between MN and PNH diagnosis was 85 (35–125) months. Three of them had a PNH diagnosis after an acute haemolytic event. The median (range) sizes of blood PNH clones, assessed by flow cytometry, were 78% (15%–100%) and 91% (20%–100%, two missing values) on granulocytes and monocytes respectively. Overall, 10 patients (50%) experienced at least one thrombotic event. This translates to a 2-year cumulative incidence of thrombosis since MN diagnosis of 42.8% (95% confidence interval [CI], 17.5%–66.1%). All patients presented evidence of haemolysis and of the 16 (80%) a total of 10 were or six became erythrocyte transfusion dependent. In all, 13 patients (13/16) were treated with eculizumab and all of them remained transfusion dependent. One additional patient (patient number 13) received eculizumab for thrombosis. The median (range) duration for eculizumab treatment was 18 (1–143) months. Remarkably, none of the patients experienced a thrombotic event during eculizumab treatment, but two patients (patients number 6 and 15) developed a thrombosis after eculizumab was discontinued. Reasons for not receiving eculizumab, despite transfusions needs, were early death from acute respiratory distress related to interstitial pneumonia for patient number 12 and very low life expectancy for two patients (patients number 9 and 14). In all, 10 patients required a MN-specific treatment without improvement of haemolysis and transfusion needs. Five patients underwent allogeneic haematopoietic stem cell transplantation (alloHSCT) after PNH diagnosis, of which four were alive in remission and without transfusion requirements at the last follow-up. After a median follow-up of 113 months, overall survival was 87.6 months (95% CI 38–not reached) and 87.6 months (95% CI 16–not reached) from MN and PNH diagnosis respectively. In total 10 patients died. The association of small PNH clones with myeloid malignancies is well known and reported; here, we report the impact of large haemolytic PNH clones on the clinical manifestation and outcome of MN. Most patients presented with haemolysis, suggesting the initial presence of a PNH clone. As MN diagnosis was the first referral to a haematologist for all patients, we were not able to rule out a pre-existing clone. To avoid inclusion of ‘missed’ cases of PNH, we excluded all patients with previous symptoms compatible with PNH or aplastic anaemia, which traditionally lead to an earlier referral than MDS. However, the question of which mutation appears first, PIGA or MN mutation, remains central. Only clonal architecture analysis may answer this question and this retrospective work was not designed to address it. Shen et al.5 have hypothesised PIGA as a first- or second-hit mutation and co-existence of both PNH and MDS clones. In our cohort, we found a high incidence of proliferation-inducing mutations, especially JAK2 mutations, in patients with delayed diagnosis of PNH, which may explain an accelerated growth of a newly mutated PNH subclone.10 In addition, when comparing molecular abnormalities found in our cohort with traditional clonal evolutions of PNH,9 we noticed that there was almost no chromosome 7 abnormalities but an increased proportion of MPN in our cohort (40% vs. 1.1%), which may confirm a very different population. Other selection bias might have occurred. Notably, we may have missed cases, as email queries identified five more patients, suggesting that some patients are not reported to the reference centre. Half of the 20 patients presented a thrombotic event. Albeit not statistically comparable, this result seems to be superior to the reported incidence of thrombosis in traditional PNH13 (10-year cumulative incidence of 30.7%, 95% CI 25.4%–35.9%), in MDS with small clones8 or in MPN14 (<10%), thus identifying a higher-risk population with the addition of both MN and PNH thrombosis risks. All of them had evidence of haemolysis at the diagnosis of their MN highlighting the contribution of haptoglobin, peripheral blood film and LDH in the screening of these patients. Therefore, systematic search for a PNH clone might be recommended in case of haemolytic anaemia or thrombosis in patients at MN diagnosis, even in patients with MPN. Remarkably, MDS IPSS-R was low, and most patients did not require immediate specific therapy. The median overall survival was >7 years, suggesting that MN, in our cohort, do not exhibit an aggressive behaviour. However, MN treatments were largely inefficient on transfusion needs, and response to treatment might then not be effectively assessed following current standards. Eculizumab therapy in eligible patients also did not show any obvious benefit on transfusion requirements. Nevertheless, no recurrence of a thrombotic event was observed in eculizumab therapy, which could support its use to prevent thrombosis in these high-risk patients. Finally, four of the five patients who underwent alloHSCT were alive without transfusion need at the last follow-up. This suggests that HSCT should be proposed for fit patients based on MN disease risk and that thrombosis prevention, with or without eculizumab, might be a key point of pre-transplant therapeutic management. Régis Peffault de Latour, Flore Sicre de Fontbrune and Aurélien Sutra del Galy designed the research study and analysed the data. All authors were in charge of patients, contributed to data collection and wrote the manuscript. Régis Peffault de Latour and Flore Sicre de Fontbrune have received research funding from, consulted for, and received honoraria from Alexion Pharmaceuticals. Edouard Forcade has received a travel grant and consulted for Alexion Pharmaceuticals. All patients have provided a non-opposition statement. Figure S1. 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.