Changing prognosis in Paroxysmal Nocturnal ... - Wiley Online Library

Changing prognosis in Paroxysmal Nocturnal Haemoglobinuria disease subcategories; an analysis of International PNH Registry

Accepted Article

Gérard Socié1, MD, PhD; Hubert Schrezenmeier2, MD, PhD; Petra Muus3, MD; Igor Lisukov4†, MD, Alexander Röth5, MD; Austin Kulasekararaj6, MD; Jong Wook Lee7, MD; David Araten8, MD; Anita Hill13, MD; Robert Brodsky9, MD; Alvaro Urbano-Ispizua10, MD; Jeffrey Szer11, MD; Amanda Wilson12, MD; and Peter Hillmen13, MD, PhD, on behalf of the PNH Registry 1

Hematology Transplantation, AP-HP, Hospital Saint Louis, University Paris VII Denis Diderot, and Inserm UMR 1160, Paris, France 2 Institute of Transfusion Medicine, University of Ulm Ulm and Institute of Clinical Transfusion Medicine and Immunogenetics, German Red Cross Blood Transfusion Service Baden-Württemberg-Hessen and University Hospital Ulm, Germany 3 Department of Hematology, Radboud UMC, Nijmegen, the Netherlands. 4 † Institution of Children Hematology and Transplantation, ( deceased), Russia 5 University Hospital Essen, Germany 6 Kings College Hospital, London, UK 7 Seoul St. Mary's Hospital, Korea 8 Laura and Isaac Perlmutter Cancer Center, NYU Langone Medical Center, New York VA Medical Center, USA 9 Johns Hopkins University Medical Center, USA 10 Hospital Clínic Barcelona, Spain 11 Royal Melbourne Hospital, Australia 12 Alexion Pharmaceuticals, New Haven, CT, USA 13 St James's University Hospital, Leeds, UK

Authors’ contribution GS designed the study, analysed the data and wrote the manuscript; HS, PM, RB, AUI, JS and PH are member of the board of the International PNH Registry; they participate to study design and analysed the data. AW performed the statistical analyses. All authors reviewed and amend the manuscript Conflict of interest disclosures GS, HS, PM, AH, AUI, JS and PH received consulting fees or speaker fees from Alexion Pharmaceutical Keywords; PNH, causes of death, PNH Registry Correspondence to: Professor Gerard SOCIE Head Division Hematology / Immunology /Oncology AP-HP, Hospital Saint Louis 1 Avenue Claude Vellefaux 75475 Paris CEDEX 10, FRANCE Email; [email protected]

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/imj.13160 This article is protected by copyright. All rights reserved.

Abstract Background: Paroxysmal nocturnal haemoglobinuria (PNH) is a rare disease. Although much progress has been made in the understanding of the pathophysiology of the disease, far less is

Accepted Article

known with respect to the clinical outcomes of the patients with PNH. Few retrospective studies provide survival estimates and even fewer have explored the clinical heterogeneity of the disease. Haemolytic and aplastic anaemia forms have been recognised as main disease categories with the haemolytic form being associated with worst prognosis by the largest studied cohort some years ago. Aims: To describe mortality and causes of death in PNH overall and by PNH Classification and to

evaluate risk factors associated with mortality. Methods: We analysed data on 2356 patients enrolled in the International PNH Registry with

multivariate analyses using time dependent covariates. Patients were classified into haemolytic, aplastic anaemia/PNH syndrome or intermediate PNH. Results: Overall, 122 (5.2%) patients died after enrolment, the incidence according to subcategories being 5.1%, 11.7%, 2.0%, and 4.8% for patients with haemolytic PNH, AA-PNH, intermediate, and insufficient data, respectively. Older age and decreased performance status also affected survival in multivariate analysis. Improved outcome of patients with haemolytic PNH suggests that eculizumab treatment in PNH may be associated with improved survival. Conclusions: A detailed analysis of clinical presentations and causes of death in patients with

PNH, overall and by disease subcategories provide evidence that, in the current era, patients with haemolytic PNH are no longer those who harbour the worst prognosis. This finding differs sharply from what has been previously reported.

_____________________________________________________________________________________

This article is protected by copyright. All rights reserved.

Introduction Paroxysmal nocturnal haemoglobinuria (PNH) is a rare, acquired disease of the hematopoietic

Accepted Article

stem cell characterised by chronic intravascular haemolysis and by marrow failure

1-6

. The

cellular abnormality in this life-threatening disease originates from a mutation in the phosphatidylinositol

glycan

class

A

(PIGA)

gene,

resulting

in

a

deficiency

of

glycosylphosphatidyl-inositol (GPI)-anchored complement regulatory proteins, including CD55 and CD59, from the surface of blood cells 7. Patients with PNH haemolysis experience a marked increased risk of thromboembolism (TE) which is the leading cause of death 8-10.

Retrospective analyses have reported that, despite best supportive care, the 10-year survival rate in patients with PNH ranged from 50% for patients diagnosed in the mid-20th century to over 70% in a more recent series 8-14. TE is the leading cause of mortality in patients with PNH,

accounting for between 40% and 67% of deaths with known causes 3,10,12-17. Patients with PNH also experience symptoms including fatigue, abdominal pain, headache, shortness of breath, dysphagia, and erectile dysfunction. These symptoms can be debilitating and significantly reduce the quality of life (QoL) of patients with PNH. PNH may develop in the absence of another bone marrow disorder or as a condition secondary to bone marrow disorders such as aplastic anaemia (AA) 4,17,18. The only potentially curative therapy for PNH is allogeneic bone marrow transplantation (BMT); however, this procedure is associated with substantial morbidity and mortality [reviewed in

4,14

] and, is not now considered an appropriate therapeutic option for most patients 16. A small

proportion of patients have been observed to experience a spontaneous remission of their disease usually many years after their initial diagnosis 19. Historically, management of PNH was


limited to the use of supportive measures such as blood transfusions and anticoagulation therapy. However, it has been reported that the risk of TE in patients with PNH remains high,

Accepted Article

even in patients who have no clinical evidence of TE or are receiving prophylactic anticoagulation, 11 which is itself associated with an increased risk of bleeding complications. In 2007, eculizumab (Soliris®, Alexion Pharmaceuticals, Inc., New Haven, CT, USA), a humanised monoclonal antibody that inhibits terminal complement activation, was approved for the treatment of patients with PNH. A phase II study (PILOT) and two phase III studies (TRIUMPH and SHEPHERD) demonstrated that eculizumab was well tolerated and provided a rapid, sustained, and clinically meaningful reduction in haemolysis, fatigue, and transfusion requirements, along with improved QoL

19-25

. Subsequent studies have reported that

eculizumab is associated with a 92% reduction in the risk of TE and suggested an improvement in patient survival 20,26.

The natural history of PNH is highly variable and has previously been investigated by retrospective analyses involving relatively small patient populations 8-14. The International PNH

Registry was implemented to gather comprehensive data on the natural history of PNH and the management of patients with PNH, and to evaluate the clinical outcomes of the disease in order to better understand its progression and variability on a global scale. The long-term aim of the registry is to improve diagnosis and therapeutic strategies, optimise patient management and outcomes, and enhance knowledge of the natural history of the disease 27. In this study our objectives were: 1) to describe mortality and causes of death in PNH; 2) to describe mortality and causes of death separately by PNH Classification and, 3) to evaluate risk factors associated with mortality in the PNH Registry


Methods The International PNH Registry is a prospective, non-interventional, observational study.

Accepted Article

Patients of any age with a clinical diagnosis of PNH (by any applicable diagnostic method and/or detectable PNH clone) were eligible for inclusion in the Registry, regardless of disease severity, comorbidities, or treatments (past, current, or planned). The Registry has been described in detail elsewhere27 and collects information from patients monitored in current medical practice irrespective of past, present, or future treatment. The Registry was approved by the institutional review boards (or equivalent) of participating centres and all patients provided written informed consent prior to inclusion. The Registry is sponsored by Alexion Pharmaceuticals, Inc., and is overseen by an independent executive committee of international PNH experts. Data captured in the registry include patients’ demographics, medical and treatment history, comorbid conditions, PNH clone size, disease characteristics and outcomes, symptoms, PNH-specific treatments, PNH-related events, morbidity, mortality, pregnancy, patient QoL, and health resource utilisation. Clinical data captured include lactate dehydrogenase (LDH) level, haemoglobin concentration, transfusion requirements, thrombotic events (identified using major adverse vascular event categories), physician-reported renal dysfunction, and other laboratory data. Patient medical information and study questionnaire data were collected at study enrolment and every six months thereafter. All patients in the PNH Registry as of December 1, 2014 were eligible for analysis. Patients were excluded if they were missing demographic data (enrolment date, date of birth, sex), or did not have any follow-up data after study enrolment. Patients were followed from disease start or


enrolment until the earliest occurrence of death, other reasons for Registry discontinuation, or the last follow-up assessment in the Registry.

Accepted Article

The PNH Classification variable was calculated at enrolment and based on history of bone marrow disease and thrombotic events, and the lab values within 6 months of enrolment. Patients were classified into four strata to allow comparisons with the largest series published today by the French Registry 8: •

Haemolytic PNH includes patients with 1) any history of thrombosis at enrolment, or 2) haemoglobin 1.5 x 109 cells/L or b) platelet count >120 x 109 cells/L

•

AA-PNH Syndrome includes patients with at least two of the following: 1) haemoglobin ≤ 100 g/L, 2) platelet count ≤80 x 109 cells/L, and/or 3) neutrophil count ≤1 x 109 cells/L

•

Intermediate PNH includes patients who did not fulfil the other strata

•

Insufficient data to be classified

The main endpoint for these analyses was mortality. Mortality was reported by the sites on the study CRF with a date of death and text to describe the probable cause of death. The lead author (GS) reviewed the text to categorise the cause of death for these analyses. The cumulative incidence of mortality was also estimated using the competing risks method of Fine and Gray (1999) with bone marrow transplant considered a competing risk. The Fine and Gray test was used to test for a difference in time to death between the different PNH Classifications 28

.


An unadjusted Cox proportional hazards regression analysis was used to estimate the hazard ratio and to test for a difference in time from disease start to death between the PNH

Accepted Article

Classifications. For this analysis, patients were followed from the date of disease onset. The data are considered left truncated, because survival time is calculated from the disease start date, so subjects do not enter the risk set until sometime after the time origin, i.e., they could not have died prior to the date of Registry enrolment. An adjusted Cox proportional hazards regression analysis of time to death was also performed with left truncated data and including eculizumab treatment as a time dependent covariate, and other covariates selected for inclusion in the model based on patient and clinical characteristics that might impact mortality, as well as the results of an exploratory stepwise model. The additional covariates included in the model were as follows: age at enrolment, Karnofsky score at enrolment (< 80, ≥ 80, or

unknown), fatigue at enrolment (yes or no), LDH ratio at enrolment (< 1.5, ≥ 1.5, or unknown), and BMT during follow-up (yes or no). All analyses were performed using SAS software version 9.2 (SAS Institute Inc., Cary, NC, USA), except that the competing risk Cox models were performed using SAS software version 9.4.

Results As of December 1st 2014 3,259 patients have been registered in the International PNH Registry. Of those 3,192 have complete demographic data. Patients with no follow-up after enrolment or presenting with associated disease other than AA were excluded, leading to a population of 2,356 patients who are the subject of this report (Supplementary Figure 1). Using the above


defined disease subcategory definition, 702, 375, and 691 patients were classified as haemolytic PNH, AA-PNH or intermediate, respectively. Due to some missing data 588 could not

Accepted Article

be strictly classified, however, 70% have a large PNH clone as assess by flow cytometry, and 75% were treated with eculizumab. Thus most patients with insufficient data are close to the haemolytic PNH category (Table 1).

Patient and disease characteristics at enrolment and their symptom history are summarised in Table 1. Patients with the AA-PNH syndrome tend to be younger and had a shorter follow-up

since disease diagnosis when compared to other categories. As expected from previous reports patients with AA-PNH had a smaller granulocyte clone size, and only 5% were treated with eculizumab. Conversely, abdominal pain, dysphagia and haemoglobinuria were more frequently reported in the haemolytic and intermediate subcategories. Table 2 summarises the outcome of this large population. The median follow-up since “disease start” was 9.4 years overall but significantly less for patients with AA-PNH. Overall 122 patients died (5.2%) after enrolment, the incidence according to subcategories being 5.1%, 11.7%, 2.0% and 4.8% for patients with haemolytic PNH, AA-PNH, intermediate and insufficient data, respectively. Table 2 also summarises other important outcomes including thrombotic events

after enrolment in the Registry, rate of transplantation and eculizumab use. As expected, more patients with AA-PNH underwent transplantation and more patients in the other categories were commenced on eculizumab after enrolment. Primary causes of death (Table 2) were categorised a priori depending on disease presentation, for example as transplantation-linked or aplastic anaemia-linked if associated with GvHD or haemorrhage respectively. (Supplementary file 1; provides detailed primary and secondary


cause of death for each patient). As a result most of the patients with AA-PNH who died were classified as dying of aplastic anaemia since nearly 60% of deaths were linked to haemorrhage

Accepted Article

or infection. Conversely in patients with other disease subcategories the causes of death were relatively heterogeneous with PNH being scored as primary cause of death in 11-25% mainly because of fatal thrombosis. Of note, malignant diseases accounted for a significant proportion of deaths with more than 8% due to leukaemia and another 8% from solid cancers. The diseasespecific cumulative incidence of mortality at 10 years (with transplantation considered as competing risk) was nevertheless relatively low (5.2%) but was significantly higher in patients with AA-PNH (18.36%) relative to all other disease subcategories (Figure 1, and Table 3). We finally analysed the risk of mortality from disease diagnosis to last follow-up by multivariate analyses (Table 4). In an unadjusted model, patients with AA-PNH had an inferior prognosis, while patients with intermediate disease severity had a better prognosis, compared to patients with haemolytic PNH. Those with insufficient data shared the same prognosis than those with classical haemolytic form. In an adjusted final model, however, only those with an intermediate form retained a better prognosis. The use of eculizumab was significantly associated with a better outcome. However, since eculizumab use was closely associated with disease subcategory we introduced interaction terms between the use of eculizumab and disease subcategory; this analysis again confirms a better outcome of the intermediate form than in those with haemolytic form, irrespective of eculizumab treatment. However, as described in Appendix 1 in the supplementary data, caution is required when interpreting these data given the number of patients per group in the interaction term.


Other significant risk factors for decreased survival included, older age at enrolment (p