The relationship between pancreatic cancer and coffee, decaffeinated coffee and tea drinking habits was evaluated using data from a hospital-based case-control study conducted in Northern Italy on 150 histologically confirmed cases and 605 controls with acute, non-neoplastic, other than digestive tract diseases unrelated to coffee consumption or to any of the known or potential risk factors for cancer of the pancreas. Compared with subjects who did not drink coffee, the multivariate relative risks were 1.7 for those who drank less than 2 cups per day, but only 1.4, and 1.1 respectively for drinkers of 3 or 4 and 5 or more cups. Likewise, there was no association with duration of consumption of coffee, decaffeinated coffee or tea. These findings were reviewed together with published evidence from other case-control studies (or cohort studies analyzed as case-control) of coffee and pancreatic cancer. When appropriate statistical methods were used to pool information, and the data from the first study which was the basis of the hypothesis were omitted, the relative risk of pancreatic cancer based on 1,464 cases was 1.2 for moderate coffee drinkers and 1.4 for heavy drinkers, and we suspect that at least part of this moderate residual association is confounded by cigarette smoking. Thus, although the present investigation and a general overview of published epidemiological evidence are compatible with a small effect of coffee on pancreatic carcinogenesis, interpretation of these findings is not obvious on account of the possibility of residual confounding and other sources of bias.

Trends in age-specific and age-standardized death certification rates from all cerebrovascular diseases and various diagnostic subcategories in Italy during the period 1955-78 have been analysed. In both sexes, a decrease in excess of 25% was evident in the overall age-standardized cerebrovascular disease mortality. However, rates were roughly stable in males up to age 50 and in females up to age 45, and slightly but consistently increasing in the younger age groups (under 40), mostly in females. The largest downward trends were for both sexes in the 55 to 74 age groups, and the declines were more marked in females, averaging 3% per year. Since death certification is most reliable in the younger age groups and it is difficult to imagine any modification of risk factors which should affect mortality in later middle age but not in younger age groups, there is no obvious and simple interpretation of this pattern of trends. A comparison with similar trends in ischemic heart disease and other causes of death suggests that the decline in overall cerebrovascular disease mortality might be partially or largely artefactual, though a between-sexes comparison indicates that at least part of the decrease registered in females may well be real. The extent of the decline, however, has been almost certainly more limited in Italy than in most other Western countries. Only in the younger age group (30-34) did rates show a larger increase in females, which might be related to increased prevalence of cigarette smoking, or the use of oral contraceptives.

Abstract Dietary factors in the aetiology of stomach cancer were investigated using data from a case‐control study conducted in Northern Italy on 206 histologically confirmed carcinomas and 474 control subjects in hospital for acute, non‐digestive conditions, unrelated to any of the potential risk factors for giistric cancer. Dietary histories concerned the frequency of consumption per week of 29 selected food Items (including ttie major sources of starches, proteins, fats, fibres, vitamins A and C, nitrates and nitrites in the Italian diet) and subjective stores for condiments and salt intake. Pasta and rice (the rajor sources of starch), polenta (a porridge made of maize) and ham were positively related with gastric cancer risk, whereas green vegetables and fresh fruit as a whole (and specifically citrus fruit) and selected fibre‐rich aliments (such aii whole‐grain bread or pasta) showed protective effects on giistric cancer risk. Allowance for major identified potential distorting factors (chiefly indicators of socio‐economic status) reduced the positive association with pasta or rice consumption, but did not appreciably modify any of the other risk estimates. When a single logistic model was fitted including all food items significant in univariate analysis, the 3 items remaining statistically significant were green vegetables (rel‐acive risk, RR = 0.27 for upper vs. lower tertile), polenta (f;R = 2.32) and ham (RR = 1.60). Indices of beta‐carotene and ascorbate intake were negatively and strongly related with giistric cancer risk, but the association with these micronu‐trients was no longer evident after simultaneous allowance for various food items. An approximately 7‐fold difference in risk was found between extreme quintiles of a scale measuring major positive and negative associations.

This study assesses the association between dietary transfatty acid (TFA) intake and the risk of selected cancers. Mailed questionnaires were completed between 1994 and 1997 in eight Canadian provinces by 1182 incident, histologically confirmed cases of the stomach, 1727 of the colon, 1447 of the rectum, 628 of the pancreas, 3341 of the lung, 2362 of the breast, 442 of the ovary, 1799 of the prostate, 686 of the testis, 1345 of the kidney, 1029 of the bladder, 1009 of the brain, 1666 non-Hodgkin's lymphomas, 1069 leukemias, and 5039 population controls. Information on dietary habits and nutrition intake was obtained using a food frequency questionnaire, which provided data on eating habits 2 years before the study. Odds ratios (OR) and 95% confidenc530e intervals (CI) were derived by unconditional logistic regression to adjust for total energy intake and other potential confounding factors. Dietary TFA were positively associated with the risk of cancers of the colon (OR: 1.38 for the highest vs. the lowest quartile), breast in premenopause (OR: 1.60), and prostate (OR: 1.42). There were a borderline association for pancreas cancer (OR: 1.38; P=0.06). No significant association was observed for cancers of the stomach, rectum, lung, ovary, testis, kidney, bladder, brain, non-Hodgkin's lymphomas, and leukemia, although the ORs for the highest quartile were above unity for all neoplasms considered, except testis. Our findings add evidence that high TFA is associated with an increased risk of various cancers. Thus, a diet low in transfat may play a role in the prevention of several cancers.

Probabilities of cancer death in Italian males and females over the calendar period 1955 to 1980 were computed from age-specific death certification rates for various neoplasms and contemporary general life tables of the whole Italian population. There were substantial increases in eventual probabilities of total cancer mortality: from 16.9 to 27.1% for males and from 15.2 to 19.3% for females. The upward trends were particularly large for lung (from 1.9 to 7.2%) and other tobacco-related sites in males. In females, the largest increases were for neoplasms of the intestines (from 1.6 to 3.0%) and breast (from 1.9 to 3.1%). Separate analyses of probabilities of cancer death in specific age intervals showed more limited changes in middle age, and some moderate decrease at younger ages, probably attributable to improved treatment. The probability estimates presented are clearly based on criticizable assumptions, since they do not allow for any potential subsequent change in mortality and are inappropriate for analyzing the evolution of cancer rates because they reflect (in opposite directions) trends in cancer mortality and in all other causes of death. Nonetheless, a correct interpretation of these estimates does provide some important information from a public health viewpoint, in terms of resource allocation and health care planning, and can help quantify the increasing demand for oncologic services and structures over the next few decades.

The relationship between use of oral contraceptives (OCs) and other contraceptive methods and the risk of ovarian cancer was examined in a combined analysis of 3 hospital-based case-control studies conducted in Italy, the United Kingdom, and Greece, for a total of 971 ovarian cancer cases and 2,258 controls under age 65. Compared with never-users, the combined multivariate relative risk (RR) for ever-users was 0.6 (95% confidence interval, CI = 0.4-0.8) and the estimates were consistent in the 3 datasets. The protection was also similar across strata of age and parity. Considering various measures of OC use, available in the Italian and British datasets only, the protection conveyed on ovarian cancer risk increased with the duration of use and persisted in the medium-long period: the RR in women reporting their last OC use greater than or equal to 15 years prior to diagnosis was 0.5 (95% CI = 0.2-1.0). The risks in ever-users were appreciably lower in those women who reported their first OC use before 25 years of age (RR = 0.3 for first use before age 25, 0.8 for first use at age 25-34 and 0.7 at 35 years or after). Such findings emerged similarly from Italian and British data. This combined analysis, besides offering further quantitative estimates of the protective effects of OCs on ovarian cancer risk in European populations, provides useful insights into the time pattern of the relationship between OC use and ovarian carcinogenesis, suggesting that the protection persists for 15 years or more after cessation of use and may be larger for use at younger age.

We assessed the prevalence of interstitial lung disease (ILD) in a cohort of smokers included in a lung cancer screening trial. Two observers independently reviewed, for the presence of findings consistent with ILD, the computed tomography (CT) examinations of 692 heavy smokers recruited by the Multicentric Italian Lung Detection (MILD) trial. Four CT patterns were considered: usual interstitial pneumonia (UIP), other chronic interstitial pneumonia (OCIP), respiratory bronchiolitis (RB) and indeterminate. Subsequently, the evolution of ILD in those subjects who had undergone a repeat CT examination after 3 yrs was assessed. The UIP pattern and the OCIP pattern were identified in two (0.3%) out of 692 and 26 (3.8%) out of 692 patients, respectively; 109 (15.7%) out of 692 patients showed CT abnormalities consistent with RB, while an indeterminate CT pattern was reported in 21 out of 692 (3%) patients. Age, male sex and current smoking status were factors associated with the presence of OCIP and UIP (combined) pattern, although the relationship did not attain statistical significance. A progression of the disease was observed in three (25%) out of 12 subjects with OCIP who underwent repeat CT after 3 yrs. Thin-section CT features of ILD, probably representing smoking-related ILD, are not uncommon in a lung cancer screening population and should not be overlooked.

Mortality rates for 21 cancer sites in 20 Italian regions have been correlated with several economic and dietary variables (including alcohol and coffee consumption), patterns of cigarette smoking and reproductive habits. In both sexes, a large number of strong correlations emerged, the most notable ones being the strong positive coefficients between cigarettes sold in the early 1950s and lung cancer mortality in middle-aged males in the early 1970s, between gross internal product or meat consumption and cancer of the intestines in both sexes, between total per caput consumption and cancer of the prostate and between mean age at first birth, gross internal product and milk consumption and cancer of the breast. Cancer of the ovary was positively correlated with mean age at the first birth, and negatively with average number of births. Among the unexpected correlations observed, the most remarkable ones were the strong positive coefficient between skin cancer mortality and latitude (which can however be explained in terms of different constitutional characteristics of skin color in various Italian regions), and the pattern of coefficients emerging for gastric cancer, showing positive correlations with gross internal product or meat and negative ones with bread, pasta or fish. These and other results (including the analysis of several first-order partial correlation coefficients) are discussed with regard to their limitations and major points of interest, and in comparison with similar studies conducted on different populations.

Authors’ Response From JOSEPH K MCLAUGHLIN,1,2 LOREN LIPWORTH,1,2 ROBERT E TARONE,1,2 CARLO LA VECCHIA,3,4 WILLIAM J BLOT1,2 and PAOLO BOFFETTA5

We evaluated the risk of breast cancer in relation to the frequency of consumption of a few selected dietary items. Data were used from a case-control study of 1,108 histologically confirmed breast cancer patients and 1,281 control subjects who were in the hospital for acute conditions unrelated to any of the established or suspected risk factors for breast cancer. Moderately elevated risk estimates were associated with higher levels of fat consumption in seasonings [butter, margarine, and oil, relative risk (RR) = 1.34, 95% confidence interval (CI) = 1.06-1.71] and meat (RR = 1.36, 95% CI = 1.12-1.65), whereas a reduced risk (RR = 0.42, 95% CI = 0.34-0.51) was associated with a more frequent green vegetable consumption. It was not possible to show that these associations were incidental, because allowance for several identified potential confounding factors, including the major identified or potential risk factors for breast cancer, did not materially modify the risk estimates. Further, no appreciable interaction emerged with age or menopausal status, because the diet-related risk estimates were similar in pre- or postmenopausal women. However, the implications of these findings in terms of specific micronutrients (e.g., retinol or beta-carotene) and biological correlates are still unclear. Alcohol consumption was significantly greater among breast cancer cases, with a multivariate risk estimate of 2.92 for the highest level. Thus, the present findings confirm that various aspects of diet may influence the risk of breast cancer, although the small amount of available knowledge does introduce serious uncertainties in any discussion of the potential implications in terms of prevention on a public health scale.