In developed areas of the world, ovarian cancer is a common neoplasm, ranking 7th and 6th most frequent for incidence and mortality, respectively.1 On a worldwide basis, the total number of cases has been estimated around 192,000, thus representing over 4% of all cancers in women and the 6th leading site for incidence.1
There are, however, large variations in the incidence of ovarian cancer in different areas of the world. The highest incidence areas are in Europe (especially the Nordic countries and the United Kingdom) and North America, with a ratio of approximately 4 separating rates in the highest and the lowest incidence countries worldwide, and over 3 between the corresponding mortality figures.1, 2, 3 Ovarian cancer is therefore an important public health issue in Western countries, although more than 50% of new cases diagnosed every year worldwide occur in developing countries.1
Incidence and mortality rates in the highest-risk areas (the United States, Canada, Scandinavia and the United Kingdom) have remained approximately constant or have declined between 1980 and 1995,3 while increasing trends were observed in previously low-risk areas, such as Japan, India and Singapore, and in southern and eastern Europe.
Several factors have probably impacted on ovarian cancer trends. Available evidence indicates that age at menarche is a risk factor, but only has a modest effect on ovarian cancer risk.4, 5, 6 Lifelong number of menstrual cycles has also been associated with ovarian cancer risk, suggesting that ovulation may be implicated in the process of ovarian carcinogenesis.4, 7 Several studies showed a direct relation between late age at menopause and the risk of ovarian cancer.4, 5, 6, 7
Nulliparity and low parity have been consistently related to ovarian cancer. Most studies showed a decline in risk associated with number of full-term pregnancies beyond the first one, thus suggesting that additional risk reduction is conferred by events accompanying each pregnancy.8, 9, 10
The protection afforded by combined oral contraceptives (OCs) is the other established, and most important from a public health perspective, feature of epithelial ovarian cancer.4, 11 The overall estimated protection is approximately 40% in ever OC users and increases with duration of use.10, 11, 12, 13, 14, 15, 16, 17, 18 The favorable effect of OCs against ovarian cancer risk seems to persist for at least 15–20 years after OC use has ceased, and it is not confined to any particular type of OC formulation.17, 18, 19 The issue of fertility drugs and ovarian cancer has also attracted lively interest, but the findings of various studies remain inconsistent.20, 21, 22, 23 Hormone therapy in menopause has also been related to increased ovarian cancer risk.24, 25, 26
Potential links between ovarian cancer and diet were originally suggested on the basis of international differences or correlation studies. Positive relations were observed with fats, proteins and total calories and are generally in the same direction as those of endometrial and breast cancer.27, 28 A relation between ovarian cancer and intake of meat and fats has also been reported from some cohort and case-control studies, whereas fruit and vegetables appear to be inversely related.27, 28, 29 Some case-control studies found direct associations between measures of fat intake and risk of ovarian cancer.27, 30 Starchy foods, and consequently diets with a high glycemic index and glycemic load, have also been related to excess ovarian cancer risk.30, 31 The possibility that milk sugar lactose, or its metabolites, have some effect on oocytes with a compensatory gonadotropic stimulation and excess ovarian cancer risk has been investigated. Several, but not all, have found excess risk with lactose consumption32 and absorption,33 but the issue remains unsettled. Studies from Greece34 and Italy35 suggested that monounsaturated fats (olive oil) and fiber intake may be favorable. The role of diet on ovarian cancer incidence and mortality rates and trends across Europe remains undefined, however.36
There have long been clinical observations suggesting familial aggregations of ovarian cancer. Besides the clustering of ovarian cancer, an excess of breast cancer and a more general excess of several cancers (including colon and endometrium),37 have been described. These patterns are consistent with an autosomal dominant gene with variable penetration.38 The estimated relative risks from case-control studies that included data on family history range from 2 to 10, but are between 3 and 5 in most studies.4, 39, 40, 41
In terms of population attributable risk, a large case-control study from Italy35 estimated that 5% of ovarian cancers were attributable to nulliparity, 12% to never OC use and 4% to a family history of breast or ovarian cancer in first-degree relatives. Among women aged ≥ 50 years, later age at menopause accounted for 16% of all ovarian cancer cases. Low intake of vegetables accounted for 24% of cases and a high fat score for 7%. All these factors together explained about half of cases, thus indicating the major influence of exposure to these known factors on the prevailing geographic and temporal variations of ovarian cancer rates.
To provide a comprehensive picture of descriptive epidemiology of ovarian cancer in Europe, we have systematically considered trends in incidence and mortality in 28 European countries and in major European geographic areas from 1953 through 2000. Some of the principal factors that may explain the observed trends are discussed hereafter.
Material and methods
Incidence data coded as ovarian cancer (ICD-10 C56) from 119 cancer registries in Europe, together with corresponding registry population files, were extracted from the EUROCIM software database42 by year of diagnosis and 5-year age group. The minimum inclusion requirement for temporal analysis was set at consecutive compilation in the last 3 volumes (6–8) of Cancer Incidence in Five Continents (CI5).43, 44, 45 This criterion was chosen as a general marker of each registry's data quality over time, given that the editorial process includes a comprehensive assessment of cancer registries' quality control procedures in relation to the comparability, completeness and validity of their incidence data. Table I provides details of the cancer registry data sets included in the analysis. In 11 countries, the registries served at the national level, while in the remaining 7 countries, a number of regional registries (see footnotes in Table I) were aggregated to obtain a proxy of the (unknown) national incidence. The varying start-up and final years available for each registry within one country led to a pragmatic selection of registries and years that ensured the same populations were used throughout the period of study. The time span of the final incidence data set varied from 11 to 47 years.
Table I. Availability of Ovarian Cancer Incidence Data for Temporal Analysis
Mortality data from ovarian cancer (ICD-9 183) were extracted from the WHO mortality database for each European country, year of death and 5-year age group, together with corresponding national population data from the same source. As with incidence, the particular limitations and potential difficulties in interpreting mortality data are well documented, although for the latter there are few indicators of data quality available in European countries. A sufficient condition for inclusion for a given country was therefore the availability of recent data spanning more than 10 years. A total of 27 countries met this criteria, with the period of availability ranging from 15 to 42 years (Table II). For Germany, a combination of data from the Federal Republic of Germany and German Democratic Republic was used for the period 1983–89; thereafter, the whole country.
Table II. Availability of Ovarian Cancer Mortality Data for Temporal Analysis
Estimated national population data comprising of females (all ages) based on most recent year available.
Number of ovarian cancer deaths for most recent year available.
Age-truncated standardized rates were calculated in 3 age strata (20–49, 50–74, 75+) for each year based on the 5-year age-specific incidence and mortality rates in each country using the world standard population.46, 47 The observed trends are smoothed and presented using moving averages of rates in 3 consecutive years, centered on the year of observation. This reduces the effects of random variation inherent in the data while still retaining the temporal characteristics of the underlying trend. In the case of 2 consecutive missing years, the average was based on adjacent years.
To enable comparisons of the relative changes in the recent trends of ovarian cancer incidence and mortality, log-linear regression models were fitted for each national population and age strata. The estimated annual percentage change (EAPC) for the period 1988–1997 was obtained from the formula 100 × [exp(β) − 1], where β is the parameter estimate obtained on fitting period of event as a continuous variable to the logarithm of the rate. A 95% confidence interval (95% CI) for each EAPC was also calculated. In addition, absolute changes between 2 5-year periods (1980–1984 and 1993–1997 for incidence, later period 1994–1998 for mortality) were also calculated.
The results are presented in graphical and tabular forms by country stratified according to their status as either 1 of the 15 member states of the European Union (EU) or otherwise (applicant state or other European country), as well as by United Nations-defined European region. The rates are plotted on a log-transformed 2-cycle ordinate (e.g., with a y-scale of 1–100) and on an abscissa covering a 40-year span (1960–2000), with a Y:X ratio scaled to be approximately 2:1. Presented with these properties, a 1% change in the rate per annum signifies a 10 degree change in the slope, a rule proposed by Devesa et al.48 to aid visual inspection and systematic comparison. All analyses were performed in Stata 8 (StataCorp, College Station, TX).
Figure 1 reports the trends in overall incidence (where available) and mortality rates of ovarian cancer in the 15 EU countries. Countries with highest rates in the 1960s and 1970s were in northern and western Europe (Nordic countries, Austria, Germany and the United Kingdom), but trends in these areas have tended to decline over more recent calendar periods, mainly for mortality. Incidence and mortality rates were approximately parallel in Sweden and France, while in other countries (i.e., Finland, The Netherlands, and the United Kingdom) recent trends in incidence have been less favorable than for mortality. In France, Italy and notably other southern European countries such as Portugal, Spain and Greece, mortality trends started low, but have since displayed upward trends with time, at least up to the early 1990s, for France and Italy. Figure 2 gives comparable data for accession and other European countries. In most central and eastern countries, rates were originally relatively low and tended to rise with time. Declines in mortality were seen in Hungary and the Czech Republic in recent years, though corresponding increases in incidence were observed for the latter country.
Incidence is further considered in Figure 3 for younger (aged 25–49) and older (aged 50–74) women by European area. The overall picture in Europe is of slowly increasing trends, particularly in older women. In several northern European countries (i.e., Denmark, Norway and the United Kingdom), incidence trends were more favorable in younger women in the last decades, while in Sweden, trends in both age groups have fallen. A differential pattern by age is not clear in most areas of Europe.
Figure 4 gives corresponding values for age-standardized mortality at ages 25–49 and 50–74 in 4 European areas. The declining trends are more evident than for incidence; in most of northern and western Europe, mortality trends were considerably more favorable in young adults. Within southern Europe, a fall in mortality was observed in Italy, Portugal and Slovenia over the last 2 decades among young women only. In eastern Europe, only Hungary and the Czech Republic showed downward trends in ovarian cancer mortality over the last 2 decades, again mainly the younger-age group.
The relative changes in Tables III and IV confirm the observed trends in the figures. Table III reports age-standardized incidence rates for all ages, as well as truncated for the age groups 25–49, 50–74 and 75 and over for 1980–1984 and 1993–1997 for 10 EU and 8 other European countries, together with the corresponding absolute and relative changes in the rates together with 95% confidence intervals for the latter. Corresponding estimates for mortality for all 15 EU countries and 11 other European countries are provided in Tables IV and V, respectively.
Table III. Age-Standardized Incidence Rates for Ovarian Cancer in Selected European Countries and Age Groups and Corresponding Change in Rates, 1980–1997
Relative change (1988–1997)
Relative change (1988–1997)
Poland: relative change based on years 1988–1996; aggregated rate (1994–1998) based on years 1994–1996.
For the age group 25–49, a number of northern countries experience some declines in incidence, although Finland and the Netherlands exhibit increasing rates of 2–3% per annum on average (Table III). Elsewhere there is notable variability in the direction of trends. Relatively stable trends are observed in the age group 50–74, although decreases seen in France and Sweden of over 1% per year are offset by increases in Italy, Slovenia and the Czech Republic of around 2% per year on average. Recent trends for the older ages are upward or stable.
As to mortality, the EAPCs in these populations confirm a widespread downward trend for the age groups 25–49 and 50–74 that is more marked in the younger-age group (Table IV). This observation excludes Bulgaria, where there are rapid increases in the 1990s, and in Greece and Spain, where there are sizable longer-term increases in ovarian mortality rates apparent among women regardless of age. In addition, in Ireland, the United Kingdom, Poland, Romania and Portugal, increases in mortality are observed in women aged 75 and over.
There are 2 major messages revealed by this comprehensive analysis of trends in ovarian cancer incidence and mortality across Europe. First, incidence rates (in selected countries) and, notably, mortality rates are declining in most northern European countries, mainly in younger-age groups. Second, ovarian cancer incidence and mortality rates are still increasing in a few southern and eastern European countries. This has led to a further leveling of ovarian cancer rates across Europe,49 although a 2.5-fold difference was still observed in the late 1990s between the highest overall mortality rate of 9.3/100,000 in Denmark, and the lowest of 3.6 in Portugal. In the late 1980s, the difference was over 3-fold, between 9.9/100,000 in Denmark and 2.9 in Spain.49
It is unlikely that problems of diagnosis and certification have played a major role in geographic differences across Europe in the young, but some improvement in diagnosis and certification has probably taken place over the last few decades, since the introduction of echography, CT scan and endoscopy, thus leading to spurious trends, mainly in the elderly. Furthermore, geographic differences and trends over time may have been influenced by changing patterns of ovariectomy together with hysterectomy across different countries and over time.
These concerns notwithstanding, the favorable ovarian cancer trends in the young in several northern and western European countries are likely to be influenced by the spread of OC use across subsequent generation of women, which has taken place earlier and to a larger extent in northern Europe.17, 18, 19, 20, 49, 50 The declining mortality trends are also in part attributable to improved treatment of germ cell neoplasms,51 which, as for testicular cancer,52, 53 is likely to have taken place earlier in western than in central and eastern Europe.
The upward trends in earlier generations observed mainly in several southern and central European countries can at least be partly attributed to reduced parity in these populations over the second half of the 20th century,54 since parity is inversely related to ovarian cancer risk.2, 7, 8
Besides reproductive and hormonal factors, it is likely that changing lifestyle habits, including nutrition, diet and physical exercise,2, 5, 27, 28, 29, 30, 31, 32, 33, 34, 35, 55, 56 have played some role in trends in ovarian cancer risk and hence on national ovarian cancer incidence and mortality rates over time. Since several of these lifestyle factors have tended to become more homogeneous across European populations, these may account for the narrowing of ovarian cancer incidence and mortality rates across Europe.
There are therefore several reasons to believe that the diverse patterns of ovarian cancer incidence and mortality trends registered in Europe over the last 4 decades are largely real and can be explained by a combination of changing risk factors for incidence and, additionally for mortality, improving treatment.
The Comprehensive Cancer Monitoring Programme in Europe (CaMon) project is undertaken in collaboration with the following European cancer registries: in Czech Republic, Czech National Cancer Registry, Prague (Dr. Marie Jechová); in Denmark, Danish Cancer Registry, Copenhagen (Dr. Hans H. Storm); in Estonia, Estonian Cancer Registry, Tallinn (Dr. Tiiu Aareleid); in Finland, Finnish Cancer Registry, Helsinki (Dr. Timo Hakulinen); in France, Registre Bas Rhinois des Cancers, Strasbourg (Dr. Michel Velten), Registre Général des Tumeurs du Calvados, Caen (Dr. J. Macé-Lesech), Registre des Tumeurs du Doubs, Besançon (Dr. Arlette Danzon), Registre du Cancer de l'Isère, Meylan (Dr. François Ménégoz), Registre du Cancer de la Somme, Amiens (Ms. Nicole Raverdy), Registre des Cancers du Tarn, Albi (Dr. Martine Sauvage); in Germany, Saarland Cancer Registry Saarbrücken (Mr. Hartwig Ziegler); in Iceland, Icelandic Cancer Registry, Reykjavik (Dr. Laufey Tryggvadottir); in Italy, Registro Tumori Toscano, Florence (Dr. Eugenio Paci), Registro Tumori Lombardia (Provincia di Varese), Milan (Dr. Paolo Crosignani), Registro Tumori della Provincia di Parma (Dr. Vincenzo De Lisi), Registro Tumori della Provincia di Ragusa, Ragusa (Dr. Rosario Tumino), Piedmont Cancer Registry, Turin (Dr. Roberto Zanetti); in the Netherlands, Eindhoven Cancer Registry, Eindhoven (Dr. Jan Willem Coebergh), Maastricht Cancer Registry, Maastricht (Dr. Miranda Dirx); in Norway, Cancer Registry of Norway, Oslo (Dr. Frøydis Langmark); in Poland, Cracow Cancer Registry, Cracow (Dr. Jadwiga Rachtan), Lower Silesian Cancer Registry, Wroclaw (Mr. Jerzy Blaszczyk), Warsaw Cancer Registry, Warsaw (Dr. Maria Zwierko); in Slovakia, National Cancer Registry of Slovak Republic, Bratislava (Dr. Ivan Plesko); in Slovenia, Cancer Registry of Slovenia, Ljubljana (Dr. Maja Primic-Zakelj); in Spain, Tarragona Cancer Registry, Reus (Dr. Jaume Galceran), Registro de Cáncer de Granada, Granada (Dr. Carmen Martínez Garcia), Registro de Cáncer de Murcia, Murcia (Dr. Carmen Navarro Sánchez), Registro de Cáncer de Navarra, Pamplona (Dr. E. Ardanaz Aicua), Zaragoza Cancer Registry, Zaragoza (Dr. Carmen Martos Jimenez); in Sweden, Swedish Cancer Registry, Stockholm (Dr. Lotti Barlow); in Switzerland, Krebsregister Basel-Stadt und Basel-Land, Basle (Dr. Gernot Jundt), Registre Genevois des Tumeurs, Geneva (Dr. Christine Bouchardy), Registre Neuchâtelois des Tumeurs, Neuchâtel (Dr. Fabio Levi), Krebsregister St. Gallen Appenzell, St. Gallen (Dr. Thomas Fisch), Registre Vaudois des Tumeurs, Lausanne (Dr. Fabio Levi), Kantonalzürcherisches Krebsregister, Zürich (Dr. Nicole Probst); and in United Kingdom, National Cancer Intelligence Centre, London (Dr. Mike Quinn), Scottish Cancer Intelligence Unit, Edinburgh (Dr. David Brewster). Supported by the Italian Association for Cancer Research and the Italian League against Cancer (to C.L.V.). The CaMon project is funded by the European Commission (Sl2.327599, 2001CVG3-512).