Cancer mortality in Europe, 1995–1999, and an overview of trends since 1960
Version of Record online: 19 FEB 2004
Copyright © 2004 Wiley-Liss, Inc.
International Journal of Cancer
Volume 110, Issue 2, pages 155–169, 10 June 2004
How to Cite
Levi, F., Lucchini, F., Negri, E., Boyle, P. and La Vecchia, C. (2004), Cancer mortality in Europe, 1995–1999, and an overview of trends since 1960. Int. J. Cancer, 110: 155–169. doi: 10.1002/ijc.20097
- Issue online: 31 MAR 2004
- Version of Record online: 19 FEB 2004
- Manuscript Accepted: 2 DEC 2003
- Manuscript Revised: 25 NOV 2003
- Manuscript Received: 3 JUL 2003
- Swiss League Against Cancer
- Italian Association for Cancer Research
- time trends;
Mortality data, abstracted from the World Health Organization database, are presented in tabular form for 26 cancer sites or groups of sites, plus total cancer mortality, in 36 European countries during the period 1995–1999. Trends in mortality are also given in graphic form for 23 major countries plus the European Union as a whole over the period 1960–1999. In the European Union, total cancer mortality declined by 7% for both sexes over the last 5 years considered. The fall since the late 1980s was 10% in both sexes, corresponding to the avoidance of over 90,000 deaths per year, as compared to the rates of the late 1980s. For the first time, over the last few years, some leveling of mortality was reported also in the Russian Federation, the Czech Republic, Poland, Hungary and other Eastern European countries, although cancer rates in those areas remain exceedingly high. The overall favorable pattern of cancer mortality over recent years is largely driven by the decline of tobacco-related cancer mortality in men. However, important components of the trends are also the persistence of substantial falls in gastric cancer, mainly in Russia and Eastern Europe, the recent decline in intestinal cancer in both sexes and of breast cancer in women, together with the long-term falls in uterine (cervical) cancer, leukemias, Hodgkin's disease and other neoplasms amenable to advancements in diagnosis and treatment. Female lung cancer mortality has been declining in the Russian Federation, but is still rising in other areas of the continent. Thus, urgent intervention is needed to bring under control the tobacco-related lung cancer epidemic in European women before it reaches the high level observed in North America. Supplementary material for this article can be found on the International Journal of Cancer website at http://www.interscience.wiley.com/jpages/0020-7136/suppmat/index.html © 2004 Wiley-Liss, Inc.
In Western Europe, cancer mortality rates peaked in the late 1980s, and the fall in cancer mortality in the European Union (EU) between 1988 and 1997 has been approximately 10%.1, 2 Over more recent years,1 mortality from lung and other tobacco–related neoplasms, but also a few other common sites, has started to level off and decline among males in Central and Eastern Europe, too.3
These trends in major European areas, however, concealed different patterns of trends for various cancer sites, countries and age groups.4, 5, 6 We systematically analyzed trends in mortality from 26 cancer sites, besides total cancer, in 24 major European countries over the period 1955–1994.6, 7
The present update report has the major aim of providing a summary description and documentation of the global pattern of trends in cancer mortality in Europe until the late 1990s. Only a few general comments, and a selected number of references, are included to assist reading and interpretation of trends for major cancer sites. Since long-term trends in cancer mortality include useful information in order to understand recent patterns and to project most likely future trends,8, 9 the present article also gives a summary overview of previous trends since 1960.
MATERIAL AND METHODS
Official death certification numbers for 36 European countries (including the Russian Federation but excluding a few smaller countries such as Andorra and Liechtenstein) were derived from the World Health Organization (WHO) database as available on electronic support (www3.who.int/whosis/menu.cfm).10 Besides the United Kingdom as a whole, data were presented also for England and Wales, Scotland and Northern Ireland. For Belgium, data were available only for 1995–1996; for Iceland, for 1995–1997; and for Denmark, for 1995–1998.
The EU was defined as the 15 member states in 1999 (Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, The Netherlands, Portugal, Spain, Sweden and the United Kingdom). Accession (in 2004) and applicant states were not included.
During the calendar period considered (1960–1999), 4 different revisions of the International Classification of Diseases (ICD) were used.11, 12, 13, 14 Classification of cancer deaths was recoded, for all calendar periods and countries, according to the ninth revision.13 To improve validity and comparability of data throughout different countries, we pooled together all intestinal sites, including rectum, all uterine cancers (cervix and endometrium), all skin neoplasms (melanoma and nonmelanomatous) and some relatively uncommon cancers (all connective and soft tissue sarcomas, all non-Hodgkin's lymphomas). Neoplasms of the brain or nerves are not presented, since it was not possible to pool together consistently subsequent revisions of the ICD.
Estimates of the resident population, generally based on official censuses, were obtained from the same WHO database. From the matrices of certified deaths and resident populations, age-specific rates for each 5-year age group (from 0–4 to 80–84 and ≥ 85 years) and calendar period were computed. Age-standardized rates per 100,000 population, at all ages and truncated 35–64 years, were computed using the direct method and based on the world standard population.15 In a few countries, data were missing for part of one or more calendar years.6 No extrapolation was made for missing data.
In the interpretation of the data presented, it is important first to consider problems related to random variation, which are clearly greater in relation to smaller populations and rarer cancers. Since this is essentially a descriptive report, no inference is made on the statistical significance of rates and trends. Secondly, and more complex, there are problems of death certification reliability and validity in various countries.16, 17 In general, for several common cancer sites, including lung, stomach, colorectum and breast, death certification is sufficiently reliable to permit meaningful inference on trends for most European countries, particularly those from Western Europe, as well as for major Central and Eastern European countries, and chiefly under the age of 65 years. Some underrecording of cancer deaths was reported for the Russian Federation in the late 1980s and 1990s due to a fall in precision of coding of causes of death. This was, however, mainly restricted to the elderly living in rural areas18 and should therefore not have materially influenced truncated rates. Greater caution is required for a few other countries (particularly Albania and Belarus, whose rates were exceedingly low, and a few new national entities, for which no long-term trend was available) and for a few cancer sites, whose diagnosis and certification may be substantially influenced by the availability of diagnostic techniques and the accuracy of death certification, reflecting variable data reliability, and different criteria and attitudes toward death certification. These include, among others, liver, bone, soft tissue sarcomas, prostate, myeloma and other lymphoid neoplasms.16, 17, 19, 20 Furthermore, we were unable to obtain consistent trends over time in various countries for neoplasms of brain and nerves.
Figure 1 includes world standardized rates (at all ages and truncated 35–64 years) for 23 major cancer sites plus total cancer mortality in the EU between 1960 and 1999 for males and females. Figure 2 gives histograms for overall age-standardized death certification rates for 11 selected cancers and total cancer mortality in 36 countries (plus the EU) over the most recent calendar quinquennium (1995–1999). Trends for the EU are given first. All countries, including EU member states and accession countries, are given subsequently in alphabetical order. Selected scales have been defined for each neoplasm in order to assist reading and interpretation.
In addition to Figures 1 and 2, for each cancer site a table and a figure are available as supplementary material on the International Journal of Cancer website at http://www.interscience.wiley.com/jpages/0020-7136/suppmat/index.html. The table includes for the period 1995–1999, for each site and sex, the observed number of deaths and the age-standardized (world standard) overall and truncated 35–64 years mortality rates. The figures include world standardized rates (at all ages and truncated 35–64 years) for (the EU and 23 countries selected among the largest ones providing long term trends) and 5-year calendar periods from 1960–1964 to 1995–1999 for males and females.
Mouth or pharynx
In the EU as a whole, mortality from oral and pharyngeal cancer peaked for males in the late 1980s at 6.3/100,000 and leveled off thereafter to reach 5.8/100,000 at all ages. The fall was even larger in middle age (from 13.4 to 12.3; −8%). Rates were substantially lower for women, but tended steadily to rise, to reach 1.1/100,000 in 1995–1999. This pattern of trends reflects the fall in oral and phyryngeal cancer mortality for males since the late 1980s, mainly in countries like Italy and mostly France, which had exceedingly high rates in the past. There was also some leveling in male oral and pharyngeal cancer mortality in countries like Austria, Germany and the United Kingdom, whose rates have been rising up to the early 1990s, but not in Belgium, Denmark or Scotland, whose rates were still substantially upwards in the late 1990s. In Eastern Europe, there was some leveling of rates in Poland, the Czech Republic as well as in Russia, but mortality was still rising in Bulgaria, Romania and, most notably, in Hungary, whose rates in middle-age men are now higher than those of the mid-1980s in France. Some recent rises in men only were observed also in Portugal, while oral cancer mortality for Spanish men tended to level off over the last few years. These variable trends reflect the pattern of consumption of tobacco, but mostly of alcohol drinking, in various European regions.21, 22, 23 Thus, the persistence of unfavorable trends in oral cancer in males from the United Kingdom (mainly Scotland), despite the fall in lung cancer in that country,3 is essentially attributable to change in alcohol drinking in the United Kingdom. Likewise, the upward and discontinuous trends in Russia reflect the dramatic change in alcohol-related mortality in that country.24, 25 Second, it remains unclear whether different types of alcoholic beverages may have differential influences on oral carcinogenesis, and whether spirits are associated to higher risks than beer or wine.26, 27, 28, 29 This may contribute to explain some of the exceedingly high rates in countries like Hungary or Slovakia, where hard, fruit-derived spirits are commonly consumed. In the late 1990s, the highest mortality rates for males were in Hungary (20.2/100,000 men) and Slovakia (17.9/100,000), the lowest ones in Greece, Finland, Iceland and Sweden (around 2/100,000). For females, relatively high rates (1.6–2.6/100,000) were observed in Hungary and Denmark, too, again reflecting drinking and smoking patterns in these populations.
In the overall EU, as well as in most major European countries, the pattern was similar for esophageal cancer, underlying the similarities of risk factors for oral, pharyngeal and (squamous cell) esophageal cancer. The rise in the United Kingdom or The Netherlands was steeper, however, for esophageal cancer, suggesting that adenocarcinomas (which are related to tobacco and obesity, but not to alcohol drinking)30 may have become a non-negligeable proportion of all esophageal cancer deaths. The trends in high mortality countries for men (i.e., France and Italy mainly in the past, Hungary and Scotland over more recent years) are still dominated by squamous cell cancer, strongly related to alcohol and tobacco consumption.31, 32 Thus, trends in the Czech Republic and Hungary were upwards up to the most recent calendar period. The highest rates for males in 1995–1999 were registered in Hungary and Scotland (9–11/100,000) and, for women, in Scotland (4.7/100,000) and other U.K. areas (around 3/100,000).
Over the last decade, gastric cancer mortality declined in the EU as a whole from 14.7 to 10.2/100,000 men, and from 6.9 to 4.7/100,000 women. Although in absolute terms the fall was smaller than in previous calendar periods, in proportional terms there was no evidence of leveling in the decline of gastric cancer rates.33 In the late 1990s, the highest gastric cancer rates (over 20/100,000 males, over 10/100,000 females) were in Eastern Europe (Russia, the Czech Republic, Hungary, Poland, etc.), but also in Portugal. The lowest gastric cancer rates in the late 1990s were registered in Denmark (5.9/100,000 men, 2.9/100,000 women). Still, the fall over the last decade was over 30% in both sexes for Denmark, too. In all countries considered, however, rates were steadily downward, likely reflecting the spread of the improvement in diet and food conservation across time over the past decades.34 Over the last 2 decades, the fall was approximately 50% in the EU, 45% in Eastern Europe and 40% in Russia. Since the rates were appreciably higher, however, in absolute terms, the declines were greater in Russia and Eastern Europe. Steady declines in gastric cancer mortality, moreover, were observed in middle-age populations as well, suggesting that they are likely to persist in the near future.33 The reasons for the generalized declines in gastric cancer rates throughout Europe are complex and not completely understood. Almost certainly, these include a more varied and affluent diet, better food conservation, including refrigeration, as well as the control of Helicobacter pylori infection,34, 35, 36, 37 and reduced tobacco smoking, at least in males.38 Whether improved diagnosis and treatment has also played some role on the favorable trends in gastric cancer throughout Europe, particularly over most recent calendar periods, remains open to evaluation.39, 40 During the 1990s, gastric cancer was the main determinant of the fall observed in cancer mortality in the European Union.2 Even assuming, as suggested by the present analysis, a persisting fall in gastric cancer mortality over the next decade, the number of deaths avoided in the first decade of the current century will probably be not greater than 15,000.33 These figures will remain appreciably larger in Eastern European countries, and mostly in the Russian Federation, where gastric cancer does remain a public health priority.
Intestine (chiefly colon and rectum)
An appreciable fall was observed in mortality from intestinal (mostly colorectal) cancer in the EU (from 20.1/100,000 men in 1960–1964 to 19.2/100,000 in 1995–1999; from 14.0 to 12.1/100,000 in women). Thus, while in previous calendar years some fall in colorectal cancer mortality rates across Europe was observed in women but not so consistently in men,41 over most recent years a fall was observed for men, too. The favorable pattern in colorectal cancer mortality for both sexes was observed in all major EU countries, as well as in Norway and Switzerland, while rates were still upwards in Hungary, Poland and Romania, but not in the Czech Republic or the Russian Federation. Consequently, the Czech Republic had the highest colorectal cancer rates over recent calendar years (35.4/100,000 men, 18.0/100,000 women), followed by Hungary and Slovakia, but Denmark also had high rates (22.8/100,000). The trends in colorectal cancer mortality across Europe over recent years likely reflect favorable modification in (dietary) risk factor exposure, as well as some potential effect of early diagnosis42, 43, 44 and possibly treatment of the disease. Hormone replacement therapy and other female hormones may also have played a role in the fall of colorectal cancer in women.41, 45, 46
Trends in liver cancer mortality are extremely difficult to understand and interpret, since the liver is one of the most common sites of secondaries, and the distinction between primary and metastatic liver cancer is variable across countries and calendar periods. Almost certainly, for instance, the changing trends in Bulgaria, the Czech Republic, Spain and Sweden reflect changes in primary liver cancer diagnosis and certification more than real changes in mortality. Still, on an EU scale, the consistent trends in all-age and truncated 35–64 year rates, but the diverging trends in the 2 sexes (upward in men, downward, or approximately stable, in women), probably reflect some real underlying trend, which has to be related to the main recognized causes of liver cancer, i.e., hepatitis viruses and alcohol.47, 48, 49, 50 In principle, these are avoidable causes by reducing alcohol drinking and extending coverage of hepatitis B virus (HBV) vaccination.23, 37, 51, 52 Tobacco is another recognized risk factor for liver cancer,53, 54 but its potential impact on national mortality rates remains difficult to evaluate. Likewise, the substantial geographic variations between the highest mortality rates in Romania, the Czech Republic, Hungary and France (8/100,000 males), Italy, Spain and Switzerland (6/100,000) and the lowest ones in England and Wales, Greece and Norway (1/100,000) are, at least in part, real. In women, the highest rates (over 3/100,000) were in Romania, Hungary and other Eastern European countries, again pointing to the role of hepatitis and alcohol, but also the lack of accuracy in liver cancer certification and distinction between primary and metastatic liver cancer in these areas of the continent.
Gallbladder and bile ducts
Mortality from cancer of the gallbladder and bile ducts is higher in women than in men, but rates have been declining in EU women since the 1970s, the fall being larger over most recent calendar years. Overall male rates in the EU have remained stable, reflecting the upward or stable trends in some countries (France, Italy and Spain), together with the falls in Scandinavian countries, Austria and Germany (i.e., some of the highest mortality areas), but also in the United Kingdom. Rates were historically high in the Czech Republic, Hungary and Poland as well, but have been declining over the last 2 decades, mostly in women. These trends may have been partly influenced by changes in diagnosis and certification, nutrition and diet, alcohol and perhaps tobacco consumption.55 The recent favorable trends in gallbladder cancer mortality, however, essentially reflect the recent rise of cholecystectomy in subjects with gallstones due to improvements in surgical techniques, including laparoscopic cholecystectomy, since having gallstones is the best-recognized factor for the disease.56, 57 In the late 1990s, the highest female mortality rates were registered in the Czech Republic, Hungary, Poland and Slovakia (4–6/100,000), and male ones in the Czech Republic and Hungary (around 3.5/100,000). The lowest mortality rates (0.4–0.7/100,000 in both sexes) were in the United Kingdom.
Mortality from cancer of the pancreas rose for both sexes in the EU up to the late 1980s, but has tended to level off thereafter. This overall pattern conceals some favorable trends, mainly for males, in the United Kingdom, the Netherlands and Sweden since the 1980s, and some rises in France, Germany and Italy, at least up to the most recent calendar years. Rates were persistently upward in Spain, Greece, Bulgaria and Romania, but tended to level off in the Czech Republic, Hungary and Poland, though at very high levels. At least part of the earlier trends may be due to improved diagnosis and certification of the disease following the introduction of ultrasound, computerized tomography, endoscopic retrograde cholangiopancreatography and fine-needle aspiration. It is unlikely, however, that these diagnostic improvements have played a major role over more recent calendar periods, in consideration also of the similar pattern in the truncated rates for population aged 35–64, i.e., when diagnosis and certification are known to be more accurate. The similar patterns of trends across Europe, in any case, weigh against a major role of changes in diagnostic accuracy in any single country. These patterns of trends reflect therefore the spread of cigarette smoking across subsequent generations of men and women in various European countries, confirming the relevant role of tobacco in pancreatic carcinogenesis.58 However, changes in rates were smaller for pancreatic cancer than for lung cancer due to its less strong association with tobacco smoking.59, 60 The higher rates in Eastern Europe than in the EU for men but not for women can also be related to the higher smoking rates in these countries, mostly in men.59 Other factors, including mainly nutrition and diet, may also have played some role on these favorable trends.60, 61 Pancreatic cancer risk, in fact, has been related to obesity, low physical activity, high intake of saturated fats and low intake of vegetables and fruit.62, 63, 64, 65 The highest rates for both sexes were registered in the late 1990s in the Czech Republic, Hungary and Iceland (around 10/100,000 men, 6–7/100,000 women), but Latvia, Estonia and Lithuania also had high rates for men.
Laryngeal cancer is much more common in men than in women, and females rates in most European countries (except Hungary, Ireland and the United Kingdom) were unremarkable. For men in the EU, there was a substantial drop in rates since the mid-1980s (from 4.6 to 3.3/100,000 at all ages and from 8.3 to 5.5/100,000 at age 35 to 64). This was due to the fall in originally high mortality countries (France, Italy, Spain), but also Germany and England and Wales (but not Scotland) showed favorable trends over recent calendar years. Trends were favorable for Switzerland, too, whereas laryngeal cancer mortality has substantially increased in the Czech Republic, Hungary, Poland, Romania and Bulgaria. In the Russian Federation, rates have increased until the early to mid-1990s and have tended to level off only in more recent years. This is another neoplasm strongly related to tobacco smoking and alcohol consumption, and its trends were in the same direction, but somewhat more favorable than those of oral, pharyngeal and esophageal cancer, likely reflecting the predominant role of tobacco on laryngeal carcinogenesis.66 In the late 1990s, the highest rates (over 7/100,000 men) were in Belarus, Croatia, Hungary, the Russian Federation, Ukraine and Poland.
Trachea, bronchus and lung
Two figures are given for lung cancer, one including both sexes, and one women only, to provide a clearer picture of trends in women. In men, lung cancer rates have been declining since the late 1980s in the whole EU. The overall fall was 13% (from 52.5 to 45.6/100,000) at all ages, and 18% (from 73.5 to 60.5/100,000) at age 35 to 64. Most countries showed trends toward leveling (e.g., France, Greece, Spain) or falls of male lung cancer mortality trends over recent calendar years, the only exceptions being Portugal and Romania. Poland and mostly Hungary had reached exceedingly high rates (125–150/100,000 at age 35–64) in the early 1990s, but mortality tended to level off in these countries since most recent years, too. Lung cancer rates in the late 1980s were extremely high in the Czech Republic, too, but tended to fall substantially over the last decade. A decrease in lung cancer mortality after 1994 was also observed in the Russian Federation, although its rates remain considerably higher than in the EU. The decline of lung cancer mortality among males in Russia has been attributed to a cohort effect that followed a shortage of cigarettes after World War II.67 The pattern of trends is largely different for women, since steady upward trends were evident in all countries, except Denmark, Iceland, Ireland and the United Kingdom, whose rates were already exceedingly high (25–40/100,000 the truncated ones, 15–25/100,000 the overall rates). Female lung cancer rates were upwards in Hungary and Poland, too, with comparatively high rates in middle-age women (38/100,000 in Hungary). Rates in females have also been declining in the Russian Federation and were lower in the late 1990s than in the early 1980s in Russia.3 Although rising in most other major European countries (e.g., France, Germany, Italy, Spain), overall female lung cancer rates were still below 10/100,000, i.e., substantially lower than in North America.68 This indicates the scope of the intervention in tobacco control to avoid a major lung cancer epidemic in European women in the near future.3, 69 Since lung is the strongest tobacco-related cancer site, these trends essentially reflect smoking habits of subsequent generations of women and men in various European countries.70 Occupational, environmental and dietary factors of lung cancer may also have exerted minor influences on these favorable trends.3, 71
Death certification from (mesothelioma of) the pleura is subject to substantial uncertainties, given the difficulties in diagnosis and certification validity. Certification of pleural cancer substantially changed between the 9th and 10th revision of ICD,13, 14 since the latter revision introduced a specific code (C45) for mesothelioma. Consequently, rates over most recent years for countries that adopted the 10th revision are not directly comparable with those of previous years. It is known, moreover, that at least part of the fall registered in the United Kingdom since the early 1990s is due to changed classification and certification criteria.72 These cautions notwithstanding, it is clear that the asbestos-related pleural cancer epidemic is not anymore expanding in major European countries, some fall in rates being observed mostly in middle-age men. In the EU as a whole, the decline since the early 1990s for the overall rate was 7% (from 1.13 to 1.05/100,000 men). This suggests that the asbestos-related pleural cancer epidemic is likely to level off in Western Europe in the near future, as did in the United States starting from the 1980s,73 reflecting the control of asbestos use in Europe since the early 1970s, and perhaps the changes in the type of asbestos used in different countries over time.74, 75, 76, 77 Leveling of mesothelioma incidence over recent years have been reported from Sweden76 and Norway.77 In the late 1990s, the highest pleural cancer rates were in Northern Ireland, Scotland, France, Norway and Italy.
Bone is one of the most common sites of secondaries. Hence trends in bone cancer mortality, showing a steady and substantial fall in rates across Europe, partly or largely reflect improvements in validity of death certification over recent calendar periods. The highest bone cancer mortality rates in the late 1990s were in Eastern Europe and Greece, probably reflecting delays in diagnosis and certification improvements in some of these areas of the continent.
Connective and soft tissue sarcomas
As for bone neoplasms, it is not easy to understand and interpret the apparent rise in mortality from connective and soft tissue sarcomas, also considering the rarity of the disease. Such a rise may be related to improvements in certification of these rare neoplasms, but may also reflect real increases in mortality, whose causes, apart from the AIDS-related epidemic of Kaposi's sarcomas, remain difficult to understand.78
Skin, including melanoma
It is not possible, on the basis of death certification, to distinguish between melanoma and nonmelanomatous (mainly squamous cell) skin cancer, although most deaths below age 65 are likely due to melanoma. After the steady rise between 1960 and 1990, overall skin cancer rates in most countries tended to level off since the mid-1990s, and most truncated rates tended to decline, the only exceptions being France, Hungary, Poland and Scotland. In the EU as a whole, the fall in the truncated rates in the 1990s was 6% in men (from 3.4 to 3.2/100,000) and 4% in women (from 2.3 to 2.2/100,000). The highest skin cancer mortality rates were still registered in Scandinavian and Central-Eastern European countries (3–4/100,000 men, over 2/100,000 women), but the overall variation was relatively limited as compared to past decades. It appears therefore that some control of excessive sunshine exposure and sunburns over the last few years has led to a leveling in previous rises in mortality from skin cancers, mainly melanomas and most clearly in the younger and middle-age population.79, 80 Better access to health care, earlier detection of the disease, with a consequent improved survival, could also explain the improvement of mortality trends for CMM.81, 82 Particularly favorable trends have been observed in some northern European countries, where prevention campaigns started earlier.81, 82, 83
The leveling and subsequent decline in breast cancer mortality in Europe, first observed in the early 1990s, has continued and has become more evident in the most recent calendar years. In the whole EU, the fall in the 1990s was 7% (between 21.6 and 20.0/100,000) at all ages and 12% (between 42.5 and 38.4/100,000) at age 35–64. The largest falls were observed in the United Kingdom, which had the highest breast cancer mortality in the past.84 However, part of the apparent fall is likely due to changes in coding and registration procedures in the mid-1990s in England and Wales, which led to a rise in deaths attributed to unspecified neoplasms and a consequent systematic fall in cause-specific rates.85 Also Switzerland, Sweden and Italy showed substantial declines. These widespread falls in breast cancer mortality are likely attributable to advancements in therapy, but some role of (mammography) screening is also possible.86 Rates were still upwards in Greece (15.3/100,000 all age women) and in some Eastern European countries with low breast cancer mortality in the past, such as Bulgaria, or Romania, as well as the Russian Federation (16.7/100,000 women in 1995–1999). As a consequence, the process of leveling of breast cancer mortality across Europe has continued,87 with most rates in the late 1990s ranging from 16 to 23/100,000. The highest rates were in Denmark, Belgium and the Netherlands, and the lowest ones in Eastern Europe and Greece.
Uterus (cervix and corpus)
Since in several countries a substantial proportion of deaths from uterine cancer are attributed to “uterus unspecified” (about 60% in France, Greece and Italy; 40% in Austria and Belgium; 30% in Germany and Spain),88 mortality data for cancer of the uterus do not make it possible to distinguish between cervical and endometrial cancer. Still, the systematic and steady falls observed over the last few decades across Europe, mostly in young and middle-age women, are likely due mainly to a decline in cervical cancer mortality, following the adoption of cervical screening. Most Western European countries have now overall uterine cancer mortality rates around or below 5/100,000, and truncated ones below 10/100,000. However, rates remain appreciably higher (between 7 and 14/100,000 at all ages) in Eastern European countries (Hungary, Poland), and in some of them (Bulgaria, the Czech Republic, Romania, as well as the Russian Federation), there is little evidence of decline over the last few years. Rates in these countries are comparable to those in the EU in the early 1970s. Cervical cancer is caused by human papilloma virus (HPV)89 and may be promoted by hormonal factors, including oral contraceptives.90, 91 It has also been associated to tobacco smoking.92 Death rates from cancer of the (cervix) uteri on a national level are essentially related to the implementation of population-based screening programs. The excess mortality in Central and Eastern European countries points to a substantial delay in the adoption of cervical screening in these areas of the continent and calls for urgent adoption of adequate screening programs to limit the consequences of this avoidable cause of death.88
After earlier rises, overall ovarian cancer mortality has been stable in the EU over the last 3 decades. However, there was a steady decline in truncated (35–64 years) mortality rates, which persisted over most recent calendar periods. Falls in middle-age rates have been observed in all major (Western) European countries, although appreciable differences remain between the truncated rates of around 15/100,000 women in Denmark and The United Kingdom, around 10/100,000 in Germany, Sweden and the Netherlands, and less than 8/100,000 in Italy and Spain. Oral contraceptives have a long-term favorable effect on ovarian cancer risk.93 The recent declines in ovarian cancer mortality for middle-age women should therefore be related to the changing prevalence of oral contraceptive use in subsequent generations of women from various countries. The role of other risk factors for ovarian cancer, as well as of changed diagnosis and treatment (for germ cell neoplasms) and consequently of survival,94 remains unquantified. The lowest ovarian cancer rates were in Mediterranean countries and Portugal, with rates of 3.6–5 (per 100,000) as opposed to 9.5 in Denmark, 8.7 in Lithuania, 8.5 in Ireland and 8.0 in the United Kingdom.
Mortality from prostate cancer has long been stable below age 65, but has been rising in the elderly (with a consequent rise in the overall age-standardized rates) until the early 1990s in most European countries. A subsequent fall in rates has been observed in most European countries, and consequently in the overall rate in the EU (from 15.5 in 1995 to 14.1/100,000 in 1999; −9%). Some persisting rise was observed in Greece, Spain and most Eastern European countries, leading to a leveling of mortality rates across Europe for prostate cancer, too. Appreciable rises in prostate cancer mortality have been observed in Russia over the last 2 decades, although Russia overall age-adjusted rates (7.7/100,000) remain appreciably lower than in most other European countries. At age 35–64, however, Russian rates (6.1/100,000) were comparable to those of Norway and other Nordic countries. An important role of changed diagnosis and certification, mainly in the elderly, remains therefore a likely explanation for at least part of the patterns and trends observed. An even larger decrease than that in the EU was observed in U.S. whites since 1992.95, 96 Since no appreciable trends have been observed below age 65, it remains unclear whether (and if so how much of) the earlier rises in prostate cancer rates were due to improved diagnosis and certification of prostate cancer, and whether some of the recent falls have been favorably influenced by earlier diagnosis of the disease, including introduction of prostate-specific antigen (PSA) in Europe since the late 1980s or early 1990s.97, 98, 99 Given the appreciable decline in rates in the elderly, improved treatment is also likely to have played a role, including earlier adoption of transurethral resection of the prostate (TURP), as well as hormone (androgen blockage) and radiotherapy for patients with locally advanced disease.100, 101, 102 In the late 1990s, the highest prostate cancer mortality rates (over 17/100,000) were in Scandinavian countries, but also in Belgium, the Netherlands, Portugal and Hungary, and the lowest ones (under 10/100,000) in Greece and several Eastern European countries.
Following the adoption of platinum-based therapy in the 1970s, testicular cancer has become one of the neoplasms most amenable to treatment.103 Despite appreciable rises in incidence, consequently, mortality from testicular cancer has more than halved over the last 3 decades in the EU, and the falls have steadily continued over the last few years, indicating that therapeutic progress is still ongoing. Over the last few years, mortality from testicular cancer has started to decline in Hungary and Poland, too, although rates remain substantially higher in these countries than in most of Western Europe. Further, little evidence of decline was evident in Bulgaria or Romania, again indicating the importance and urgency of adopting modern treatments for testicular cancer in these areas of the continent, too.103 Thus, over the last few years, the highest rates (over 0.6/100,000) were systematically observed in Central and Eastern European countries, whereas rates were below 0.3/100,000 in the United Kingdom and Southern Europe.
Bladder cancer is another neoplasm strongly related to tobacco smoking, aside from occupational exposure to aromatic amines, mainly in the past.104 Consequently, its mortality rates, with recent declines for males across Western Europe, well reflect the patterns of tobacco smoking for subsequent generations of European men. Moreover, a variable role in the falls has probably been played by reduced occupational exposure to carcinogens, too.104 Male bladder cancer rates have been rising up to the early 1990s in the Czech Republic, Hungary and Poland, leveling off only more recently, whereas they are still rising in Bulgaria and Romania. In the late 1990s, the highest bladder cancer rates were in Denmark, Spain and Poland (8–9/100,000), and the lowest ones in Finland and Sweden (around 4/100,000). Bladder cancer in women is rarer, and its rates show no clear pattern of trends over time. As for lung cancer, the highest rates (around or over 2/100,000) were observed in Scotland, other U.K. areas and Denmark, i.e., where tobacco smoking has long been widespread in women.
Kidney and other urinary organs
Kidney cancer is also related to tobacco smoking, though less strongly than respiratory or bladder cancers.105 It is also related to overweight and hypertension.106 After earlier rises, mostly in men, kidney cancer mortality rates have been leveling off and somewhat declining in middle-age men over the last few years most of Western Europe, while rates are still upwards in Poland, Hungary and Bulgaria. These trends may therefore be related to an earlier control of smoking and hypertension in Western Europe. Over the most recent calendar period, the highest kidney cancer rates were around 10/100,000 in the Czech Republic and 8/100,000 in Estonia, Lithuania and Latvia, and the lowest ones (between 2 and 3/100,000) in Romania, Portugal, Greece and Spain. Although absolute rates were low, the geographic pattern was similar in women.
Although very difficult to interpret, due to the rarity of the disease, mortality rates from eye cancer tended steadily to decline over the last few decades. Incidence of ocular melanoma has also declined in the United States between 1974 and 1998, probably following better eye protection from UV radiation and improvements in the treatment of these neoplasms.107 As with skin cancer, the highest eye cancer rates were in Northern Europe.
Mortality from thyroid cancer, which is more common in women than in men, tended to decline over the last 2 decades, although it remains unclear how much of the decline is due to better certification of benign thyroid nodules, which are the best recognized risk factor for thyroid cancer,108 and how much is attributable to earlier diagnosis and improved treatment of thyroid neoplasms. The falls tended to be larger in women. The range of variation of female mortality rates in the late 1990s was between 0.6–0.9 in Hungary, Latvia, Lithuania and Slovakia and 0.3/100,000 in the United Kingdom. In males, the highest rates (around 0.5/100,000) were in Austria, Hungary, Lithuania and Germany, i.e., for both sexes high rates were in areas with iodine deficiency in the past.109
Hodgkin's disease is another neoplasm highly amenable to treatment, and consequently its mortality rates have substantially and steadily declined in Western Europe during the last few decades. The persistence of the favorable trends over the most recent calendar years indicates that continuing progress in integrated treatment is still in course.110 As for other neoplasms amenable to treatment, no similar trends were observed in Eastern Europe, some falls being evident only over the last few years, particularly in the Czech Republic. Consequently, Eastern European countries have now appreciably higher rates than most Western countries (higher or around 1/100,000 men, over 0.5/100,000 women). Thus, over 1,000 potentially avoidable deaths per year from Hodgkin's disease, largely at younger ages, were still taking place in Central and Eastern Europe in the late 1990s, again indicating the importance and urgency of improving cancer treatment and control in these areas of the continent.110
In contrast with rates for Hodgkin's disease, mortality from non-Hodgkin's lymphomas has been increasing across Europe during the last 4 decades, these being the major group of cancers showing steady upward trends until the most recent calendar years.111, 112, 113 Between the early and the late 1990s, the rise in the EU has been 5% for men (between 4.2 and 4.4/100,000) and 8% for women (between 2.6 and 2.8/100,000). Non-Hodgkin's lymphomas are much less amenable to treatment than Hodgkin's disease, but the steady rises registered almost certainly reflect real increases in disease incidence, although the causes of these neoplasms remain largely undefined. The subsequent phases of the AIDS epidemic, the spread of hepatitis C virus (HCV)114 and the increased number of transplantations can explain only a limited proportion of the observed rises. In the late 1990s, the highest mortality rates were in Finland, Ireland, Norway and the United Kingdom (over 5/100,000 men, over 3/100,000 women), and the lowest ones in Eastern Europe and Greece (below 2/100,000 in both sexes).
Mortality from myeloma has tended steadily to rise over the last few decades across Europe. Since the upward trends were larger in the elderly and were not observed in some originally high mortality countries, such as Sweden, Denmark, The Netherlands and Norway, these upward trends were at least in part attributed to advancements in the diagnosis of the disease.115 Whether and to what extent the unfavorable trends in mortality also reflect real rises in the incidence remain unclear. In any case, a systematic tendency has been observed toward leveling of rates around high values (3/100,000 men, 2/100,000 women) over recent calendar periods, although mortality rates from myeloma remain relatively low in countries like Austria, France, Greece, Italy and Portugal, as well as most of Eastern Europe. The highest rates in both sexes were still registered in Scandinavian countries and the United Kingdom.
Leukemias include a heterogeneous group of neoplasms.116 Some of them, such as acute lymphoblastic leukemias of childhood, are amenable to treatment117 and consequently contribute to the overall favorable trends in leukemia mortality across Europe. The fall in the EU over the last calendar period has been 5% in men (from 5.6 to 5.3/100,000) and 6% in women (from 3.5 to 3.3/100,000). The falls were earlier and steeper in Nordic countries, started only in the 1970s in Italy, in the 1980s in Spain, and, as with other cancers amenable to treatment, are less evident in Eastern Europe as well as in the Russian Federation. Consequently, the highest leukemia rates in the late 1990s were in Hungary and other eastern countries (over 6/100,000 men, over 4/100,000 women), and the lowest ones in Scandinavian countries, the United Kingdom and Switzerland. This again points to the inadequate adoption of modern integrated therapy schemes for cancers amenable to treatment in several Central and Eastern European countries.116
All neoplasms (benign and malignant)
Overall cancer mortality declined by 7% for both sexes (from 181.2 to 168.5/100,000 men, from 103.2 to 96.2/100,000 women) in the EU over the last 5-year calendar period. The fall since the late 1980s was about 10% in both sexes, corresponding to the avoidance of over 90,000 deaths per year in the EU, as compared to the rates of the late 1980s.2 This fall was of comparable magnitude to that observed in the United States since 1992,118 although somewhat larger for females in the EU. For the first time, over the last few years, some leveling of mortality was registered also in Southern Europe, including Spain and Portugal, as well as the Russian Federation, Poland, Hungary and other Eastern European countries, although cancer rates in these areas remain exceedingly high.
The overall favorable pattern of cancer mortality over recent years is largely driven by the decline of tobacco-related cancer mortality in men. However, significant components of the trends are the persistence in substantial fall in gastric cancer, the recent decline in intestinal cancer in both sexes and of breast cancer in women, together with the long-term falls in uterine cancer, leukemias, Hodgkin's disease and other neoplasms amenable to advancements in diagnosis and treatment. Over recent decades, falls in mortality from stomach, uterine cancer, lung (in males), prostate, female breast, colorectum, pancreas, leukemias and ovary were observed in the United States, too.119
However, there remains an approximately 2-fold variation in male cancer mortality across Europe, between the rates of 270 (per 100,000) in Hungary and 230 in Croatia, and those of 137 in Finland and 124 in Sweden. Likewise, the range of variation is approximately 2-fold in females, between 138–140/100,000 in Denmark and Hungary and 77–78/100,000 in Greece and Spain (but also 84/100,000 in Finland). Thus, countries with lowest cancer mortality rates in Europe for males are those that earlier integrated effective measures to control tobacco smoking. In both sexes, cancer mortality is also comparatively low in some southern European countries, pointing to a favorable role or effect of some components of the so-called Mediterranean diet on cancer rates.120, 121, 122
The maintenance, and potential improvement, of favorable trends in cancer mortality in the near future would therefore require an integrated strategy focusing on control of tobacco, alcohol abuse23 and other major risk factors, including avoidance of obesity,123, 124, 125, 126 taking up of physical activity,127 favorable changes in diet (e.g., increasing daily intake and variety of vegetables and fruit120, 128, 129), avoiding excessive sun and UV exposure.52, 80 Early diagnosis for selected neoplasms can also have a relevant impact, including cervix for women from 25 years of age, breast for women from 50 years of age and perhaps colorectum for men and women from 50 years of age.52 Further, adoption of therapeutic advancements on a population scale in various areas of the continent may appreciably contribute to reduce cancer mortality burden, particularly in Central and Eastern Europe.130, 131, 132, 133
- 7Trends of cancer mortality in Europe, 1955–1989: I, digestive sites; II, respiratory tract, bone, connective and soft tissue sarcomas, and skin; III, breast and genital sites; IV, urinary tract, eye, brain and nerves, and thyroid; V, lymphohaemopoietic and all cancers. Eur J Cancer 1992; 28: 132–235; 514–99; 927–98; 1210–81; 1509–81., , , , , .
- 10World Health Organization. Statistical information system: WHO mortality database. Geneva: World Health Organization, 2003.
- 11World Health Organization. International classification of disease: 7th rev. Geneva: World Health Organization, 1957.
- 12World Health Organization. International classification of disease: 8th rev. Geneva: World Health Organization, 1967.
- 13World Health Organization. International classification of disease: 9th rev. Geneva: World Health Organization, 1977.
- 14World Health Organization. International statistical classification of disease and related health problems: 10th rev. Geneva: World Health Organization, 1992.
- 15Comparison between registries: age-standardized rates. In: WaterhouseJAH, MuirCS, ShanmugaratnamK, PowellJ, PeachamD, WhelanS, eds. Cancer incidence in five continents. vol. 4. Lyon: IARC, 1982. 671–5., .
- 22IARC. Monographs on the evaluation of carcinogenic risk to humans. Alcohol drinking. vol. 44. Lyon: IARC, 1988.
- 34World Cancer Research Fund. Food, nutrition, and the prevention of cancer: a global perspective. Washington, DC: American Institute for Cancer Research, 1997.
- 55Biliary tract cancer. In: SchottenfeldD, FraumeniJJr, eds. Cancer epidemiology and prevention. New York: Oxford University Press, 1996. 794–812., , , .
- 59World Health Organization. Tobacco or health: a global status report. Geneva: World Health Organization, 1997.
- 60Pancreatic cancer. In: SchottenfeldD, FraumeniJJr, eds. Cancer epidemiology and prevention. New York: Oxford University Press, 1996. 725–71., , .
- 80Pigmentary traits, modalities of sun reaction, history of sunburns, and melanocytic nevi as risk factors for cutaneous malignant melanoma in the Italian population: results of a collaborative case-control study. Cancer 2000; 88: 2703–10., , , , .
- 85Office for National Statistics. Mortality statistics: review of the registrar general of deaths by cause, sex and age in England and Wales, 1993 (revised) and 1994. London: HMSO, 1996.
- 97Prostate cancer: creation and control of an epidemic. J Epidemiol Biostat 1996; 1: 65–7..
- 105Renal cancer. In: SchottenfeldD, FraumeniJJr, eds. Cancer epidemiology and prevention. New York: Oxford University Press, 1996. 1142–55., , ,
- 109Thyroid cancer. In: SchottenfeldD, FraumeniJJr, eds. Cancer epidemiology and prevention. New York: Oxford University Press, 1996. 1000–21..
- 114B-cell non-Hodgkin lymphoma and hepatitis C virus infection: a systematic review of 66 studies. Int J Cancer, in press., , , , .
- 117Acute lymphoblastic leukemia. N Engl J Med 1998; 338: 605–15., .
- 127IARC. Handbook of cancer prevention. Weight control and physical activity. vol. 6. Lyon: IARC, 2002.
- 129IARC. Handbook of cancer prevention. Fruit and vegetables. vol. 8. Lyon: IARC, 2003.
|suppmat_figures_levi.zip||3616K||Supporting Information file suppmat_figures_levi.zip|
|suppmat_levi.xls||107K||Supporting Information file suppmat_levi.xls|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.