- Email: [email protected]
Improved survival of Hodgkin's patients in South-East Netherlands since 1972
European Journal of Cancer Vol. 33, No. 3, pp. 436-441, 1997 9 1997 Elsevier Science Ltd. All rights reserved Printed in Great Bntam 0959-8049197 $17.00+0.00
Pergsmon PII: SO959-8049(96)00485-6
Original Papers Improved Survival of Hodgkin’s Patients in South-East Netherlands Since 1972 D.J. van Spronsen,’ I.M. Dijkema,’ L.W. Vrints,2’3 G. Hofstra,4 M.A. Crommelin,3,5 F.L.G. Erdkamp,3,6 J.W.W. Coebergh3a4and W.P.M. Breed’a3 ‘Department of Internal Medicine; ‘Department of Pathology, Catharina Hospital, Michelangelolaan 2, 5623 EJ, Eindhoven, The Netherlands; ‘On behalf of HOMAAN, the Haemato-Oncology study group of the “Department of Epidemiology and Biostatistics, Comprehensive Cancer Center South (I.K.Z.), Eindhoven; Erasmus University, Rotterdam; ‘Department of Radiotherapy, Catharina Hospital, Eindhoven; and 6Department of Internal Medicine, Maasland Hospital Sittard, The Netherlands
In the past 30 years, staging and treatment of Hodgkin’s disease have changed dramatically, and prolonged remission can now be induced in the majority of patients. Our purpose was to assess improvement in long-term survival, previously reported for specific patient groups, among unselected patients diagnosed and treated between 1972 and 1993 in general hospitals in South-East Netherlands. Data on all 345 Hodgkin’s patients were derived from the population-based Eindhoven Cancer Registry; histopathology and clinical records were reviewed. Follow-up was attained up to 1994. Relative survival rates, i.e. the ratio of observed to expected rates, were 80% after 5, 70% after 10 and 67% after 15 years. Independent prognostic factors for lower overall survival were (in decreasing order of significance): advanced age, histology (lymphocyte depletion), advanced stage and earlier period of diagnosis. Distribution of age and stage did not change over the study period, but there was a modest increase in the incidence of the nodular sclerosis histological subtype. Crude S-year survival rates improved from 60% in the period 1972-1976 to 81% in the period of 1987-1992 (P < 0.005). The largest improvement occurred in the 1970s and was most prominent among those aged over 50 years. As previously reported, cured Hodgkin’s patients exhibit a higher mortality rate, which can be explained by treatment-related long-term complications such as second malignancies and cardiovascular diseases. The relatively high survival rates compared to other population-based studies may be attributable to the existence of a regional network within the framework of a comprehensive cancer centre. Better staging, new combinations of chemotherapy, improved radiation technology, advances in supportive care as well as more frequent intensive treatment of the elderly could explain the improvement in prognosis. 0 1997 Elsevier Science Ltd. All rights reserved. Key words: Hodgkin’s disease, prognosis, survival, population-based, EurJ Cancer, Vol. 33, No. 3, pp. 436-441,
1997
INTRODUCTION IN THE past 30 years, staging and treatment
histology, registry
dality has proved to be adequate for patients with earlystage disease, the long-term survival rates being 80-90%. In
of Hodgkin’s
disseminated disease, combination chemotherapy has the potential of inducing complete remission in 75590% of patients, with 50-63% being cured, while primary combined modality therapy is appropriate in only a minority of cases.
disease have changed dramatically. Prolonged remission can now be induced in the majority of patients with chemotherapy and/or irradiation [l, 21. Radiotherapy as a single mo-
Successful treatment, also after long-term follow-up, has largely been described for patients treated in specialised hospitals or selected patients enrolled in clinical trials [3-61.
Correspondence to D. J. van Spronsen. Received 11 Jul. 1996; revised 17 Sep. 1996; accepted 18 Sep. 1996. 436
Improved
Survival of Hodgkin’s
Only a few reports describe survival rates for unselected patients [7-l 11. An improvement in outcome in the last decades has been reported for subcategories, i.e. patients with advanced disease [ 121 or early disease treated with irradiation only [ 131, or for patients treated in specialised centres [3-51. The improvement in prognosis due to newer treatment strategies can be assumed to be smaller for unselected patients, since the general population is likely to comprise more elderly and more untreated patients with a poor performance status. The purpose of our study, based on data from the cancer registry in South-East Netherlands, was to investigate whether the survival of all patients with Hodgkin’s disease, diagnosed between 1972 and 1993, has improved. We also estimated the prognostic value of presenting symptoms for overall survival. PATIENTS AND METHODS The study comprised all patients aged 15 years and older with Hodgkin’s disease diagnosed between 1972 and 1993 and followed until 1994. Data was collected from the Eindhoven Cancer Registry (IKZ), which covers a densely populated area in South-East Netherlands with approximately 900 000 inhabitants, almost 7% of the Dutch population. Between 1972 and 1992, 372 patients with Hodgkin’s disease were registered. There was insufficient data on 6 patients and 1 patient was diagnosed and treated outside the region. Review of the medical records revealed that 4 patients had non-Hodgkin’s lymphoma (1%) and 1 had diphantoin lymphadenopathy. The available histology was reviewed and classified according to the Rye classification system by an experienced haematopathologist (LWV) [ 141. Review of histology revealed non-Hodgkin’s lymphoma in another 15 cases (4%). Nodular sclerosis was subdivided into Grade I and Grade II according to the criteria of the British National Lymphoma Investigation Group
USI. Patients were diagnosed and treated in ten general hospitals and the regional Radiotherapy Department, where linear accelerator equipment was installed in 1973. Staging occurred according to the Ann Arbor classification system [16], without a uniform diagnostic protocol. However, since 1979, the therapeutic policy for each patient has, in most cases, been discussed in a regional multidisciplinary group; this group has participated in the EORTC Lymphoma Group since 1983, and 36% of eligible patients are enrolled in clinical trials. Patients with stages IA or IIA generally received mantle field or inverted Y-field irradiation (40 Gy); patients with stages IB or IIB underwent more extensive irradiation with or without chemotherapy. Patients with Table
437
Patients
stages III or IV received chemotherapy, usually MOPP (mechlorethamine, vincristine, procarbazine, prednisone), ABVD (doxorubicin, bleomycin, vinblastine, dacarbazine) or a hybrid scheme, frequently followed by iceberg irradiation to a dose of 20 Gy. All medical records were reviewed, using the cancer registry as a sampling frame. Primary therapeutic modalities and presenting symptoms were documented, as were response to therapy, relapses, second malignancies, vital status until 1994 and cause of death. Follow-up could not be completed in two cases because the patients moved abroad. Expected survival rates were calculated from life tables for the regional population (supplied by The Netherlands Central Bureau for Statistics) [17]. Relative survival was calculated as the ratio of the observed to the expected actuarial rates, using a programme provided by the Finnish Cancer is given by the difference Registry [ 181. Excess mortality between observed and expected mortality. Actuarial survival curves were computed according to period of diagnosis (1972-1976, 1977-1981, 1982-1986, 1987-1992) and age (~50 and >50 years) [19]. The log-rank test was used to assess differences between survival curves. For 323 patients with complete data, the Cox proportional hazard model was used to perform multivariate analysis of the prognostic value for overall survival of the following factors: age, stage (including presence of Bsymptoms), period of diagnosis, histology and gender. Age, stage, histology and therapeutic modality for given time periods were compared by means of the chi-squared test. RESULTS A total of 345 patients (203 males (59%) and 142 females (41%)) was studied, corresponding to a crude incidence of 1.9 per 100 000 person-years. Distribution of patients according to age, stage and histology in the four time periods is given in Table 1. Mean age and stage distribution did not change over time. The incidence of the nodular sclerotic subtype increased modestly, while the incidence of lymphocyte depletion and lymphocyte predominance subtypes decreased slightly, over time. Stages III and IV occurred in 35% of patients younger than 50 years and 44% of patients 50 years and older (data not shown). The proportion of patients receiving chemotherapy is shown in Table 2. Whereas MOPP chemotherapy was the most commonly used drug regimen in the 197Os, MOPPI ABVD alternating or hybrid schemes were mainly used in the late 1980s. The percentage of patients receiving some kind of irradiation remained constant over time, but application of involved field (iceberg) irradiation increased and
1. Trends in distribution of age, stage and histology for 345 patients with Hodgkin’s South-East
Netherlands,
disease in a population-based
study in
1972-1992
Histology* Period 1972-1976 1977-1981 1982-1986 1987-1992 1972-1992 (fl)
Iz 77 83 87 98 345
Mean age (years)
Stage I & II
LP
NS
MC
LD
39 40 41 38 39
58% 65% 55% 59% 59%
1 1% 15% 5% 8% 10%
53% 61% 59% 65% 60%
28% 24% 32% 25% 27%
8% 0% 4% 2% 3%
(31)
(193)
(88)
(11)
* Histology remained unknown in 22 cases because histological sections were not available or of poor quality. LP, lymphocyte predominance; NS, nodular sclerosis; MC, mixed cellularity; LD, lymphocyte depletion.
438
et al.
D.J. van Spronsen
Table 2. Trends in treatment of 345 patients with Hodgkin’s disease in South-East Netherlands, 1972-1992 Period 1972-1976 1977-1981 1982-1986 1987-1992 1972-1992 (n)
Chemotherapy Other? MOPP/ABVD$
MOW* 50% 49% 31% 10% 34% (116)
14% 2% 15% 22% 14% (48)
3% 2% 8% 40% 14% (50)
Irradiation IFII -I-NIP
Othefl
Both modalities
No treatment
8% 6% 22% 43% 21% (72)
14% 30% 30% 13% 22% (76)
41% 33% 32% 45% 38% (131)
3% 1% 9% 2% 4% (13)
50% 40% 17% 14% 29% (101)
* Includes six cycles of MOW. t Mainly ABVD and, in the late 198Os, EBVP (etoposide, bleomycine, vinblastine, prednisone). $ Includes alternating cycles of MOW and ABVD or a hybrid scheme of MOIWABVD. §Total nodal irradiation (TNI). IIIceberg irradiation (= involved field irradiation of initially bulky disease foci). 7 Mainly mantle field and inverted Y-field irradiation. Table 3. Trends in response to treatment, relapse rates and causes of death for 345 patients with Hodgkin’s disease treated in South-East Netherlands, 1972-1992
Period
Complete Total
1972-1976 1977-1981 1982-1986 1987-1992 1972-1992
72% 89% 76% 89% 82%
Deaths
50% 34% 33%
total n
52 (67%) 28 (34%) 29 (33%) 12 (12%) 121 (35%)
7%3
28%
* Treatment-related deaths such as sepsis, pulmonary percentage due to short duration of follow-up.
Causes of death Complications*
Hodkin’s
remission With relapse
fibrosis
use of total nodal irradiation decreased (Table 2). Complete remission was achieved in 82% of all patients, 28% of whom suffered one or more relapses. The percentage of complete remissions increased over time (Table 3). A total of 121 (35%) patients died: 50 (15%) due to active Hodgkin’s disease, 25 (7%) because of complications (treatment related sepsis, pneumonitis and pulmonary fibrosis) and 46 (13%) from unrelated or unknown causes. The percentage of deaths from active Hodgkin’s disease decreased over time (Table 3). Crude 5-, lo- and 15-year survival rates for all patients were 76, 63 and 57%, respectively. The corresponding relative survival rates were 80%, 70% and 67%. A significant improvement in survival was seen, with crude 5-year survival rates of 60, 86, 73 and 81% for the periods 1972-1976, 1977-1981, 1982-1986 and 1987-1992 (PC 0.005); the corresponding relative 5-year survival rates were 62, 89, 79 and 84%, respectively. This improvement was most prominant at the end of the 197Os, the greatest benefit being for patients 50 years of age and older (Figure 1). The Cox regression model was used to assess the prognostic value of presenting symptoms; patients presenting with stage IA and nodular sclerosis grade I were considered to represent a standard hazard risk ratio of 1.0. Independent prognostic factors adversely influencing overall survival were (in order of decreasing significance): advanced age (X0 years, H.R. 4.1), histology (H.R. for lymphocyte depletion 3.7), advanced stage (IIIB-IVB, H.R. 2.1) and initial diagnoses in the 1970s (1972-1980, H.R. 1.9) (Table 4). Survival for the 203 (59%) patients who never relapsed after achieving complete remission revealed a higher mortality compared to the reference population (Figure 2). A total of 22 patients developed a second malignancy, with a corresponding lo-year cumulative incidence of 8% for all malignancies and 3% for leukaemia. A 2-fold decrease in
(%)
(%)
27 13 15
14 6 7 3 7
26 15 11 4 13
5$ 15
and pneumonitis.
Unrelatedt
(%)
t Unrelated
deaths
and deaths
of unknown
cause.
lo-year cumulative incidence for all malignancies observed in the 1980s (10% for the period 1972-1980 pared with 5% for 1981-1992; P= 0.1).
*Low
was com-
(a>
\
--__-__ -. -_-___
-
_.
--------.
. .
1987-1992
5
10
Years
after diagnosis
(b)
E ,” E
‘\ \ _ _ _ _ -a.. _ ~.~_.___...‘“““..... -.
40-
a
.-..-...._
1987-1992 1982-1986
-_-
1977-1981
----
1972-1976
‘\
-. .__-----*
.--,
-
______.
20 0
I 5
’ 0 Years
after
I 10 diagnosis
Figure 1. Relative survival of patients with Hodgkin’s disease (a) C 50 years of age; or (b) 2 50 years of age.
Improved
Survival of Hodgkin’s
Patients
439
Table 4. Overall survival for 323 patients with Hodgkin’s disease diagnosed in South-East Netherlands in the period 1972-1992, according to major prognostic factors Univariate survival Log-rank P value 5 year* (%)
Comparison
Multivariate survival P value Hazard ratio
Age <50 years >50 years Sex Female Male Stage IA-IIIA IIIB-IVB Period 1981-1992 1972-1980 Histology LP MC NSI NSII LD * Crude
100
0.0001
1.0 4.1
81 77
>0.2
1.0 1.4
0.13
84 59
0.0001
1.0 2.1
0.001
83 74
0.04
1.9
0.007 20.2 >0.2
0.0001
0.7 0.7 1.0 2.1 3.7
0.0001
1.0
89
82 84 65 36
0.02 0.002
5-year survival.
- .-.-...
_
*......._.___
---..
..____!I)% Excess . . . . . . . . . . .yy+_ty .I_?”
..
80 -
..-..-..
“0
84 58
Expected survival Observed
survival
5
Years
10
15
after diagnosis
Figure 2. Observed versus expected survival of patients cured of Hodgkin’s disease.
DISCUSSION Survival of patients with Hodgkin’s disease improved markedly in this population-based study for patients both older and younger than 50 years; the most prominent improvement occurred in the 1970s. This improvement cannot be attributed to a change in patient characteristics, since distribution of age, stage and histology did not change significantly over time. Nor can the improvement in survival be explained by an increase in life expectancy for the general population, since the improvement in relative survival is adjusted for the expected mortality for the general population. The improved survival should be explained by major advances in the management of Hodgkin’s disease. Staging procedures have changed due to the availability of computed tomography, causing a decline in the need for staging laparotomies. As indicated by the numerous randomised clinical trials, an individualised approach to the treatment of each patient is an essential skill for oncologists. Individual courses of action have been discussed since 1979 in a regional multidisciplinary group, which has participated in the EORTC Lymphoma Group since 1983. The ABVD or hybrid schemes of combination chemotherapy, which are
presumed to be less toxic and leukaemogenic compared to the MOW regimen, were administered mainly in the late 1980s. Radiation therapy has improved due to fractionation, the higher energy beams produced by the machines and better delineation of radiation fields. Patients with poor performance status or advanced disease and, especially, the elderly may have been treated more frequently and aggressively, as a result of advances in supportive care and treatment of comorbidity. Although survival has improved markedly in the past 20 years, a temporary decrease in survival was seen in the early 1980s in both age groups. This may have been coincidental or it could be related to the relatively large percentage patients presenting with advanced stage, the relatively large number of patients who did not receive cancer therapy or learning effects of new therapies. The prognostic value of presenting symptoms has been described in detail and is a tool for clinical strategy [20], which has been applied in our region since the early 1980s. After adjustment for period of diagnosis and gender, prognostic factors adversely influencing overall survival were advanced age, histology (lymphocyte depletion) and advanced stage. Period of diagnosis also had a prognostic value independent of age, histology, stage and gender, with a nearly 2-fold reduction in hazard rate for patients diagnosed in the 1980s. As previously reported, cured patients exhibited an increased mortality rate compared to the general population. This increased mortality is due mainly to long-term treatment-related second malignancies observed, in particular, after combination therapy [21] and coronary heart disease associated with mantle field irradiation which can provoke increased atherosclerosis [22]. The 2-fold decrease in the cumulative incidence of a second malignancy during the 1980s was probably related to the choice of less toxic treatment strategies, such as the ABVD regimen. Improved survival has been described in several hospitalbased studies (Table 5) [3-51. In a study of 570 patients treated at a single centre in Bologna, the crude 5-year
440
D.J. van Spronsen et al. Table 5. Comparison of studies on 5-year survival of Hodgkin’s
Study IDHD
[3]
Type
Period
Hospital
1960-69 1970-79 19801970-80 1980-87 1960-70 1970-88 1973-77 1983-87 1978-80 1983-85 1977-82 1973-77 1978-82 1983-87 1972-76 1977-81 1982-86 1987-92
Bologna
[4]
Hospital
Florence
[5]
Hospital
Denmark
[7]
Population
Eurocare
[S]
Population
CCPDS [9] SEER [lo]
Hospital Population
This study
Population
* Estimated
from figures.
t Males.
$ Relative
n 14 225
570 1121
4677 2476 9418
345
disease Crude
5-year survival 73%* SO%* 82%* 75%* 90%* 65%* 85%* 57%? 64%t 58% 67% 77% 70%$ 73%$ 77%$ 60% 86% 73% 81%
survival.
survival rate improved from approximately 75% in the 1970s to 90% in the 1980s [4]. In a series of 1121 patients treated at one radiation institution in Florence, the crude 5-year survival rate improved from approximately 65% for patients treated before 1970 to 85% for patients treated between 1970 and 1988 [5]. In a large series of 14 225 clinical stage I-IV patients treated in 15 cancer centres or five co-operating groups all over the world, 5year overall survival rates improved from approximately 73% in the 1960s to 82% in the early 1980s [3]. With crude 5-year survival rates improving from 60% for the period 1972-1976 to 81% for the period 1987-1992, the results of our population-based study do not seem to differ much from those of most hospital-based studies performed in the same period. Although clinical trials are essential for establishing the efficacy of therapeutic advances, improvement in population-based survival due to the use of new and more effective treatment regimes in the general health care environment provide additional evidence of their effectiveness in general (Table 5). In a Danish registry-based study, the crude 5-year survival rates for males improved from 57% for the period 1973-1977 to 64% for the period 1983-1987 [7]. In a study of the results of several cancer registries throughout Europe, the crude 5-year survival rates improved from 58% for those diagnosed in the period 1978-1980 to 67% for the period 1983-1985 [8]. In the United States for the period 1977-1982, the crude 5-year survival rate was 62% for the population-based SEER-programme compared to 77% for the hospitalbased CCPDS-programme [9]. In another study of the SEER-programme, an improvement in the relative 5-year survival rate from 70% in 1973-1977 to 77% in 1983survival rates 1987 was reported [lo]. Population-based are lower compared to hospital-based series, since population-based studies are not biased by selection of patients and may comprise more elderly patients, patients with poor performance status, and possibly also some patients
with non-Hodgkin’s lymphoma (up to 17% in some series) [23]. A review of the medical records and histology revealed non-Hodgkin’s lymphoma in 5% of cases in our study. The improved prognosis for patients with Hodgkin’s disease is also demonstrated by the decrease in mortality rates in Scotland since 1970 [24]. Similarly, a significant reduction in annual mortality rates has been reported by the National Center for Health Statistics in the United States for the period 1979-1988 [25]. Also, in our region, mortality rates, derived from the cause-of-death register, have decreased since the 1960s from 2.4 to 0.3 per 100 000 person-years in the late 1980s for males, while the age-adjusted incidence decreased slightly [26]. It is likely that the decline was overestimated as a result of incorrect causes of death. In conclusion, population-based survival rates for patients with Hodgkin’s disease diagnosed and treated in general hospitals in South-East Netherlands in the past 20 years are more or less similar to those demonstrated in centre-based studies. The marked improvement in survival found for both the young and the elderly was independent of age, stage, histology and gender and thus reflects improved treatment results. Nonetheless, patients cured of Hodgkin’s disease exhibit an increased mortality risk due to treatmentrelated malignancies and, presumably, coronary heart disease.
Weinshel EL, Peterson BA. Hodgkin’s disease. Cancer 3 Clin 1993, 43, 327-346. Urba WJ, Longo DL. Hodgkin’s disease. N Engl 3 Med 1992, 326, 678-687. Henry-Amar M, Somers R. Survival outcome after Hodgkin’s disease: a report from the International Data base on Hodgkin’s Disease. Semin Oncol 1990, 17, 758-768. Mazia P, Bocchia M, Zinzani PL, et al. Hodgkin’s disease: summary of twenty years’ experience. Haematologi~a 1992, 17, 487-493.
Improved
Survival of Hodgkin’s
5. Biti G, Cellai E, Magrini SM, Papi MG, Ponticelli I’, Boddi V. Second solid tumors and leukemia after treatment for Hodgkin’s disease: an analysis of 1121 patients from a single Int J Radiation Oncol Biol Phys 1994, 29, 25-3 1. institution. C, Dekker A, et al. De 6. Van Rijswijk R, Verbeek J, Haanen resultaten van de behandeling van de ziekte van Hodgkin: 15 jaar ervaring van de universiteitsklinieken van Nijmegen en Utrecht. Ned Tijdchr Geneeskd 1987, 131, 1866-1870. J. Survival of Danish cancer patients 7. Storm HH, Clemmensen 1943-87. APMZS 1993, 189-193. R, Hakulinen T, 8. Berrino F, Sant M, Verdecchia A, Capocaccia Esteve J. Survival of Cancer Patients in Europe: the EUROCARE Study, 1978-85. Lyon, Scientific Publications, 1995, 363-372. 9. Davis S, Dahlberg S, Myers MH, Chen A, Steinhorn SC. Hodgkin’s disease in the United States: a comparison of patients characteristics and survival in the Centralised Cancer Patient Data System and the Surveillance Epidemiology and End Results Program. J Nat1 Cancer Inst 1987, 78, 471-478. 10. Medeiros LJ, Greiner TC. Hodgkin’s disease. Cancer 1995, 75, 357-369. 11. Erdkamp FLG, Wijlhuizen ThJ, van Dam FE. De ziekte van Hodgkin: een regionaal, retrospectief onderzoek. Ned Tzjdschr Geneeskd 1988, 132, 1801-1806. 12. Friedenberg WR, Dirks I’, Beltaos E, et al. Improved survival in the treatment of advanced Hodgkin’s disease at a nonuniversity institution (1970-79). Cancer 1986, 57, 12-17. 13. Hoppe RT, Hanlon AL, Hanks GE, Owen JB. Progress in the treatment of Hodgkin’s disease in the United States, 1973 versus 1983: the patterns of care study. Cancer 1994, 74, 31983203. 14. Lukes RJ, Butler JJ. The pathology and nomenclature of Hodgkin’s disease. Cancer Res 1966, 26, 1063-1081. 15. Bennet MH, Maclennan KA, Easterling MJ, Hudson BV, Jeliffe AM, Hudson GV. The prognostic significance of cellular subtypes in nodular sclerosing Hodgkin’s disease: an analysis of 271 non-laparotomised cases (BNLI report no 22). Clin Radio1 1983, 34, 497-501. 16. Carbone PI’, Kaplan HS, Musshoff K, et al. Report of the committee on Hodgkin’s disease staging classification. Cancer Res 1971, 31, 1860-1861. 17. Hakulinen T. Cancer survival corrected for heterogeneity in patients withdrawal. Biometrics 1982, 38, 933-942.
Patients
441
18. Hakulinen T, Tenkanen L. Regression analysis of relative survival rates. Appl Stat 1987, 36, 309-317. 19. Berkson J, Gage RI’. Calculation of survival rates for cancer. Proc Staff Meet Mayo Clin 1950, 25, 270-286. 20. Gobbi PG, Comelli M, Grignani GE, Piersca C, Bertoloni D, Ascari E. Estimate of expected survival at diagnosis in Hodgkin’s disease: a means of weighting prognostic factors and a tool for treatment choice and clinical research. A report from IDHD. Haematologica 1994, 79, 241-255. 21. Van Leeuwen FE, Klokman WJ, Hagenbeek A, et al. Second cancer risk following Hodgkin’s disease: a 20 year follow-up study. ‘3 Clin Oncol 1994, 12, 312-325. 22. Boivin JF, Hutchison GB, Lubin JH, Mauch I’. Coronary artery disease mortality in patients treated for Hodgkin’s disease. Cancer 1992, 169, 1241-1247. 23. Hartge I’, Devesa SS, Fraumeni JF Jr. In Doll R, Fraumeni JF Jr, Muir CS, eds. Trends in Cancer Incidence and Mortality. New York, Cold Spring Harbor Laboratory Press, 1994, 20, 423453. 24. Boyle I’, Soukop M, Scully C, et al. Improving prognosis of Hodgkin’s disease in Scotland. Eur 3 Cancer Clin Oncol 1988, 24, 229-234. 25. Hooper WC, Holman RC, Stine TW, Chorba TL. Hodgkin’s disease mortality in the United States. Cancer 1992, 70, 11661171. 26. Van Spronsen DJ, van der Heijden LH. Hodgkin’s disease. In Coebergh JWW, van der Heijden L, Janssen-Heijnen MLG, eds. Cancer Incidence and Survival in the South-East Netherlands Eindhoven, Comprehensive Cancer Center South, 1955-94. 1995, 82-83.
Acknowledgements-We thank Louis van der Heijden for data management, Theo Stijnen for his statistical expertise, the Central Bureau of Statistics for supplying regional mortality data and the hospitals participating in the registry of the Comprehensive Cancer Center South (Eindhoven, Geldrop, HelmondDeurne, Roermond, Veghel, Veldhoven, Venlo and Venray) for allowing access to the medical records.
Pergsmon PII: SO959-8049(96)00485-6
Original Papers Improved Survival of Hodgkin’s Patients in South-East Netherlands Since 1972 D.J. van Spronsen,’ I.M. Dijkema,’ L.W. Vrints,2’3 G. Hofstra,4 M.A. Crommelin,3,5 F.L.G. Erdkamp,3,6 J.W.W. Coebergh3a4and W.P.M. Breed’a3 ‘Department of Internal Medicine; ‘Department of Pathology, Catharina Hospital, Michelangelolaan 2, 5623 EJ, Eindhoven, The Netherlands; ‘On behalf of HOMAAN, the Haemato-Oncology study group of the “Department of Epidemiology and Biostatistics, Comprehensive Cancer Center South (I.K.Z.), Eindhoven; Erasmus University, Rotterdam; ‘Department of Radiotherapy, Catharina Hospital, Eindhoven; and 6Department of Internal Medicine, Maasland Hospital Sittard, The Netherlands
In the past 30 years, staging and treatment of Hodgkin’s disease have changed dramatically, and prolonged remission can now be induced in the majority of patients. Our purpose was to assess improvement in long-term survival, previously reported for specific patient groups, among unselected patients diagnosed and treated between 1972 and 1993 in general hospitals in South-East Netherlands. Data on all 345 Hodgkin’s patients were derived from the population-based Eindhoven Cancer Registry; histopathology and clinical records were reviewed. Follow-up was attained up to 1994. Relative survival rates, i.e. the ratio of observed to expected rates, were 80% after 5, 70% after 10 and 67% after 15 years. Independent prognostic factors for lower overall survival were (in decreasing order of significance): advanced age, histology (lymphocyte depletion), advanced stage and earlier period of diagnosis. Distribution of age and stage did not change over the study period, but there was a modest increase in the incidence of the nodular sclerosis histological subtype. Crude S-year survival rates improved from 60% in the period 1972-1976 to 81% in the period of 1987-1992 (P < 0.005). The largest improvement occurred in the 1970s and was most prominent among those aged over 50 years. As previously reported, cured Hodgkin’s patients exhibit a higher mortality rate, which can be explained by treatment-related long-term complications such as second malignancies and cardiovascular diseases. The relatively high survival rates compared to other population-based studies may be attributable to the existence of a regional network within the framework of a comprehensive cancer centre. Better staging, new combinations of chemotherapy, improved radiation technology, advances in supportive care as well as more frequent intensive treatment of the elderly could explain the improvement in prognosis. 0 1997 Elsevier Science Ltd. All rights reserved. Key words: Hodgkin’s disease, prognosis, survival, population-based, EurJ Cancer, Vol. 33, No. 3, pp. 436-441,
1997
INTRODUCTION IN THE past 30 years, staging and treatment
histology, registry
dality has proved to be adequate for patients with earlystage disease, the long-term survival rates being 80-90%. In
of Hodgkin’s
disseminated disease, combination chemotherapy has the potential of inducing complete remission in 75590% of patients, with 50-63% being cured, while primary combined modality therapy is appropriate in only a minority of cases.
disease have changed dramatically. Prolonged remission can now be induced in the majority of patients with chemotherapy and/or irradiation [l, 21. Radiotherapy as a single mo-
Successful treatment, also after long-term follow-up, has largely been described for patients treated in specialised hospitals or selected patients enrolled in clinical trials [3-61.
Correspondence to D. J. van Spronsen. Received 11 Jul. 1996; revised 17 Sep. 1996; accepted 18 Sep. 1996. 436
Improved
Survival of Hodgkin’s
Only a few reports describe survival rates for unselected patients [7-l 11. An improvement in outcome in the last decades has been reported for subcategories, i.e. patients with advanced disease [ 121 or early disease treated with irradiation only [ 131, or for patients treated in specialised centres [3-51. The improvement in prognosis due to newer treatment strategies can be assumed to be smaller for unselected patients, since the general population is likely to comprise more elderly and more untreated patients with a poor performance status. The purpose of our study, based on data from the cancer registry in South-East Netherlands, was to investigate whether the survival of all patients with Hodgkin’s disease, diagnosed between 1972 and 1993, has improved. We also estimated the prognostic value of presenting symptoms for overall survival. PATIENTS AND METHODS The study comprised all patients aged 15 years and older with Hodgkin’s disease diagnosed between 1972 and 1993 and followed until 1994. Data was collected from the Eindhoven Cancer Registry (IKZ), which covers a densely populated area in South-East Netherlands with approximately 900 000 inhabitants, almost 7% of the Dutch population. Between 1972 and 1992, 372 patients with Hodgkin’s disease were registered. There was insufficient data on 6 patients and 1 patient was diagnosed and treated outside the region. Review of the medical records revealed that 4 patients had non-Hodgkin’s lymphoma (1%) and 1 had diphantoin lymphadenopathy. The available histology was reviewed and classified according to the Rye classification system by an experienced haematopathologist (LWV) [ 141. Review of histology revealed non-Hodgkin’s lymphoma in another 15 cases (4%). Nodular sclerosis was subdivided into Grade I and Grade II according to the criteria of the British National Lymphoma Investigation Group
USI. Patients were diagnosed and treated in ten general hospitals and the regional Radiotherapy Department, where linear accelerator equipment was installed in 1973. Staging occurred according to the Ann Arbor classification system [16], without a uniform diagnostic protocol. However, since 1979, the therapeutic policy for each patient has, in most cases, been discussed in a regional multidisciplinary group; this group has participated in the EORTC Lymphoma Group since 1983, and 36% of eligible patients are enrolled in clinical trials. Patients with stages IA or IIA generally received mantle field or inverted Y-field irradiation (40 Gy); patients with stages IB or IIB underwent more extensive irradiation with or without chemotherapy. Patients with Table
437
Patients
stages III or IV received chemotherapy, usually MOPP (mechlorethamine, vincristine, procarbazine, prednisone), ABVD (doxorubicin, bleomycin, vinblastine, dacarbazine) or a hybrid scheme, frequently followed by iceberg irradiation to a dose of 20 Gy. All medical records were reviewed, using the cancer registry as a sampling frame. Primary therapeutic modalities and presenting symptoms were documented, as were response to therapy, relapses, second malignancies, vital status until 1994 and cause of death. Follow-up could not be completed in two cases because the patients moved abroad. Expected survival rates were calculated from life tables for the regional population (supplied by The Netherlands Central Bureau for Statistics) [17]. Relative survival was calculated as the ratio of the observed to the expected actuarial rates, using a programme provided by the Finnish Cancer is given by the difference Registry [ 181. Excess mortality between observed and expected mortality. Actuarial survival curves were computed according to period of diagnosis (1972-1976, 1977-1981, 1982-1986, 1987-1992) and age (~50 and >50 years) [19]. The log-rank test was used to assess differences between survival curves. For 323 patients with complete data, the Cox proportional hazard model was used to perform multivariate analysis of the prognostic value for overall survival of the following factors: age, stage (including presence of Bsymptoms), period of diagnosis, histology and gender. Age, stage, histology and therapeutic modality for given time periods were compared by means of the chi-squared test. RESULTS A total of 345 patients (203 males (59%) and 142 females (41%)) was studied, corresponding to a crude incidence of 1.9 per 100 000 person-years. Distribution of patients according to age, stage and histology in the four time periods is given in Table 1. Mean age and stage distribution did not change over time. The incidence of the nodular sclerotic subtype increased modestly, while the incidence of lymphocyte depletion and lymphocyte predominance subtypes decreased slightly, over time. Stages III and IV occurred in 35% of patients younger than 50 years and 44% of patients 50 years and older (data not shown). The proportion of patients receiving chemotherapy is shown in Table 2. Whereas MOPP chemotherapy was the most commonly used drug regimen in the 197Os, MOPPI ABVD alternating or hybrid schemes were mainly used in the late 1980s. The percentage of patients receiving some kind of irradiation remained constant over time, but application of involved field (iceberg) irradiation increased and
1. Trends in distribution of age, stage and histology for 345 patients with Hodgkin’s South-East
Netherlands,
disease in a population-based
study in
1972-1992
Histology* Period 1972-1976 1977-1981 1982-1986 1987-1992 1972-1992 (fl)
Iz 77 83 87 98 345
Mean age (years)
Stage I & II
LP
NS
MC
LD
39 40 41 38 39
58% 65% 55% 59% 59%
1 1% 15% 5% 8% 10%
53% 61% 59% 65% 60%
28% 24% 32% 25% 27%
8% 0% 4% 2% 3%
(31)
(193)
(88)
(11)
* Histology remained unknown in 22 cases because histological sections were not available or of poor quality. LP, lymphocyte predominance; NS, nodular sclerosis; MC, mixed cellularity; LD, lymphocyte depletion.
438
et al.
D.J. van Spronsen
Table 2. Trends in treatment of 345 patients with Hodgkin’s disease in South-East Netherlands, 1972-1992 Period 1972-1976 1977-1981 1982-1986 1987-1992 1972-1992 (n)
Chemotherapy Other? MOPP/ABVD$
MOW* 50% 49% 31% 10% 34% (116)
14% 2% 15% 22% 14% (48)
3% 2% 8% 40% 14% (50)
Irradiation IFII -I-NIP
Othefl
Both modalities
No treatment
8% 6% 22% 43% 21% (72)
14% 30% 30% 13% 22% (76)
41% 33% 32% 45% 38% (131)
3% 1% 9% 2% 4% (13)
50% 40% 17% 14% 29% (101)
* Includes six cycles of MOW. t Mainly ABVD and, in the late 198Os, EBVP (etoposide, bleomycine, vinblastine, prednisone). $ Includes alternating cycles of MOW and ABVD or a hybrid scheme of MOIWABVD. §Total nodal irradiation (TNI). IIIceberg irradiation (= involved field irradiation of initially bulky disease foci). 7 Mainly mantle field and inverted Y-field irradiation. Table 3. Trends in response to treatment, relapse rates and causes of death for 345 patients with Hodgkin’s disease treated in South-East Netherlands, 1972-1992
Period
Complete Total
1972-1976 1977-1981 1982-1986 1987-1992 1972-1992
72% 89% 76% 89% 82%
Deaths
50% 34% 33%
total n
52 (67%) 28 (34%) 29 (33%) 12 (12%) 121 (35%)
7%3
28%
* Treatment-related deaths such as sepsis, pulmonary percentage due to short duration of follow-up.
Causes of death Complications*
Hodkin’s
remission With relapse
fibrosis
use of total nodal irradiation decreased (Table 2). Complete remission was achieved in 82% of all patients, 28% of whom suffered one or more relapses. The percentage of complete remissions increased over time (Table 3). A total of 121 (35%) patients died: 50 (15%) due to active Hodgkin’s disease, 25 (7%) because of complications (treatment related sepsis, pneumonitis and pulmonary fibrosis) and 46 (13%) from unrelated or unknown causes. The percentage of deaths from active Hodgkin’s disease decreased over time (Table 3). Crude 5-, lo- and 15-year survival rates for all patients were 76, 63 and 57%, respectively. The corresponding relative survival rates were 80%, 70% and 67%. A significant improvement in survival was seen, with crude 5-year survival rates of 60, 86, 73 and 81% for the periods 1972-1976, 1977-1981, 1982-1986 and 1987-1992 (PC 0.005); the corresponding relative 5-year survival rates were 62, 89, 79 and 84%, respectively. This improvement was most prominant at the end of the 197Os, the greatest benefit being for patients 50 years of age and older (Figure 1). The Cox regression model was used to assess the prognostic value of presenting symptoms; patients presenting with stage IA and nodular sclerosis grade I were considered to represent a standard hazard risk ratio of 1.0. Independent prognostic factors adversely influencing overall survival were (in order of decreasing significance): advanced age (X0 years, H.R. 4.1), histology (H.R. for lymphocyte depletion 3.7), advanced stage (IIIB-IVB, H.R. 2.1) and initial diagnoses in the 1970s (1972-1980, H.R. 1.9) (Table 4). Survival for the 203 (59%) patients who never relapsed after achieving complete remission revealed a higher mortality compared to the reference population (Figure 2). A total of 22 patients developed a second malignancy, with a corresponding lo-year cumulative incidence of 8% for all malignancies and 3% for leukaemia. A 2-fold decrease in
(%)
(%)
27 13 15
14 6 7 3 7
26 15 11 4 13
5$ 15
and pneumonitis.
Unrelatedt
(%)
t Unrelated
deaths
and deaths
of unknown
cause.
lo-year cumulative incidence for all malignancies observed in the 1980s (10% for the period 1972-1980 pared with 5% for 1981-1992; P= 0.1).
*Low
was com-
(a>
\
--__-__ -. -_-___
-
_.
--------.
. .
1987-1992
5
10
Years
after diagnosis
(b)
E ,” E
‘\ \ _ _ _ _ -a.. _ ~.~_.___...‘“““..... -.
40-
a
.-..-...._
1987-1992 1982-1986
-_-
1977-1981
----
1972-1976
‘\
-. .__-----*
.--,
-
______.
20 0
I 5
’ 0 Years
after
I 10 diagnosis
Figure 1. Relative survival of patients with Hodgkin’s disease (a) C 50 years of age; or (b) 2 50 years of age.
Improved
Survival of Hodgkin’s
Patients
439
Table 4. Overall survival for 323 patients with Hodgkin’s disease diagnosed in South-East Netherlands in the period 1972-1992, according to major prognostic factors Univariate survival Log-rank P value 5 year* (%)
Comparison
Multivariate survival P value Hazard ratio
Age <50 years >50 years Sex Female Male Stage IA-IIIA IIIB-IVB Period 1981-1992 1972-1980 Histology LP MC NSI NSII LD * Crude
100
0.0001
1.0 4.1
81 77
>0.2
1.0 1.4
0.13
84 59
0.0001
1.0 2.1
0.001
83 74
0.04
1.9
0.007 20.2 >0.2
0.0001
0.7 0.7 1.0 2.1 3.7
0.0001
1.0
89
82 84 65 36
0.02 0.002
5-year survival.
- .-.-...
_
*......._.___
---..
..____!I)% Excess . . . . . . . . . . .yy+_ty .I_?”
..
80 -
..-..-..
“0
84 58
Expected survival Observed
survival
5
Years
10
15
after diagnosis
Figure 2. Observed versus expected survival of patients cured of Hodgkin’s disease.
DISCUSSION Survival of patients with Hodgkin’s disease improved markedly in this population-based study for patients both older and younger than 50 years; the most prominent improvement occurred in the 1970s. This improvement cannot be attributed to a change in patient characteristics, since distribution of age, stage and histology did not change significantly over time. Nor can the improvement in survival be explained by an increase in life expectancy for the general population, since the improvement in relative survival is adjusted for the expected mortality for the general population. The improved survival should be explained by major advances in the management of Hodgkin’s disease. Staging procedures have changed due to the availability of computed tomography, causing a decline in the need for staging laparotomies. As indicated by the numerous randomised clinical trials, an individualised approach to the treatment of each patient is an essential skill for oncologists. Individual courses of action have been discussed since 1979 in a regional multidisciplinary group, which has participated in the EORTC Lymphoma Group since 1983. The ABVD or hybrid schemes of combination chemotherapy, which are
presumed to be less toxic and leukaemogenic compared to the MOW regimen, were administered mainly in the late 1980s. Radiation therapy has improved due to fractionation, the higher energy beams produced by the machines and better delineation of radiation fields. Patients with poor performance status or advanced disease and, especially, the elderly may have been treated more frequently and aggressively, as a result of advances in supportive care and treatment of comorbidity. Although survival has improved markedly in the past 20 years, a temporary decrease in survival was seen in the early 1980s in both age groups. This may have been coincidental or it could be related to the relatively large percentage patients presenting with advanced stage, the relatively large number of patients who did not receive cancer therapy or learning effects of new therapies. The prognostic value of presenting symptoms has been described in detail and is a tool for clinical strategy [20], which has been applied in our region since the early 1980s. After adjustment for period of diagnosis and gender, prognostic factors adversely influencing overall survival were advanced age, histology (lymphocyte depletion) and advanced stage. Period of diagnosis also had a prognostic value independent of age, histology, stage and gender, with a nearly 2-fold reduction in hazard rate for patients diagnosed in the 1980s. As previously reported, cured patients exhibited an increased mortality rate compared to the general population. This increased mortality is due mainly to long-term treatment-related second malignancies observed, in particular, after combination therapy [21] and coronary heart disease associated with mantle field irradiation which can provoke increased atherosclerosis [22]. The 2-fold decrease in the cumulative incidence of a second malignancy during the 1980s was probably related to the choice of less toxic treatment strategies, such as the ABVD regimen. Improved survival has been described in several hospitalbased studies (Table 5) [3-51. In a study of 570 patients treated at a single centre in Bologna, the crude 5-year
440
D.J. van Spronsen et al. Table 5. Comparison of studies on 5-year survival of Hodgkin’s
Study IDHD
[3]
Type
Period
Hospital
1960-69 1970-79 19801970-80 1980-87 1960-70 1970-88 1973-77 1983-87 1978-80 1983-85 1977-82 1973-77 1978-82 1983-87 1972-76 1977-81 1982-86 1987-92
Bologna
[4]
Hospital
Florence
[5]
Hospital
Denmark
[7]
Population
Eurocare
[S]
Population
CCPDS [9] SEER [lo]
Hospital Population
This study
Population
* Estimated
from figures.
t Males.
$ Relative
n 14 225
570 1121
4677 2476 9418
345
disease Crude
5-year survival 73%* SO%* 82%* 75%* 90%* 65%* 85%* 57%? 64%t 58% 67% 77% 70%$ 73%$ 77%$ 60% 86% 73% 81%
survival.
survival rate improved from approximately 75% in the 1970s to 90% in the 1980s [4]. In a series of 1121 patients treated at one radiation institution in Florence, the crude 5-year survival rate improved from approximately 65% for patients treated before 1970 to 85% for patients treated between 1970 and 1988 [5]. In a large series of 14 225 clinical stage I-IV patients treated in 15 cancer centres or five co-operating groups all over the world, 5year overall survival rates improved from approximately 73% in the 1960s to 82% in the early 1980s [3]. With crude 5-year survival rates improving from 60% for the period 1972-1976 to 81% for the period 1987-1992, the results of our population-based study do not seem to differ much from those of most hospital-based studies performed in the same period. Although clinical trials are essential for establishing the efficacy of therapeutic advances, improvement in population-based survival due to the use of new and more effective treatment regimes in the general health care environment provide additional evidence of their effectiveness in general (Table 5). In a Danish registry-based study, the crude 5-year survival rates for males improved from 57% for the period 1973-1977 to 64% for the period 1983-1987 [7]. In a study of the results of several cancer registries throughout Europe, the crude 5-year survival rates improved from 58% for those diagnosed in the period 1978-1980 to 67% for the period 1983-1985 [8]. In the United States for the period 1977-1982, the crude 5-year survival rate was 62% for the population-based SEER-programme compared to 77% for the hospitalbased CCPDS-programme [9]. In another study of the SEER-programme, an improvement in the relative 5-year survival rate from 70% in 1973-1977 to 77% in 1983survival rates 1987 was reported [lo]. Population-based are lower compared to hospital-based series, since population-based studies are not biased by selection of patients and may comprise more elderly patients, patients with poor performance status, and possibly also some patients
with non-Hodgkin’s lymphoma (up to 17% in some series) [23]. A review of the medical records and histology revealed non-Hodgkin’s lymphoma in 5% of cases in our study. The improved prognosis for patients with Hodgkin’s disease is also demonstrated by the decrease in mortality rates in Scotland since 1970 [24]. Similarly, a significant reduction in annual mortality rates has been reported by the National Center for Health Statistics in the United States for the period 1979-1988 [25]. Also, in our region, mortality rates, derived from the cause-of-death register, have decreased since the 1960s from 2.4 to 0.3 per 100 000 person-years in the late 1980s for males, while the age-adjusted incidence decreased slightly [26]. It is likely that the decline was overestimated as a result of incorrect causes of death. In conclusion, population-based survival rates for patients with Hodgkin’s disease diagnosed and treated in general hospitals in South-East Netherlands in the past 20 years are more or less similar to those demonstrated in centre-based studies. The marked improvement in survival found for both the young and the elderly was independent of age, stage, histology and gender and thus reflects improved treatment results. Nonetheless, patients cured of Hodgkin’s disease exhibit an increased mortality risk due to treatmentrelated malignancies and, presumably, coronary heart disease.
Weinshel EL, Peterson BA. Hodgkin’s disease. Cancer 3 Clin 1993, 43, 327-346. Urba WJ, Longo DL. Hodgkin’s disease. N Engl 3 Med 1992, 326, 678-687. Henry-Amar M, Somers R. Survival outcome after Hodgkin’s disease: a report from the International Data base on Hodgkin’s Disease. Semin Oncol 1990, 17, 758-768. Mazia P, Bocchia M, Zinzani PL, et al. Hodgkin’s disease: summary of twenty years’ experience. Haematologi~a 1992, 17, 487-493.
Improved
Survival of Hodgkin’s
5. Biti G, Cellai E, Magrini SM, Papi MG, Ponticelli I’, Boddi V. Second solid tumors and leukemia after treatment for Hodgkin’s disease: an analysis of 1121 patients from a single Int J Radiation Oncol Biol Phys 1994, 29, 25-3 1. institution. C, Dekker A, et al. De 6. Van Rijswijk R, Verbeek J, Haanen resultaten van de behandeling van de ziekte van Hodgkin: 15 jaar ervaring van de universiteitsklinieken van Nijmegen en Utrecht. Ned Tijdchr Geneeskd 1987, 131, 1866-1870. J. Survival of Danish cancer patients 7. Storm HH, Clemmensen 1943-87. APMZS 1993, 189-193. R, Hakulinen T, 8. Berrino F, Sant M, Verdecchia A, Capocaccia Esteve J. Survival of Cancer Patients in Europe: the EUROCARE Study, 1978-85. Lyon, Scientific Publications, 1995, 363-372. 9. Davis S, Dahlberg S, Myers MH, Chen A, Steinhorn SC. Hodgkin’s disease in the United States: a comparison of patients characteristics and survival in the Centralised Cancer Patient Data System and the Surveillance Epidemiology and End Results Program. J Nat1 Cancer Inst 1987, 78, 471-478. 10. Medeiros LJ, Greiner TC. Hodgkin’s disease. Cancer 1995, 75, 357-369. 11. Erdkamp FLG, Wijlhuizen ThJ, van Dam FE. De ziekte van Hodgkin: een regionaal, retrospectief onderzoek. Ned Tzjdschr Geneeskd 1988, 132, 1801-1806. 12. Friedenberg WR, Dirks I’, Beltaos E, et al. Improved survival in the treatment of advanced Hodgkin’s disease at a nonuniversity institution (1970-79). Cancer 1986, 57, 12-17. 13. Hoppe RT, Hanlon AL, Hanks GE, Owen JB. Progress in the treatment of Hodgkin’s disease in the United States, 1973 versus 1983: the patterns of care study. Cancer 1994, 74, 31983203. 14. Lukes RJ, Butler JJ. The pathology and nomenclature of Hodgkin’s disease. Cancer Res 1966, 26, 1063-1081. 15. Bennet MH, Maclennan KA, Easterling MJ, Hudson BV, Jeliffe AM, Hudson GV. The prognostic significance of cellular subtypes in nodular sclerosing Hodgkin’s disease: an analysis of 271 non-laparotomised cases (BNLI report no 22). Clin Radio1 1983, 34, 497-501. 16. Carbone PI’, Kaplan HS, Musshoff K, et al. Report of the committee on Hodgkin’s disease staging classification. Cancer Res 1971, 31, 1860-1861. 17. Hakulinen T. Cancer survival corrected for heterogeneity in patients withdrawal. Biometrics 1982, 38, 933-942.
Patients
441
18. Hakulinen T, Tenkanen L. Regression analysis of relative survival rates. Appl Stat 1987, 36, 309-317. 19. Berkson J, Gage RI’. Calculation of survival rates for cancer. Proc Staff Meet Mayo Clin 1950, 25, 270-286. 20. Gobbi PG, Comelli M, Grignani GE, Piersca C, Bertoloni D, Ascari E. Estimate of expected survival at diagnosis in Hodgkin’s disease: a means of weighting prognostic factors and a tool for treatment choice and clinical research. A report from IDHD. Haematologica 1994, 79, 241-255. 21. Van Leeuwen FE, Klokman WJ, Hagenbeek A, et al. Second cancer risk following Hodgkin’s disease: a 20 year follow-up study. ‘3 Clin Oncol 1994, 12, 312-325. 22. Boivin JF, Hutchison GB, Lubin JH, Mauch I’. Coronary artery disease mortality in patients treated for Hodgkin’s disease. Cancer 1992, 169, 1241-1247. 23. Hartge I’, Devesa SS, Fraumeni JF Jr. In Doll R, Fraumeni JF Jr, Muir CS, eds. Trends in Cancer Incidence and Mortality. New York, Cold Spring Harbor Laboratory Press, 1994, 20, 423453. 24. Boyle I’, Soukop M, Scully C, et al. Improving prognosis of Hodgkin’s disease in Scotland. Eur 3 Cancer Clin Oncol 1988, 24, 229-234. 25. Hooper WC, Holman RC, Stine TW, Chorba TL. Hodgkin’s disease mortality in the United States. Cancer 1992, 70, 11661171. 26. Van Spronsen DJ, van der Heijden LH. Hodgkin’s disease. In Coebergh JWW, van der Heijden L, Janssen-Heijnen MLG, eds. Cancer Incidence and Survival in the South-East Netherlands Eindhoven, Comprehensive Cancer Center South, 1955-94. 1995, 82-83.
Acknowledgements-We thank Louis van der Heijden for data management, Theo Stijnen for his statistical expertise, the Central Bureau of Statistics for supplying regional mortality data and the hospitals participating in the registry of the Comprehensive Cancer Center South (Eindhoven, Geldrop, HelmondDeurne, Roermond, Veghel, Veldhoven, Venlo and Venray) for allowing access to the medical records.