Epidemiological studies have suggested that aspirin may have a protective effect against prostate cancer, but the evidence is still limited and inconsistent. The role of aspirin in prostate cancer risk was analysed in a multicentric case-control study conducted in Italy between 1991 and 2002, including 1261 incident cases of carcinoma of the prostate and 1131 hospital controls. A total of 115 (9.1%) cases versus 90 (8.0%) controls reported regular aspirin use. The multivariate odds ratio (OR) for regular aspirin users was 1.10 (95% confidence interval (95% CI) 0.81-1.50). No relation was found with duration of use (OR = 1.03 for <5 years, and 1.17 for > or =5 years) and time since first use (OR = 1.03 for <10 years, and 1.35 for > or =10 years). These findings do not support a protective role of regular aspirin use on prostate cancer risk.

To the Editors: In the Health Professional Follow-up Study, Michaud et al. ([1][1]) found a 55% increased risk of bladder cancer for subjects with a history of gastric ulcer. Previous epidemiologic investigations on the issue are limited to patients who had undergone gastric surgery ([2][2], [3][3

The trends in the incidence of various morphological types of cutaneous malignant melanoma (CMM) in the Swiss Cancer Registries of Vaud and Neuchatel (total population 786,000 in 2000) were analysed over the period 1978-2002. Overall, the age-standardized (world) incidence of CMM increased from 5.7/100,000 to 16.8/100,000 for men (+195%) and from 7.9/100,000 to 18.7/100,000 for women (+137%). However, there was no consistent trend for nodular melanoma and other or unspecified morphological types, and the upward trends were observed for lentigo maligna melanoma (from 0.2/100,000 to 1.5/100,000 for men and from 0.5/100,000 to 2.1/100 000 for women) and mainly superficial spreading melanoma [from 2.5/100,000 to 10.2/100,000 for men (+308%) and from 3.3/100,000 to 12.3/100,000 for women (+273%)]. This has implications for aetiology, as the various morphological types of CMM may have different determinants, and for prevention, as the rise in superficial spreading melanomas, which are comparatively thin, may explain the increased survival rates registered for CMM over the last three decades.

Milk has been considered a risk factor for atherosclerosis and coronary heart disease (CHD) because of its content in cholesterol, saturated fatty acids, and possibly lactose. However, epidemiological data are scanty and inconclusive. A Scottish prospective study found a relative risk of 0.92 for CHD in men with higher milk intake, and the preliminary analyses of the Caerphilly study found a protective association.1 No significant relation with CHD was found in other prospective studies, including the British Regional Heart study conducted in men (relative risk, RR, 0.88 for drinkers compared with non-drinkers),2 a British study on vegetarians (increased risk in high milk consumers and reduced in moderate consumers),3 and the Iowa study in postmenopausal women (RR 0.94 for the highest tertile of intake compared with the lowest),4 although absolute numbers of events were limited. An Italian case-control study of acute myocardial infarction (AMI) in women found …

Besides tobacco and alcohol, diet has been thought to be associated with laryngeal cancer risk. We thus analyzed the role of various food groups, as well as specific seasoning fats, in a case-control study conducted in Northern Italy and the Swiss Canton of Vaud from 1992 to 2000. Our study included 527 incident, histologically confirmed cases and 1,297 frequency-matched controls, selected among patients admitted to the same hospitals as cases for acute, nonneoplastic conditions, unrelated to smoking, alcohol consumption and long-term modifications of diet. The subjects' usual diet was investigated through a validated food frequency questionnaire, including 78 foods and beverages. Odds ratios (OR) and 95% confidence intervals (CI) were estimated using unconditional multiple logistic regression models. After adjustment for major confounding factors, a significant trend of increasing risk was observed for eggs (OR = 1.7 for the highest compared to the lowest quintile), red meat (OR = 3.1), processed meat (OR = 1.7), fish (OR = 1.6) and sugars (OR = 1.6). Significant inverse associations were observed for pulses (OR = 0.7), raw vegetables (OR = 0.2), cooked vegetables (OR = 0.3), citrus fruit (OR = 0.6) and other fruit (OR = 0.5). In regard to seasoning fats, a significant reduction of cancer risk was observed for olive oil (OR = 0.4) and specific seed oils (OR = 0.6), while mixed seed oils were directly associated with laryngeal cancer risk (OR = 2.2). Our study suggests that increasing vegetables and fruit, decreasing meat consumption and perhaps substituting olive oil or specific seed oils for other types of seasoning lipids might help reduce laryngeal cancer risk.

A comparative study of trends in colorectal cancer incidence in the Miyagi Prefecture of Japan and Connecticut, USA, from 1959 to 1992 showed substantial rises for both colon and rectal cancer in Japan. For Japanese men born around 1930, values were comparable to—or even higher than—those of Connecticut white men.1 Colorectal cancer incidence, however, is not only due to changes in underlying rates of the disease, but is also influenced by advancements in screening and early diagnosis.2, 3, 4 National mortality rates are less influenced by any such modification and provide therefore less biased information on long-term trends. In a previous study, we considered trends in mortality from colorectal cancer in Japan from 1955 to 1989.5 Overall age-standardized rates increased from 7.0 to 14.6/100,000 in men and from 6.0 to 9.5/100,000 in women. However, a tendency toward a leveling off and a subsequent decline of rates in the younger age groups were observed since the early 1980s.5, 6 From these age-specific trends, we inferred therefore that—in the absence of major differences in aetiologic and biologic correlates of colorectal cancer at young and elderly ages—colorectal cancer mortality in Japan was unlikely to further increase.6 A recent analysis on colon cancer incidence and mortality in Japan from 1950 to 2000 in relation to dietary habits reported increasing trends up to the early 1990s, with a subsequent leveling off thereafter, probably reflecting the changes in dietary factors, including meat consumption, as well as other lifestyle factors in the Japanese population.7 To provide updated information on this issue, we have therefore considered trends in colorectal cancer mortality in Japan over the last 3 decades, using data from the World Health Organization (WHO). Official death certification data for colorectal cancer were abstracted from the WHO database over the period 1970–2000.8 During the calendar period considered, 4 different revisions of the International Classification of Diseases (ICD) were used.9, 10, 11, 12 Thus, classifications of cancer deaths were recoded, for all calendar periods, according to the Ninth Revision of the ICD.11 To improve validity and comparability of data throughout different calendar periods, we considered all intestinal sites combined (ICD IX 152–154 and 159.0). Estimates of the resident population, based on official censuses, were obtained from the same WHO database.8 From the matrices of certified deaths and resident populations, age-specific rates for each 5-year age group and calendar periods were computed. Age-standardized death rates per 100,000 at all ages, truncated 35–64 years and for 3 different age groups (25–44, 45–64, ≥65 years), were computed using the direct method and based on the world standard population13 in order to allow comparison with mortality rates from North America and other countries of the world. The joinpoint regression model was used to describe changes in trends over successive periods of time. This type of nonlinear regression model allows the identification of the best fitting point(s)—called “joinpoint(s)”—where the rates significantly change.14 The estimated annual percent change (EAPC) was then calculated for each of the identified trends by fitting a regression line to the natural logarithm of the rates using calendar year as a regressor variable (i.e., given y=a + bx, where y = ln(rate) and x = calendar year, the EAPC is estimated as 100*(eb−1)). The joinpoint analysis was performed using the Joinpoint software from the Surveillance Research Programme of the U.S. National Cancer Institute.15 Figure 1 shows the trends in age-standardized death rates for cancer of the intestines (mainly colon and rectum) in men and women at all ages and truncated 35–64 years (Fig. 1a) and in 3 different age groups (Fig. 1b–d) across subsequent quinquennia between 1970 and 1999 and year 2000. Overall mortality rates from colorectal cancer increased in Japanese men from 10.20/100,000 in 1970–74 to 17.32/100,000 in 1995–99 but leveled off to 16.32/100,000 in 2000. Death rates in women increased from 8.00/100,000 in 1970–74 to 9.95/100,000 in 1990–94 and stabilized thereafter (9.74/100,000 in 2000). Truncated rates in men aged 35–64 years showed an upward trend up to the mid 1990s (from 13.24/100,000 in 1970–74 to 21.29/100,000 in 1990–94), then leveled off in the late 1990s and declined to a value of 20.67/100,000 in 2000. Similarly, truncated death rates in women increased from 10.93/100,000 in 1970–74 to 13.84/100,000 in 1990–94 and showed a tendency to stabilize or decrease thereafter (12.69/100,000 in 2000). In the younger age group (25–44 years), mortality from colorectal cancer was about 2–3/100,000 in men and 1–2/100,000 in women and tended to decline since the early 1980s in both sexes (Fig. 1a). At age 45–64 years, colorectal cancer mortality started to level off in the early 1990s and to decline since the mid 1990s. At age 65 or more, rates leveled off only since the late 1990s for men, while rates in women have been moderately upward over the last 2 decades. Trends in age-standardized (world population) mortality rates per 100,000 men and women from colorectal cancer, overall and in various age groups, Japan, 1970–2000. (a) All ages and truncated at age 35–64 years; (b) 25–44 years; (c) 45–64 years; (d) ≥65 years. Solid line with crosses, men; solid line with squares, women; dashed line with crosses, men, truncated (35–64 years); dashed line with squares, women, truncated (35–64 years). A further investigation of trends in colorectal cancer mortality for the period 1970–2000 using the joinpoint analysis (Table I) indicated that overall rates significantly increased up to 1996 (with a peak of 17.53/100,000) in men and up to 1992 (with a peak of 10.09/100,000) in women and then declined. Truncated rates in men tended to level off from 1990 to 1996 and showed a downward trend thereafter; in women they started to decline in 1992. Earlier trends were observed in the younger age group, with a significant downward trend since the early 1970s in men and since 1975 in women. This updated analysis of trends in mortality from colorectal cancer in Japan indicates that after a steady rise of rates up to the middle 1990s, death rates tended to level off in the last years in both sexes. The favorable trends in the younger age group, where declining rates were observed since the early 1970s, and the reversal of trends in middle-aged adults since the early 1990s confirm the presence of cohort effects, which are likely to continue in the near future, suggesting a likely further decrease in colorectal cancer mortality in Japan.16 The rapid increase in colorectal mortality in Japan over the past decades has generally been attributed to the introduction of a “western” diet typical of North America and Western Europe, characterized by a high intake of animal fats and meat (particularly red meat) and a low consumption of cereals and fibers.17 Other lifestyle changes related to the economic development and westernization of Japan, such as a decrease in energy expenditure and a consequent increase in body mass index, could also have played a role in earlier unfavorable trends in colorectal cancer.18 The reasons for the decline in colorectal cancer mortality observed over recent years in Japan, as well as in the USA5 and several European countries,19 are complex and not completely understood. The observed trends, particularly in the young Japanese generations, likely reflect favorable modifications in (dietary) risk factors exposure, including a stabilization in the consumption of fats and meat and a wider availability of vegetables.7, 20 An early diagnosis3, 4, 21 and an improved treatment of the disease22 could also have had an effect on the decline in colorectal mortality, although their role in colorectal cancer mortality remains open to debate.23 Diagnosis and certification of cancer of all intestinal sites, including colon and rectum, have long been sufficiently reliable and have not substantially changed in the last 3 decades, thus permitting meaningful inference of trends. Although the distinction between colon and rectum may pose some problems, since a large proportion of cancers arise in the recto-sigmoid junction,13, 24 this should not have affected our estimates based on all intestinal cancers combined. Thus, it is unlikely that changes in diagnostic and certification practices for colorectal cancer can explain the recent decline in mortality trends in Japan. Although recent colorectal mortality rates are higher in Japan than in the USA—where a decline started about 15–20 years before5—they are still considerably lower than in several countries from central and northern Europe.19 Furthermore, the recent reversal of trends in colorectal cancer mortality, particularly in young generations, suggests that they are likely to further decrease in the near future. Yours sincerely, The authors thank Mrs. M.P. Bonifacino for editorial assistance. Cristina Bosetti, Matteo Malvezzi, Liliane Chatenoud, Eva Negri, Fabio Levi, Carlo La Vecchia.

The favourable trends in men are partly or largely due to the recent declines in the prevalence of smoking in European men, together with reduced occupational exposure to occupational carcinogens. The decreases in women are more difficult to explain. Better control of urinary tract infections has probably played a role, while the role of diet and other potential urinary tract carcinogens remains undefined.

Cancers of the upper aero-digestive tract (i.e., oral cavity, pharynx, larynx and oesophagus) are largely attributable to smoking and drinking habits, but the correct estimation of the dose-response relationship between alcohol and cancer risk is challenging. Step functions are widely used to estimate risks and to evaluate trends of continuous exposure. However, results are influenced by the selection of the reference category and cutpoints. More flexible models, like spline regression and fractional polynomial models, may be an attractive alternative for avoiding strict assumptions about the dose-response relationship. Data from a large series of hospital-based case-control studies conducted in Italy and the Swiss Canton of Vaud in the last 2 decades were reassessed to compare findings from logistic regression spline models and standard step function analysis. For all examined cancers, the risk increased to the consumption of 150 grams of ethanol per day (1.5 litre/day of wine), with a possible threshold effect emerging for cancer of the pharynx and larynx (<50 grams of ethanol per day) only. For higher consumptions, the risks flattened. Our study suggests that regression spline models can be useful to estimate the pattern of risk of a continuous exposure variable, such as alcohol consumption, and provide more accurate estimates than categorical analysis when ORs within each interval, especially in the reference category, are not homogeneous.

In our integrated series of case-control studies conducted in Italy and Switzerland (324 oral, 397 pharyngeal, 271 oesophageal, 506 laryngeal cancers and 3,263 controls), individuals who also drank alcoholic beverages outside meals showed an increased risk compared to those who drank at meals only. At any alcohol intake level, subjects also drinking between meals showed a more elevated risk of developing an upper aero-digestive tract cancer than subjects drinking only at meals. After adjustment for potential covariates, and, after allowance for the number of daily drinks to adjust for different alcohol-intake levels, the odds ratios for subjects reporting drinking outside meals were 1.5 (95% confidence interval [CI]: 1.0-2.2) for oral, 1.8 (95% CI: 1.3-2.5) for pharyngeal, 1.7 (95% CI: 1.2-2.5) for oesophageal and 1.2 (95% CI: 0.9-1.7) for laryngeal cancers. Our findings show that drinking pattern with respect to food consumption may influence alcohol carcinogenesis in the upper digestive and respiratory tract. An "alcohol washing effect" by chewing and swallowing is suggested.

The role of coffee in the aetiology of hepatocellular carcinoma has raised great interest. In Italy, coffee consumption is high, thus allowing the investigation of the topic over a broad range of consumption. A hospital-based case-control study was conducted in Italy in 1999-2002, including 185 incidents, histologically confirmed cases of hepatocellular carcinoma aged 43-84 years. Controls were 412 subjects admitted to the same hospitals' networks for acute, non-neoplastic diseases unrelated to diet. Coffee and tea consumption were assessed using a validated food-frequency questionnaire. Odds ratios (ORs) and corresponding the 95% confidence intervals (CI) were computed using unconditional multiple logistic regression, adjusting for hepatitis viruses seropositivity, alcohol intake, smoking habits and other potential confounding factors. Compared to people who drunk <14 cups/week of coffee, the risk of hepatocellular carcinoma decreased for increasing levels of consumption (OR=0.4, 95% CI: 0.2-1.1 for >or=28 cups/week, p for trend = 0.02). In the present study, inverse relations were observed across strata of hepatitis C and, B virus infections and alcohol drinking. No significant association emerged with consumption of decaffeinated coffee (OR=0.7, 95% CI=0.2-2.5) or tea (OR=1.4, 95% CI=0.8-2.7). The present study supports the hypothesis of a favourable effect of coffee, though not decaffeinated coffee and tea, on the risk on hepatocellular carcinoma.

Abstract We analysed the relation between family history of cancer in first‐degree relatives and risk of prostate cancer (PC) and benign prostatic hyperplasia (BPH) using data from a multicentric case‐control study conducted in Italy from 1991 to 2002 on 1,294 cases of incident, histologically confirmed PC, 1,369 cases of BPH and 1,451 men admitted to the same network of hospitals for acute, nonneoplastic conditions. Unconditional logistic regression was used to estimate odds ratios (OR) of PC and BPH, adjusted for age and other confounders. Men with a family history of PC had an OR of PC of 4.0 (95% confidence interval [CI] 2.5–6.5), and the risk was higher when the proband was younger, when 2 or more relatives were affected or when the affected relative was a brother. The risk of PC was also increased in men with a family history of cancer of the ovary (OR = 6.2, 95% CI 1.2–32), bladder (OR = 3.5, 95% CI 1.6–7.4) and kidney (OR = 3.1, 95% CI 1.1–8.5). An involvement of breast/ovarian cancer predisposition genes in a small proportion of PCs was suggested by the cluster of these cancers in female relatives of a few PC cases. The risk of BPH was increased in men with a family history of bladder cancer (OR = 2.2, 95% CI 1.0–5.0) but not PC (OR = 1.2, 95% CI 0.7–2.2). Our study adds further information on the association of family history of cancer and risk of PC and is, to our knowledge, the first comprehensive epidemiologic information on family history of cancer and risk of BPH. © 2004 Wiley‐Liss, Inc.

The pattern of trends in CaP rates observed across Europe is consistent with a favorable role of improved diagnosis, but mainly of advancements of therapy (including more widespread adoption of radical prostatectomy and androgen blockage) on CaP mortality in Western Europe.