Analysis of patterns of food intake in nutritional epidemiology: food classification in principal components analysis and the subsequent impact on estimates for endometrial cancer.

OBJECTIVE: To assess the effect of different methods of classifying food use on principal components analysis (PCA)-derived dietary patterns, and the subsequent impact on estimation of cancer risk associated with the different patterns. METHODS: Dietary data were obtained from 232 endometrial cancer cases and 639 controls (Western New York Diet Study) using a 190-item semi-quantitative food-frequency questionnaire. Dietary patterns were generated using PCA and three methods of classifying food use: 168 single foods and beverages; 56 detailed food groups, foods and beverages; and 36 less-detailed groups and single food items. RESULTS: Classification method affected neither the number nor character of the patterns identified. However, total variance explained in food use increased as the detail included in the PCA decreased (approximately 8%, 168 items to approximately 17%, 36 items). Conversely, reduced detail in PCA tended to attenuate the odds ratio (OR) associated with the healthy patterns (OR 0.55, 95% confidence interval (CI) 0.35-0.84 and OR 0.77, 95% CI 0.49-1.20, 168 and 36 items, respectively) but not the high-fat patterns (OR 0.95, 95% CI 0.57-1.58 and OR 0.85, 0.51-1.40, 168 and 36 items, respectively). CONCLUSIONS: Greater detail in food-use information may be desirable in determination of dietary patterns for more precise estimates of disease risk.

Diet in the epidemiology of endometrial cancer in western New York (United States).

OBJECTIVES: We examined diet and risk of endometrial cancer among women in the Western New York Diet Study (1986-1991). METHODS: Self-reported frequency of use of 172 foods and beverages during the 2 years before the interview and other relevant data were collected by detailed interviews from 232 endometrial cancer cases and 639 controls, frequency-matched for age and county of residence. Odds ratios (OR) and 95% confidence intervals (CI) were estimated by unconditional logistic regression, adjusting for age, education, body mass index (BMI), smoking history, hypertension, diabetes, age at menarche, parity, oral contraceptive use, menopausal status, menopausal estrogen use, and energy. RESULTS: Risks were reduced for women in the highest quartiles of intake of protein (OR 0.4, 95% CI: 0.2-0.9), dietary fiber (OR 0.5, 95% CI: 0.3-1.0), phytosterols (OR 0.6, 95% CI: 0.3-1.0), vitamin C (OR 0.5, 95% CI: 0.3-0.8) folate (OR 0.4, 95% CI: 0.2-0.7), alpha-carotene (OR 0.6, 95% CI: 0.4-1.0), beta-carotene (OR 0.4, 95% CI: 0.2-0.6), lycopene (OR 0.6, 95% CI: 0.4-1.0), lutein + zeaxanthin (OR 0.3, 95% CI: 0.2-0.5) and vegetables (OR 0.5, 95% CI: 0.3-0.9), but unrelated to energy (OR 0.9, 95% CI: 0.6-1.5) or fat (OR 1.6, 95% CI: 0.7-3.4). CONCLUSIONS: Our results support previous findings of reduced endometrial cancer risks associated with a diet high in plant foods.

Risk of premenopausal breast cancer in association with occupational exposure to polycyclic aromatic hydrocarbons and benzene.

OBJECTIVES: This study examined the relationship between risk of premenopausal breast cancer and occupational exposure to benzene and polycyclic aromatic hydrocarbons (PAH) and whether the proposed relationship between PAH and breast cancer differed by tumor estrogen receptor (ER) status. METHODS: In a case-referent study of premenopausal breast cancer, occupational histories and other information were obtained through interviews, and job-exposure matrices were used to assess exposure to PAH and benzene. RESULTS: A dose-response relationship for the probability of exposure to benzene [low: odds ratio (OR) 1.64, 95% confidence interval (95% CI) 0.64-4.21; high: OR 1.95, 95% CI 1.14-3.33) and to PAH (low: OR 1.56, 95% CI 0.78-3.12; high: OR 2.40, 95% CI 0.96-6.01). Risk increased with duration of exposure to benzene, but not to PAH. A dose-response relationship was not evident for the intensity of exposure to benzene or to PAH. When analyses were stratified by tumor ER status, PAH exposure was related to a greater increase in the risk of ER-positive (OR 2.27, 95% CI 1.14-4.54) than ER-negative (OR 1.12, 95% CI 0.47-2.64) breast cancer. Risk of ER-positive, but not ER-negative, tumors increased with the probability of exposure to PAH. CONCLUSIONS: The findings suggest an association between risk and occupational exposure to benzene. Although it was difficult to study PAH independently of benzene, there was some suggestion of an association between PAH exposure and ER-positive tumors. These data should be interpreted with caution because of the limitations of this study, including low-response rates and small numbers of exposed persons.

Risk of premenopausal breast cancer and patterns of established breast cancer risk factors among teachers and nurses.

BACKGROUND: Contrasting results have been published regarding the risk of breast cancer among teachers and nurses. Confounding by reproductive factors may explain the increased risk observed among women in these occupations as information on those factors were not available in most studies. METHODS: We examined the risk of premenopausal breast cancer among teachers and nurses using occupational histories in a case-control study where information on established risk factors was available. RESULTS: Having ever held a teaching job was not related to breast cancer (OR = 0.74, 95% CI = 0.44-1.28) and women who worked for 10 years or less in this occupation had a non-significant deficit of risk (OR = 0.52, 95% CI = 0.27-1.02). No elevation in risk was found in association with having ever been a nurse (OR = 0.85, 95% CI = 0.45-1.61) or with duration of nursing. Although direct comparison of established risk factors among teachers and nurses and other women in the study showed some evidence of differential distribution, especially when comparing teachers to other women, adjustment for reproductive variables and other breast cancer risk factors did not change the results of this study. CONCLUSION: These findings suggest that teachers and nurses are not at an increased risk of breast cancer. This study also suggests that established risk factors for premenopausal breast cancer may not explain the elevation of risk found in other studies of teachers and nurses. However, this conclusion is limited by the fact that in the present study teachers and nurses had lower than expected breast cancer risk with or without adjustment for established risk factors. Limitations of this study such as low response rates and limited statistical power should be considered in the interpretation of these findings.

Breast cancer risk and lifetime occupational history: employment in professional and managerial occupations.

OBJECTIVE: In this case-control study, occupational histories were used to assess the relation between risk of breast cancer and employment in professional and managerial occupations while adjusting for reproductive and other risk factors. METHODS: Incident, primary, female cases of breast cancer diagnosed between 1986 and 1991, and randomly selected controls were interviewed to obtain detailed medical, reproductive, and occupational histories. Mantel-Haenszel crude odds ratios (OR) and 95% confidence intervals (95% CIs) were used to estimate risk of breast cancer related to the job of longest duration. Unconditional logistic regression was used to estimate crude and adjusted ORs and 95% CIs associated with having ever been employed and duration of employment in a professional or managerial occupation. RESULTS: A non-significant threefold increase in risk was found among premenopausal women whose major job was in the occupational category of precision production, craft, and repair (95% CI 0.90 to 20.35). No increase in risk was found for premenopausal women whose major job was a managerial or professional occupation. However, an inverse relation between risk of premenopausal breast cancer and having ever held a professional or managerial job was observed (OR 0.53, 95% CI 0.34 to 0.82). This relation was strongest for women who worked one to 10 years (OR 0.47, 95% CI 0.29 to 0.77). Postmenopausal breast cancer was not related to professional and managerial employment. CONCLUSIONS: In this population, employment in professional and managerial occupations is not associated with postmenopausal risk of breast cancer, but seems to be related to a reduction in risk of premenopausal breast cancer. Methodological limitations of this study including response rates are discussed.

Diet and alcohol consumption and lung cancer risk in the New York State Cohort (United States)

The relationship between diet and alcohol and lung cancer was evaluated among participants of the New York State Cohort (United States), comprising 27,544 men (395 cases) and 20,456 women (130 cases) who completed a brief mailed questionnaire in 1980. Participants were followed up through 1987 with the assistance of the New York State Department of Health's Vital Statistics Section and Cancer Registry. Among men, inverse relationships with vitamin C, folate, and carotenoids, and positive associations with total fat, monounsaturated and saturated fat were observed after adjusting for age, education, cigarettes/day, years smoking, and total energy intake. The relationships observed with folate and saturated fat were stronger for heavy smokers. Also, the effect of folate, total fat, and monounsaturated fat seemed to be limited to squamous cell carcinomas. We found no indication that cholesterol or polyunsaturated fat was associated with lung cancer. Diet did not appear to exert a major role on lung cancer risk among women. Although diet modification should never be considered a substitute for smoking cessation, its role as an additional strategy in lung cancer prevention deserves attention.

Cigarette smoking, N-acetyltransferase 2 genetic polymorphisms, and breast cancer risk.

OBJECTIVE: To determine if N-acetyltransferase 2 (NAT2) polymorphisms result in decreased capacity to detoxify carcinogenic aromatic amines in cigarette smoke, thus making some women who smoke more susceptible to breast cancer. DESIGN: Case-control study with genetic analyses. DNA analyses were performed for 3 polymorphisms accounting for 90% to 95% of the slow acetylation phenotype among whites. SETTING AND PARTICIPANTS: White women with incident primary breast cancer (n=304) and community controls (n=327). RESULTS: Neither smoking nor NAT2 status was independently associated with breast cancer risk. There were no clear patterns of increased risk associated with smoking by NAT2 status among premenopausal women. In postmenopausal women, NAT2 strongly modified the association of smoking with risk. For slow acetylators, current smoking and smoking in the distant past increased breast cancer risk in a dose-dependent manner (odds ratios [95% confidence intervals] for the highest quartile of cigarettes smoked 2 and 20 years previously, 4.4 [1.3-14.8] and 3.9 [1.4-10.8], respectively). Among rapid acetylators, smoking was not associated with increased breast cancer risk. CONCLUSIONS: Our results suggest that smoking may be an important risk factor for breast cancer among postmenopausal women who are slow acetylators, demonstrate heterogeneity in response to carcinogenic exposures, and may explain previous inconsistent findings for cigarette smoking as a breast cancer risk factor.

Premenopausal breast cancer risk and intake of vegetables, fruits, and related nutrients.

BACKGROUND: Given the international variations in breast cancer incidence rates and the changes in breast cancer incidence among migrant populations, it has been hypothesized that diet is a factor influencing risk of this disease. Many studies indicate that a diet high in vegetables and fruits may protect against breast cancer. PURPOSE: We conducted a case-control study of diet, including the intake of non-food supplements, and premenopausal breast cancer risk. We evaluated in detail usual intake of vegetables and fruits (each measured as the total reported grams consumed for all queried vegetables and fruit), vitamins C and E, folic acid, individual carotenoids, and dietary fiber with its components. METHODS: Case patients (n=297) were identified through pathology records from hospitals in Erie and Niagara counties in western New York. They consisted of premenopausal women 40 years of age or oder who were diagnosed with breast cancer from November 1986 through April 1991. Control subjects (n=311), frequency-matched to case patients on the basis of age and county of residence, were randomly selected from New York State Department of Motor Vehicles records. In-person interviews included detailed reports of usual diet in the period 2 years before the interview. Unconditional logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs). RESULTS: There was a reduction in risk associated with high intake of several nutrients. With the lowest quartile of intake as the referent, adjusted ORs for the highest quartile of intake for specific nutrients were as follows: vitamin C (OR=0.53; 95% CI=0.33-0.86), alpha-tocopheral (OR=0.55; 95% CI=0.34-0.88), folic acid (OR=0.50; 95% CI=0.31-0.82), alpha-carotene (OR=0.67; 95% CI=0.42-1.08) and beta-carotene (OR=0.46; 95% CI=0.28-0.74), lutein + zeaxanthin (OR=0.47; 95% CI=0.28.0-77), and dietary fiber from vegetables and fruits (OR=0.48; 95% CI=0.30-0.78). No association with risk was found for beta-cryptoxanthin, lycopene, or grain fiber. Fruits were weakly associated with a reduction in risk (fourth quartile OR=0.67; 95% CI=0.42-1.09). No association was found between breast cancer risk and intake of vitamins C and E and folic acid taken as supplements. A strong inverse association between total vegetable intake and risk was observed (fourth quartile OR=0.46; 95% CI=0.28-0.74). This inverse association was found to be independent of vitamin C,alpha-tocopherol, folic acid, dietary fiber, and alpha-carotene. Adjusting for beta-carotene or lutein + zeaxanthin somewhat attenuated the inverse association with vegetable intake. CONCLUSIONS: In this population, intake of vegetables appears to decrease premenopausal breast cancer risk. This effect may be related, in part, to beta-carotene and lutein + zeaxanthin in vegetables. It appears, however, that, of the nutrients and food components examined, no single dietary factor explains the effect. Evaluated components found together in vegetables may have a synergistic effect on breast cancer risk; alternatively, other unmeasured factors in these foods may also influence risk.

Reproducibility of nutrient intake in a food frequency questionnaire used in a general population.

This study evaluates the reproducibility of nutrient intake in a 45-item food frequency questionnaire (FFQ). The FFQ was mailed in 1980 to persons eligible to participate in a large cohort study on diet and cancer risk; a follow-up version with 75 food items was mailed in 1988 to selected original participants. A random sample of 500 men and 500 women from the New York State general population was selected from individuals who responded to both waves of the study. The subjects' 1988 responses were compared with their original 1980 responses; Pearson's correlations ranged from 0.25 for retinol to 0.55 for vitamin C with or without supplements and vitamin E with supplements in women. Reproducibility of nutrient intake in this questionnaire indicates that brief FFQs may be a useful tool to study nutrient intake and chronic disease relationships, although they are subject to substantial measurement error and dietary change.

Cytochrome P4501A1 and glutathione S-transferase (M1) genetic polymorphisms and postmenopausal breast cancer risk.

Polycyclic aromatic hydrocarbons, possible human breast carcinogens, are metabolized by cytochrome P4501A1 (CYP1A1) and glutathione S-transferase (GSTM1). A CYP1A1 polymorphism (isoleucine to valine substitution in exon 7) or the null allele for GSTM1 may affect the mutagenic potential of polycyclic aromatic hydrocarbons. We examined polymorphisms in GSTM1 and CYP1A1 in relation to breast cancer risk. Included were 216 postmenopausal Caucasian women with incident breast cancer and 282 community controls. DNA analyses suggested no increased breast cancer risk with the null GSTM1 genotype [odds ratio (OR) = 1.10; CI, 0.73-1.64], although there was some indication that the null genotype was associated with risk among the youngest postmenopausal women (OR = 2.44; CI, 0.89-6.64). Slightly elevated risk was associated with the CYP1A1 polymorphism (OR = 1.61; CI, 0.94-2.75) and was highest for those who smoked up to 29 pack-years (OR = 5.22; CI, 1.16-23.56). Statistical power to detect an effect may be limited by small numbers, and larger sample sizes would be required to corroborate these suggestive findings.

Alcohol, tobacco, diet and the risk of oral cancer: a pooled analysis of three case-control studies.

This combined analysis of data from three large case-control studies of oral cancer confirms the important effect of tobacco in the aetiology of the disease. The studies have been conducted in the United States, Italy and China and results for risks associated with tobacco smoking were generally consistent across centres, while those for alcohol were not; increased risks amongst alcohol drinkers were evident in two centres but not in the study conducted in Turin, Italy. In addition, the combined analysis had large enough numbers to analyse the risk of tobacco consumption in non-drinkers. In females these showed increased risks while in males the effect of tobacco alone was weaker. Given the popularity of tobacco smoking, and its consequent high attributable risk in terms of oral cancer it is reassuring, in terms of public health, that cessation will result in a substantial reduction in risk; a 30% reduction in risk for those stopping smoking between 1 and 9 years, and a 50% reduction for those stopping more than 9 years. Although encouraging smokers to stop should be the principal aim, decreases in risk for everyone could be achieved by encouraging high fruit and vegetable consumption.

Occupational exposure to extreme temperature and risk of testicular cancer.

A study of 250 patients with pathologically confirmed testicular cancer diagnosed between January 1977 and June 1980 and 250 population controls matched for age and residence was carried out to investigate risk factors for testicular cancer. This paper reports results of risk associated with occupational exposure to extreme (< or = 60 degrees F or > or = 80 degrees F), high (> or = 80 degrees F), and low (< or = 60 degrees F) temperature. Interviews of approximately 1.5 h duration were conducted by trained male interviewers, using a standardized interview schedule. Mantel-Haenszel methods and logistic regression models were employed to estimate these temperature effects on risk of testicular cancer. The estimated adjusted odds ratios of testicular cancer, when 16 potential confounders were controlled for, were 1.71 (95% Cl: 1.13-2.60) for occupational exposure to extreme temperatures; 1.70 (1.04-2.78) for low temperature; and 1.20 (0.80-1.80) for high temperature. The findings suggest that occupational exposure to extreme, low, and high temperature may increase risk of testicular cancer, independent of other potential risk factors.

Reproducibility of food intake in a food frequency questionnaire used in a general population.

This study evaluates the reproducibility of food intake reports in a 45-item self-administered food frequency questionnaire (FFQ) used for a diet study in a large cohort in New York State. The first FFQ was mailed in 1980, and a followup version containing 75 food items was mailed in 1988 to the eligible original participants. Reproducibility was evaluated by comparing the subjects' 1988 responses with their original 1980 responses. Spearman's correlations ranged from 0.27 (nuts other than peanuts) to 0.56 (summer squash) in men and from 0.25 (tomatoes) to 0.51 (summer squash) in women. Analyses stratified by age, education, marital status, body mass index, cigarette smoking, alcohol drinking, and special diet revealed few substantial differences in correlations. The moderate reproducibility of foods in this questionnaire indicates that brief FFQs may be a useful tool to study food intake and chronic disease relationships in this population.

Diet in the epidemiology of postmenopausal breast cancer in the New York State Cohort.

A number of authors have presented evidence that high dietary fat increases the risk of breast cancer, and a number have presented evidence to the contrary. In this study, dietary histories were obtained in 1980 from 18,586 postmenopausal women in New York State. These women were followed through 1987 to ascertain their incidence of breast cancer and other cancers and deaths from all causes, as registered in the New York State Tumor Registry and Office of Vital Statistics. Survival analysis revealed that the incidence of breast cancer increased with age, was higher among the nulliparous, was higher for those with a late (> 26 years) age at first pregnancy, and increased with increasing socioeconomic status--all risk factors discovered before for breast cancer. No increase in risk was related to the ingested amount of calories, vitamins A, C, or E, dietary fiber, or fat. Although dietary fat has been found to be associated with higher risk of cancer at a number of other sites, e.g., the lung, colon, and rectum, and although some previous writers have suggested an association with risk of breast cancer, the findings in three cohort studies as well as in eight substantial case-control studies are negative and suggest that a relation is far from established.

Smoking, alcohol, dentition and diet in the epidemiology of oral cancer.

This matched case-control study was conducted in Western New York. The smoking, alcohol consumption, dental hygiene and diet of 290 cases were compared with those of 290 sex-, age-, and neighbourhood-matched controls. The results confirm earlier findings that cigarette smoking and alcohol consumption impart substantial risk of oral cancer. The results also confirm that poor oral hygiene increases the risk of oral cancer, although this effect is much smaller than those of cigarette smoking and alcohol consumption. The results suggest that, of macronutrients, intake of fat is more likely than those of protein or carbohydrate to be related to risk. Of micronutrients, calcium, sodium, riboflavin and retinol are associated with risk, while thiamin, niacin, and dietary fibre are associated with decreased risk. Although patterns of dietary effects are discernable, these effects are in general much weaker than are those of smoking and alcohol consumption.

Alcohol consumption and lung cancer in white males.

Experimental and epidemiologic investigations in alcoholic and nonalcoholic populations have suggested a role of alcohol in lung carcinogenesis. The association between alcohol consumption and lung cancer was investigated among 280 White males with histologically confirmed, primary lung cancer and 564 White male controls, participants in the Western New York Diet Study (United States). Among heavy smokers (over 40 pack-years), total alcohol consumption was associated with an increased risk of lung cancer with adjustment for age, years of education, pack-years of cigarette smoking, and intake of carotenoids and fat. In this group, the odds ratio for drinkers of more than 24 drinks per month was 1.6 compared with those who drank less. Drinkers of more than 12 beers per month were 1.6 times more likely to develop lung cancer than nondrinkers of beer after controlling for age, years of education, and cigarette smoking (95 percent confidence interval = 1.0-2.4, P for trend = 0.003). Occupational and dietary factors did not seem to explain these findings. Although cigarette smoking is the major cause of lung cancer, the role of alcohol, independent or in interaction with cigarette smoking, deserves further investigation.

Dietary patterns and colon cancer in western New York.

Seven dietary patterns were identified among control subjects in the Western New York Diet Study (1975-1986) by application of principal components analysis to data from a 95-item food frequency interview. The results of case-control analyses of colon cancer risk for these patterns are presented. Cases were matched with neighborhood controls on the bases of age and sex; 205 colon case-control male and 223 female pairs were obtained. The dietary patterns and intakes of energy, total fat, and dietary fiber were examined with logistic regression for their individual contributions to risk. In males, three of these dietary patterns were associated positively with fat and energy consumption; they elevated risk for colon cancer and accounted for more risk than did the specific nutrients. Control for energy and fat intakes allowed the protective influences of additional dietary patterns to be expressed. No patterns elevated risk in women; two patterns were protective for colon cancer. Controlling for energy and fat intake enhanced the protection afforded by one of these patterns but had no influence on that of the other. Measures of foods rather than single nutrients may be more inclusive of dietary exposures to risk as well as being related more directly to underlying health behaviors. Therefore they may be better able to account for risk in diseases with multiple causation.

Use of electric blankets and risk of postmenopausal breast cancer.

Chronic exposure to 60-Hz electromagnetic fields has been hypothesized to increase breast cancer risk by suppressing the normal nocturnal rise in pineal melatonin. From 1987 to 1989 in western New York the authors investigated the use of electric blankets as a risk factor for breast cancer in a case-control study of postmenopausal women aged 41-85 years. A study population of 382 cases and 439 randomly selected community controls was queried regarding use in the previous 10 years, the frequency of use by season, and the mode of use. After adjusting for age and education, the odds ratio (OR) for use of an electric blanket in the past 10 years (33% of cases, 35% of controls) was 0.89 (95% confidence interval (95% CI) 0.66-1.19), and the risk did not differ in a dose-response fashion by the number of years used. The risk associated with daily use relative to nonuse was 0.97 (95% CI 0.70-1.35). Use sometimes to warm the bed and sometimes throughout the night was not associated with risk (OR = 0.64, 95% CI 0.40-1.05). The risk of breast cancer among those who used the blanket continuously throughout the night was 1.31 (95% CI 0.88-1.95). Those who reported daily use in season, continuously throughout the night for the past 10 years did not have significantly elevated risk (OR = 1.25, 95% CI 0.73-2.16). Adjustment for the Quetelet index and reproductive risk factors did not alter the results. These findings do not support the hypothesis that electric blanket use is associated with an increased risk for breast cancer. The slightly elevated estimate of risk for the most frequent electric blanket users and the potential public health significance of electromagnetic field exposure suggest that further inquiries be undertaken.