Global patterns and trends in ovarian cancer incidence: age, period and birth cohort analysis
BMC Cancer volume 19, Article number: 984 (2019)
Ovarian cancer (OC) is the seventh most common malignancy worldwide and the most lethal gynaecological malignancy. We aimed to explore global geographical patterns and temporal trends from 1973 to 2015 for 41 countries in OC incidence and especially to analyse the birth cohort effect to gain further insight into the underlying causal factors of OC and identify countries with increasing risk of OC.
OC data were drawn from the Cancer Incidence in Five Continents databases and online databases published by governments. The joinpoint regression model was applied to detect changes in OC trends. The age–period–cohort model was applied to explore age and birth cohort effects.
The age-standardized rate of OC incidence ranged from 3.0 to 11.4 per 100,000 women worldwide in 2012. The highest age-standardized rate was observed in Central and Eastern Europe, with 11.4 per 100,000 women in 2012. For the most recent 10-year period, the increasing trends were mainly observed in Central and South America, Asia and Central and Eastern Europe. The largest significant increase was observed in Brazil, with an average annual percentage change of 4.4%. For recent birth cohorts, cohort-specific increases in risk were pronounced in Estonia, Finland, Iceland, Lithuania, the United Kingdom, Germany, the Netherlands, Italy, Malta, Slovenia, Bulgaria, Russia, Australia, New Zealand, Brazil, Costa Rica, Ecuador, India, Japan, the Philippines and Thailand.
Disparities in the incidence and risk of OC persist worldwide. The increased risk of birth cohort in OC incidence was observed for most countries in Asia, Central and Eastern Europe, and Central and South America. The reason for the increasing OC risk for recent birth cohorts in these countries should be investigated with further epidemiology studies.
Ovarian cancer (OC) is the seventh most common malignancy worldwide, with 238,719 newly diagnosed cases in 2012 . The incidence of OC has appreciable geographic variation worldwide . OC is more frequently diagnosed at an advanced stage, and its prognosis is poor, which makes this cancer the most lethal gynaecological malignancy . Thus, understanding the aetiology of OC and identifying the causal factors and populations at high risk are essential for primary prevention.
To better understand the effect of lifestyle factors or reproductive patterns on OC incidence, we conducted an age–period–cohort analysis to explore the effect of birth cohort [3,4,5,6]. Birth cohort effects can reflect the long-established generational effect of causal factors in cancer incidence [3,4,5,6]. For example, a recent age–period–cohort analysis in Japan, the Republic of Korea and Singapore indicated that the increased risk of OC in younger birth cohorts was caused by changes in reproductive patterns and a shift towards a westernized lifestyle and dietary factors . To date, no study has examined global trends in OC by an age–period–cohort analysis, which has been found to be more useful than a conventional cross-sectional analysis in evaluating trends .
We aimed to explore global geographical patterns of OC incidence and temporal trends from 1973 to 2015 for 41 countries. In particular, we aimed to analyse the birth cohort effect to examine the importance of changes in lifestyle or reproductive patterns, identify countries with increasing risk of OC and highlight trends that deserve closer attention by public health and cancer prevention specialists.
OC incidence data were categorized according to the International Classification of Disease for Oncology (ICD-O), 3rd edition (C56). The data for the age-standardized rate (ASR) of OC incidence in 2012 for regions and 184 countries worldwide were drawn from the GLOBOCAN 2012 database , which formulated national estimates of cancer incidence from the best available data source (often based on data from Cancer Incidence in Five Continents (CI5)) and weighted averages of regional data in each country, with variable levels of accuracy depending on the extent and validity of locally available data . We also extracted long-term data from the CI5, Volumes IV–X, the CI5plus database, and online databases published by governments. CI5 is the main source of high-quality global cancer incidence data for validity, completeness, and comparability . Incidence data for the United States (USA) were extracted from the Surveillance, Epidemiology and End Results Program (SEER), which represents the most reliable data source of cancer incidence in the USA . Although these databases, including GLOBOCAN, CI5 and SEER, have been used extensively in studies examining the global patterns and trends in the incidence of various cancers [12,13,14], there are some possible heterogeneities of these databases and different registries in our study. For example, GLOBOCAN includes simulated figures different from the active data collection of CI5 and SEER. The discrepancy between simulated figures of GLOBOCAN and the accurate data of CI5 and SEER is likely to be greater in countries with lower quality cancer registry data. The accuracy of data in these databases also varies from region to region and in different registries, especially for developing countries. Caution should be taken when interpreting the findings of developing countries.
To examine the trends in OC incidence, the inclusion requirement in our study was a continuous data of at least 15 years and containment in the last volume of the CI5 series (Volume X) to avoid statistical instability and ensure the quality of the data [6, 15]. Finally, 41 countries were selected. Of these countries, the incidence data for 26 countries were at the national level. For the remaining countries with two or more cancer registries, we pooled the cases and population data in all registries to cover the largest geographic area with an estimated national level .
We calculated a summary ASR using direct standardization with the world standard population . To examine the geographic diversity in OC incidence, ASR by region and 184 countries in 2012 were plotted. To graphically present the trend in OC ASR, we performed locally weighted scatterplot smoothing (LOWESS) regression to fit smoothed lines . To examine the changes in ASR, we performed a joinpoint regression model to calculate the annual percent change and the average annual percent change (AAPC) .
We conducted age-period-cohort analyses in all 41 countries. We subtracted the midpoints of 5-year age groups (20–24, 25–29, …, 80–84) from the corresponding 1-year calendar periods of diagnosis to obtain birth cohorts. Finally, we described the magnitude of the rates λ(a, p) as a function of age (a), period (p) and birth cohort (c) using a log-linear model, with Poisson distribution and with the log of the person-years at risk defined as an offset [3, 5, 6]:
We applied a full age–period–cohort model to estimate birth cohort effects with incidence rate ratio (IRR) relative to the reference birth cohort. To overcome the non-identifiability problem of the linear dependence between three factors, we constrained the linear component of the period effect to have a zero slope, assuming that the linear changes in OC incidence resulted from cohort-related factors [3, 5, 6]. This statistical method has been widely applied in many published papers about global trends in the incidence of other cancers [3, 5, 6].
The global map was depicted by using ArcGIS (version 10.2). The figures were drawn by using Sigma Plot (version 12.5). Joinpoint regression models were performed by the Joinpoint Regression Program (version 184.108.40.206). The age–period–cohort model analyses and graphs were conducted using APCfit in Stata (version 13.0).
There were an estimated 238,719 incident cases of OC and an ASR of 6.1 per 100,000 women worldwide in 2012 (Table 1). Approximately 9.2 per 100,000 women of ASR occurred in more developed regions and 5 per 100,000 women in less developed regions. The highest ASR was observed in Central and Eastern Europe, with 11.4 per 100,000 women, while the lowest ASR was observed in Micronesia, with 3.0 per 100,000 women (Fig. 1 and Fig. 2).
Table 2 shows temporal trends of ASR from 1973 onward as well as the AAPC values for the last 10-year period. The scatter plots with LOWESS regression curves are shown in Fig. 3, Additional file 1: Figure S1 and S2 (nine exemplary countries are shown in the text and the remaining countries are shown in supplemental figures to make the figure clearer). Over the entire study period, the ASR increased continually in Brazil, Colombia, Ecuador, Costa Rica, Germany, Estonia, Latvia, Lithuania, the United Kingdom (UK), Spain, Bulgaria, Poland, Slovakia, India, Japan and Thailand. For the most recent 10-year period, increasing trends were mainly observed in Central and South America, Asia and Central and Eastern Europe (Additional file 1: Figure S3). Larger significant increases were observed in Brazil (AAPC = 4.4%), the Republic of Korea (AAPC = 2.1%) and Japan (AAPC = 1.7%), whereas larger decreases were found in Austria (AAPC = − 3.7%), Israel (AAPC = − 3.2%) and the Czech Republic (AAPC = − 2.8%) (Additional file 1: Figure S3).
Figure 4, Additional file 1: Figure S4 and S5, shows the 5-year age-specific OC incidence rates by birth cohort. The non-parallel appearance of the observed incidence rates versus the birth cohort across age groups indicates a strong cohort effect, as seen in almost all countries. For countries with age-specific differences, the incidence rate increased in the recent birth cohorts with age > 70 years in Denmark and Germany, age > 65 years in Finland, age > 50 years in Poland and Thailand, age > 40 years in Ecuador, age > 35 years in Latvia and age > 30 years in Korea. The other age groups of these countries show a decreasing trend. The incidence rate increased in the recent birth cohorts with age < 30 years in the Netherlands, Ireland and France, age < 35 years in Norway, age < 45 years in Russia, New Zealand and Singapore, age < 50 years in the UK, Slovenia and Estonia, and age < 70 years in Japan. The incidence rate decreased in the Netherlands between 30 and 65 years old, in France between 30 and 70 years old, and in the UK between 50 and 60 years old.
Figure 5, Additional file 1: Figure S6 and S7, depicts the graphs for the age and cohort effects. The incidence rates increased sharply with age in most countries. We observed four patterns of IRR for birth cohort effects. First, a continuous increase was found in Bulgaria, Estonia, Germany, Italy, Lithuania, Malta, Russia, Slovakia, Slovenia, the UK, Australia, Brazil, Costa Rica, Ecuador, India, Japan, Singapore and Thailand. Second, a trend in IRR analogous to a v-shaped curve was observed in Iceland, the Netherlands, New Zealand, Finland and the Philippines. The IRR in the Philippines decreased among birth cohorts from 1900 to 1920 and increased rapidly from 1920 onwards, while that in Iceland decreased slightly before the 1960–1970 birth cohort and increased from 1970 onwards, and that in the Netherlands, Finland and New Zealand increased after the 1970–1980 birth cohort. Third, a trend in IRR similar to an inverted v-shaped curve was observed in several countries. For example, the IRR of cohort effects in the USA and the USA White population increased among birth cohorts from 1890 to 1920 and decreased from 1920 onwards; that in France increased among birth cohorts from 1890 to 1930 and decreased from 1930 onwards; that in Norway and Switzerland increased among birth cohorts from 1890 to 1940 and decreased from 1940 onwards; that in Czech Republic and Ireland decreased from 1950 onwards; that in Croatia, Latvia, Poland, Canada, Israel and the USA Black populations decreased from 1960 onwards; and that in China, Spain, Colombia and the Republic of Korea decreased from 1970 onwards. Fourth, a consistent decrease in IRR was detected in Austria, Denmark and Sweden. For recent generations of the 41 countries, the IRR increased in Estonia, Finland, Iceland, Lithuania, the UK, Germany, the Netherlands, Italy, Malta, Slovenia, Bulgaria, Russia, Australia, New Zealand, Brazil, Costa Rica, Ecuador, India, Japan, the Philippines and Thailand.
In this study, we observed a large global variation in the ASR of OC. During the most recent 10-year period, increasing trends were mainly observed in Central and South America, Asia and Central and Eastern Europe. We also found visibly increasing IRR among recent birth cohorts in most countries worldwide.
The birth cohort effect indicates that individuals born in the same time period tend to adopt similar lifestyles that may influence their carcinogenic risks in specific ways. Changing lifestyle habits, including cigarette smoking, diet and oral contraceptive pills (OCPs), and obesity could have crucial impacts on the birth cohort risk of OC and hence influence national OC incidence trends.
There is a biological plausibility between smoking and OC; that is, researchers have discovered adducts of benzo(a) pyrene in the ovarian follicular cells of women who have ever smoked, and these adducts would increase DNA damage risk through a direct carcinogenic effect . Women who have ever smoked had a 6% higher risk of OC than did those who have never smoked . The global smoking prevalence among females (≥15 years old) declined from 10.6% in 1980 to 6.2% in 2012, with a decline of 1.7% every year . From 1980 to 2012, smoking prevalence among females mainly declined in the Americas and Oceania countries, especially in North America, Australia, and New Zealand, while prevalence started to decline moderately during recent years in European countries and has changed little in Asian countries . For example, the smoking prevalence in Canada, the USA, Denmark, Iceland, Norway, Sweden, the United Kingdom and Israel was higher than 20% in 1980 and declined by greater than 2% annually, whereas the prevalence in Bulgaria, which was similar to that of the above-mentioned countries in 1980, increased with statistical significance since 1980 . The trends in smoking prevalence could partly explain OC trends in these regions and countries for the study period, especially for the most recent 10 years. In 2012, the greatest smoking prevalence among females occurred in European regions, followed by Oceania and North America, while smoking prevalence in Asian and African countries was low; that is, smoking prevalence across high-income countries in 2012 varied greatly, from less than 15% in Canada, Iceland, Israel, Japan, Sweden, and the USA to greater than 26% in Bulgaria and France, while in many middle-income countries, the smoking prevalence never exceeded 5% . The global smoking prevalence pattern in 2012 also seems to explain the global patterns in OC incidence in 2012 from our study that the highest ASR for OC was observed in Central and Eastern Europe, Northern Europe, Southern Europe, North America, and Oceania. Generation-specific smoking prevalence data can also help explain the trends of OC better. In Japan, smoking prevalence continuously increased in the 1930s–1970s birth cohort among women . In Germany, the smoking prevalence increased from 20% in the 1926–1930 birth cohort to approximately 50% in the 1966–1970 birth cohort among women . This finding seems to be consistent with the increasing birth cohort risk of OC in these countries in our study.
Mechanistically, red meat and processed meat are sources of iron, high salt content, saturated fats, and several mutagens, including N-nitroso and nitrosamine compounds, heterocyclic amines and polycyclic aromatic hydrocarbons, which are associated with DNA damage and an increased risk of OC . A healthy dietary pattern was associated with a reduced 14% risk of OC, and a western-style dietary pattern, such as high intakes of red meat and processed meat, was associated with an 19% increased risk of OC . In 2010, the average red meat intake met the recommended standards of ≤1 (100 g) serving/week in only 5 of 187 countries (representing 20.3% of the world’s population). Global average red meat consumption increased by 1.5 g/day from 1990 to 2010, while the consumption significantly increased by 8.3 g/day in East Asia . The greatest increases in intake occurred in Latvia (+ 15.2 g/day) and the Republic of Korea (+ 13.4 g/day), while a large decline was observed in Canada (− 7.1 g/ day), the Netherlands (− 6.1 g/day) and the USA (− 4.7 g/day) . The average processed meat intake met the recommended standards of ≤1 (50 g) serving/week in 55 of 187 countries (representing 38.5% of the world’s population) in 2010 . Some countries consume more red meat and processed meat (e.g., American nations, such as Colombia; and European nations, such as Poland, Russia, Latvia and Lithuania) . The change in red meat intake and the dietary pattern worldwide also seems to partly explain the pattern and trends in OC incidence in the regions and countries examined in our study. Notably, the birth cohort effect implies the importance of diet in early life. Diet may explain part of the increased OC risk of recent birth cohorts in Japan, since Japanese dietary patterns have shifted greatly to Western-style meals over recent decades . The increasing risk for the cohort born after the 1920s in Korea from our study might be partly explained by the gradual westernization of lifestyles and dietary pattern towards increased meat and fat consumption in Korea .
Compared with normal weight women, obese women would reduce serum progesterone levels because of an increase in anovulatory cycles, while progesterone has a protective impact on ovarian carcinogenesis . In addition, obesity increased insulin and insulin–like growth factor–1 levels, which would increase OC risk . Hence, overweight women had a 7% higher risk of OC, and obese women had a 28% higher risk . A population-based study indicated that the estimated population attributable fraction of OC cases in 2012 associated with excess body mass index (defined as 25 kg/m (2) or greater) is 33% for Eastern Europe, 30% for Northern Europe, 30% for Southern Europe and 34% for North America . These findings may partly explain the highest ASR of OC in Europe and North America in 2012. The proportion of adults with body mass index ≥25 kg/m2 increased from 29.8 to 38.0% for females from 1980 to 2013 worldwide . The proportion of women with a high body mass index increased even faster than the global average in the high-income countries included in our study (except Japan) . The increasing prevalence of obesity due to lifestyle changes is likely to cause a birth cohort effect. In Australia, a quasi–V-shaped obesity trend was observed for females for birth cohorts from 1915 to 1980 , which seems to be consistent with the changes in the birth cohort risk of OC in our study. The prevalence of obesity among women increased from 5.0 to 10.1% from 1993 to 2009 in China  and from 8.0 to 16.5% from 1987 to 2012 in Spain . Such an increase in obesity may explain in part the increasing birth cohort risk of OC in these countries.
OCPs are a well-established protective factor for OC. The biological mechanisms underlying this association include that OCPs could suppress ovulation, lower follicle-stimulating hormone, eliminate the midcycle surge of luteinizing hormone and reduce stromal cell reactivity, all of which would reduce the risk of OC . The risk of OC decreased by 20% for each 5 years of OCP use . Furthermore, the reduction in OC risk persisted for more than 30 years after OCP use had ceased . In Northern and Western Europe and North America, where the use of OCPs was earlier and more widespread, the favourable trends in OC can partly be attributed to its long-term protection. OCPs were introduced in Europe in the early 1960s . The estimated proportion of women aged 15–45 taking OCPs was 20 to 30% in the UK, Demark and Sweden in the mid-1970s and approximately 30 to 40% of women aged 15–45 in Northern Europe by the late 1980s and early 1990s [40,41,42]. In France, the proportion of women aged 20–44 who regularly take OCPs increased from 28.3% in 1978 to 45.4% in 2000 . By 2010, 79% of women aged 15–29 were taking OCPs in France . As a consequence, the differences in the introduction time of OCPs and the prevalence of OC may partly explain why birth cohorts after the 1920s–1940s in most countries of Northern and Western Europe and North America showed a decreasing trend in birth cohort risk. Our study indicated that the IRR of OC increased until the cohort born approximately 1918 in the USA and 1923 in Australia, and these individuals were the first generation to use OCPs . In France, the change in birth cohort risk in our study can be explained by the reduction in the cumulative risk in the cohorts born from 1930 onwards, corresponding to the advent of OCPs among the female population . Furthermore, in Bulgaria of Central and Eastern Europe, the continuous and rapidly increasing OC birth cohort trends may be explained in part by the low oral contraceptive use (6.2% in 2007) . In Asia, oral contraceptive use was also very low; that is, the prevalence of OCP use in India, China, and the Republic of Korea was less than 4% before 2010 , which might explain in part the continuous increase in OC rate in most Asian countries due to the low prevalence of OCP use. In Japan, oral contraceptives were released for general use only in 1999 , and the prevalence of OCP among women aged 15–49 years was only 1.1% in 2015 , which may also explain in part the increasing trends and birth cohort risk in OC incidence in Japan. The prevalence of OCP use in Central and Eastern Europe and Asian countries was markedly lower than those in western countries, such as the USA (16.0%) and France (39.5%) . Thus, the decreasing trends and birth cohort risk of OC incidence are large in most countries of Northern and Western Europe, North America and Oceania, while the increasing trends and birth cohort risk of OC incidence occurred in Central and Eastern Europe and Asian countries.
Pregnancy could reduce the lifetime number of ovulatory cycles , lower gonadotropin secretion and subsequent oestrogen stimulation of the ovarian surface epithelium , and clear precancerous cells from the ovary , all of which would reduce the risk of OC. The level of protection increases with the number of childbirths (relative risk per child, 0.90), and compared to nulliparous women, the risk of OC among parous women decreased 30% . Reproductive factors also seem to be important in influencing OC trends worldwide. The total fertility rates per woman in Western and Northern European countries, such as France, Denmark, Iceland, Ireland and Norway, were 2.0 and have slightly increased in recent decades . The fertility rates in North American countries also slightly increased from 1.7 in 1975 to 1.9 in 2015, while the total fertility rates in Oceanian countries slightly decreased from 2.7 to 2.4 . In addition, downward trends in OC incidence from the 1970s to the 1990s in western countries could be partly explained by the increasing fertility rates after World War II, since females of the baby boom generation reached child-bearing ages . However, with the influence of family planning and western culture, the parity has substantially decreased in most Asian countries, some Southern American countries and Central and Eastern European countries since 1965 (from an average of 6 to < 3 by 2000) . Indeed, over the last 4 decades (1975–2015) in Asia, the total fertility rates per women decreased from 5.0 to 2.3 in India, from 5.5 to 2.9 in the Philippines and from 3.9 to 1.5 in Thailand . In China, the mean parity decreased from 4.9 to 1.1 for urban women and from 5.9 to 1.4 for rural women born between 1930 and 1974 . In South America, the total fertility rates per woman decreased from 4.3 to 1.7 in Brazil and from 5.1 to 2.4 in Ecuador from 1975 to 2015 . In Central and Eastern Europe, the total fertility rates per woman in Bulgaria, Poland and Russia decreased from 2.5 in 1975 to 1.3 in 2015 . The parity of fertile women between 15 and 49 years old in these regions and countries decreased by almost half during the half century, which may partly explain why the successive birth cohort effect from 1900 to 1935 onwards in these countries shows an upward trend, especially for Asian and Southern American countries with larger decreases in fertility rates.
Oestrogen stimulation of ovarian tissue via various pathways after hormone replacement therapy (HRT) use may support the positive biological association between HRT and OC . Compared with never-use HRT women, HRT users had a 37% higher risk of OC . In the USA and the UK, the prevalence of HRT substantially increased before 2002 and then halved abruptly in 2002 following Women’s Health Initiative reports of increased risks of OC or other diseases [56, 57]. The patterns of HRT in Western and Northern Europe and Australia are similar to those observed in the USA and the UK . Our results were consistent with these findings, suggesting an accelerated decline in OC incidence rates for middle- and older-aged women in these areas and countries after 2002. We speculated that the decreasing trends in OC may be due to the large reduction in the use of HRT after 2002.
Notably, the incidence trends of OC may also be partially influenced by changes in diagnostic facilities and disease classifications, particularly in high-income countries. Some improvement in the diagnosis and certification of OC in western countries has taken place from the 1980s to the 1990s since the echography, CT scan and endoscopy were introduced , thus leading to partial increasing incidence trends in OC in those countries during that period. In addition, in Central and Eastern European countries, the delayed introduction of advanced diagnosis and management method of OC  may have caused the recent increasing incidence trends of OC in these countries from our study. In India, new diagnostic equipment was introduced, and diagnostic methods were concentrated in some large-scale hospitals from the mid-1980s to mid-1990s , so these changes may have instant effects on OC rates and led to the observed increasing incidence trends of OC during that period in our study. The declining trends in OC incidence after 2000 may be partly due to changes in the disease classification. Compared to ICD-O 2nd edition published by the World Health Organization in 1990, ICD-O 3rd edition, published in 2000 no longer considers ovarian neoplasms with borderline malignancy or low malignant potential as ovarian malignant tumours, while borderline tumours accounted for approximately 15% of all OC cases . For example, a recent study indicated that the slight decreasing trends in OC incidence in Denmark after 2000 may be attributed to diagnostic changes in pathological standards for ovarian carcinomas versus ovarian borderline tumours and the different assessments among pathologists . In addition, diversities in cancer screening practices could influence comparisons of OC rates between countries in our study. However, our study could not evaluate how these factors exactly influence the trends in OC incidence, which is one of the major limitations in our study and should be explored further with studies in the future.
Ecological fallacy is another major limitation in our study. The included databases in our study, such as the GLOBOCAN 2012, CI5 and SEER databases, were derived from different cancer registries at different ecological levels, and these regional registry data were aggregated to obtain an estimate of national incidence. In our study, ecological fallacy occurs when these regional registry data are aggregated, and conclusions drawn on the basis of a group-level analyses differ from those that would have been drawn based on individual-level data. Thus, ecological fallacy must be considered as a possible interpretation of the findings in our study, and these findings must be interpreted with caution when conclusions drawn on country-level were inferred towards an individual level from our study, especially for the above-mentioned factors.
There also exists other limitations in our study. First, in developing countries, regular medical records could be incomplete, and population estimates could be imprecise, which could affect the accuracy of OC incidence rates. In developing countries, cancer registries are usually established in metropolises, and the prevalence of unhealthy lifestyle behaviours is higher than that in the general population. Second, there may be an underestimate of the regional cancer registries with limited resources. However, there may be an overestimate of the high-quality regional cancer registries with long-term OC prevention strategies. Third, random variation in statistical analysis could occur in low-risk populations on account of small numbers. Finally, our study did not include African countries because of a lack of historical data in CI5. As a result, the findings of our study should be interpreted with caution.
Despite the above-mentioned limitations, this study systematically assessed geographical variations and long-term trends in OC incidence worldwide, focusing on the differences between countries and birth cohort effects. The brilliant highlights are that we detected the geographical disparities in regions and 41 countries worldwide and explored birth cohort trends, which is a more systematic analysis than that in previous studies. In addition, we adopted the CI5 database with rigorous data-quality standards.
The major finding of our study is the increasing IRR of OC observed by birth cohort for most countries in Asia, Central and Eastern Europe, and Central and South America. The increasing trends may be due to the increased prevalence of smoking, the westernized dietary patterns, obesity, and the decreased prevalence of parity, while the decreasing trends may be due to the increased prevalence of OCPs and the decreased prevalence of HRT. Public health and cancer prevention specialists should pay more attention to countries with increasing risk of OC. The reason for the increasing OC risk for recent birth cohorts in these countries should be investigated with further epidemiology studies.
Availability of data and materials
National OC incidence estimates in 2012 for 184 countries were extracted from the GLOBOCAN database (http://globocan.iarc.fr). For the 41 countries studied, national incidence data were available for 26 countries, and registry incidence data were available for 15 countries.
National incidence data sources: Austria (http://eco.iarc.fr/EUREG/AnalysisT.aspx), Croatia (https://www.hzjz.hr/sluzba-epidemiologija-prevencija-nezaraznih-bolesti/publikacije-odjel-za-maligne-bolesti/), Czech Republic (http://www.svod.cz/?sec=aktuality&lang=en#), Denmark (http://www-dep.iarc.fr/NORDCAN/english/frame.asp), Finland (http://www-dep.iarc.fr/NORDCAN/english/frame.asp), France (http://invs.santepubliquefrance.fr/Dossiers-thematiques/Maladies-chroniques-et-traumatismes/Cancers/Surveillance-epidemiologique-des-cancers/Estimations-de-l-incidence-de-la-mortalite-et-de-la-survie/Estimation-de-l-incidence-et-de-la-mortalite-par-cancer-en-France-metropolitaine-entre-1980-et-2012-Tumeurs-solides), Iceland (http://www-dep.iarc.fr/NORDCAN/english/frame.asp), Ireland (http://www.ncri.ie/data/incidence-statistics), Malta (http://eco.iarc.fr/EUREG/AnalysisT.aspx), the Netherlands (http://www.cijfersoverkanker.nl/selecties/dataset_1/img594e1bb7eb0e9), Norway (http://www-dep.iarc.fr/NORDCAN/english/frame.asp), Russia (http://www.oncology.ru/service/statistics/malignant_tumors/), Slovakia (http://www.nczisk.sk/Publikacie/Pages/Edicia-analytickych-publikacii.aspx), Slovenia (http://www.onko-i.si/dejavnosti/epidemiologija_in_register_raka/registri_raka/), Sweden (http://www-dep.iarc.fr/NORDCAN/english/frame.asp), Australia (http://www.aihw.gov.au/acim-books/), New Zealand (http://www.health.govt.nz/nz-health-statistics/health-statistics-and-data-sets/cancer-data-and-stats), Japan (http://ganjoho.jp/en/professional/statistics/table_download.html), the Republic of Korea (http://ncc.re.kr/cancerStatsList.ncc?searchKey=total&searchValue=&pageNum=1), Bulgaria (http://ci5.iarc.fr/Default.aspx), Costa Rica (http://ci5.iarc.fr/Default.aspx), Estonia (http://ci5.iarc.fr/Default.aspx), Israel (http://ci5.iarc.fr/Default.aspx), Latvia (http://ci5.iarc.fr/Default.aspx), Singapore (http://ci5.iarc.fr/Default.aspx).
Registry incidence data sources: Canada, Brazil, Colombia, Ecuador, Germany, Switzerland, the United Kingdom, Italy, Spain, Poland, China, India, the Philippines, Thailand (data of these countries were drawn from the Cancer Incidence in Five Continents database: http://ci5.iarc.fr/Default.aspx), and the United States (data were drawn from the Surveillance, Epidemiology and End Results Program: www.seer.cancer.gov).
Average annual percent change
Cancer Incidence in Five Continents
Hormone replacement therapy
International Classification of Disease for Oncology
Incidence rate ratio
Locally weighted scatterplot smoothing
Oral contraceptive pills
Surveillance, Epidemiology and End Results
Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A. Global cancer statistics, 2012. CA Cancer J Clin. 2015;65(2):87–108.
Jayson GC, Kohn EC, Kitchener HC, Ledermann JA. Ovarian cancer. Lancet (London, England). 2014;384(9951):1376–88.
Luo G, Zhang Y, Guo P, Wang L, Huang Y, Li K. Global patterns and trends in stomach cancer incidence: age, period and birth cohort analysis. Int J Cancer. 2017;141(7):1333–44.
Wang J, Lv H, Xue Z, Wang L, Bai Z. Temporal trends of common female malignances on breast, cervical, and ovarian Cancer mortality in Japan, Republic of Korea, and Singapore: application of the age-period-cohort model. Biomed Res Int. 2018;2018:5307459.
Erdmann F, Lortet-Tieulent J, Schuz J, Zeeb H, Greinert R, Breitbart EW, Bray F. International trends in the incidence of malignant melanoma 1953-2008--are recent generations at higher or lower risk? Int J Cancer. 2013;132(2):385–400.
Petrick JL, Braunlin M, Laversanne M, Valery PC, Bray F, McGlynn KA. International trends in liver cancer incidence, overall and by histologic subtype, 1978-2007. Int J Cancer. 2016;139(7):1534–45.
Rosenberg PS, Anderson WF. Age-period-cohort models in cancer surveillance research: ready for prime time? Cancer Epidemiol Biomarkers Prev. 2011;20(7):1263–8.
Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F. GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 11. Lyon: International Agency for Research on Cancer; 2013. Available from: http://globocan.iarc.fr, Accessed 1 May 2017
Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer. 2015;136(5):E359–86.
Bray F, Ferlay J, Laversanne M, Brewster DH, Gombe Mbalawa C, Kohler B, Pineros M, Steliarova-Foucher E, Swaminathan R, Antoni S, et al. Cancer incidence in five continents: inclusion criteria, highlights from volume X and the global status of cancer registration. Int J Cancer. 2015;137(9):2060–71.
Surveillance, Epidemiology, and End Results Program. SEER*Stat database: incidence—SEER 9 regs research data, 2016 Sub (1973–2014). www.seer.cancer.gov.
Antoni S, Ferlay J, Soerjomataram I, Znaor A, Jemal A, Bray F. Bladder Cancer incidence and mortality: a global overview and recent trends. Eur Urol. 2017;71(1):96–108.
Arnold M, Sierra MS, Laversanne M, Soerjomataram I, Jemal A, Bray F. Global patterns and trends in colorectal cancer incidence and mortality. Gut. 2017;66(4):683–91.
DeSantis CE, Bray F, Ferlay J, Lortet-Tieulent J, Anderson BO, Jemal A. International variation in female breast Cancer incidence and mortality rates. Cancer Epidemiol Biomarkers Prev. 2015;24(10):1495–506.
Coburn SB, Bray F, Sherman ME, Trabert B. International patterns and trends in ovarian cancer incidence, overall and by histologic subtype. Int J Cancer. 2017;140(11):2451–60.
Segi M, Fujisaku S, Kurihara M. Geographical observation on cancer mortality by selected sites on the basis of standardised death rate. Gan. 1957;48(2):219–25.
Znaor A, Lortet-Tieulent J, Laversanne M, Jemal A, Bray F. International variations and trends in renal cell carcinoma incidence and mortality. Eur Urol. 2015;67(3):519–30.
Zenzes MT, Puy LA, Bielecki R. Immunodetection of benzo [a] pyrene adducts in ovarian cells of women exposed to cigarette smoke. Mol Hum Reprod. 1998;4(2):159–65.
Beral V, Gaitskell K, Hermon C, Moser K, Reeves G, Peto R. Ovarian cancer and smoking: individual participant meta-analysis including 28,114 women with ovarian cancer from 51 epidemiological studies. Lancet Oncol. 2012;13(9):946–56.
Ng M, Freeman MK, Fleming TD, Robinson M, Dwyer-Lindgren L, Thomson B, Wollum A, Sanman E, Wulf S, Lopez AD, et al. Smoking prevalence and cigarette consumption in 187 countries, 1980-2012. JAMA. 2014;311(2):183–92.
Marugame T, Kamo K, Sobue T, Akiba S, Mizuno S, Satoh H, Suzuki T, Tajima K, Tamakoshi A, Tsugane S. Trends in smoking by birth cohorts born between 1900 and 1977 in Japan. Prev Med. 2006;42(2):120–7.
Schulze A, Mons U. Trends in cigarette smoking initiation and cessation among birth cohorts of 1926-1970 in Germany. Eur J Cancer Prev. 2005;14(5):477–83.
Kolahdooz F, van der Pols JC, Bain CJ, Marks GC, Hughes MC, Whiteman DC, Webb PM. Meat, fish, and ovarian cancer risk: results from 2 Australian case-control studies, a systematic review, and meta-analysis. Am J Clin Nutr. 2010;91(6):1752–63.
Wang HF, Yao AL, Sun YY, Zhang AH. Empirically derived dietary patterns and ovarian cancer risk: a meta-analysis. Eur J Cancer Prev. 2018;27(5):493–501.
Micha R, Khatibzadeh S, Shi P, Andrews KG, Engell RE, Mozaffarian D. Global, regional and national consumption of major food groups in 1990 and 2010: a systematic analysis including 266 country-specific nutrition surveys worldwide. BMJ Open. 2015;5(9):e008705.
Tamakoshi K, Kondo T, Yatsuya H, Hori Y, Kikkawa F, Toyoshima H. Trends in the mortality (1950-1997) and incidence (1975-1993) of malignant ovarian neoplasm among Japanese women: analyses by age, time, and birth cohort. Gynecol Oncol. 2001;83(1):64–71.
Park B, Park S, Kim TJ, Ma SH, Kim BG, Kim YM, Kim JW, Kang S, Kim J, Kim TJ, et al. Epidemiological characteristics of ovarian cancer in Korea. J Gynecol Oncol. 2010;21(4):241–7.
Risch HA. Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the role of androgens and progesterone. J Natl Cancer Inst. 1998;90(23):1774–86.
Calle EE, Kaaks R. Overweight, obesity and cancer: epidemiological evidence and proposed mechanisms. Nat Rev Cancer. 2004;4(8):579–91.
Liu Z, Zhang TT, Zhao JJ, Qi SF, Du P, Liu DW, Tian QB. The association between overweight, obesity and ovarian cancer: a meta-analysis. Jpn J Clin Oncol. 2015;45(12):1107–15.
Arnold M, Pandeya N, Byrnes G, Renehan PAG, Stevens GA, Ezzati PM, Ferlay J, Miranda JJ, Romieu I, Dikshit R, et al. Global burden of cancer attributable to high body-mass index in 2012: a population-based study. Lancet Oncol. 2015;16(1):36–46.
Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, Mullany EC, Biryukov S, Abbafati C, Abera SF, et al. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: a systematic analysis for the global burden of disease study 2013. Lancet. 2014;384(9945):766–81.
Allman-Farinelli MA, Chey T, Bauman AE, Gill T, James WP. Age, period and birth cohort effects on prevalence of overweight and obesity in Australian adults from 1990 to 2000. Eur J Clin Nutr. 2008;62(7):898–907.
Xi B, Liang Y, He T, Reilly KH, Hu Y, Wang Q, Yan Y, Mi J. Secular trends in the prevalence of general and abdominal obesity among Chinese adults, 1993-2009. Obes Rev. 2012;13(3):287–96.
Basterra-Gortari FJ, Bes-Rastrollo M, Ruiz-Canela M, Gea A, Martinez-Gonzalez MA. Prevalence of obesity and diabetes in Spanish adults 1987-2012. Med Clin. 2017;148(6):250–6.
Ness RB, Grisso JA, Vergona R, Klapper J, Morgan M, Wheeler JE. Oral contraceptives, other methods of contraception, and risk reduction for ovarian cancer. Epidemiology. 2001;12(3):307–12.
Cibula D, Gompel A, Mueck AO, La Vecchia C, Hannaford PC, Skouby SO, Zikan M, Dusek L. Hormonal contraception and risk of cancer. Hum Reprod Update. 2010;16(6):631–50.
Beral V, Doll R, Hermon C, Peto R, Reeves G. Ovarian cancer and oral contraceptives: collaborative reanalysis of data from 45 epidemiological studies including 23,257 women with ovarian cancer and 87,303 controls. Lancet. 2008;371(9609):303–14.
Bray F, Loos AH, Tognazzo S, La Vecchia C. Ovarian cancer in Europe: cross-sectional trends in incidence and mortality in 28 countries, 1953-2000. Int J Cancer. 2005;113(6):977–90.
La Vecchia C, Levi F, Lucchini F, Negri E, Franceschi S. Descriptive epidemiology of ovarian cancer in Europe. Gynecol Oncol. 1992;46(2):208–15.
Kjaerbye-Thygesen A, Huusom LD, Frederiksen K, Kjaer SK. Trends in the incidence and mortality of ovarian cancer in Denmark 1978-2002. Comparison with other Nordic countries. Acta Obstet Gynecol Scand. 2005;84(10):1006–12.
Xiang F, Guan Q, Liu X, Xiao H, Xia Q, Liu X, Sun H, Song X, Zhong Y, Yuan CH, et al. Distribution characteristics of different human papillomavirus genotypes in women in Wuhan, China. J Clin Lab Anal. 2018;32(8):e22581.
Tretarre B, Molinie F, Woronoff AS, Bossard N, Bessaoud F, Marrer E, Grosclaude P, Guizard AV, Delafosse P, Bara S, et al. Ovarian cancer in France: trends in incidence, mortality and survival, 1980-2012. Gynecol Oncol. 2015;139(2):324–9.
Webb PM, Green AC, Jordan SJ. Trends in hormone use and ovarian cancer incidence in US white and Australian women: implications for the future. Cancer Causes Control. 2017;28(5):365–70.
Yoshida H, Sakamoto H, Leslie A, Takahashi O, Tsuboi S, Kitamura K. Contraception in Japan: current trends. Contraception. 2016;93(6):475–7.
Department of Economic and Social Affairs, Population Division. Trends in Contraceptive Use Worldwide 2015. New York: Department of Economic and Social Affairs, Population Division; 2015.
Fathalla MF. Incessant ovulation--a factor in ovarian neoplasia? Lancet. 1971;2(7716):163.
Cramer DW, Welch WR. Determinants of ovarian cancer risk. II. Inferences regarding pathogenesis. J Natl Cancer Inst. 1983;71(4):717–21.
Adami HO, Hsieh CC, Lambe M, Trichopoulos D, Leon D, Persson I, Ekbom A, Janson PO. Parity, age at first childbirth, and risk of ovarian cancer. Lancet. 1994;344(8932):1250–4.
Wentzensen N, Poole EM, Trabert B, White E, Arslan AA, Patel AV, Setiawan VW, Visvanathan K, Weiderpass E, Adami HO, et al. Ovarian Cancer risk factors by histologic subtype: an analysis from the ovarian Cancer cohort consortium. J Clin Oncol. 2016;34(24):2888–98.
Europe the continent with the lowest fertility. Human reproduction 2010, 16(6):590–602.
United Nations, Department of Economic and Social Affairs, Population Division (2015). World Population Prospects: the 2015 Revision, Key Findings and Advance Tables. Working Paper NO. ESA/P/WP.241. Available from: https://esa.un.org/unpd/wpp/Publications/Files/Key_Findings_WPP_2015.pdf.
Vaccarella S, Lortet-Tieulent J, Plummer M, Franceschi S, Bray F. Worldwide trends in cervical cancer incidence: impact of screening against changes in disease risk factors. Eur J Cancer. 2013;49(15):3262–73.
Lewington S, Li L, Murugasen S, Hong LS, Yang L, Guo Y, Bian Z, Collins R, Chen J, He H, et al. Temporal trends of main reproductive characteristics in ten urban and rural regions of China: the China Kadoorie biobank study of 300 000 women. Int J Epidemiol. 2014;43(4):1252–62.
Greiser CM, Greiser EM, Doren M. Menopausal hormone therapy and risk of ovarian cancer: systematic review and meta-analysis. Hum Reprod Update. 2007;13(5):453–63.
Beral V, Gaitskell K, Hermon C, Moser K, Reeves G, Peto R. Menopausal hormone use and ovarian cancer risk: individual participant meta-analysis of 52 epidemiological studies. Lancet. 2015;385(9980):1835–42.
Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ, Kooperberg C, Stefanick ML, Jackson RD, Beresford SA, Howard BV, Johnson KC, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women's Health Initiative randomized controlled trial. JAMA. 2002;288(3):321–33.
La Vecchia C, Rota M, Malvezzi M, Negri E. Potential for improvement in cancer management: reducing mortality in the European Union. Oncologist. 2015;20(5):495–8.
Dhillon PK, Yeole BB, Dikshit R, Kurkure AP, Bray F. Trends in breast, ovarian and cervical cancer incidence in Mumbai, India over a 30-year period, 1976-2005: an age-period-cohort analysis. Br J Cancer. 2011;105(5):723–30.
Skirnisdottir I, Garmo H, Wilander E, Holmberg L. Borderline ovarian tumors in Sweden 1960-2005: trends in incidence and age at diagnosis compared to ovarian cancer. Int J Cancer. 2008;123(8):1897–901.
Or Knudsen A, Schledermann D, Nyvang GB, Mogensen O, Herrstedt J. Trends in gynecologic cancer among elderly women in Denmark, 1980–2012. Acta Oncol. 2016;55(Suppl 1):65–73.
This work was supported by the National Natural Science Foundation of China [grant number: 81473064 and 81773543]. The funders had no role in study design, data collection, data analysis, data interpretation, in writing the manuscript, and decision to publish.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Fig. S1 Temporal trends in age-standardized (world 1960 Segi population) ovarian cancer incidence rates per 100,000 women for European countries for all ages from 1973 to 2015. Fig. S2 Temporal trends in age-standardized (world 1960 Segi population) ovarian cancer incidence rates per 100,000 women for non-European countries for all ages from 1973 to 2015. Fig. S3 Average annual percentage change in ovarian cancer incidence rates for all ages in the most recent 10-year period available. * Average annual percent change is significantly different from zero. Fig. S4 Ovarian cancer incidence rates per 100,000 women by year of birth for European countries. For each graph, the rates in 5-year age groups (e.g., 20–24, 25–29, …, 80–84) are plotted. Fig. S5 Ovarian cancer incidence rates per 100,000 women by year of birth for non-European countries. For each graph, the rates in 5-year age groups (e.g., 20–24, 25–29, …, 80–84) are plotted. Fig. S6 Fitted age-specific ovarian cancer incidence rates per 100,000 women (left) and incidence rate ratios by birth cohort (right) in European countries. The default is for the reference points at the median value (with respect to the number of cases) for the cohort to be variable. Fig. S7 Fitted age-specific ovarian cancer incidence rates per 100,000 women (left) and incidence rate ratios by birth cohort (right) in non-European countries. The default is for the reference points at the median value (with respect to the number of cases) for the cohort to be variable.
About this article
Cite this article
Zhang, Y., Luo, G., Li, M. et al. Global patterns and trends in ovarian cancer incidence: age, period and birth cohort analysis. BMC Cancer 19, 984 (2019). https://doi.org/10.1186/s12885-019-6139-6