Antonella Gentile

Family history and the risk of stomach and colorectal cancer.

Background. The role of a family history of selected neoplasms in first‐degree relatives in the risk of gastrointestinal cancers has been investigated, but requires further quantification.

Nutrition and diet in the etiology of endometrial cancer

The risk of endometrial cancer in relation to nutrition and frequency of consumption of a few selected dietary items was evaluated in a case‐control study of 206 patients with endometrial cancer and 206 control subjects with acute conditions unrelated to any of the established or potential risk factors for endometrial cancer. Obesity was strongly and positively associated with the risk of endometrial cancer, and several conditions related to body weight, such as early menarche, diabetes mellitus, or hypertension were more common in cases. The risk of endometrial cancer was elevated in subjects reporting (on a subjective basis) greater fat (butter, margarine, and oil) intake (relative risk estimate for the higher compared to the lower scores equals 5.65, with 95% confidence interval of 2.76–11.55). Cases reported less frequent intake of green vegetables, fruit, and whole‐grain foods: thus, the risk of endometrial cancer appeared inversely related to indices of beta‐carotene and fiber intake. Furthermore, cases consumed milk, liver and fish less frequently than controls. No significant difference was noted between cases and controls in the frequency of intake of carrots, meat, eggs, ham, and cheese. Alcohol consumption was somewhat larger among the cases, but this trend in risk was not significant. Dietary information collected in this study probably is too limited and inconsistent to permit analysis of biologic correlates of these findings or discussion of their potential implications in terms of prevention on a public health scale. Nonetheless, the mere existence of differences in reported diet between endometrial cancer cases and controls is of interest, and may warrant further, more detailed investigation.

Sexual factors, venereal diseases, and the risk of intraepithelial and invasive cervical neoplasia.

The relation between major indicators of sexual habits (age at first intercourse and total number of sexual partners), history of selected venereal diseases, and cervical neoplasia was investigated using data from a case‐control study of 206 cases of cervical intraepithelial neoplasia compared with 206 age‐matched outpatient controls, and of 327 cases of invasive cancer compared with 327 control subjects in hospital for acute conditions unrelated to any of the established or suspected risk factors for cervical cancer. The relative risks increased with decreasing age at first intercourse and increasing number of sexual partners both for intraepithelial and for invasive cancers. The effects of these two variables were independent, since they were only marginally affected by reciprocal adjustment, or by allowance for several other identified potential distorting factors. The negative association with age at first intercourse was particularly strong in the case of invasive cancers, with risk estimates over five‐fold elevated for women reporting their first intercourse before age 18 compared with those aged over 22 years. This relation might be discussed in terms of multistage models of carcinogenesis, which predict that the incidence of epithelial carcinomas is a function of duration of exposure. In fact, when age was allowed for, the relative risks of cervical neoplasia were positively and strongly related with the total duration of the interval between age at diagnosis/interview and age at first intercourse. Clinical histories of several sexually transmitted diseases were positively associated with the risk of intraepithelial neoplasia. In particular, genital warts were reported by nine cases but no control subject. No such association, however, emerged for invasive carcinomas. Thus, the current findings confirm that, although intraepithelial neoplasia and invasive cervical cancer appear to share several important epidemiological features, the specific (infectious) agents implicated in dysplastic lesions probably differ to some extent from those causing invasive cancer.

Menstrual factors and the risk of epithelial ovarian cancer.

The relationship between menstrual factors and the risk of ovarian cancer was evaluated in the framework of a hospital-based case-control study conducted in the greater Milan area on a total of 634 cases of ovarian cancer and 1626 control subjects with a spectrum of acute conditions unrelated to any of the established or potential risk factors for ovarian cancer. Compared with women whose menarche occurred at age 13 or later, the risk of ovarian cancer was moderately (and not significantly) elevated for earlier menarche (multivariate relative risk, RR = 1.13, 95% confidence interval, CI = 0.97-1.37). There was a positive association with age at menopause, the multivariate relative risk, compared with women aged 45 years or less at menopause, being 1.25 for those aged 45-49 years, 1.40 for 50-53 and 1.58 for 54 or over (chi 2(1) trend = 8.86, p = 0.003). A lifelong irregular menstrual pattern (defined as frequent menstrual-like episodes of bleeding less than 21 or more than 35 days apart) was negatively associated with the risk of ovarian cancer (multivariate RR = 0.45, 95% CI = 0.31-0.65). The effect of age at menopause and lifelong menstrual pattern was similar in different strata of age, whereas there was a positive association with early menarche among younger women which disappeared with advancing age. The present findings confirm the influence of various menstrual factors on the risk of epithelial ovarian cancer.(ABSTRACT TRUNCATED AT 250 WORDS)

Dietary vitamin A and the risk of intraepithelial and invasive cervical neoplasia

The risk of invasive and intraepithelial cervical neoplasia in relation to the frequency of intake of the major sources of preformed vitamin A (retinol) and beta-carotene in the Italian diet was analyzed in a study of 392 cases of invasive cancer compared with 392 age-matched controls hospitalized for acute conditions unrelated to any of the established or suspected risk factors for cervical cancer, and of 247 cases of cervical intraepithelial neoplasia compared with 247 controls found to have normal smears at the same screening clinics where cases had been identified. Women with invasive cancer consumed milk, green vegetables, and carrots less frequently, but no significant relation was noted for meat or liver. Consequently, estimated beta-carotene, but not retinol, intake was inversely and strongly related to the risk of invasive cervical cancer. Compared with women whose intake was over 150,000 international units (IU) per month, the relative risks were 3.0 for 100 to 149,999 and 4.7 for less than 100,000 IU. It was not possible to show that these relationships were incidental, since allowance for several identified potential distorting factors, including indicators of social status and the major risk factors for cervical cancer, did not materially modify the risk estimates. In contrast, no association emerged between any of the food items and vitamin A estimates considered and intraepithelial neoplasia. Thus, the results of this study can be interpreted in one of two ways: either some residual uncontrolled bias was responsible for the strong dietary correlates of invasive cervical cancer risk or beta-carotene (or any other correlate of a vegetable-rich diet) has effect on one of the later stages of the process of carcinogenesis, thus influencing the risk of invasive carcinoma but not of its precursors.

Alcohol and breast cancer: Update from an Italian case-control study

The analysis of alcohol and breast cancer risk from an Italian case-control study of breast cancer has been updated to include a total of 2402 women with histologically confirmed breast cancer and 2220 controls with acute conditions unrelated to any of the established or potential risk factors for breast cancer, admitted to a network of teaching and general hospitals in the greater Milan area. Compared with non-drinkers, the estimated relative risk for ever drinkers was 1.4 (95% CI = 1.2-1.6). The multivariate risks were 1.3 for less than 1 drink (e.g. 10 g of ethanol) per day, 1.3 for 1 less than 2 drinks, 1.4 for 2-3 and 2.2 for over 3 drinks per day. The positive association between alcohol and breast cancer was consistent across strata of age, socio-demographic variables, smoking, menstrual, reproductive, and hormonal factors, family history of breast cancer, nutrition and diet indicators. In particular, this study indicated that the alcohol-breast cancer relationship is probably not materially different in younger and older women, that alcohol drinking at an early age is not a particularly clear indicator of subsequent risk, nor that the effect of alcohol is enhanced in thin women. Thus, the findings of this study are both internally consistent and in agreement with most previous evidence, since the relative risks of 1.3-1.4 for a consumption of 10-30 g of alcohol per day compare well with the corresponding highest intake levels in most American studies. This study is of particular interest, since it provides data on higher levels of alcohol consumption, on which very little information is available. There are nontheless at least three elements which leave open the question of causality: the absence of duration-risk relationship, the inconsistency with evidence from descriptive epidemiology, and the lack of plausible biological mechanisms.

Anthropometric indicators of endometrial cancer risk

The relation between selected anthropometric indicators and the risk of endometrial cancer was evaluated using data from a case-control study conducted in Northern Italy on 562 cases and 1780 controls in hospital for acute, non-neoplastic or hormone-related disease. There was no appreciable association between height and endometrial cancer: compared with the lowest quintile, the multivariate relative risks (RR) were 0.9, 0.9, 0.7 and 0.8 for each subsequent quintile. Weight was directly associated with risk (RR = 2.7 for top vs. bottom quintile), and the positive association was even stronger when indices of body mass which make allowance for height were considered: the relative risks for extreme quintiles were 3.4 for W/H2 (Quetelets index, weight and height), 3.8 for W/H1.5 and 3.5 for W0.33/H. Surface area, which was positively correlated both with height and weight, showed a weaker direct association (RR = 2.4 between extreme quintiles). The relations with measures of body weight were apparently stronger in postmenopause, but the point estimates for the upper quintile were also around 2 in premenopausal women. Although the major findings of this study are not new, they provide more detailed information that was hitherto available on the relation between various anthropometric indicators and endometrial cancer risk. In relation to height, with the sample size of this study it was possible to exclude, at the conventional 95% probability, relative risks above 1.0 for the fourth and above 1.1 for the fifth as compared with the lowest quintile. This provides indirect evidence against the hypothesis that nutritional status early in life is related to the subsequent development of endometrial cancer.

Alcohol and epithelial ovarian cancer

The relationship between alcohol consumption and the risk of epithelial ovarian cancer was analysed using data from a case-control study of 801 histologically confirmed epithelial ovarian cancers and 2114 controls in hospital for acute, non-neoplastic, gynecological, or hormone-related conditions, admitted to a network of teaching and general hospitals in the greater Milan area, northern Italy, i.e. a region with comparatively frequent alcohol consumption by women. Compared to alcohol abstainers, the multivariate relative risks (RRs) were 1.0 [95% confidence interval (CI), 0.7 to 1.4] for less than one, 1.1. (95% CI 0.9 to 1.6) for one to two, 1.2 (95% CI 1.0 to 1.5) for two to three and 1.3 (95% CI 0.9 to 1.8) for three or more drinks per day. A significant direct trend in risk with dose emerged. This finding chiefly derived from an association between ovarian cancer risk and consumption of wine (which accounts for over 90% of alcohol intake in this female population). Although no significant interaction between the effect of alcohol consumption and various womens characteristics emerged, there was a hint that the adverse influence of alcohol consumption is more marked in middle-age and less educated women. Thus, the results of this study suggest that relatively elevated alcohol intake (of the order of 40 g per day or more) may cause a modest increase of epithelial ovarian cancer risk.

Determinants of oral contraceptive use in northern Italy

Socio-demographic characteristics, general lifestyle habits, reproductive and medical histories were compared of 328 women who had ever used oral contraceptives and 2306 never users from a case-control surveillance conducted in Northern Italy. Oral contraceptive use was positively and strongly related with the level of education and indicators of social class. The rate ratio of ever use (adjusted for age and diagnostic category) was 3.3 for women with 12 years of education or more compared with less than 7 years, and 3.0 for women of highest compared with lowest social class. The frequency of pill use was lower among never married women, and significantly elevated among smokers (rate ratio = 2.4 for heavy smokers). In contrast, no relation was evident between alcohol or coffee consumption and pill use. Likewise, ever users of oral contraceptives were not significantly different from women who had never used the pill with regard to major reproductive factors (parity and age at first pregnancy) or several medical variables of potential interest (e.g., diabetes, hypertension, hyperlipidemia), with the only possible exclusion of obesity which was less frequent among pill users. Thus, this study indicates that the major determinants of the persistently low frequency of oral contraceptive use in this Northern Italian population are social rather than reproductive or medical factors. These findings have important implications for epidemiological research on oral contraceptive and disease in this population, and underline the importance of selection and screening of oral contraceptive users on the basis of knowledge of medical factors and lifestyle habits.

Breast cancer and combined oral contraceptives: An Italian case-control study

The risk of breast cancer in relation to use of oral contraceptives was evaluated using data from a hospital-based case-control study from Northern Italy on 1517 cases below age 60 and 1351 controls admitted for acute diseases unrelated to any of the known or potential risk factors for breast cancer. The multivariate relative risk for ever vs. never users was 1.3 (95% confidence interval = 1.0-1.7). However, the risk was not related to duration of use: indeed the highest risk was observed among short-term users (less than 2 years), and the point estimate was 0.9 among users for 5 years or more. The elevated risk among short-term users, if not due to residual confounding or selection mechanisms, is probably explainable in terms of recall bias (i.e. more careful report of short or very short use by cases). No definite pattern was observed in relation to latency or recency of use, and the point estimates were 0.8 for women who had ever used the pill before age 25 and 0.8 for those who had ever used the pill before first full-term pregnancy. Thus, the study presents further reassuring information on the oral contraceptive/breast cancer debate. Its major limitation lies in the low prevalence of oral contraceptive users in Italy, with a consequently reduced statistical power, although, with the number of cases involved, it was possible to exclude a relative risk of 1.4 for long-term use or for ever use before first birth.