Breast cancer mortality patterns and time trends in 10 new EU member states: Mortality declining in young women, but still increasing in the elderly
Article first published online: 18 AUG 2004
Copyright © 2004 Wiley-Liss, Inc.
International Journal of Cancer
Volume 112, Issue 6, pages 1056–1064, 20 December 2004
How to Cite
Tyczynski, J. E., Plesko, I., Aareleid, T., Primic-Zakelj, M., Dalmas, M., Kurtinaitis, J., Stengrevics, A. and Parkin, D. M. (2004), Breast cancer mortality patterns and time trends in 10 new EU member states: Mortality declining in young women, but still increasing in the elderly. Int. J. Cancer, 112: 1056–1064. doi: 10.1002/ijc.20514
- Issue published online: 12 OCT 2004
- Article first published online: 18 AUG 2004
- Manuscript Accepted: 28 MAY 2004
- Manuscript Received: 2 FEB 2004
- breast cancer;
- European Union;
- risk factors
Breast cancer is the cancer diagnosed most frequently in women worldwide. In Europe it is the most common cancer in the female population, with approximately 350,000 new cases diagnosed each year including 130,000 deaths. Incidence rates are increasing in the majority of European countries, whereas a decline in mortality rates has been observed in many West European countries since the late 1980s and early 1990s. Our study examines breast cancer mortality patterns and time trends in the new European Union (EU) member states and compares them with the situation in current EU member states. A Joinpoint regression analysis was used to assess temporal changes in mortality rates and the trends examined in the light of known risk factors, screening programs and advances in treatment. In the majority of the countries analyzed, a deceleration in the increase of mortality rates appeared, followed by a decrease of mortality in many of them in the second half of the 1990s. The declining tendency was visible primarily in young women, and to a lesser extent in middle-aged women, whereas in elderly women a continuing increase of mortality was observed. Analysis of mortality data, information from previous publications, as well as analysis of known factors influencing breast cancer risk suggest that changes observed are due mainly to recent advances in treatment rather than changes in lifestyle risk factors or the result of screening programs. Early detection and a shift toward more favorable stage distribution could have played the leading role for mortality decline in younger patients. © 2004 Wiley-Liss, Inc.
Worldwide, breast cancer is the most frequent cancer in women. The highest incidence rates are observed in North America, whereas the lowest risk of breast cancer is observed in Asia and Africa.1 Breast cancer is also the most common cancer in females in Europe. It is estimated that in the year 2000 there were 350,000 new breast cancer cases in Europe and the number of deaths from breast cancer was estimated at 130,000.2 Breast cancer is responsible for 26.5% of all new cancer cases among women in Europe and 17.5% of cancer deaths.
There are significant differences in the frequency of breast cancer incidence and mortality within Europe. The highest rates are observed in Western and Northern Europe whereas Eastern Europe has the lowest incidence and mortality rates. There are also differences in the direction of mortality time trends among European countries.3 A decline in mortality has occurred in the 1980s and the 1990s in some countries, e.g., the Netherlands, UK and Sweden. There has been a steady increase in mortality in countries such as Spain, Romania and Hungary since the 1950s and 1960s.
Time-trends in breast cancer mortality in Western European countries have been analyzed and described several times.3, 4, 5 To date, however, only fragmentary data have been published on patterns and trends in other European countries (including new EU member states).6, 7 In addition, several Western European countries have introduced national or regional breast cancer screening programs, whereas no such programs have been implemented in Central and Eastern European countries.3 Comparison of breast cancer patterns and trends between Western Europe and other parts of the continent may therefore allow the assessment of the impact of screening programs on mortality patterns, as well as the impact of other factors (including advances in treatment procedures, e.g., use of tamoxifen). There are several lifestyle and behavioral factors that are associated with the risk of breast cancer such as age at menarche and menopause, childbearing, breast-feeding, hormonal status, consumption of alcohol and type of diet, obesity, radiation and genetic susceptibility. Our study assesses the impact of these factors on breast cancer mortality in the new member states of the EU.
MATERIAL AND METHODS
Mortality data for malignant neoplasms of the breast (ICD-9 code 174) were extracted from the World Health Organization (WHO) Mortality Databank for 10 European countries (new members of the European Union), by 5-year age group (http://www-depdb.iarc.fr/who/menu.htm). Data suitable for the analysis of time-trends were available for 9 countries (Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovakia and Slovenia). Corresponding population data, by age and year, were also extracted. Estimates of cancer mortality for the year 2000 for all 10 accession countries (the above 9 plus Cyprus) were derived from the Globocan2000 software and database.2
Age-standardized mortality rates (ASRs) were calculated for all ages combined and for the following age groups: 30–49, 50–64 and 65+. The World Standard Population was used for age-adjustment.8 A Joinpoint regression was fitted to provide estimated annual percentage change (EAPC) and to detect points in time where significant changes in the trends occur.9, 10 For each EAPC estimate we also calculated the corresponding 95% confidence interval (95%CI). The Joinpoint regression model describes continuous changes in rates and uses the grid-search method to fit the regression function with unknown joinpoints. Thus, Joinpoint is a useful way to summarize trends in cancer rates and allows assessment of recent changes in trend.
There were an estimated 130,000 breast cancer deaths in Europe in 2000, nearly 80,000 deaths in the EU member countries and nearly 12,000 in the 10 new EU member states (Table I). The average age-standardized mortality rate (ASR) in the new member states was 19.2/105, lower than the average for the 15 EU members (22.5/105) and the average for the whole of Europe (20.2/105). The highest ASRs among the 10 analyzed countries were observed in Malta (28.4/105), Hungary (25.3/105) and the Czech Republic (21.0/105) (Table I). The lowest mortality rates were found in Cyprus (16.7/105), Poland (16.8/105) and Latvia (18.1/105) (Table I). The relative frequency of breast cancer among all deaths due to cancer in females varied from 14.2% in Poland to 27.7% in Cyprus (Table I).
|Country||Breast cancer - females|
|ASR1||Number of deaths||% of all cancer deaths|
|EU new member||19.2||11,840||15.3|
Mortality time trends in individual countries
The term “the youngest age group” refers to the age category 30–49 years of age, “middle-aged” refers to 50–64 years of age and “the oldest age group” will refers to age category 65 years of age or greater.
No data were available to analyze changes in breast cancer mortality over time.
For all ages combined mortality was relatively stable between 1986–1993 (EAPC of 0.2%), but a significant decline in mortality rates occurred after 1994 (EAPC −2.4%) (Table II, Fig. 1). Examining the changes by age group (since 1986) shows a significant decline in mortality in young women (30–49 years) by 5.0% per year, and a significant decrease in middle-aged women (50–64 years) by 0.7% per year. In the oldest age category (age 65 and more) mortality seems to be relatively constant (Fig. 2).
|Country||Number of joinpoints||Year of joinpoint||Period||EAPC1||95% CI|
In Estonia, overall mortality from breast cancer has been increasing in the period 1981–1999 (EAPC of 1.6%) (Table II, Fig. 1). In young women (30–49 years) mortality rates were stable over the whole period, whereas in middle-aged and older women, a significant increase of mortality by 2.1% per annum was observed (Fig. 3).
The increase in all ages mortality rates was faster before 1975; 3.4% per annum than in 1975–1993 (1.2%). Since 1994 a non-significant decline has been noted (−0.4%) (Table II). In the oldest age category (65+) the rate of increase slowed from 5.9% per year to 1.4% after 1971 (Fig. 4). A similar slowing was observed in middle-aged women (50–64 years) from 3.8% to 0.9% (Fig. 4). In young women (30–49 years) the increase of mortality of 1.7% per year before 1989 changed to a significant decline of −1.4% per year thereafter (Fig. 4).
A continuous significant increase has been observed in all-age mortality (of 1.5% per year) (Table II), as well as in the oldest age category (2.4% per year) (Fig. 5). In middle-aged women a non-significant decrease in mortality was observed in the second half of the 1980s and in the 1990s (EAPC of −0.6%). Among the youngest women a small (non-significant) increase of mortality (of 1.3% per annum) was noted throughout the period (Fig. 5).
Overall mortality rates from breast cancer were increasing by 2.0% per year from 1981 until the mid-1990s, but since 1996 a non-significant drop in mortality has occurred (Table II). Mortality rates in both middle-aged and the oldest age categories were increasing over the whole period, by 1.6% and 3.2% per year respectively (Fig. 6). In the youngest women mortality rates were stable until mid 1990s, and then declined (Fig. 6).
There were significant increases in all-ages mortality in the 1960s and 1970s of 2.7% per year, followed by a non-significant decline in the 1980s and the 1990s (Table II). Mortality rates in the oldest age group showed marked fluctuations (Fig. 7). In middle-aged females, no significant changes in mortality have been observed since the beginning of the 1970s, whereas in the youngest age category a significant decline of 2.1% per annum has been noted since about 1980 (Fig. 7).
The rate of increase in mortality slowed from 2.2% per year in the period 1965–1979, to 0.7% in the period 1980–1995, to a decrease (−2.1%) after 1995 (Table II). This deceleration was also visible in specific age categories: from 2.2–0.8% in elderly women, from 2.2–0.1% in middle-aged females, and from 2.1% to −4.9% in the youngest age group (Fig. 8).
There was an increase in mortality rates (EAPC of 2.0%) until 1985, followed by a plateau (−0.2%) (Table II). Trends in specific age classes were heterogeneous: in the youngest and middle-aged women, a non-significant decline in mortality has taken place since the mid-1980s (−2.0% and −0.7% respectively) (Fig. 9). In the oldest age category mortality rates increased by 2.0% per annum throughout the entire period (1968–2000) (Fig. 9).
After an increase in mortality between 1985–1992 (EAPC of 1.6%), rates have fluctuated in the period 1993–1999 (Table II). In specific age categories, changes in mortality rates were consistent only in the youngest age group, where a significant decline in mortality was observed throughout the period analyzed (EAPC of −1.7%) (Fig. 10).
Our study presents breast cancer patterns and trends in the new member states (NMS) of the European Union (EU) and compares them with the situation in other parts of Europe, in particular with the 15 current member states of the EU (CMS). Possible explanations for the observed temporal changes in mortality in the NMS are examined below.
Our study has some limitations. Mortality rates have been used to study breast cancer patterns because not all of the NMS have had a nation-wide cancer reporting system (there are no such national incidence data from Hungary, Cyprus or Poland). Moreover, the quality and completeness of incidence data from several NMS have not been satisfactory. A second limitation is the varying time periods for which data were available (different for particular countries, although available for all for the second half of the 1980s and the whole of the 1990s).
In general, the level of breast cancer mortality in the NMS is lower than in Europe as a whole and the current EU member states (Table I). There is, however, significant heterogeneity in the level of mortality between NMS. Hungary and Malta have mortality rates higher than EU average, whereas Poland and Cyprus have mortality rates nearly 20% lower than the European average value.
Declines (or a plateau) in breast cancer mortality have been observed in several West European countries since the 1980s (e.g., United Kingdom and the Netherlands).3, 5 A decline in mortality has also been observed in the United States Caucasian population since the late 1980s.13 Mammography-based screening programs were implemented in several countries of Western Europe in the late 1980s and the beginning of the 1990s.14, 15 There is, however, a lag between the implementation of a program and its measurable outcome (i.e., decline in mortality), estimated by randomised prospective screening trials as approximately 10 years.16 It is difficult to distinguish between the effects of screening on mortality and other influences (e.g., improved methods of treatment).17, 18 For example, it has been estimated that less than one-third of the 21.3% decline in mortality of women 50–69 years of age in the United Kingdom (those involved in screening program) in the period 1990–1998 was due to screening, whereas the remainder was due to other factors (mainly the effect of improved treatment).19 The recent sudden declines in breast cancer mortality in other developed countries may also be due mainly to advances in treatment rather than to screening programs.4, 16
In several NMS a decline in mortality (or a leveling off) has been observed in the second half of the 1990s (e.g., Czech Republic, Hungary, Malta, Poland, Slovakia), although breast cancer screening program have not yet been implemented in any of these countries. Moreover, the decreasing tendency appeared mainly in young women, 30–49 years of age, who are outside the age group targeted by screening. The only country where a significant decline in mortality in middle-aged women (50–64 years) occurred was the Czech Republic, whereas in elderly women (65 years of age and more), an increase of mortality was observed in all countries.
One of the factors that may play an important role in diminishing mortality from breast cancer is adjuvant tamoxifen treatment. It has been shown that tamoxifen significantly reduces mortality among breast cancer patients.20 The effect is more marked in subjects with estrogen-receptor (ER)-positive tumors (about 70% of cases in Europe), and in older patients (over 50 years of age).21 As noted earlier, however, the decline in mortality in the NMS has been observed mainly in young women (30–49 years). It suggests that in the NMS adjuvant tamoxifen treatment does not play a leading role in the process of mortality decline. Because the decline is observed mainly in young women it is likely that earlier diagnosis and shift toward more favorable stage distribution are the contributors for decrease in mortality. It may be a result of increased breast cancer awareness in populations of the NMS. Recent observations in Slovakia support that hypothesis.22
Changing exposure to other factors that are known to affect the risk of breast cancer must be considered as possible explanations. They include childbearing, breast-feeding, type of diet and obesity, use of alcohol and tobacco, oral contraceptives use and hormonal replacement therapy.23, 24 Mortality from breast cancer may also be influenced by prevention (e.g., chemoprevention using tamoxifen or raloxifene) and more directly by improvement of treatment procedures. Unfortunately, there is little information available on exposures to these factors in NMS.
It is known that reproductive factors play an important role in influencing breast cancer risk,25, 26 an association that has been shown in studies from NMS (Estonia, Poland, Slovakia, Slovenia).27, 28, 29, 30, 31 It is possible that differences in the mortality rates among the NMS, as well as changes in mortality trends may be (in part) due to differences in childbearing among women in these countries. Table III shows that fertility rates (average number of children born to women in reproductive age), although different across the countries, were relatively stable in the 1970s and 1980s in the majority of countries. Between 1990–2000, however, a steep decline was noted in all analyzed countries. In Slovenia, the average number of children per woman decreased from 4.0 (birth cohort 1882–1886) to 1.8 (birth cohort 1947–1951) and is still declining.6 It has also been shown that the incidence of breast cancer increased by 70% in Slovenia from 1971–1993. The percentage of nulliparous women and the average family size explained 38% of the variation in the cohort effects.6
|Country||Total fertility rate|
Data on the percentage of live births under age 20 and over 35 (Table IV) show that in several countries (i.e., Czech Republic, Hungary, Slovenia) there has been a drop in the proportion of children born by women below age 20. It suggests that in these countries an average age at the time of first birth has increased over time. This observation is supported by data from Poland, where the average age of women at first birth increased from 22.8 years in 1970, to 23.4 in 1980, 23.5 in 1990 and 24.2 years in 2000.7 It therefore seems unlikely that changes in childbearing habits in the NMS have had a significant (favorable) impact on breast cancer mortality changes in the 1900s.
|Country||% of all livebirths under age 20||% of all livebirths at age 35+|
Another life style factor, which probably plays important role in breast cancer risk, is breast-feeding. It has been shown that breast-feeding may decrease the risk of developing breast cancer,32, 33 especially in premenopausal women.34 Available data suggest that in many NMS, an increase in prevalence of breast-feeding took place within recent decades.11 In the Czech Republic, for instance, prevalence of breast-feeding (at 6 months of age) increased from 9.0% in 1993 to 28.4% in 2001, in Estonia from 14.8% in 1992 to 39.1% in 2001, in Hungary from 12.7% in 1980 to 44.5% in 2000 and in Slovakia from 24.6% on 1991 to 32.3% in 2001.11 It is, however, to early to expect that change in breast-feeding habits in the 1990s would have an important influence on breast cancer mortality patterns.
Moreover, if changes in breast-feeding would have had a beneficial impact on breast cancer it should have been visible in incidence rates first. An increase of overall and age-specific incidence rates in several NMS (Czech Republic, Estonia, Slovakia, and Slovenia) has been reported by Botha et al.,3 whereas in the majority of these countries a decline (or plateau) of mortality in the second half of the 1990s has been observed. Data from other NMS published in the EUROCIM data set indicate an increase in incidence rates (Latvia, Lithuania, and some Polish regional cancer registries). It may suggest that lifestyle risk factors, such as reproductive factors and breast-feeding do not play a leading role in diminishing breast cancer mortality in NMS because it should be observed first in decreasing incidence rates.
Mortality from a given cancer is a function of incidence and survival. It has been reported by Sant et al.35 that, based on the results of the EUROCARE-2 study, 5-year relative survival rates of breast cancer patients in Estonia and Poland in the period 1978–1989 were lower than in Western and Northern Europe (Fig. 11). It was also reported by the same authors that increase of survival rates in the period 1978–1989 in Eastern Europe was much more rapid than in other European regions (e.g., Eastern Europe, increase by 24.8%; Western Europe, by 8.6%; Northern Europe, by 5.1%).35 If the first period (1978–1980, for which the survival for Poland was surprisingly low compared to subsequent periods) was taken out of the calculations, however, the change in survival rates (i.e., between periods 1981–1983 and 1987–1989) would be similar for all European regions. For example, for Western Europe the change was 8.3%, whereas for Eastern Europe it was 7.9%. Moreover, in Estonia, 5-year relative survival changed only moderately between 1978–1980 and 1987–1989 (range = 57–62%).36 The recent results of the EUROCARE-3 study demonstrated no improvement in survival of breast cancer patients in Estonia also for 1990–1994 (5-year rate of 62%).37 It seems to explain an absence of an overall decline in breast cancer mortality in Estonia. In Poland, decline in mortality in the second half of 1990s (EAPC of −2.1% estimated in our analysis) was similar to that observed by Botha et al.3 since the end of the 1980s in Scotland (−2.0%), and higher than observed in Finland (−0.2%), Iceland (−0.5%) and Sweden (−1.2%).
It has been suggested by Aareleid and Brenner38 that the main reason for the big difference between Estonia and the developed European countries in breast cancer survival is a late, mainly symptom-based, diagnosis (particularly in elderly women), and the low proportion of small tumors (T1N0M0) diagnosed. The latter was also observed by Sant.39 It is also intriguing that differences in survival between some NMS and other parts of Europe vary in particular age groups. Our calculations based on the data from Quinn et al.36 show that in Estonia, Poland and Slovenia relative difference (compared to an European average) in survival in young women (15–44 years) and in middle-aged women are smaller than in older age groups (Table V). Moreover, in Estonia in the period 1978–1989, survival rates increased in young and middle-aged women, but declined in elderly.12 It resulted in increased difference in survival between the youngest and the oldest women in Estonia (difference increased from 15% in 1978–1980 to 32% in 1987–1989). At the same time the tendency for the whole Europe was reverse: the difference between the youngest and the oldest age groups declined from 25% to 10% (survival were increasing faster in the elderly women). It may suggest that in the several NMS beneficiaries of the improvements in detection and treatment of breast cancer were mainly young and middle-aged women. Changes in mortality time-trends by age in the NMS seem to support this hypothesis.
In view of the knowledge (or lack of it) about recent changes in incidence, survival and risk factors, it is difficult to unequivocally assess why mortality from breast cancer has started to decline in some NMS in the 1990s but not in others. An increase of incidence observed by other authors,3 accompanied by decline in mortality observed mainly in young age groups and lack of organized screening programs, suggest that the leading role in mortality changes may be due to more successful treatment. This could be implementation of better techniques and medications (tamoxifen), as well as earlier diagnosis (due to better access to health care facilities). It is supported by the observation of other authors indicating that recent declines in breast cancer mortality are due mainly to improvements in treatment procedures and their accessibility.3, 4, 16, 19 Early detection and a shift toward more favorable stage distribution could have play a leading role for mortality decline in younger patients.
Undoubtedly, there is a need for a more comprehensive Europe-wide analysis of breast cancer patterns and temporal changes in light of the risk factors and preventive measures known to influence the risk of breast cancer. It requires good quality data not only on breast cancer incidence, survival and mortality, but also good quality information on known risk factors. This is unlikely to be achieved without a comprehensive Europe-wide surveillance system for cancer and its determinants.
- 2GLOBOCAN 2000: Cancer incidence, mortality and prevalence worldwide. Version 1.0. IARC CancerBase No. 5. Lyon: IARC Press, 2001., , , .
- 7A plateau of breast cancer mortality in Poland—an impact of time periods or birth generations? A joinpoint regression and APC analysis of recent time trends. Nowotwory J Oncol 2003; 4: 398–404..
- 10National Cancer Institute. Joinpoint regression program. Version March 2002.
- 11WHO-EURO. Health for all database. Copenhagen: WHO Euro, 2003.
- 12Survival of cancer patients in Europe: the EUROCARE-2 study. no. 151. Lyon: IARC, 1999., , , , , , , .
- 20Early Breast Cancer Trialists' Collaborative Group (EBCTCG). Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy. Lancet 1992; 339: 1–15, 71–85.
- 22Epidemiological aspects of in situ female breast cancer in Slovakia. Klinicka onkologie 2003; 16: 227–30., , .
- 29East-European Study Group of Breast Cancer Epidemiology (EESGBCE). Comparative study of breast cancer risk factors in Estonia and Slovakia. Neoplasma 1990; 37: 97–104.
- 33Collaborative Group on Hormonal Factors in Breast Cancer (CGHFBC). Breast cancer and breast-feeding: collaborative reanalysis of individual data from 47 epidemiological studies in 30 countries, including 50,302 women with breast cancer and 96,973 women without the disease. Lancet 2002; 360: 187–95.
- 37EUROCARE-3: survival of cancer patients diagnosed 1990-94—results and commentary. Ann Oncol 2003; 14( Suppl): v6–118., , , , , , , , , , , , et al.