Rising Incidence of Pancreatic Cancer in France
Bouvier, Anne-Marie MD, PhD; David, Mathilde MD; Jooste, Valérie PhD; Chauvenet, Marion MD; Lepage, Côme MD, PhD; Faivre, Jean MD, PhD
From the Registre Bourguignon des Cancers Digestifs; Centre Hospitalier Universitaire de Dijon; National Institute for Health and Medical Research; and Université de Bourgogne, Dijon, France.
Received for publication November 3, 2009; accepted March 24, 2010.
Reprints: Anne-Marie Bouvier, MD, PhD, Registre Bourguignon des Cancers Digestifs (INSERM U866), Faculté de Médecine, BP 87 900, 21079 Dijon Cedex, France (e-mail: Anne-Marie.Bouvier@u-bourgogne.fr).
Objectives: Time trends in the incidence of pancreatic cancer vary considerably between countries. The aim of this study was to provide time trends in incidence during a 25-year period in a well-defined French population.
Methods: The cancer registry in Burgundy (France) was used to study time trends between 1981 and 2005 by sex, age, subsite, and histology. They were analyzed using an age-period cohort model.
Results: Age-standardized incidence rates increased from 5.7 (1981-1985) to 7.9 per 100,000 (2001-2005) in men and from 2.6 to 4.6 in women. The mean percentage of variation by the 5-year period was +9.9% (P < 0.001, 95% confidence interval, 6.2%-13.6%) and +13.4% (P = 0.004, 95% confidence interval, 9.4%-17.5%), respectively. The increase in incidence was higher for cancers of the tail and corpus than for cancers of the head of pancreas and for malignant pancreatic endocrine tumors than for adenocarcinomas. The cumulative risk of developing a pancreatic cancer rose from 0.51% for men born in 1900 to 1.13% for those born in 1950. It was 0.34% and 0.55% for women, respectively.
Conclusions: The incidence of pancreatic cancer has increased sharply in France both by period and by birth cohort. Etiological studies are required to explain the increase in the incidence.
Despite advances in diagnostic procedures over time, only minor changes have occurred in the management and survival of patients with pancreatic cancer.1 Pancreatic cancer is currently generating particular interest among clinicians and epidemiologists because it is a major challenge in oncology. The symptoms of pancreatic cancer are nonspecific and appear relatively late, and currently, there is no effective method for early detection. Pancreatic cancer accounted for 11.5% of diagnosed digestive cancers in France in 2005.2 The number of new cases was estimated to be 7218. Studies from population-based cancer registries have shown that patterns for pancreatic cancer incidence differ from one area to another.3-7 The incidence was shown to be increasing in some areas, whereas there was no change or a decrease in others. Most pancreatic cancers are adenocarcinomas. However, little is known about trends by histological type. Secular trends in the incidence by period and by birth cohort could provide clues crucial to a better understanding of the disease. However, few studies have taken into account the effects of age, the period of diagnosis, and the birth cohort.8-11 The aim of this study was to provide data on trends in the incidence of pancreatic cancers in a well-defined French population during a 25-year period and to specifically disentangle the influence of age, period of diagnosis, and birth cohort.
MATERIALS AND METHODS
The incidence data originated from the population-based digestive cancer registry, which covers the resident population of 2 well-defined administrative areas in Burgundy, France (Côte d'Or and Saône-et-Loire). These areas have a combined population of 1,050,000 according to the 1999 census. Information is regularly obtained from pathologists, hospital and private physicians (gastroenterologists, surgeons, and oncologists), general practitioners, and death certificates. Because they are somewhat unreliable, death certificates are only used to identify the missing cases. The quality and exhaustiveness of the registry are certified every 4 years by an audit of the National Institute for Health and Medical Research and of the National Public Health Institute.
The diagnosis of pancreatic cancer (International Classification of Diseases, 10th Revision, C25) either was made histologically or was based on medical imaging. Histological diagnosis was classified according to the International Classification of Diseases for Oncology.12 There were 1304 adenocarcinomas, 78 malignant pancreatic endocrine tumors, and 114 cases with other histological diagnosis (56 undifferentiated carcinomas, 36 cystadenocarcinomas, 10 malignant lymphomas, 7 epidermoid carcinomas, and 5 sarcomas).
The population data used to calculate incidence rates were estimated by interpolation and extrapolation of the results of the 4 French censuses (1975, 1982, 1990, and 1999). Incidence rates were standardized by the direct method using the world standard population. Time trends were analyzed using the age-period-cohort model. Polynomial functions were used to smooth any age, period, and cohort effect. Because of the relative rarity of pancreatic cancers, polynomial models were fitted as generalized linear models with Poisson-distributed error and logarithmic link. Models were compared using likelihood ratio tests at the 0.05 level. The goodness of fit was used to assess the adequacy of the model. The best-fitting age-cohort model was used to estimate the relative risk by birth cohort. Each cohort included all patients born during a 5-year period and was defined by the central birth year. For example, the 1900 cohort corresponds to subjects born between 1898 and 1902, and so on. Predicted relative risks by birth cohort were presented for the 1900 cohort to the 1950 cohort. For each individual, the relative risk is the risk of developing the disease during the 0- to 74-year age span compared with that of the central birth cohort, that is, from 1928 to 1932. Time trends were analyzed with Stata 10 software (StatCorp LP, College Station, Tex).
A total of 2740 pancreatic cancers were diagnosed between 1981 and 2005. Age was tabulated into 6 classes: younger than 45 years (n = 60), between 45 and 54 years (n = 210), between 55 and 64 years (n = 473), between 65 and 74 years (n = 836), between 75 and 84 years (n = 839), and older than 84 years (n = 320; 2 subjects with age unknown). The time of diagnosis was divided into 5 periods of 5 years: 1981 to 1985 (n = 330), 1986 to 1990 (n = 479), 1991 to 1995 (n = 529), 1996 to 2000 (n = 665), and 2001 to 2005 (n = 738).
Time Trends in Incidence by Period of Diagnosis
The incidence of pancreatic cancer increased for both sexes over time. Age-standardized rates varied from 5.7 per 100,000 for the 1981-1985 period to 7.9 per 100,000 for the 2001-2005 period in men and from 2.6 to 4.7 per 100,000 in women. The corresponding mean percentage of variation by a 5-year period was +9.9% (P < 0.001, 95% confidence interval [CI], 6.2%-13.6%) in men and +13.4% (P = 0.004, 95% CI, 9.4%-17.5%) in women.
Trends in incidence rates by age classes for men and women are given in Figure 1. For both sexes, incidence increased with periods. The mean variations by 5-year period and sex are given in Table 1. Variations were higher for women. It varied for men between 8.7% for the 55- to 64-year-old age group (P = 0.034) and 14.7% for the 85 years or older age group (P = 0.061). For women, the mean variation by 5-year period ranged from 7.9% for the 85 years or older age group (P = 0.124) to 16.8% for the 65- to 74-year-old age group (P = 0.001).
For both sexes, the increase in incidence was higher for cancers located in the tail and the corpus than for cancers located in the head of the pancreas (Table 1). The increase in the incidence of endocrine tumors was more marked than that of adenocarcinomas. There was a decrease in the incidence of other rare histological tumors.
Time Trends in Incidence by Birth Cohort
The cohort's cumulative risk of developing pancreatic cancer for the age range of 0 to 74 years by sex is given in Figure 2. Trends in relative risk of pancreatic cancer were different for men and women: there was a sharp increase in risk throughout the studied cohorts for men. In women, after an increase in risk until the 1930 birth cohort, the cumulative risk tended to stabilize throughout the more recent cohorts. The cumulative risk rose from 0.51% for men born around 1900 to 1.13% for those born around 1950, a more than 2-fold increase. The corresponding values in women were 0.34% and 0.55%, a 1.6-fold increase.
In Figure 3, the incidence rates of pancreatic cancer are plotted by sex and age for selected cohorts. Age-specific incidence rates increased with advancing birth cohorts, and this was more marked in men than in women. At every age, the incidence rates were 3 times higher for men born in 1950 than for those born in 1900. In women, there was a 2-fold increase. The incidence in people aged 85 years and older was 15 times that in the 50- to 54-year-old age group, and this was true in men as well as in women.
Although mortality data are available in almost all countries, they are not the best indicator to measure the trends in incidence of pancreatic cancer. Death certificates are not particularly reliable for this cancer, and variations in mortality rates over time may partly reflect differences in the quality of death certificates rather than differences in the incidence.13 Population-based incidence statistics, although they are not available everywhere, represent the best tool to measure trends in the incidence of pancreatic cancer. Worldwide incidence data are available from cancer registries in particular through the 5 volumes of Cancer Incidence in Five Continents covering the period analyzed in this study.3-7
Differences in trends in the incidence of pancreatic cancers over the world are striking. In Europe, an increase in the incidence was reported by other cancer registries in France and in Slovenia.3-7 In contrast, in most European countries, incidence rates during the past 25 years have been relatively stable. A decrease in the incidence in men was observed only in Nordic countries,14 in England,8,15 and in Switzerland.3-7 The highest incidence rates for pancreatic cancer are reported in Japan, where incidence has been increasing slightly. Rates are also high in the United States, although without any significant variation during the past 25 years.11,16 Incidence rates have been relatively stable in most other areas. However, incidence rates were higher in North America, Oceania, and Europe than in Central and South America and most Asian countries.
Very few publications have dealt with trends in incidence according to age group. Data from the SEER program showed a decrease in incidence from 1973 to 2001 both in patients younger than 60 years and in older patients in 7 of the 9 studied areas.17 In the same way, the incidence decreased for all age groups after the 1980s in Sweden10 and in England and Wales between 1975 and 2000.15 In contrast, incidence rates increased in all age groups in our study. It can be concluded that different trends by age group are reported in the available results. No data on trends in incidence by subsite are available. In our study, the increase in incidence by period of diagnosis was more marked for cancers of the corpus and the tail than for cancers of the head. These differences in time trends suggest that the etiology of the disease can at last be partly different according to the site of the cancer. These data should be taken into account in etiological studies. Not surprisingly, most pancreatic cancers with histological proof were adenocarcinomas. Their incidence increased, slightly faster than that of cancers with no histologic verification. Although malignant digestive endocrine tumors remained rare, the increase in their incidence was more marked than for adenocarcinomas. In this case, better knowledge of these tumors can at least partly explain the reported increase in incidence.
Much fewer data on trends in incidence by birth cohort are available. In our study, there was a more than 2-fold increase in the incidence in men born around 1950 than among those born around 1900 and a 1.6-fold increase in women. A similar trend was reported in Connecticut.11 In Denmark, the increase in incidence by birth cohort was seen only in men,14 whereas in England and in Wales, rates decreased in men and were stable in women.8 The period effect potentially reflects changes in registration practices and/or exposure to risk factors that affect all ages during the same period. Pancreatic cancer became easier to diagnose with the development of abdominal ultrasound examination and then of computed tomography. It can be hypothesized that the improvement in diagnostic technologies affected time trends in the incidence of pancreatic cancer. However, the introduction of these facilities would not have accounted for the different time trends between age groups, sex, and neighboring countries. Changes in diagnostic strategies may partly account for the observed trends. The birth cohort effect suggests exposure to early stage carcinogens with long-term latency periods. For instance, the increase in the prevalence of obesity is more likely to cause a birth cohort effect. The period effect observed both in men and in women and in all age groups can be related to the exposure to environmental risk factors affecting different ages and different generations during the same period. In particular, exposure to late-stage carcinogens can produce a period effect. They can modify the cancer risk during a relatively short period.
The reasons for the increased incidence of pancreatic cancers in France are not readily apparent. Tobacco smoking is the most well-established risk factor, with a 2- to 4-fold higher incidence in smokers, and this accounts for approximately 25% of pancreatic cancers.18-20 However, trends in smoking habits cannot explain the upward trend in incidence in both sexes. In France, the proportion of smokers by birth cohort and of tobacco-related cancers fell in men.21 A Finish study failed to show strong differences in trends in the incidence of pancreatic cancers in men according to social classes with different exposures to tobacco.22 There has been a considerable debate about whether obesity increases the risk of pancreatic cancer. A meta-analysis of 14 studies concluded that there was a weak, but consistent association between increasing body mass index and pancreatic cancer.23 The prevalence of overweight has increased in France during the past 2 decades and may at least partly explain the observed cohort effect. Obesity is associated with abnormal glucose metabolism and type 2 diabetes, both of which are implicated in the development of pancreatic cancers. Our understanding of the etiology of pancreatic cancer remains too weak. High fruit and vegetable consumption, which can be protective, cannot explain the reported trends.
In conclusion, the major finding of this study was the increase in the incidence of pancreatic cancer observed both by calendar period and by birth cohort. It may be partly attributable to the increasing prevalence of overweight. The association with other potential exposures needs to be clarified from further epidemiologic work. The fact that incidence patterns differ from one country to another remains poorly explained.
1. David M, Lepage C, Jouve JL, et al. There was little change in the management and prognosis of pancreatic cancer over a 30-year period. Br J Cancer
2. Belot A, Velten M, Grosclaude P, et al. Estimation Nationale de l'Incidence et de la Mortalité par Cancer en France Entre 1980 et 2005
. St Maurice, France: National Public Health Institute; 2008. Available at: www.invs.sante.fr
3. Muir C, Waterhouse J, Mack T, et al. Cancer Incidence in Five Continents
. Vol V (88 vol). Lyon, France: International Agency for Research on Cancer; 1987.
4. Parkin DM, Muir CS, Whelan SL, et al. Cancer Incidence in Five Continents
. Vol VI (120 vol). Lyon, France: International Agency for Research on Cancer; 1992.
5. Parkin D, Whelan S, Ferlay J, et al. Cancer Incidence in Five Continents
. In: IARC Scientific Publication Vol VII (143 vol). Lyon, France: International Agency for Research on Cancer; 1997.
6. Parkin DM, Whelan S, Ferlay J, et al. Cancer Incidence in Five Continents
. In: IARC Scientific Publication Vol VIII (143 vol). Lyon, France: International Agency for Research on Cancer; 2002.
7. Curado MP, Edwards B, Shin HR, et al. Cancer Incidence in Five Continents
. In: IARC Scientific Publication Vol IX (160 vol). Lyon, France: International Agency for Research on Cancer; 2007.
8. Fitzsimmons D, Osmond C, George S, et al. Trends in stomach and pancreatic cancer incidence and mortality in England and Wales, 1951-2000. Br J Surg
9. Imamura Y, Sobue T. Cancer statistics digest. Mortality trend of colon, rectal, liver, "gallbladder and biliary tract" and pancreas cancer in Japan by birth cohort. Jpn J Clin Oncol
10. Luo J, Adami HO, Reilly M, et al. Interpreting trends of pancreatic cancer incidence and mortality: a nation-wide study in Sweden (1960-2003). Cancer Causes Control
11. Zheng T, Holford TR, Ward BA, et al. Time trend in pancreatic cancer incidence in Connecticut, 1935-1990. Int J Cancer
12. World Health Organisation. ICD-02. International Classification of Diseases for Oncology
. Geneva, Switzerland: World Health Organisation; 1992.
13. Riboli E, Stanta G, Delendi M, et al. Comparison between diagnoses of cancers of the stomach, colon, rectum, gall-bladder, liver and pancreas on death certificates and at autopsy in Trieste, 1970-84. IARC Sci Publ
14. Nagenthiraja K, Ewertz M, Engholm G, et al. Incidence and mortality of pancreatic cancer in the Nordic countries 1971-2000. Acta Oncol
15. Wood HE, Gupta S, Kang JY, et al. Pancreatic cancer in England and Wales 1975-2000: patterns and trends in incidence, survival and mortality. Aliment Pharmacol Ther
16. Shaib YH, Davila JA, El-Serag HB. The epidemiology of pancreatic cancer in the United States: changes below the surface. Aliment Pharmacol Ther
17. Zhang J, Dhakal I, Yan H, et al. Trends in pancreatic cancer incidence in nine SEER Cancer Registries, 1973-2002. Ann Oncol
18. Silverman DT, Dunn JA, Hoover RN, et al. Cigarette smoking and pancreas cancer: a case-control study based on direct interviews. J Natl Cancer Inst
19. Zheng W, McLaughlin JK, Gridley G, et al. A cohort study of smoking, alcohol consumption, and dietary factors for pancreatic cancer (United States). Cancer Causes Control
20. Lowenfels AB, Maisonneuve P. Epidemiology and prevention of pancreatic cancer. Jpn J Clin Oncol
21. Beck F, Guilbert P, Gautier A. Health Attitudes and Behaviours, 2005 Health Barometer
. Saint-Denis, France: INPES; 2008:76-110.
22. Weiderpass E, Pukkala E. Time trends in socioeconomic differences in incidence rates of cancers of gastro-intestinal tract in Finland. BMC Gastroenterol
23. Berrington de Gonzalez A, Sweetland S, Spencer E. A meta-analysis of obesity and the risk of pancreatic cancer. Br J Cancer
This article has been cited 2 time(s).
Plos OneRegional Differences in Islet Distribution in the Human Pancreas - Preferential Beta-Cell Loss in the Head Region in Patients with Type 2 DiabetesPlos One
Plos OneQuantitative Analysis of Pancreatic Polypeptide Cell Distribution in the Human PancreasPlos One
pancreatic cancers; cancer registry; incidence; trends
© 2010 Lippincott Williams & Wilkins, Inc.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.
Readers Of this Article Also Read