Hodgkin disease survival in Europe and the U.S.

Prognostic significance of morphologic groups

Authors

  • Claudia Allemani PhD,

    Corresponding author
    1. Epidemiology Unit, Department of Preventive and Predictive Medicine, Istituto Nazionale per lo Studio e la Cura dei Tumori, Milano, Italy
    • Unit of Epidemiology, Department of Preventive and Predictive Medicine, Istituto Nazionale per lo Studio e la Cura dei Tumori, Via Venezian 1, I-20133, Milan, Italy
    Search for more papers by this author
    • Fax: (011) 39 0223903516

  • Milena Sant MD,

    1. Epidemiology Unit, Department of Preventive and Predictive Medicine, Istituto Nazionale per lo Studio e la Cura dei Tumori, Milano, Italy
    Search for more papers by this author
  • Roberta De Angelis ScD,

    1. National Center for Epidemiology, Surveillance, and Health Promotion, Department of Cancer Epidemiology, Istituto Superiore di Sanita, Rome, Italy
    Search for more papers by this author
  • Rafael Marcos-Gragera MD,

    1. Unitat d'Epidemiologia i Registre de Cancer de Girona, Institut Catala d'Oncologia, Spain
    Search for more papers by this author
  • Jan Willem Coebergh MD,

    1. Comprehensive Cancer Center South, Eindhoven Cancer Registry, Eindhoven, the Netherlands
    Search for more papers by this author
  • and the EUROCARE Working Group

    Search for more papers by this author
    • The EUROCARE Working Group for this article was: Austria: W. Oberaigner (Tyrol Cancer Registry); Denmark: H. H. Storm (Danish Cancer Society, Department of Cancer Prevention and Documentation); Estonia: T. Aareleid (Estonian Cancer Registry); Czech Republic: M. Jechova and M. Rousarova (Institute of Health Information and Statistics and West Bohemia Cancer Registry); Finland: T. Hakulinen (Finnish Cancer Registry); France: G. Hedelin (Bas-Rhin Cancer Registry), I. Tron and E. Le Gall (Bretagne Childhood Cancer Registry), G. Launoy (Calvados Digestive Cancer Registry), J. Mace-Lesec'h (Calvados General Cancer Registry), J. Faivre (Cote d'Or Digestive Cancer Registry), G. Chaplain (Cote d'Or Gynaecologic Cancer Registry), P.-M. Carli (Cote d'Or Malignant Haemopathies Registry), A. Danzon (Doubs Cancer Registry), B. Tretarre (Herault Cancer Registry), M. Colonna (Isere Cancer Registry), B. Lacour (Lorraine Childhood Cancer Registry), N. Raverdy (Somme Cancer Registry), C. Berger and B. Freycon (Rhone-Alpes Childhood Registry), P. Grosclaude (Tarn Cancer Registry), and J. Esteve (University of Lyon); Germany: P. Kaatsch (German Childhood Cancer Registry), H. Ziegler (Saarland Cancer Registry), and D. Holzel and G. Schubert Fritschle (Munich Cancer Registry); Iceland: L. Tryggvadottir (Icelandic Cancer Registry); Italy: F. Berrino (Project Leader), C. Allemani, P. Baili, L. Ciccolallo, P. Crosignani, G. Gatta, A. Micheli, M. Sant, E. Taussig, and S. Sowe (Istituto Nazionale per lo Studio e la Cura dei Tumori, Lombardy Cancer Registry); S. Ferretti (Ferrara Cancer Registry); E. Conti (Latina Cancer Registry); M. Vercelli and A. Quaglia (Liguria Region Cancer Registry-Department of Oncology, Biology and Genetics, University GE, IST Genova); F. Pannelli (Macerata Cancer Registry, Marche Childhood Cancer Registry); M. Federico and M. E. Artioli (Modena Cancer Registry); M. Ponz De Leon and P. Benatti (Modena Colorectal Cancer Registry), V. De Lisi and L. Servente (Parma Cancer Registry), R. Zanetti and S. Patriarca (Piedmont Cancer Registry), C. Magnani and G. Pastore (Piedmont Childhood Cancer Registry); L. Gafa and R. Tumino (Ragusa Cancer Registry); F. Falcini (Romagna Cancer Registry); M. Budroni (Sassari Cancer Registry); E. Paci and E. Crocetti (Tuscan Cancer Registry); P. Zambon and S. Guzzinati (Venetian Cancer Registry); and R. Capocaccia, E. Carrani, R. De Angelis, P. Roazzi, M. Santaquilani, A. Tavilla, F. Valente, and A. Verdecchia (Istituto Superiore di Sanita, Rome); Malta: M. Dalmas (Malta National Cancer Registry); Norway: F. Langmark and A. Andersen (Cancer Registry of Norway, Institute of Population-Based Cancer Research); Poland: J. Rachtan (Cracow Cancer Registry), M. Bielska-Lasota and Z. Wronkowski (Warsaw Cancer Registry); Slovakia: I. Plesko and A. Obsitnikova (National Cancer Registry of Slovakia); Slovenia: V. Pompe-Kirn (Cancer Registry of Slovenia); Spain: I. Izarzugaza (Basque Country Cancer Registry), C. Martinez-Garcia (Granada Cancer Registry), I. Garau (Mallorca Cancer Registry), C. Navarro and M. D. Chirlaque (Murcia Cancer Registry), E. Ardanaz and C. Moreno (Navarra Cancer Registry), J. Galceran (Tarragona Cancer Registry), A. Torrella (Childhood Tumor Registry of Valencia), and R. Peris-Bonet (National Childhood Cancer Registry and Instituto Lopez Pinero, Valencia); Sweden: L. Barlow and T. Moller (Cancer Registry of Sweden); Switzerland: G. Jundt (Basel Cancer Registry), J. M. Lutz, and C. Bouchardy (Geneva Cancer Registry); the Netherlands: J. W. W. Coebergh (Eindhoven Cancer Registry), A. van der Does-van den Berg (Dutch Childhood Oncology Group), and O. Visser (Amsterdam Cancer Registry); United Kingdom–England: S. Godward (East Anglian Cancer Registry), M. P. Coleman (London School of Hygiene and Tropical Medicine), E. M. I. Williams (Merseyside and Cheshire Cancer Registry), D. Forman (Northern and Yorkshire Cancer Registry and Information Service), M. J. Quinn (Office for National Statistics), M. Roche and S. Edwards (Oxford Cancer Intelligence Unit), C. Stiller (Childhood Cancer Research Group, Oxford), J. Verne (South West Cancer Intelligence Services), H. Moller and J. Bell (Thames Cancer Registry), J. L. Botha (Trent Cancer Registry), and G. Lawrence (West Midlands Cancer Intelligence Unit); UK–Scotland: R. Black (Scottish Cancer Intelligence Unit); and UK–Wales: J. A. Steward (Welsh Cancer Intelligence and Surveillance Unit).


Abstract

BACKGROUND.

The survival of patients with Hodgkin disease (HD) varies markedly across Europe and generally is shorter than the survival of patients in the U.S. To investigate these differences, the authors compared population-based HD survival in relation to morphologic type among populations in Europe and the U.S.

METHODS.

The authors analyzed 6726 patients from 37 cancer registries that participated in EUROCARE-3 and 3442 patients from 9 U.S. Surveillance, Epidemiology, and End Results (SEER) registries. Patients were diagnosed during 1990 to 1994 and were followed for at least 5 years. The European registries were grouped into EUROCARE West, EUROCARE UK, and EUROCARE East. Morphologic groups were nodular sclerosis, mixed cellularity, lymphocyte depletion, lymphocyte predominance, and not otherwise specified (NOS). The influence of morphology on geographic differences in 5-year relative survival was explored by using multiple regression analysis.

RESULTS.

In the model that was adjusted by age, gender, and years since diagnosis, the relative excess risk (RER) of death was 0.93 (95% confidence interval [95% CI], 0.81–1.05) in EUROCARE West, 1.15 (95% CI, 1.04–1.28) in EUROCARE UK, and 1.39 (95% CI, 1.21–1.60) in EUROCARE East (compared with the SEER data). When morphology was included, EUROCARE UK and SEER no longer differed (RER, 1.06; 95% CI, 0.95–1.18). Morphology distribution varied markedly across Europe and much less in the U.S., with nodular sclerosis less common in Europe (45.9%) than the U.S.(61.7%). The RER data showed that patients who had lymphocyte depletion, NOS, and mixed cellularity had a significantly worse prognoses compared with patients who had nodular sclerosis, whereas patients who had lymphocyte predominance had the best prognosis.

CONCLUSIONS.

The current results provide population-based evidence that morphology strongly influences the prognosis of patients with HD. However differences in the morphologic case mix explains only some of the geographic variations observed in survival. Cancer 2006. © 2006 American Cancer Society.

In the EUROCARE-3 study, considerable variations in survival were observed among patients with hematologic malignancies across Europe. The 5-year relative survival for patients with Hodgkin disease (HD) ranged from 62% to 88% in participating registries,1 even though effective treatments for the disease have been available since the 1970s2 and should have disseminated throughout Europe. Overall 5-year relative survival was higher in the U.S. (83%; Surveillance, Epidemiology, and End Results [SEER] Program)3 than in Europe (78%).1

Various morphologic subtypes of HD are recognized and are associated with differing prognoses.4 Intercountry differences in survival may be caused by different access to and availability of appropriate therapies, different stage at diagnosis, or different morphologic case mix. The objective of this study was to explore at the population level the influence of morphology on the geographic differences in 5-year relative survival for patients with HD, as reported by EUROCARE and SEER, taking account of differences in the gender and age distribution at diagnosis of the various populations and the years since diagnosis.

MATERIALS AND METHODS

We analyzed 6726 patients who were diagnosed with HD from 1990 through 1994 in 37 of the 67 population-based cancer registries that participate in EUROCARE-31 and 3442 patients who were diagnosed with HD over the same period in 9 U.S. cancer registries (the San Francisco-Oakland Standard Metropolitan Statistical Area, Connecticut, Detroit-Metropolitan, Hawaii, Iowa, New Mexico, Seattle [Puget Sound], Utah, and Atlanta-Metropolitan) for which data on survival and treatment were available in the SEER public-use data base.3 All patients were followed for at least 5 years. The EUROCARE-3 registries that were included in the study were those that provided information on morphology according to the International Classification of Diseases for Oncology, 2nd edition (ICDO-2)5 and had <35% of patients with “not otherwise specified” (NOS) morphology.

The European registries were grouped geographically: EUROCARE West included registries from France, Germany, Italy, the Netherlands, Spain, Switzerland and the national registries of Iceland and Malta; EUROCARE UK included the English registries of East Anglia, Mersey, Oxford, South West, Trent, Yorkshire and the national registry of Scotland; and EUROCARE East included the Polish registry of Warsaw and the national registries of the Czech Republic, Estonia, Slovakia, and Slovenia. The SEER cancer registries were considered together, because their survival rates and case mixes are relatively homogeneous.3

One hundred nineteen patients (EUROCARE: 106 patients, 1.6%; SEER: 13 patients, 0.4%) who were known to registries by death certificate only (DCO) or who were diagnosed only at autopsy were excluded from the current analyses. HD morphology was classified according to the ICDO-2 in EUROCARE registries, while for the SEER cancer registries the ICDO-3 classification was available.6 For HD, the changes in the new edition are trivial coding changes. Therefore, we transformed ICDO-2 codes to ICDO-3 codes (codes 9657 and 9658 to code 9651, code 9666 to code 9665, and code 9660 to code 9659).

We grouped morphology codes into 5 categories: NOS (codes 9650 and 9661–9662); mixed cellularity (code 9652), lymphocyte depletion (codes 9653–9655), lymphocyte predominance (codes 9651 and 9659),7 and nodular sclerosis (codes 9663–9665 and 9667). This classification, which was proposed 40 years ago, is used widely without substantial modification in all countries.8 We used 4 age categories (0–14 years, 15–44 years, 45–64 years, and 65–99 years).

Five-year relative survival was estimated by using the Hakulinen method9 with 95% confidence intervals (95% CIs) calculated from the standard error according to Greenwood's method.10 Relative survival is an estimate of the probability of cancer survival after adjusting for competing causes of death determined from general population life tables, which are specific for each country or registry. For Europe, general population life tables that were specific for each country or registry territory were used. For the U.S., only race/ethnicity-specific national life tables were available in the SEER data set. To obtain a national life table for all races/ethnicities, we weighted the race/ethnicity-specific national life tables according to the proportions of whites, blacks and others present among the SEER HD patients.

Differences in 5-year relative survival by country were modeled by using a multiple regression approach based on generalized linear models and adopting the Poisson assumption for the observed number of deaths.11 The relative excess risks (RERs) of death derived from these models quantified the extent to which the risk of death in a given group differed from that in the reference category after taking into account the background risk of death in the general population of each country or region.

RESULTS

Table 1 shows the total numbers of patients by registry and geographic grouping, with relative quality indicators (percentages of patients lost to follow-up, autopsy/DCO diagnoses, microscopically confirmed diagnoses, and patients with NOS morphology). Patients who were lost to follow-up included 1.0% of patients (range, 0.0–13.0%) in EUROCARE West, 0.1% of patients (range, 0.0–1.3%) in EUROCARE UK, 1.6% of patients (range, 0.0–9.0%) in EUROCARE East, 0.7% of patients in EUROCARE overall, and 0.0% of patients in SEER. Patients with autopsy/DCO diagnoses included 1.0% of patients (range, 0.0–6.7%) in EUROCARE West, 0.3% of patients (range, 0.0–0.9%) in EUROCARE UK, 5.7% of patients (range, 1.7–8.7%) in EUROCARE East, 1.6% of patients in EUROCARE overall, and 0.4% in SEER (range 0.0–1.0). Patients with microscopically verified diagnoses included 97.1% of patients (range, 84.5–100.0%) in EUROCARE West, 95.8% of patients (range, 89.7–100.0%) in EUROCARE UK, 96.8% of patients (range, 92.1–100.0%) in EUROCARE East, 97.1 of patients in EUROCARE overall, and 99.2% of patients (range, 95.8–99.8%) in SEER. Note that most of the “autopsy/DCO” diagnoses, in fact, were autoptic and therefore were verified microscopically. Thus, the proportions of microscopically verified and autopsy/DCO diagnoses are not necessarily mutually exclusive.

Table 1. Numbers of Patients, with Quality Indicators, for the EUROCARE and Surveillance, Epidemiology, and End Results Cancer Registries Included in the Current Study
Country/AreaRegistryNo. of patientsLost to follow-up (%)Autopsy/DCO (%)Microscopically verified (%)NOS (%)
  • DCO indicates death certificate only; NOS, not otherwise specified; SEER, Surveillance, Epidemiology, and End Results.

  • *

    Registries that cover the entire country.

  • Categories that incorporate the registries listed above, except in United States (which includes the San Francisco-Oakland Standard Metropolitan Statistical Area, Connecticut, Detroit-Metropolitan, Hawaii, Iowa, New Mexico, Seattle [Puget Sound], Utah, and Atlanta-Metropolitan).

FranceBas-Rhin1281.60.0100.014.1
 Calvados580.00.0100.08.6
GermanySaarland1340.02.297.820.1
IcelandIceland*300.06.7100.026.7
ItalyFerrara573.51.798.229.8
 Genoa1150.00.092.217.4
 Macerata470.00.095.729.8
 Modena995.03.096.013.1
 Parma480.00.0100.033.3
 Ragusa440.00.0100.027.3
 Romagna500.00.0100.024.0
 Sassari420.00.097.64.8
 Turin1452.10.098.614.5
 Tuscany2000.02.084.525.5
 Varese1362.20.797.87.3
MaltaMalta*290.00.0100.013.8
NetherlandsAmsterdam3120.00.3100.011.9
 Eindhoven990.00.0100.06.1
SpainBasque2550.02.395.77.4
 Granada490.00.0100.014.3
 Mallorca690.00.0100.04.3
 Navarra620.01.6100.08.1
 Tarragon640.01.698.420.3
SwitzerlandBasel175.90.0100.017.6
 Geneva5413.01.898.113.0
EUROCARE West 23431.01.097.115.0
 East Anglia2331.30.089.76.9
 Mersey2300.00.997.426.5
 Oxford3240.00.0100.027.8
 Scotland6700.10.696.330.7
 South West7650.00.097.023.7
 Trent4790.00.690.628.4
 Yorkshire4340.00.597.714.5
EUROCARE UK 31350.10.395.824.0
Czech RepublicCzech*1310.05.3100.016.0
EstoniaEstonia*1981.04.599.018.2
PolandWarsaw1779.01.792.126.5
SlovakiaSlovakia*5410.08.795.714.2
SloveniaSlovenia*2011.02.599.520.9
EUROCARE East 12481.65.796.817.9
EUROCARE overall 67260.71.697.119.7
USASEER34420.00.499.210.3

Patients with NOS morphology included 15.0% of patients (range, 4.3–33.3%) in EUROCARE West, 24.0% of patients (range, 6.9–30.7%) in EUROCARE UK, 17.9% of patients (range, 14.2–26.5%) in EUROCARE East, 19.7% of patients in EUROCARE overall, and 10.3% of patients (range, 5.8–16.2%) in SEER.

Table 2 shows morphology by country and geographic grouping. Nodular sclerosis was identified in 48.6% of patients (range, 31.0–73.2%) in EUROCARE West, 48.2% of patients (range, 39.0–50.7%) in EUROCARE UK, 35.1% of patients (range, 19.2–52.0%) in EUROCARE East, 45.9% of patients in EUROCARE overall, and 61.7% of patients in SEER. Mixed cellularity was identified in 23.6% of patients (range, 9.7–31.2%) in EUROCARE West, 14.8% of patients (range, 14.7–15.1%) in EUROCARE UK, 26.6% of patients (range, 14.1–31.8%) in EUROCARE East, 20.1% of patients in EUROCARE overall, and 19.6% of patients in SEER. Lymphocyte depletion was identified in 4.8% of patients (range, 0.0–17.2%) in EUROCARE West, 4.4% of patients (range, 4.0–5.7%) in EUROCARE UK, 6.1% of patients (range, 1.0–9.9%) in EUROCARE East, 4.8% of patients in EUROCARE overall, and 2.0% of patients in SEER. Lymphocyte predominance was identified in 8.0% of patients (range, 3.3–20.7%) in EUROCARE West, 8.6% of patients (range, 8.4–9.5%) in EUROCARE UK, 14.3% of patients (range, 2.0–29.3%) in EUROCARE East, 9.5% of patients in EUROCARE overall, and 6.4% of patients in SEER. Finally, NOS morphology was identified in 15.0% of patients (range, 9.4–26.7%) in EUROCARE West, 24.0% of patients (range, 22.2–30.7%) in EUROCARE UK, 17.9% of patients (range, 14.2–26.5%) in EUROCARE East, 19.7% of patients in EUROCARE overall, and 10.3% of patients in SEER.

Table 2. Distribution (Percentage) of Each Morphologic Group by Country (European) and Geographic Grouping
Country/AreaMorphologic group
No. of patientsNodular sclerosis (%)Mixed cellularity (%)Lymphocyte depletion (%)Lymphocyte predominance (%)NOS (%)
  1. SEER indicates Surveillance, Epidemiology, and End Results; NOS, not otherwise specified.

  2. * EUROCARE West consists of registries from France, Germany, Italy, the Netherlands, Spain, Switzerland, and the national registries of Iceland and Malta; EUROCARE UK consists of the English registries of East Anglia, Mersey, Oxford, South West, Trent, Yorkshire, and the national registry of Scotland; and EUROCARE East consists of the Polish registry of Warsaw and the national registries of the Czech Republic, Estonia, Slovakia, and Slovenia.

France18647.331.23.75.412.4
Germany13436.629.96.76.720.1
Iceland3060.010.00.03.326.7
Italy98342.326.44.08.219.1
Malta2931.017.317.220.713.8
Netherlands41173.29.71.74.910.5
Spain49946.125.78.410.49.4
Switzerland7138.029.65.612.714.1
EUROCARE West234348.623.64.88.015.0
England246550.714.74.08.422.2
Scotland67039.015.15.79.530.7
EUROCARE UK313548.214.84.48.624.0
Czech Republic13129.026.09.919.116.0
Estonia19819.224.29.129.318.2
Poland17752.014.13.44.026.5
Slovakia54131.431.86.915.714.2
Slovenia20149.726.41.02.020.9
EUROCARE East124835.126.66.114.317.9
EUROCARE overall672645.920.14.89.519.7
SEER344261.719.62.06.410.3

The distribution according to morphologies in the SEER registries was more homogeneous than that of the EUROCARE registries. Furthermore, the overall proportion of patients with NOS morphology in the SEER registries was approximately equal to that of the EUROCARE-3 registries with the lowest proportions of patients with NOS morphology.

Table 3 shows the distribution of morphologic groups by age in the EUROCARE and SEER data sets. The proportions NOS morphology and lymphocyte depletion were higher in the EUROCARE population than in the SEER population for all age categories, whereas the proportion with mixed cellularity was higher in the EUROCARE population for all age groups except for ages 45 to 64 years, and lymphocyte predominance was higher in the EUROCARE population for the age groups 15 to 44 years and 65 to 99 years. The proportion of nodular sclerosis was higher in the SEER population for all age groups.

Table 3. Distribution (Percentage) of Each Morphologic Group by Age in the EUROCARE and Surveillance, Epidemiology, and End Results Registry Pools
Morphologic groupDistribution (%)
Ages Birth-14 yearsAges 15–44 yearsAges 45–64 yearsAges 65–99 years
EUROCARESEEREUROCARESEEREUROCARESEEREUROCARESEER
  1. SEER indicates Surveillance, Epidemiology, and End Results; NOS, not otherwise specified.

NOS18.83.515.58.521.812.332.210.3
Mixed cellularity21.318.215.914.527.031.325.719.6
Lymphocyte depletion2.30.03.21.06.42.09.32.0
Lymphocyte predominance13.615.49.35.110.610.17.56.4
Nodular sclerosis44.062.956.170.934.244.325.361.7
Total100.0100.0100.0100.0100.0100.0100.0100.0
Total no. of patients3091433908227714025621107460

Figure 1 shows 5-year relative survival, overall and by morphologic group, in the EUROCARE groupings and SEER registries. Overall survival was 80% (95% CI, 78–82) in EUROCARE West, 77% (95% CI, 75–79%) in EUROCARE UK, 75% (95% CI, 72–78%) in EUROCARE East, and 83% (95% CI, 82–84%) in SEER. The highest survival rate was observed for lymphocyte predominance (EUROCARE West, 88% [95% CI, 83–93%]; EUROCARE UK, 100% [95% CI, 96–104%]; EUROCARE East, 72% [95% CI, 64–80%]; SEER, 88% [95% CI, 83–93%]), followed by nodular sclerosis (EUROCARE West, 86% [95% CI, 84–88%]; EUROCARE UK, 82% [95% CI, 80–84%]; EUROCARE East, 87% [95% CI, 84–90%]; SEER, 87% [95% CI, 85–89%]). The lowest survival rate was observed for lymphocyte depletion (EUROCARE West, 54% [95% CI, 43–65%]; EUROCARE UK, 53% [95% CI, 44–62%]; EUROCARE East, 55% [95% CI, 42–68%]; SEER, 50% [95% CI, 36–64%]). Mixed cellularity (EUROCARE West, 73% [95% CI, 69–77%]; EUROCARE UK, 74% [95% CI, 69–79%]; EUROCARE East 71% [95% CI, 65–76;%]; SEER, 75% [95% CI, 71–79%]) and NOS (EUROCARE West, 72% [95% CI, 67–77%]; EUROCARE UK, 66% [95% CI, 62–70%]; EUROCARE East, 62% [95% CI, 54–70%]; SEER, 70% [95% CI, 64–76%]) were characterized by intermediate survival.

Figure 1.

Five-year relative survival is illustrated according to morphologic group in EUROCARE West, EUROCARE UK, EUROCARE East, and the Surveillance, Epidemiology, and End Results (SEER) Program.

In general, there were no significant differences between the 4 geographic groupings (3 in Europe and 1 in the U.S.) in terms of survival within each morphologic group, except for patients who had lymphocyte predominant morphology, for whom 5-year survival was significantly longer in EUROCARE UK than in all other groupings.

Table 4 shows 5-year RERs of death by geographic grouping adjusted by years since diagnosis, age at diagnosis and gender (Model 1) and also for each morphologic group (Model 2), in each case taking into account the influence of the other variables. The SEER was used as the reference category.

Table 4. Relative Excess Risks by Area Adjusted for Years since Diagnosis, Age at Diagnosis, Gender (Model 1) and Morphologic Group (Model 2)
VariableModel 1Model 2
RER95% CIRER95% CI
  • RER indicates relative excess risk; 95% CI, 95% confidence interval; SEER, Surveillance, Epidemiology, and End Results; NOS, not otherwise specified.

  • *

    EUROCARE West consists of registries from France, Germany, Italy, the Netherlands, Spain, Switzerland, and the national registries of Iceland and Malta; EUROCARE UK consists of the English registries of East Anglia, Mersey, Oxford, South West, Trent, Yorkshire, and the national registry of Scotland; and EUROCARE East consists of the Polish registry of Warsaw and the national registries of the Czech Republic, Estonia, Slovakia, and Slovenia.

Area*
 SEER (USA)1 1 
 EUROCARE West0.930.81–1.050.890.79–1.00
 EUROCARE East1.391.21–1.601.321.15–1.52
 EUROCARE UK1.151.04–1.281.060.95–1.18
Years since diagnosis
 1 y1 1 
 2 y0.540.49–0.610.540.49–0.61
 3 y0.380.33–0.430.380.33–0.43
 4 y0.320.28–0.370.320.28–0.37
 5 y0.250.21–0.300.250.21–0.30
Age, y
 Birth-141 1 
 15–442.511.66–3.802.501.65–3.78
 45–647.344.84–11.146.914.55–10.48
 65–9922.2114.67–33.6219.1012.62–28.92
Gender
 Male1 1 
 Female0.790.72–0.860.810.74–0.88
Morphologic group
 Nodular sclerosis  1 
 NOS  1.831.63–2.06
 Mixed cellularity  1.271.13–1.43
 Lymphocyte depletion  2.422.05–2.86
 Lymphocyte predominance  0.760.62–0.92

For Model 1 the excess risk of death was significantly higher than reference in EUROCARE East (RER, 1.39; 95% CI, 1.21–1.60) and EUROCARE UK (RER, 1.15; 95% CI, 1.04–1.28) and was nonsignificantly lower than reference for EUROCARE West (RER, 0.93; 95% CI, 0.81–1.05). The excess risk of death was highest in the first year after diagnosis (reference category), and RER decreased to 0.25 (95% CI, 0.21–0.30) in the 5th year. RER of death increased with increasing age at diagnosis (reference category, ages 0–14 years), and the excess risk of death was 20% lower in females (RER, 0.79; 95% CI, 0.72–0.86) compared with males.

When morphologic groups were included in the model (Model 2), the RERs did not change substantially for EUROCARE East or EUROCARE West, but the difference between EUROCARE UK and SEER no longer was present (RER, 1.06; 95% CI, 0.95–1.18). The RER by morphologic group was 2.42 (95% CI, 2.05–2.86) for lymphocyte depletion, 1.83 (95% CI, 1.63–2.06) for NOS, 1.27 (95% CI, 1.13–1.43) for mixed cellularity, and 0.76 (95% CI, 0.62–0.92) for lymphocyte predominance; nodular sclerosis was used as the reference category.

We also tested interactions between area and follow-up, morphologic group and follow-up, and morphologic group and age. All interactions were significant, but their inclusion in the model only slightly improved the goodness of fit of the model and only slightly altered excess risks by area. Therefore, interactions are not presented but are available on request.

DISCUSSION

Although clinical studies have found that the various HD morphologies have differing prognoses,4, 12–14 only a few population-based studies have addressed this issue.15, 16 The current study, in which we investigated large European and American populations by using multivariable analysis, provides strong evidence that morphology has a considerable influence on the survival of patients with HD. In particular, we substantiated previous data showing that patients with lymphocyte predominance had the best prognosis, and patients with lymphocyte depletion had the poorest prognosis.4, 13, 14

The other major (and related) finding of our study is that differences in morphologic case mix may explain a considerable part of the geographic differences we observed in survival. Thus, our overall survival analysis showed that 5-year relative survival in EUROCARE West did not differ significantly from survival in the other 3 areas, whereas survival in EUROCARE UK and EUROCARE East were similar, and both were significantly lower than survival in SEER (Fig. 1). Furthermore, when Model 1 was applied, which adjusted for age, gender, and years from diagnosis, the excess risks of death in EUROCARE East and EUROCARE UK were similar and significantly higher than in SEER. However, when Model 2 was applied, which also adjusted for morphology, the excess risk of death for EUROCARE UK became indistinguishable from that of SEER (RER, 1.06; 95% CI, 0.95–1.18) and was considerably lower than that for EUROCARE East.

Thus, differences in morphologic case mix accounted for the apparent similarities in overall survival between EUROCARE East and EUROCARE UK (Fig. 1, Table 4). It is unlikely that disease stage at diagnosis, which was not available in our study, would modify this finding; because, for patients who have cancers such as HD, that are treated with chemotherapy, disease stage does not have the same importance as a prognostic indicator that it has for patients who have cancers that are treated by surgery: In the EUROCARE-1 study,17 survival in England and Wales was close to the European average survival for cancers like HD, that are treated mainly by chemotherapy or combined approaches; whereas, for cancers that are treated mainly by surgery (stomach, colon-rectum, kidney and breast), survival was lower than the European average.18 Furthermore, disease stage at diagnosis is related to morphology in patients with HD: Lymphocyte predominant HD typically is diagnosed at an earlier stage than lymphocyte depletion HD.19

It is also worth noting that the RER of death in EUROCARE West versus SEER was 0.93 (95%CI, 0.81–1.05) with Model 1. Following adjustment for morphology (Model 2), the RER did not change significantly. This finding suggests that the diagnostic-therapeutic approach to HD in the countries of EUROCARE West is similar to that of the SEER areas. By contrast, the countries of EUROCARE East were characterized by significantly higher excess risk of death than the other European groupings and SEER in both models.

It is important to address issues of data comparability and quality that potentially may affect the robustness of our findings. We excluded several European registries from our analysis because of incomplete information on morphology: The proportion of patients with NOS morphology still was higher in Europe than in the U.S., although the European 5-year relative survival for each morphologic group was similar to survival in the SEER pool of registries, suggesting that the quality of the EUROCARE data set we used is comparable to that of the SEER data set.

Another important issue is the comparability of HD diagnoses and morphologic classifications between different countries and registry areas. This is an ongoing concern of population-based series.20 In the study by Glaser et al.,20 HD diagnoses and morphologies among women recorded by cancer registries were compared with those obtained by subsequent expert review. Those authors observed that, for the most common morphology (nodular sclerosis), diagnoses were largely concordant, but diagnoses were less concordant for the other morphologies. However, although these discordances had a slight effect on incidence rates, they did not influence 5-year observed survival overall or by age.

A strength of the current study is that the effects of competing causes of death were controlled for by estimating relative survival. A population-based study on the SEER data base15 that analyzed the relative risk of death for HD in relation to morphology did not estimate survival in this way but used disease-specific survival from death certificates to control for competing causes of death. Unfortunately, causes of death are not always reported accurately on death certificates,21–23 whereas relative survival does not require information on cause of death. The use of relative survival is particularly important for analyzing HD data, because an HD incidence peak occurs in individuals older than age 50 years, for whom competing causes of death are important. Conversely, for childhood HD (the other age peak in HD incidence), observed and relative survival overlap.

Consistent with previous studies,15, 24–26 we observed that survival in older patients was worse than in younger patients. It is known that immune responses decline with increasing age and may contribute to poorer outcomes in the elderly.27 The inability to tolerate adequate treatment also seems to be an important determinant of poorer overall survival in older patients with HD, particularly because comorbidities, which are common among patients in this age group, may hamper the administration of multimodal systemic treatments, especially chemotherapy.24, 26, 28–30 We found that the distribution of morphologic types also varied with age: Nodular sclerosis occurred more frequently in young individuals, and mixed cellularity and lymphocyte depletion were more common in older patients. Thus, the presence of more aggressive morphologies in the elderly also may contribute to their poorer prognoses; nevertheless, age was an independent prognostic factor.

Similar to previous studies,12, 14, 15, 31–34 we observed that females had a lower excess risk of death than males (RER, 0.79; 95% CI, 0.72–0.86). Neither differences in morphology nor differences in age were able to account for the better prognoses in women.

In conclusion, differences in excess risk of death between the geographic regions diminished when corrected for morphology, indicating that differences in morphologic case mix are an important determinant of regional survival differences for HD. Nevertheless, survival differences did not disappear completely with this adjustment, and it is likely that differences in type of treatment and disease stage at diagnosis also contributed. In particular, more advanced stage at diagnosis and less adequate treatment may explain the worse outcomes in EUROCARE East. Unfortunately, information on stage at diagnosis and treatment is not available in the EUROCARE data set, and we were unable to assess their well known effects on prognosis35 at the population level. To address this issue, we have initiated a cancer registry-based project called HAEMACARE to standardize the morphologic classification of hematologic malignancies across European countries and to monitor care.

Acknowledgements

We thank Don Ward for help with the English and Samba Sowe for secretarial assistance

Ancillary