Comparative Epidemiology of Cancer Between the United States and Italy1

[CANCER RESEARCH 48. 7285-7293. December 15, 1988] Comparative Epidemiology of Cancer between the United States and Italy1

Carlo La Vecchia,2 Randall E. Harris, and E. L. Wynder3 "Mario Negri" Institute for Pharmacological Research, Via Eritrea, 62, 20157 Milan, Italy [C. L. VJ; Institute of Social and Preventive Medicine, University of Lausanne, 1005 Lausanne, Switzerland [C. L. V]; and the American Health Foundation, Division of Epidemiology, New York, New York 10017 Â¡R.E. H., E. L. W.]

ABSTRACT Among other developed countries, Italy provides one of the best opportunities for comparative studies, since its vital statis Available statistics on smoking, alcohol, food supply, reproductive tics are satisfactorily reliable and show substantial differences history, and other lifestyle habits from the U. S. and Italy were compared from those of several other western countries over recent cal and related to mortality rates of common neoplasms over the period 1955 endar periods (4, 5). Thus, in this paper, we present and discuss to 1980. Per capita cigarette consumption has declined in the U. S. since patterns and trends in lifestyle and site-specific cancer mortality the early 1960s but continues to rise in Italy, chiefly due to the recent increase in cigarette smoking among Italian women. Alcohol consumption in Italy and the U. S. over the last three decades. has increased in both countries, being persistently about 40% higher in Italy. Changes were relatively limited in the American diet, but substan MATERIALS AND METHODS tial for the Italian one which had particularly marked increases in meat, milk, and fat consumption. Fertility rates have declined in both countries Smoking habits in Italy and the U. S. were examined based on sales but more sharply in the U. S. These lifestyle changes were reflected by data (6, 7) and interview-based national surveys (8, 9). Alcohol and distinctly divergent trends in cancer mortality rates between the two food consumption data were derived from disappearance statistics countries. In Italian males, mortality rates of urinary bladder cancer and included in the Food Balance Sheets compiled by the United Nations alcohol-related neoplasms of the aerodigestive tract (oral cavity, larynx, Food and Agricultural Organization (FAO) (10, 11). Cancer mortality and esophagus) increased in a similar manner and were persistently rates, age-standardized to the 1960 world population were obtained for elevated relative to American males. Similarly, Italian lung cancer rates, the period 1955 to 1980 from various sources (4, 6, 12-14). Trends in while starting from lower values, rose steadily to overtake American mortality were computed over separate calendar quinquennia and in rates in the younger and middle age groups of both sexes, and neoplasms two different age groups (35-49 and 50-64 years), since changes in of the intestines, breast, and ovary, starting from considerably lower lifestyle and their implications for cancer risk probably differ at younger values, tended to approach the American rates over the 25-year period and older ages. Since there were large differences in cancer mortality considered. Within Italy, mortality rates of most common neoplasms between various geographical areas in Italy (5), Italian mortality rates were substantially elevated in the North of the country relative to the for the period 1975 to 1977 were also compared in four different areas South, thereby paralleling the distinct North to South gradient in socio- defined on the basis of latitude (North, North Center, South Center, economics, diet, and affluent lifestyle which exists in the country. In our South). opinion, most of these trends are real, and their explanation should be Site-specific cancer mortality rates for Italy were considered satisfac sought, partly or largely, in the changes in tobacco and alcohol use, and torily reliable and comparable with those of the U. S. since similar the reproductive and dietary patterns described. The evidence presented registration techniques have been utilized in both countries throughout underlies the importance of this kind of exercise to formulate and test the period considered, i.e., the same revisions of the International etiological hypotheses of human diseases, which may be overlooked in Classification of Disease (ICD) were employed, death certificates were studies based on populations with more homogeneous lifestyle habits or signed by physicians and reviewed centrally, and the proportion of environmental exposures. unclassified neoplasms or ill defined causes of death under age 65 was small (i.e., about 5% of all cancer deaths and less than 1% of all deaths, respectively, in Italy) (4, 5). Furthermore, internal and external checks INTRODUCTION have shown satisfactory comparability in cancer mortality data for similar Italian populations, and in particular for younger age groups Studies of the geographical variation in cancer mortality wherein death certification is known to be more reliable (15). Moreover, provide a basic framework for formulating and testing etiolog death certification and cancer registration data show high concordance ical hypotheses. Indeed, the concept that most human cancers (over 75% for the third digit of the ICD in Italian Cancer Registries) are due to environmental factors and, hence, are in principle (16). preventable is based on the marked international variation in site-specific cancer incidence and mortality (1). Therefore, com RESULTS AND COMMENTS parative analyses of countries with divergence of cancer rates and lifestyle habits but reliability of vital statistics are clearly Tobacco of major interest. Trends in average numbers of cigarettes sold per adult in the Recent studies of Japan and the U. S. reveal that as the Japanese diet and lifestyle have become more "westernized," two countries between 1920 and 1980 are summarized in Fig. 1. In the U. S., the major increases occurred in the first half of there has been a gradual convergence in the rates of breast, the century, and sales have declined since the mid 1960s (6). In ovarian, corpus uterine, prostate, pancreatic, and colon cancers Italy, steady rises have been observed up to 1980, although with for the two countries (2, 3). a delay of two to three decades compared to American levels. Received 2/17/88; revised 6/30/88, 9/14/88; accepted 9/20/88. Recent Italian data show a plateau in cigarette sales in the The costs of publication of this article were defrayed in part by the payment 1980s (9), thus confirming the approximately 2-decade delay of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. between Italian and American trends. Upward trends in Italian 1This work was supported by the CNR (Italian National Research Council) cigarette sales were particularly pronounced in three decades: Applied Project Oncology (contract 85.00549.58), the Italian League Against Tumors, Milan, Italy, and by Public Health Service Grant CA32617 from the the 1920s and the 1950s (due to increases in cigarette smoking National Cancer Institute and Grant SIG-8 from the American Cancer Society. among men after World War I and II, respectively), and the 2Supported in part by a travel fellowship from the Italian Association for 1970s (due to increased smoking among women) (7). Cancer Research, Milan, Italy. 3To whom requests for reprints should be addressed, at American Health Table 1 gives the proportion of cigarette smoking in the two Foundation, 320 East 43rd Street, New York, NY 10017. countries on the basis of representative sample-based surveys 7285

g Â«ooH in the U. S. in both calendar years, particularly for females. These data, however, may be conservative since information on smoking obtained from interview-based surveys are known to 4000 - U.S.A. include substantial underestimates (up to 30% in comparison

"g 3500 - with sales data). The sales weighted average tar yield of Amer ican cigarettes dropped from 37 mg in 1955 to 14 mg in 1980 (22). In Italy, filter-tipped cigarettes were introduced later and 3000 - -o became popular in the 1960s. Thus, the reduction in tar yield 01 occurred later in Italy with the average value still being 18 mg 3 2500 M of tar per cigarette in 1980 (7). Co Alcohol. On the basis of sales data (10, 11, 23-26), alcohol <â€¢Â»2000 ti consumption has been increasing in both countries over the last o three decades, remaining about 40% higher in Italy than in the o 1500 U. S. (Fig. 4). The profile of alcoholic beverages was also o 01 substantially different in the two countries, beer and spirits U 1000 ^o being the major sources of alcohol in the U. S., whereas wine was by far the most common type of alcoholic beverage in Italy. fl> 500 _0 The estimated intake of alcohol in Italy (over 15 liters of ethanol E per adult per year in 1980) was among the highest in the world. Diet. A summary of national trends in food supply per capita 1920 1930 1940 1950 1960 1970 1980 is given in Table 2. Based on these data, temporal changes in Year American supplies between the late 1940s and the late 1970s Fig. 1. Number of cigarettes consumed per adult per year in the U. S. and were relatively limited, the major modifications being restricted Italy. 1920-1979 (Kristein (6) and La Vecchia (7)J. to decreases in potatoes and starchy foods (â€”37%), fruits (-32%), and vegetables (-18%) and increases in meats exclud Table 1 Estimateli proportion of current cigarette smokers, adults, U. S. and ing fish (+40%), sugar and sweets (+29%). The changes were Italy. 1955-1985 much more marked in Italy where supplies decreased for cereals Sources: Surgeon General, (17, 18); La Vecchia (21). but rose substantially for sugar and sweets (+188%), vegetables Calendar (+51%), fruits (+96%), meat (319%), eggs (+94%), milk year1955 S.Italy24.5 (+53%), and fats or oils (+119%). Although these trends are 6.2 probably partly artefactual and reflect the decreased proportion 1965 51.1 60.353.2 33.3 7.7 of home-produced foods by small farmers in Italy, the overall 1975 41.9 32.0 16.3 1980 40.9 54.3 33.216.727.9 pattern of trends from the two countries is clearly different. 45.0Â°FemalesU. 1985U.S.52.6 33.2MalesItaly65.0 17.7Â° Nevertheless, in more recent calendar periods, Italian supply Â°1983. values were still considerably higher than the U. S. for cereals (348 versus 198 g per day), but substantially lower for meat (7, 9, 17-21). Between 1955 and the mid 1980s, the proportion (176 versus 314 g per day), eggs (31 versus 44 g per day), and of current smokers among males has been persistently about milk (395 versus 476 g per day). 10% higher in Italy, although in both countries an approximate Table 3 gives total fat supply estimates and major fat sources 20% decrease was observed over this time period. The propor in the two countries. In the U. S., changes wee comparatively tion of smokers among American women rose from 25 to 33% small, and restricted to increases in lipids from meat and fats between 1955 and 1965, and remained relatively constant until or oils and decreases in those from milk resulting a 10% increase 1980, then falling to 28%. In Italy, smoking was extremely in total fat. In Italy, upward trends were much more marked uncommon among women in 1955 (6.2% were current smokers) for all the common sources resulting in an 88% increase in but steady increases have followed (reaching 17% in the 1980s). total fat. Data from recent National Health Surveys in the two coun A summary of trends in major fat consumption in Italy tries were used to estimate the prevalence of cigarette smoking between 1951 and 1977 is given in Table 4 (25), and major in successive cohorts after appropriate correction for excess differences in consumption of selected foods between broad mortality (8, 9) and the findings are plotted in Fig. 2. The Italian geographical areas based on National Household surveys results reflect distinct cohort effects in that successive genera (26) are summarized in Table 5. Historical differences still tions show different patterns of starting and stopping smoking. influence the type of fats used in different Italian Regions, In both countries, the highest smoking prevalence for males butter and other animal fats being used in Northern Italy, while occurred in the generations born in the 1920s, but subsequent olive oil is by far the most common seasoning in Southern Italy. declines have occurred later and at lower rates in Italy than in Estimated consumption levels of meat, milk, cheese, eggs, the U. S. Among females, the rise in smoking prevalence sugar, and coffee were greater in Northern regions but those of occurred earlier in the U.S., with a peak rate of about 45% for bread, pasta, fish, and oil were lower than in Southern regions the 1931-1940 generation. In contrast, cigarette smoking was (26). Over the last decade, differences between the North and extremely uncommon among Italian females born prior to the South of Italy have tended to become smaller. 1920; however, later female cohorts show rapid acceleration of Table 6 presents the trends in estimated total calorie intake the smoking habit, and the 1950 cohorts of both Italy and the and percentage of calories attributable to fat, based on disap U. S. have strikingly similar patterns in smoking prevalence pearance statistics. The latter proportionate measure is of par with about one third of females being current smokers in 1980. ticular interest since it is probably less influenced than absolute Fig. 3 compares the average number of cigarettes smoked per values by changes in food distribution and marketing. Between day in the two countries in 1965 and 1980. Values were higher 1955 and 1980, calorie intake increased in both countries (from 7286

70-,

60-

50-

Â¿0 C 4l 30

0> Q. 20

1900 1910 1920 1930 1940 1950 1960 1970 1980 1900 1910 1920 1930 1940 1950 1960 1970 1980 Years Years

70 ; 1920-29 1910-19 60- 1900-09 1890-99 o, 50. O> â€¢¿o o e 4Â°- oÂ»IOC0o*0.70-160-50-40-30-20-10-0-D1890-99, (U u Ã 30- v CL 1950- 59 20-

10- 1960-69 1920- .1960-69 1910-19 1900-09 o 1â€”¿Â» H 1910 1920 1930 1940 1950 1960 1970 1980 1910 1920 1930 1940 1950 1960 1970 1980 Years Years Fig. 2. Changes in the prevalence of cigarette smoking among successive birth cohorts of white U. S. men (A) and women (A), and Italian men (C). and women (O) [Harris (8). LaVecchia el al. (9)].

FEMALES ITALY

10-

Fig. 3. Average number of cigarettes smoked per day by current smokers in the U. S. and Italy [U. S. Office on Smoking and Health (17, 18). Anonymous 1960 1970 1980 1960 1970 1980 (19), Todd (20), La Vecchia (21)]. D, U. S.; â€¢¿.Italy.

Fig. 4. Annual per capita absolute alcohol consumption in the U. S. and Italy 3150 to 3630 in the U. S. and from 2440 to 3230 in Italy). The for persons 18 years and over. [U. S. Bureau of Census (23), 1STAT 1981 (24, percent of total calories attributable to fats rose by about 10% 26)'-D-wine;â€¢¿â€¢beer;â€¢¿â€¢spirits;"'totalalcohoL in the U. S. between the mid 1950s and 1965 levelling off thereafter, whereas in Italy, more marked increases occurred were still present in the last calendar periods considered (i.e., totaling 48% for the entire period. Although most of the total calorie supply and fat proportion were 11 and 21 % lower, changes in the composition of the Italian food supply were in respectively, in Italy, chiefly on account of lower supplies of the direction of bringing the Italian diet closer to that of the meat, milk, and eggs). U. S. (and other western countries), appreciable differences Long-term changes in consumption of selected foods 7287

Table 2 Net food supply per capila: ÃŒ948-1979 Values expressed in grams per day; Â°forU. S. 1968 only; * potatoes and other starchy foods;c for Italy 1977 only [FAO (IO, 11), ISTAT (26)]. 963- 1965178360127127131692226302403241272276985026171567236958521968-1969Â°1783571221271397423263143942233082991295026171867138763571975-1979'19834890110165922225259336201298314176443119314763958259 CerealsPotatoes,

etc.*Sugars

andsweetsPulses,

seedsVegetablesFruitsMeatEggsFishMilkFatsnuts, and

and oils (fat content)U.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.Italy1948-1950210410143105128322427319223293152224425916141164525854271954-1956189369134132126452235307263285190252586221121267829452321960-1962181369131126128632225304348267221261845226171467334356451

Table 3 Nel fai supply per capita: 1948-1969 Values expressed as grams per day [FAO (10. 11)].

CerealsPotatoes,

etc."Pulses,

seedsVegetablesFruitsMeatEggsFishMilkFatsnuts and

andoilsTotal

fatU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.ItalyU.S.Italy1948-19502.55.40.10.15.82.30.60.50.90.545.55.06.21.71.10.924.48.054.327.2141.451.61954-19562.25.10.10.15.32.10.60.50.80.748.67.06.12.21.00.924.7.556.635.1146.063.21960-19622.15.10.20.15.72.20.60.70.80.848.69.95.42.71.21.122.511.056.545.2143.678.91963-19652.24.90.20.16.22.30.70.80.71.050.111.45.22.81.21.221.911.658.552.1146.888.11968-19692.25.00.30.17.32.50.70.80.81.253.814.95.22.71.21.420.512.462.856.1154.897.8

' Potatoes and other starchy foods.

Table 4 Trends in major fat supply in Italy, Â¡951-1977 Table 5 Estimated consumption (in kg or liters per year) of selected food items in different Italian geographical areas, 1973-1985 Calendar year Data from National Household Surveys |ISTAT (26)). Type of fat" 1951 1970 1977 areaFood Year and geographical 5.5 6.0 Butter 4.1 item 13.7 29.0 27.4 Olive oil (per capita Other oils and margarine 15.2 22.5 25.6 kg or " Values expressed as kilograms per capita per year [ISTAT (25)]. liters/year)Breadcenter69.7 center66.7 center62.9

Oil 21.8 24.0 22.6 24.7 22.8 24.9 throughout the current century in Italy are plotted in Fig. 5 in Pasta 32.0 48.3 30.7 46.9 26.3 43.1 comparison to the amount consumed in 1906 to 1910 (27). Meat 56.2 35.6 58.4 45.6 49.9 43.4 Fish 5.8 12.7 6.1 13.0 9.5 16.3 Changes were relatively small in the first half of the century, Milk 84.6 63.0 82.5 66.4 81.6 69.6 but marked increases occurred after the second world war. Cheese 13.8 10.8 13.4 11.7 12.2 11.0 Trends in Cancer Mortality Rates. A summary comparison of Sugar 17.8 13.7 21.4 19.7 17.1 14.0 Wine1973North/120.2South99.775.61980North/109.2South88.373.21985North/80.4South83.860.0 death certification rates from major cancer sites in the two countries in 1978 is shown in Fig. 6 for males and Fig. 7 for females (28). The overall age-standardized cancer mortality was in Italy versus 5.9 in the U. S. for males, 11.3 versus 2.9 for higher in Italy than in the U. S. for males (175.3 versus 159.5/ females). Rates of alcohol-related neoplasms (mouth, pharynx, 100,000), but lower for females (98.1 versus 106.6). Major esophagus, liver and larynx) were higher in Italian males, as differences were registered in the rates of gastric cancer (23.2 were those of bladder cancer in males and uterine cancer (in- 7288

Table 6 Estimated calorie intake and percentage of total calories attributable to FEMALES fat, by year, in the V. S. and Italy, 1955-1980 Sources: U. S. Bureau of Census (23); Mariani-Costantini (27). ALL NEOPLASMS Â»O Calendar year Percent change Â» â€”¿O MOUTH OR PHARYNX 1955 1960 1965 1970 1975 1980 1980/1955 â€¢¿OESOPHAGUS Estimated calorie intake O â€¢¿STOMACH U.S. 3150 3160 3230 3220 3540 3630 +15.2 Italy 2440 2600 2850 2950 3160 3230 +32.4 â€¢¿-0INTESTINES

Percentage of calories attributable to fat U.S. 39% 41% 43% 43% 43% 43% +10.3 Italy 23% 25% 28% 30% 32% 34% +47.8

â€”¿â€¢UTERUS I CERVIX 8. CORPUS i

OVARY

â€¢¿OBLADDER

OÂ« LEUKEMIAS

1 10 100

Fig. 7. Age-standardized death certification rates per 100,000 females for selected cancer sites, U. S. white and Italian, 1978 [Kurihara et al. (28)[. O, U. S. white; Â».Italian.

100 cereols two separate age groups, (35-49 and 50-64 years).

1910 1920 1930 1940 1950 1960 1970 1980 Lung Cancer and Other Aerodigestive Neoplasms. For males, the Italian rates of lung cancer and other aerodigestive neo Year plasms (mouth or pharynx, larynx, and esophagus) have risen Fig. 5. Trends in consumption of selected foodstuffs in Italy compared with the 1906 to 1910 values posed as 100% [Mariani-Costantini (27)]. faster than U. S. rates in both age groups. The rise in male Italian lung cancer mortality was particularly pronounced in the late 1970s, having overtaken American rates for the 1980s. MALES Over the 25-year period considered, lung cancer mortality rates in American males increased only in late middle age whereas ALL NEOPLASMS trends were stable or slightly downward for other tobacco- MOUTH OR PHARYNX related cancers and age groups. For females, lung cancer rates

ESOPHAGUS have increased in both countries and age groups, but the in creases were greater in the U. S. These trends in lung and aerodigestive cancer rates are con â€¢¿-OINTESTINES sistent with the previously described changes in tobacco and alcohol use in the two countries, i.e., the higher prevalence of cigarette smoking and alcohol consumption among Italian men, PANCREAS and the delayed increase in smoking among Italian women. Stomach Cancer. Stomach cancer rates decreased in males and females of both age groups in both countries, although the Italian rates were consistently higher than U. S. rates. It has PROSTATE been suggested that improved food processing and preservation play an important role in the declining stomach cancer trends

â€¢¿OLEUrEMIAS (29). Pancreatic Cancer. In the 1950s and 1960s, Italian mortality

1 10 100 rates of pancreatic cancer were among the lowest in the world Fig. 6. Age-standardized death certification rates per 100,000 males for se (even lower than Japanese rates) (30), but steady and substantial lected cancer sites, U. S. white and Italian, 1978 [Kurihara et al. (28)). O, U. S. upward trends followed in males and females of both age white; â€¢¿,Italian. groups. Since cancer of the pancreas is particularly difficult to diagnose, the rising Italian rates may be heavily influenced by eluding cervix and corpus) in females. Overall standardized improved diagnosis and certification. Nevertheless, we believe lung cancer death certification rates were slightly higher in that the rising Italian trends in pancreatic cancer were partly or American males (53.1 versus 48.7), and markedly higher in largely real, particularly since they were observed in both age American females (15.4 versus 5.5). Other neoplastic sites with groups of each sex. rates higher in the U. S. included intestines and pancreas in Intestinal Cancer. Intestinal cancer mortality rates for 1955- both sexes, prostate, breast, and ovary. 1980 are higher in the U. S. than in Italy; however, the Italian More detailed information is presented in Fig. 8, which shows rates are continuing to rise at a rapid pace, presumptively in trends in age-standardized death certification rates for major response to increasing consumption of dietary fat and decreas cancer sites in the two countries between 1955 and 1980 for ing consumption of dietary fiber in Italy. These trends support 7289

BUCCAL CAVITY & STOMACH.MALES PHARYNX-MALES STOMACH-FEMALES

Â£ S loo

S so

ms 1965 197S 19Â» 1955 1965 I97S 1980

LARYNX-MALES _. ESOPHAGUS-MALES

1965 1975 1980 1955 1965 1975 191 CALENDAR YEAR CALENDAR YEAR

ESOPHAGUS-FEMALES

Â§

'9SS '965 1975 1980 1955 I96S 1975 I98D CALENDAR YEAR CALENDAR YEAR

BREAST-FEMALES UTERUS I All porn I

"" 1965 1975 1980 1955 1965 1975 I9J( CALENDAR YEAR CALENDAR YEAR

INTESTINES-MALES INTESTINES-FEMALES

Fig. 8. Age-standardized mortality trends for selected cancer sites for U. S. whites and Italian, 1955-1980. O, U. S. white 35-49; D, U. S. white 50-64; â€¢¿.Italian35-49; â€¢¿Italian 0 50-64. 8 10

1955 1965 1975 I9SS 1965 I97S 1980 CALENDAR YEAR CALENDAR YEAR

PANCREAS-MALES PANCREAS-FEMALES 1955 I96S 1975 19S SS 1965 1975 1980 CALENDAR YEAR CALENDAR YEAR

LYwPHOMAS-MALES LEUKEMIAS-MALES

1955 1965 197S 11 19SS 1965 1970 198C CALENDAR YEAR CALENDAR YEAR

URINARY BLADDER ;0 URINARY BLADDER 19SS IMS I97S 1*10 I9SS IMS 197S 1910 n MALES 1 FEMALES

LYMPHOMAS-FEMALES ;c LEUKEMIAS-FEMALES S s <>

955 1965 1975 1980 1955 1965 1975 1980 195ÃŒ 1965 1975 1980 1955 196S 1975 1980 CALENDAR YEAR CALENDAR YEAR CALENDAR YEAR CALENDAR YEAR 7290

the thesis that colorectal cancer risk is increased by high intake Table 7 Comparison ofU. S. and Italian fertility rates, 1951-1980 of dietary fat and, conversely, decreased by high intake of Fertility rate/1000 women aged 15-44 dietary fiber. Urinary Bladder Cancer. In the U. S., the mortality rates for cancer of the urinary bladder have declined among older and Calendaryear1951 (whites)107.7 younger age groups of both sexes. In contrast, the rates for 1961 112.0 83.3 Italian males are higher and rising, and the rates for females 1971 77.3 79.3 also appear to be increasing in recent years. This strikingly 1980U.S. 64.7Italy77.9 54.8 divergent pattern of bladder cancer mortality between the U. S. and Italy is very similar to the patterns for tobacco and alcohol- ity rates in the U. S. were about 35% higher than Italy in the related neoplasms (lung, oral cavity, esophagus, and larynx) for 1950s and early 1960s, possibly as a consequence of the post the two countries. Since wine is the predominant alcoholic war "baby boom," but this difference has largely diminished beverage in Italy, the implication is that wine, in addition to due to rapidly declining U. S. birthrates after 1970 (40, 41). tobacco, should be considered as a possible risk factor in cancer The marked reduction in U. S. fertility (40% between 1951 and of the urinary bladder. 1980) and trends toward delayed childbearing (40, 41), in the Uterine Cancer. A downward trend was observed for rates of face of relatively low Italian rates which have remained stable uterine (corpus and cervix) cancer rates in both countries, over time, contrasts sharply with the evolving inverse pattern although Italian rates have remained higher. Uterine cancer of breast cancer mortality in the two countries, i.e., breast cancer trends may be related to changes in sexual habits and genital rates which are stable in the U. S. but increasing steadily in hygiene, and relatively late adoption of effective screening by Italy. Furthermore, in the late 1970s and 1980s, menarche the Papanicolaou test for cervical cancer in Italy (31). occurred earlier (12.7 versus 13.1 years) and menopause later Breast Cancer. Mortality rates of female breast cancer were (51.8 versus 48.9 years) in American women relative to Italians substantially higher in the U. S. than Italy in 1955, but the (42-45). However, the more recent fertility rates and menstrual U. S. rates have remained relatively stable or declined, whereas differences would not be expected to impact on age-adjusted Italian rates have increased markedly to approach American mortality rates of breast cancer for many years and comparisons values in the 1970s. Interestingly, the elevation in American with future rates will be necessary to better evaluate their effects. Mortality occurred predominantly in the postmenopausal years, Ovarian Cancer. Ovarian cancer mortality patterns for the which is clearly shown by the 1980 age-specific ratios plotted U. S. and Italy resemble the breast cancer patterns, suggesting on a logarithmic scale in Fig. 9. that the risk factors for neoplasms of both sites may be similar. Classical breast cancer risk factors include late reproductive In the U. S., rates have declined slightly since the mid-1970s, a history, early menarche, late menopause, and genetic or familial trend which may be, in part, due to better diagnosis and treat factors (32, 33), but there is also supportive evidence that diet ment (3). A protective effect of oral contraceptives has also may play a significant role in breast cancer etiology (34, 35). been suggested, which is consistent with the observation that On an international (36) and intranational scale (37), incidence they were accepted earlier and more widely by American women and mortality rates of breast cancer are positively correlated than by Italians (it is estimated that about one in four American with measures of total calories, fat, meat, and dairy products, women aged 20-49 years used oral contraceptives in 1976 and experimental data on rodents suggests that diets rich in compared to one in 25 Italians) (45, 46). calories and fat (with the possible exception of monounsatu- Other Neoplasms. Differences in the rates of other neoplasms rated fats, which are a major component of olive oil) raise the were modest. American rates of prostate cancer and lymphoma incidence of mammary tumors (38, 39). were slightly higher than Italian rates in the older but not the Table 7 contrasts American and Italian fertility rates during younger age group, and no important secular trends were ap the period 1951-1980. These data show that the average fertil- parent for either type of malignancy. Rates of leukemia were similar and also showed stability over time. 100Ã› Oâ€”OU.S. While Women Cancer Mortality Trends within Italy. Table 8 gives geo â€¢¿â€”â€¢ITALIANWomen graphic-specific standardized cancer mortality ratios (SMR) relative to total Italy (SMR = 100) for selected cancer sites and all sites combined. Notably, the rates for most common sites o are considerably elevated in the North of the country and show cz> o a distinct North to South gradient (5, 15). Furthermore, mi grants from Southern to Northern Italy tend to maintain lower mortality rates for most cancer sites (48). This peculiar distri bution is not readily accounted for by systematic biases in death certification in various areas of Italy, and although Northern Italy is more industrialized, it is unlikely that specific conse quences of industrialization per se are responsible for the pat tern, since cancer mortality is similarly elevated in rural as well as highly industrialized Northern areas (15). It is therefore noteworthy that estimated food consumption patterns within Italy reflect diets which are higher in fat, sugar, and wine, and lower in fiber in the North Central area of Italy compared to 80 the South (Table 5). Also, cigarette sales in the early 1950s were about 40% lower in the South than in the North-Central Fig. 9. Age-specific mortality rate for breast cancer in the U. S. and Italy in area and Southern women had a higher average number of 1980 [Pickle et al. (47), Decarli and La Vecchia (14)]. births (105 versus 68 per 1000 women of fertile age in the 7291

Table 8 Standardized mortality ratios for selected cancers or groups of cancers in broad Italian geographical areas defined according to latitude only From Cislaghi et al. (5). MalesOral

cavityDigestiveRespiratoryBoneSkinGenital

organsUrinaryNervousEndocrineLymphaticLipTonguePharynxEsophagusStomachColon

160169178123124111137149132112124118121127107127102126112106123109125CN89118103111111112106107110108879587981291259111489104150107103112121101120107107102115109107111CS69878089939490997198101696557869097807381639891938392849671971028710085S566170837968868985791105352395554955570706962806860955993848376738167N12211212696111107118105121111961451201471121169713113012711013211998133114124106131115107118109114FemalesCN103118106106981051091119410746919610313312290113941041341091049712010511710989109114114101110CS9091871011109391978394176719269909199808087879187928295879278849286too91S737068977689728490829267{82 . CN T576111359 North)â€”SÃ RS" rectumLiverPancreasLarynxLungPleuraMel. (Centre__^Vgj>^outh)Ã–lVy?rWCS _806761547810648806089858483708572Legend-AS^feSÃ¬;

"iÃ/^Sf/J ^VAÂ¿Ã•

skinBreastUterusOvaryProstateTest SouthCf^-^yC/^$rXVS

isBladderKidneyBrainThyroidHodgkin'sNon-HodgkinM.

myelomaLeukemiasAll

neoplasmsN16312613311211212111210312311199

1980s) and lower average age at first birth (25.6 versus 26.4 Although data limitations may be important, they can hardly years) (49). These differences in tobacco and alcohol use, diet, eclipse, in our view, the major findings emerging from this and fertility appear to parallel the intracountry cancer rates of comparative study. Italians and Americans manifest markedly Italy, and have tended to diminish in recent calendar periods divergent lifestyles that appear to antedate and be associated with a corresponding gradual convergence in site-specific cancer with their substantial differences in mortality from many com rates (50-54). mon neoplasms. Moreover, as lifestyle factors with etiological portent become either more similar or more dissimilar in the two populations, neoplastic rates will respond accordingly. DISCUSSION Empirical comparisons of this kind are useful in formulating This report documents changes and trends in Italian and and scrutinizing hypotheses of the causes of human diseases American patterns of lifestyle which appear to be associated which may be overlooked in studies of populations with rela with divergent trends in site-specific cancer mortality. It is our tively homogeneous lifestyle habits and environmental expo expectation that Italian rates of lung and aerodigestive cancer sures. Through appropriate international studies in metabolic will continue to accelerate in the 1990s, due primarily to epidemiology, we are currently pursuing some of the interesting increased smoking, particularly among recent female cohorts, leads provided by this comparison of Italian and American and high alcohol consumption for which Italy ranks among the populations. highest of any developed nation. Declining rates of these neo plasms in the U. S. are expected to parallel the decreases in ACKNOWLEDGMENTS these same risk factors. It will be of special interest to monitor the diverging rates of urinary bladder cancer for sustained The authors thank Judy Baggott, Antonella Palmiero, and the G. A. increases in Italy and decreases in the U. S. A continuation of Pfeiffer Memorial Library staff for editorial assistance. this pattern would suggest that tobacco and wine should be considered as possible risk factors. Diminishing differences in REFERENCES diet and other lifestyle habits are expected to produce conver gent patterns in rates of other malignancies, particularly neo 1. Doll, R., and Peto, R. The causes of cancer: quantitative estimates of avoidable risks of cancer in United States today. J. Nati. Cancer Inst., 66: plasms of the intestines, pancreas, breast, ovary, and possibly 1191-1308,1981. the prostate. Upward trends in the occurrence of these neo 2. Wynder, E. L.. and Hirayama, T. Comparative epidemiology of cancers of plasms are expected to continue in Italy as the (Southern) the United States and Japan. Prev. Med.. 6: 567-594, 1977. Italian diet becomes more "westernized" with higher fat and 3. Wynder, E. L., and Hiyama, T. Comparative epidemiology of cancer between the United States and Japan: preventive implications. Gann Monograph on lower fiber content. Cancer Res., 33: 183-191, 1987. Int. J. Epidemiol., in press. 7292

4. Decarli. A., and La Vecchia, C. Cancer mortality in Italy, 1955-78. La epithelial and invasive cervical neoplasia. Cancer, 58: 935-941, 1986. mortalitÃ per tumori in Italia, 1955-78. Tumori, 70 (Suppl.).- 579-742,1984. 32. Kelsey, J. L. A review of the epidemiology of human breast cancer. Epidemiol. 5. Cislaghi, C., Decarli, A., La Vecchia, C., Laverda, N., Mezzanotte, G., and Rev., 7:74-109, 1979. Smans. M. Dati, Indicatori e Mappe di MortalitÃ Tumorale. Data, Statistics 33. Franceschi, S., La Vecchia, C., Helmrich, S. P., Mangioni, C., and Tognoni, and Maps on Cancer Mortality, Italia 1975-1977. Bologna: Pitagora Edi G. Risk factors for epithelial ovarian cancer in Italy. Am. J. Epidemiol., 7/5: trice, 1986. 714-719, 1982. 6. Kristein, M. M. 40 years of U. S. cigarette smoking and heart disease and 34. La Vecchia, C., Decarli, A., Franceschi, S., Gentile, A., Negri, E., and cancer mortality rates. J. Chronic. Dis., 37: 317-323, 1984. Parazzini, F. Dietary factors and the risk of breast cancer. Nutr. Cancer, 10: 7. La Vecchia, C. Smoking in Italy, 1949-1983. Prev. Med., 75:274-281,1986. 205-214, 1987. 8. Harris, J. E. Cigarette smoking among successive birth cohorts of men and 35. La Vecchia, C., Decarli, A., Negri, E., Parazzini. F., Gentile, A., Cecchetti, women in the United States during 1900-80. J. Nati. Cancer Inst., 71: 473- G., Fasoli, M., and Franceschi, S. Dietary factors and the risk of epithelial 479, 1983. ovarian cancer. J. Nati. Cancer Inst., 79: 663-669, 1987. 9. La Vecchia, C., Decarli, A., and Pagano, R. Prevalence of cigarette smoking 36. Armstrong, B., and Doll, R. Environmental factors and cancer incidence and among subsequent cohorts of Italian males and females. Prev. Med., 15: mortality in different countries, with special reference to dietary practices. 606-613, 1986. Int. J. Cancer, 15: 617-631, 1975. 10. Food and Agriculture Organization of the United Nations. Production Year 37. Decarli, A., and La Vecchia, C. Environmental factors and cancer mortality Book. 1981. Rome: United Nations Publication, 1982. in Italy: correlational exercise. Oncology, 43: 116-126, 1986. 11. Food and Agriculture Organization of the United Nations. Production Year 38. Carroll, K. K. Experimental evidence of dietary factors and hormone-de Book, 1955. Rome: United Nations Publication, 1956. pendent cancers. Cancer Res., 35: 3374-3383, 1975. 12. McKay, F. W., Hanson, M. R., and Miller, R. W. Cancer Mortality in the 39. Cohen, L. A., Thompson, D. O., Maeura, Y., Choi, K., Blank, M. E., and United States: 1950-1977. MCI Monogr., 59:1982. Rose, D. P. Dietary fat and mammary cancer. I. Promoting effects of different 13. Decarli, A., and La Vecchia, C. Cancer mortality in Italy, 1979. Tumori, 71: dietary fats on A'-nitrosomethylurea-induced rat mammary tumorigenesis. J. 519-528, 1985. Nati. Cancer Inst., 77: 33-42, 1986. 14. Decarli, A., and La Vecchia, C. Cancer mortality in Italy, 1980. Tumori, 72: 40. Istituto Centrale di Statistica. Indagine sulla FeconditÃ della Donna. Note e 231-240, 1986. Relazioni, 50: 1974. 15. Mezzanotte, G., Cislaghi, C., Decarli, A., and La Vecchia, C. Cancer mor 41. White, E. Projected changes in breast cancer incidence due to the trend tality in broad Italian geographical areas, 1975-1977. Tumori, 72:145-152, toward delayed childbearing. Am. J. Public Health, 77:495-497, 1987. 1986. 42. U. S. National Center for Health Statistics. Age at Menarche, United States, 16. /uni-ili. R., ViganÃ²,C., De Molli, S., Colombo, A., and Cislaghi, C. Com Vital and Health Statistics Series 11, 133: 1976. parative completeness and correspondence of cancer mortality data as col 43. La Vecchia, C., Decarli, A., Parazzini, F., Gentile, A., Negri, E., Cecchetti, lected by 1STAT and Cancer Registries. Tumori, 68: 457-463, 1982. G., and Franceschi, S. General epidemiology of breast cancer in northern 17. U. S. Office on Smoking and Health. Smoking and Health. Report of the Italy. Int. J. Epidemiol., 16: 347-355, 1987. Surgeon General. Washington, DC: Government Printing Office, 1979. 44. Kaufman, D. W., Slone, D., Rosenberg, L., Miettinen, O. S., and Shapiro, 18. U. S. Office on Smoking and Health. The Health Consequences of Smoking. S. Cigarette smoking and age at natural menopause. Am. J. Public Health, Cancer and Chronic Lung Disease in the Workplace. Report of the Surgeon 70:420-422, 1980. General. Washington, DC: Government Printing Office, 1985. 45. La Vecchia, C., Decarli, A., Fasoli, M., Franceschi, S., Gentile, A., Negri, 19. Cigarette smoking in the United States, 1986. Morbidity Mortality Weekly E., Parazzini, F., and Tognoni, G. Oral contraceptives and cancers of the Report, 36: 581-585, 1987. breast and of the female genital tract. Interim results from a case-control study. Br. J. Cancer, 54: 311-317, 1986. 20. Todd, G. F. Cigarette consumption per adult of each sex in various countries. J. Epidemici. Commun. Health, 32: 289-293, 1978. 46. Shapiro, S., Slone, D., and Neff, E. R. Age-specific secular changes in oral 21. La Vecchia, C. Patterns of cigarette smoking and trends in lung cancer contraceptive use. Am. J. Epidemiol., 114:604, 1981. mortality in Italy. J. Epidemici. Commun. Health, 39: 157-164, 1985. 47. Pickle, L. W., Mason, T. J., Howard, N., Hoover, R., and Fraumeni, J. F. Jr. Atlas of U. S. Cancer Mortality Among Whites: 1950-1980. U. S. 22. U. S. Office on Smoking and Health. The Health Consequences of Smoking. Washington, D. C.: Department of Health and Human Services, 1987. The Changing Cigarette. Report of the Surgeon General. Washington, DC: 48. Vigotti, M. A., Cislaghi, C., Balzi, D., Giorgi, D., La Vecchia, C., Marchi, Government Printing Office, 1981. M., Decarli, A., and Zanetti, R. Cancer mortality in migrant populations 23. U. S. Bureau of Census. Statistical Abstract of the United States. Washington, within Italy. Tumori, 74: 107-128, 1988. DC: Government Printing Office, 1984. 24. Istituto Centrale di Statistica. Sommario di Statistiche Storiche dell'Italia 49. La Vecchia, C., and Pampallona, S. Age at first birth, dietary practices and breast cancer mortality in various Italian regions. Oncology, 43: 1-6, 1986. 1861-1975. Roma: ISTAT, 1976. 50. La Vecchia, C., and Franceschi, S. Italian lung cancer death rates in young 25. Istituto Centrale di Statistica. Italia 1978. Alcuni Dati sugli Aspetti Demo males. Lancet, 7:406, 1984. grafici, Sociali ed Economici. Roma: ISTAT, 1981. 51. Decarli, A., La Vecchia, C., Cislaghi, C., Mezzanotte, G., and Marubini, E. 26. Istituto Centrale di Statistica. Le Regioni in Cifre. Roma: ISTAT, 1981. Descriptive epidemiology of gastric cancer in Italy. Cancer (Phila.), 58:2560- 27. Mariani-Costantini, A. Dietary trends in western Europe. Prev. Med., 12: 2569, 1986. 218-221, 1983. 52. Cislaghi, C., Decarli, A., Morosini, P., and Puntoni, R. Atlante della Mor 28. Kurihara. M., Aoki, K., and Tominaga, S. (eds.). Cancer Mortality Statistics talitÃ per Tumori in Italia. Triennio 1970-1972. Roma: Lega Italiana per la in the World. Nagoya: University of Nagoya Press, 1984. Lotta Contro i Tumori, 1978. 29. Howson, C. P., Hiyama, T., and Wynder, E. L. The decline in gastric cancer: 53. Facchini, U., Camnasio, M., Cantaboni, A., Decarli, A., and La Vecchia, C. epidemiology of an unplanned triumph. Epidemiol. Rev.,

7293

Carlo La Vecchia, Randall E. Harris and E. L. Wynder

Cancer Res 1988;48:7285-7293.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/48/24_Part_1/7285

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/48/24_Part_1/7285. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.