Elsevier

The Lancet

Volume 348, Issue 9027, 31 August 1996, Pages 573-577
The Lancet

Articles
Paroxysmal nocturnal haemoglobinuria: long-term follow-up and prognostic factors

https://doi.org/10.1016/S0140-6736(95)12360-1Get rights and content

Summary

Background

Paroxysmal nocturnal haemoglobinuria (PNH) is a rare acquired disorder of haematopoietic stem cells. Although knowledge about the pathophysiology of the disease is increasing, no multivariate analysis of factors influencing survival has been undertaken, mainly because the disease is rare. We undertook such an investigation.

Methods

Data were collected on 220 patients with PNH diagnosed over a 46-year period (1950–1995) from participating French centres. Diagnosis of the disease required, at least, an unequivocally positive Ham's test.

Findings

The Kaplan-Meier survival estimate was 65% (SE 4) at 10 years and 48% (6) at 15 years after diagnosis. 8-year cumulative incidence rates of the main complications (pancytopenia, thrombosis, and myelodysplastic syndrome) were 15% (3), 28% (4), and 5% (2), respectively. Demographic data, presenting features, initial treatment, complications, and causes of death were similar to those previously reported. In multivariate analysis, seven factors were significantly associated with survival in patients with PNH. Poor survival was associated with the occurrence of thrombosis as a complication (relative risk 10·2 [95% CI 6-17], p<0·0001), evolution to pancytopenia (5·5 [2·8-11], p<0·0001), myelodysplastic syndrome or acute leukaemia (19·1 [7·3-50], p<0·001), age over 55 years at diagnosis (4 [2·4-6·9], p<0·0001), need for additional treatment (2·1 [1·3-3·6], p<0·003), and thrombocytopenia at diagnosis (2·2 [1·3-3·8, p<0·003). Better survival was shown for patients in whom aplastic anaemia antedated PNH (0·32 [0·14-0·72], p<0·02). Factors associated in multivariate analysis with a high risk of thrombosis during the disease course were thrombosis at diagnosis (5·1 [2·5-10·6], p=0·0002), age over 54 years (2·6 [1·5-4·6, p=0·0014), and infection at diagnosis (2·6 [1·3-5·2], p=0·0099). The risk factors for progression to pancytopenia were absence at diagnosis of anaemia (4·03 [1·3-12·2], p=0·03) and neutropenia (2·45 [1·1-5·7], p=0·03). The risk factors for development of myelodysplastic syndrome or acute leukaemia were abdominal pain crisis at presentation (10·5 [2·5-44·0], p=0·004) and year of diagnosis after 1983 (8·45 [1·8-40·7], p=0·004).

Interpretation

This large number of cases permitted a detailed analysis of prognostic factors for the first time, in this rare disease. Estimates of PNH prognostic factors may serve as baseline data in the assessment of current and future treatments for this disease.

Introduction

Paroxysmal nocturnal haemoglobinuria (PNH) is an acquired chronic haemolytic anaemia, often associated with recurrent nocturnal exacerbations, neutropenia, thrombocytopenia, and episodes of venous thrombosis.1, 2 The biological distinguishing characteristic of PNH cells is the deficiency in several types of cell-surface proteins, which have the common structural feature of being anchored to the cell surface by a glycosyl-phosphatidylinositol (GPI) moiety.3 A gene, PIG-A, has been identified by expression cloning,4, 5 somatic molecular alterations of which cause the defect in GPI-anchored proteins in all PNH patients studied so far.7

The clinical course of PNH is highly variable.2 Although allogeneic bone-marrow transplantation is the only available curative therapy for PNH,2, 7, 8, 9 this procedure is itself associated with substantial morbidity and mortality. Advances in the knowledge of the molecular mechanisms of PNH have raised the possibility of a gene therapy approach. However, most of the reported clinical studies were based on a limited number of patients. Because the disease is so rare there are still no prognostic criteria to define indications for intensive therapeutic measures. Our study, of the largest group of patients with PNH to date, aimed to describe the long-term outcome of the disease and to search for prognostic factors that affect survival.

Section snippets

Patients and methods

The population of patients consists of two parts. Data on some of the first group were reported in 1985.10 From this group (1950–81) we decided to include patients who met the following criteria: a positive Ham's test reported by a physician from a French centre and demographic data available to allow clinical follow-up. The second source of data was from a questionnaire sent to all members of the French Society of Haematology in November, 1994, which requested information on all patients with

Results

The median age at presentation was 33 years (range 6-82). Initial biological and clinical symptoms are summarised in table 1. Among the 204 (93%) patients with peripheral blood abnormalities 196 had anaemia (median haemoglobin concentration [known in 189 patients] 80 g/L [range 4–12]), 90 patients had neutropenia (absolute neutrophil count [known in 218 patients], of 0·9×109L [0·1–1·5], and 112 had thrombocytopenia (platelet count [known in 219 patients], 0·4×1011/L [0·01–1·35]).

Red-blood-cell

Discussion

This retrospective study involved 220 patients with PNH diagnosed in France between 1950 and 1995, the largest number in any published study. As far as we know, no previous study7, 8, 9, 11, 16, 17, 18, 19, 20, 21, 22, 23 provided statistical estimates of prognostic criteria, mainly because of the rarity of the disease. The demographic data of our patients are similar to those of other large studies.11, 18, 19 PNH is mainly a disease of adults, however, as described by Ware and colleagues,22

References (28)

  • JV Dacie et al.

    Paroxysmal nocturnal hemoglobinuria: clinical manifestation, haematology, and nature of the disease

    Series Haematol

    (1972)
  • R Peto et al.

    Design and analysis of randomized trials requiring prolonged observations of each patient

    Br J Cancer

    (1977)
  • DP Byar

    Identification of prognostic factors

  • DG Altman et al.

    Practical problems in fitting a proportional hazards model to data with updated measurements of the covariates

    Stat Med

    (1994)
  • Cited by (423)

    View all citing articles on Scopus

    Study participants listed at end of paper

    View full text