Case-control study of breast cancer in south east England: nutritional factors.

OBJECTIVE

To explore dietary risk factors, in particular fat intake, for breast cancer, using an approach to reduce recall bias of subjects and so provide a more reliable estimate of dietary intake than previous similar studies.

METHODS

A case-control study of women aged 50-65 years attending the breast assessment clinics of the breast screening programme in Southampton and Portsmouth, southern England. Data were analysed for all women requiring further clinical procedures; all women recalled to have an early rescreen; and a random sample of women found to be normal and referred for a routine rescreening appointment (standard recall).

METHODS

An interview obtained information on various lifestyle characteristics including smoking and alcohol intake, weight, waist, and hip measurements were also taken at the clinic. Women were given a detailed questionnaire on food intake to complete at home and return by post.

RESULTS

1577 women were included in the study: 220 with breast cancer (cases); 179 with benign breast disease; 353 early rescreen and 825 given a standard recall appointment. There were few differences in nutritional intake between the four groups. Logistic regression analyses were carried out comparing the dietary intake of cases with that of each control group adjusting for important demographic and reproductive factors. Results for the case and standard recall comparison are presented. The only non-calorific nutrient to reach significance was iron, which was negatively associated with risk (p = 0.03). For fat intake, the odds decreased with increasing polyunsaturated fat (p = 0.15), showed no trend with monounsaturated fat (p = 0.37) and increased (p = 0.10) with increasing saturated fat. No pattern was clear for the other calorie providing nutrients.

CONCLUSIONS

In line with recent cohort studies, this study has shown no evidence to support the hypothesis that dietary fat is an important contributor to breast cancer rates. Biases should have been reduced by studying subjects from the screening programme who were at an early stage of disease.