British journal of Huemutology. 1 9 9 2 , 82, 358-367

Age-related incidence and other epidemiological aspects of myelodysplastic syndromes C. Aur., N.GATTERMANXA N D W. SCHNEIDER Department of Internal Medicine, Haematology and Oncology Division, Heinrich Heine University, Diisseldorf

Received 10 December 1991: accepted for publication 26 May 1 9 9 2

Summary. Although most haematologists perceive a rising prevalence and incidence of myelodysplastic syndromes (MDS), reliable epidemiological data on these disorders are largely lacking. The bone marrow register of the University of Diisseldorf allowed us to assess among other epidemiological features the incidence of MDS. which was compared to that of acute myeloid leukaemia (AML). Among a total of 18 4 1 6 different patients registered between 1975 and 1990, 584 cases of MDS ( 3 . 2 % ) and 506 cases of AML (2.8%) were identified. Over the study period. the percentage of newly diagnosed MDS rose from 1.3% to 4.50/,,while there was no upward trend for AML. Among all patients undergoing bone marrow biopsy, the proportion of those over 60 years of age increasedfrom41,9%in 1975 to 54.1Xin 1990. Wefounda strong correlation between the proportion of elderly patients and the relative frequency of MDS diagnoses. Thirty-one patients ( 5 . 3 % ) were classified as a secondary MDS because of previous treatment with cytotoxic chemotherapy and/or irradiation for a variety of malignancies. Twelve patients

were identified in whom occupational exposure to organic solvents could not be ruled out. For calculating age-specific incidence rates, the analysis was confined to the town district of Dusseldorf (575 000 inhabitants). because exact demographical data were available for this population. In the last quinquennium of the study period ( 1986-90), myelodysplastic syndromes were more frequent than AMI, in the age group 50-70 years ( 4 . 9 v 1~8/100000/year).In patients over 70, the incidence of MDS was more than 3 times that of AML (22.8 v 6 . 7 / 100 000). In this group, men had a higher incidence of MDS (33.9/100000) than women (18/100000). Crude annual incidence (all age groups) was also higher for MDS (4.1/ 1 0 0 0 0 0 ) than for AML (2.1/100000) in recent years. We conclude that MDS are relatively common haematological neoplasias. The rising incidence in recent years is probably not due to changes in aetiological factors, but may reflect increased awareness on the part of physicians and extended use of diagnostic procedures in elderly patients.

Despite an increasing number of investigations into the myelodysplastic syndromes. there is still a lack of reliable epidemiological data. In particular. only rough estimates are usually made of the incidence of these disorders. Early reports suggested that MDS were rare. For example Koath ~t nl (1964) failed to identify any preleukaemic phase in 580 leukaemia patients at Manchester Royal Infirmary in the early sixties. Meanwhile. the picture had changed. The French Registry of Acute Leukaemia and Myelodysplastic Syndromes, which covers 3 7 university centres. identified 579 new MDS patients in 198 3 in France (Groupe Franqais de Morphologie Hematologique, 1987). In the United States. Linman & Bagby ( 1 978) estitnated that about 1 5 0 0 cases of MDS occur annually. Although these and other studies

(Verwilghen & Boogaerts, 1987: Keizenstein & Dabrowski. 1991 ) indicate that MDS is more common than previously assumed. they have not yielded exact incidence rates. Large-scale epidemiological studies have been impeded by incongruent or even lacking definitions and classifications of MDS (Galton. 1986). As part of these difficulties, official morbidity and mortality statistics are inadequate tools for assessing the magnitude of the MDS problem. For instance, MUS is not listed in the ninth revision of the widely used International Classification of Disease (TCD) (World Health Organization. 1977). Instead, cases have to be assigned to various other diagnoses which are not always well defined. However. even if MDS were included in a revised version of the ICD code, this would not necessarily lead to a more reliable estimate of their frequency. In view of the wellknown difficulties in making the correct diagnosis. particularly during the early phases of disease, it cannot be expected that MDS patients are reliably identifed on a large scale. At

Correspondence: Dr C. A d . Department of Internal Medlcine. Haematology and Oncology Division. Heinrich Heine University. Moorenstrasse 5 , 4000 Diisseldorf 1. Germany.

3 58

Epidemiology of MDS present, accuracy of diagnosis and completeness of case registration seem to be confined to specialized registers such as regional cancer surveys or hospital-based statistics. Provided they can draw on a reference population that is demographically well defined, these registers can yield valuable epidemiologicaldata, including age-specific disease incidences. In this paper we present the epidemiology of MDS as reflected by the bone marrow register of the University of Dusseldorf. PATIENTS AND METHODS

Catchment area. The Dusseldorf bone marrow register draws on a population of approximately 1200 000. In order to obtain meaningful data on disease incidence, the catchment area must be recognized as having two different dimensions. On the one hand, there is the town district of Dusseldorf, a clearly defined area for which exact demographical data are available from the Statistical Office of the Land North Rhine-Westphalia. Some 575 000 inhabitants live in this region. As far as MDS incidence is concerned, we can state with confidence that hardly any patient from this area will be given a diagnosis of MDS without his bone marrow smear being seen in our laboratory. This is ensured by very close cooperation with all the hospitals in town. The few private physicians who perform bone marrow biopsies in their practice also avail themselves of our services. The town district of Diisseldorf constituted the reference population for which age-specific incidence rates of MDS and AML were calculated. The total catchment area, on the other hand, is much bigger than the town district. Among 43 hospitals in this area, the university medical centre has the only specialized haematology department and receives bone smears from virtually all cases of suspected haematological malignancy for cytological evaluation. Unfortunately, we are not able to define this wider area of patient recruitment in terms of administrative regions. This precludes the employment of demographical data and hence the calculation of incidence rates in this larger population. Patient recruitment and diagnostic criteria. The study covers a 16-year period from 1975 to 1990. Since 1983 our laboratory has used the French-American-British (FAB) classification of myelodysplastic syndromes (Bennett et aI, 1982). Retrieval of MDS cases before 1983 included the diagnoses of panmyelopathy with hypercellular marrow, sideroblastic anaemia, chronic myelomonocytic leukaemia (CMML) and smouldering leukaemia. MDS cases are entirely comparable over the study period of 16 years. This was achieved by reevaluation of all MDS bone marrow specimens from 1975 to 1990, with uniformity of morphological criteria applied. Subclassificationof MDS was made according to FAB criteria. Pure sideroblastic anaemia (PSA),which we recently characterized as a separate entity (Gattermann et aI, 1990). was included in the RARS category for the purpose of this study. Although PSA is perhaps not preleukaemia, its inclusion is in keeping with the FAB classification of MDS. Our minimal haematological criteria for the diagnosis of MDS were: (1) in the peripheral blood: single or multiple

3 59

cytopenias or just macrocytosis without anaemia: and (2) in the bone marrow: clear morphological evidence of dysplasia in one or more cell lines (this criterion being fulfilled by one or more of the following: micromegakaryocytes, megakaryocytes with multiple separate nuclei, percentage of myeloblasts between 5% and 30%, hypogranulation of promyelocytes and myelocytes, pseudo-Pelger cells, abnormal monocyte precursors, dyserythropoiesis. ring sideroblasts). We excluded patients with hypocellular bone marrow smears, because hypoplastic bone marrow is rare in MDS (Yoshida et al, 1988) and distinction between hypocellular MDS, hypocellular AML, and aplastic anaemia can be very difficult. We did not always have a histological specimen available to rule out aplastic anaemia. Furthermore, if hypocellular smears are due to marrow fibrosis, interpretation of myelodysplasia is complicated by cytological distortion of cells that may be due to the fibrosis itself (Ost & Reizenstein, 1992). The patients’ records were traced in the University clinic and other hospitals, and reviewed for several clinical and laboratory features. In particular, we looked for exclusion criteria for primary MDS, such as vitamin Bl2 and folic acid deficiency, alcoholism, hypersplenism, antibody-mediated cytopenia, paroxysmal nocturnal haemoglobinuria. solid tumours, chronic inflammatory diseases, severe metabolic disorders, and acute toxicity from myelotoxic substances or ionizing radiation. Analysis of the charts led to exclusion of 12%of patients whose cytological specimens had suggested a diagnosis of MDS. We tried to differentiate between primary and secondary MDS. Myelodysplastic disorders occurring as late complication after antineoplastic chemotherapy and/or radiation therapy were considered ‘secondary’MDS, whereas all other cases in which such exogenous noxae could not be established were classified as ‘primary’ MDS. Besides MDS patients, all cases of acute myeloid leukaemia diagnosed between 1975 and 1990 were retrieved from the database and reviewed. The AML patients were chosen as a reference group for comparison, because the epidemiology of AML is much better defined than that ofMDS (Linet & Devesa, 1990).

RESULTS MDS and AML cases in the register Among a total of 18 41 6 different patients entered into the register between 1975 and 1990, 584 cases of MDS (3.2%) and 506 cases of AML (2.8%) were identified (Table I). Case registration over the study period differed considerably between MDS and AML (Fig 1).As regards AML, there was no consistent increase or decline in newly diagnosed cases, which varied between 16 and 50 per year. On the other hand, MDS was diagnosed with increasing frequency. This increase was most pronounced during the last 8 years of study, with annual case numbers rising from 1 7 to 79. As the annual number of bone marrow specimens examined in our cytology laboratory has grown continuously, we looked whether the increase in MDS was simply due to an expansion of diagnostic services. This was not the case. As shown in Table 11, the proportion of MDS cases rose almost constantly from 1’3%

360

C. Aul, N . Gattermatin and W . Srhneider 80 70

60

c

50

-

40

v1 *

.-5

&J 30 20 10 0 '75

50

-

2

40

'76

'77

'78

'79

'80

'81

'82

'83

'84

'85

'86

'87

'88

'89

'90

'77

'78

'79

'80

'81

'82

'83

'84

'85

'86

'87

'88

'89

'90

AML

Z 30 E

2 e"

20 10 0 '75

'76

Fig 1 . Ani ial case numbers of neuly diagnosed \IDS and r\\lL in the Dusseldorf bone marrow register ( 1 975-90).

-

50

z .k

40

C

n

. d

30 20 10 0 '75

'76

'77

'78

'79

'80

'81

'82

'83

'84 '85

'86

'87

'811

'89

'90

Fig 2. Distribution of MDS cases among FAB types over the study period.

during the first year of the survey to 4.5% in 1990, which clearly indicates that the increase in MDS surpassed the increase in study material. A similar trend was not seen in the AMl. group.

MIX subtypes One hundred and twenty-three MDS patients had refractory anaemia (RA) ( 2 1 . lx),142 had RA with ringed sideroblasts (KAKS) (24.3%),1 3 3 had KA with excess of blasts (RAER)

Epidemiology of MDS

3 61

Table I. Numbers and FAB type distribution of MDS and AML cases in the Dusseldorf bone marrow register (1975-90). No. of cases

Diagnosis

FAB type

Total

Wider catchment

Diisseldorf

RA RARS RAEB RAEB/T CMML

584 123 142 133 91 95

364 74 88 80 57 65

220 49 54 53 34 30

506 128 189 19 101 35

303 79 111 12 65 19 9 2 6

203 49 78 7 36 16 7 5 5

Myelodysplastic syndromes

Acute myeloid leukaemia M1 M2 M3 M4 M5 M6 M7 Not specified

16

7 11

Abbreviations: FAB, French-American-British Cooperative Group: RA, refractory anaemia: RARS, RA with ring sideroblasts: RAEB, RA with excess of myeloblasts: RAEB/T, RAEB in transformation: CMML, chronic myelomonocytic leukaernia: M1, acute myeloblastic leukaemia without maturation: M2, acute myeloblastic leukaemia with maturation: M3. acute promyelocytic leukaemia: M4, acute myelomonocytic leukaemia: M5, acute monocytic leukaemia: M6. acute erythroleukaernia: M7, acute megakaryoblastic leukaemia.

Table 11. Relative frequencies of MDS and AML in the Diisseldorf bone marrow register (1 975-90).

Year 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1975-90

No. of bone marrow specimens

MDS

AML

(%I

(%I

557 (612*) 756 (819) 657 (728) 694 (792) 730 (780) 787 (937) 983 (1083) 966 (1253) 920 (1351) 1326 (1650) 1642 (1883) 1735 (2007) 1611 (2009) 1698 (2041) 1703 (2109) 1651 (2258)

1.3 1.5 1.8 1.6 2.7 2.2 2.8 2.4 1.9 2.9 2.5 3.1 4.7 4.5 4.6 4.5

3.6 2.3 4.0 4.2 2.3 2.0 2.8 3.5 4.1 2.4 2.5 2.4 3.1 2.4 1.9 2.7

18416 (22312)

3.2

2.8

* Repeat bone marrow aspirates included.

(22.8%), 9 1 had RAEB in transformation (RAEB/T) ( 1 5.6%) and 95 had CMML (16.3%).All FAB types contributed to the rise in annual case numbers (Fig 2). Of interest, most of our CMML patients were diagnosed after 1982. Table 111 shows the FAB type distribution in younger MDS patients as compared to population groups over 50, RAEB/T seems to be over-represented in younger patients, but this impression is not supported by statistical analysis. The sex ratio (men/women) was almost balanced within morphological subgroups (RA 1.02, RARS 1.00, RAEB 0.87, and RAEB/T 1.12).However, CMML was characterized by a high male-to-female ratio of 1.57. The series as a whole yielded a sex ratio of 1.06.

Follow-up At present, follow-up data are available for 452 MDS patients. Median follow-up is 20 months (range 0-165 months). 131 patients (29%) are still alive, whereas 321 patients (71%) have died. Causes of death were infection in 73 (22.7%),haemorrhagein 46 (14.3%)and transformation to AML in 8 4 patients (26.2%). 8 3 patients (25.9%) died of causes unrelated to MDS, and in 35 patients (10.9%) the precise cause of death could not be ascertained. In none of the cases was spontaneous recovery from a myelodysplastic

C. Aul, N. Gatterrnann and W . Schneider

362

I40 1'

I

Age (years) 1

Age (years) Fig 3 . Age distribution of hlDS and AML patients.

1 3 5 consecutive MDS patients from our register have recently been published (Aul et al, 1992).

0 0

ooxv

,8'

XH

VIJ

Age

0 RO

8 0

XI

R4

79

0

0

R