Background: Although the results of animal studies and cross-cultural comparisons generally support a role for dietary fat in the etiology of breast cancer, results of analytic epidemiology studies are equivocal.

Purpose: The association between dietary fat and subsequent breast cancer was examined in a cohort of 34,388 postmenopausal women from Iowa.

Methods: Dietary habits were assessed by a food-frequency questionnaire mailed in January 1986. Through December 31, 1989, 459 incident cases of breast cancer occurred in this cohort. Proportional hazards regression was used to examine the dietary fat-breast cancer association while adjusting for potential confounders. The effects on this association of four analytic approaches to adjustment for energy intake were also considered.

Results: After adjustment for known determinants of breast cancer, a modest positive association of total fat intake with risk of breast cancer was seen. Polyunsaturated fat intake was also positively associated with breast cancer (relative risk from lowest to highest intake, 1.0, 1.25, 1.31, and 1.49; P for trend = .052). Different approaches to adjustment for energy intake, however, provided different impressions of the dietary fat-breast cancer association. One method, involving categorization of crude fat intake and inclusion of total energy intake in regression analysis, gave relative risk estimates from low to high fat intake of 1.0, 1.17, 1.25, and 1.38 (P for trend = .18). Another method, based on categorization of fat intake residuals in which the variation in fat due to total energy intake was removed, gave corresponding estimates of 1.0, 1.24, 1.30, and 1.16 (P for trend = .29). The former suggests increasing breast cancer risk with increasing fat intake; the latter suggests no association.

Conclusions: These results are consistent with other cohort studies that have shown a weak association or no association between dietary fat and breast cancer. They are also consistent with studies suggesting that fat intake is a determinant of breast cancer, particularly after accounting for inaccuracies in dietary assessment. The effects of different energy-adjustment methods may account in part for the varying interpretations of four previous cohort studies of dietary fat and breast cancer.

Implications: Further work is needed to clarify not only the nature of the dietary fat-breast cancer association, but also the impact of different analytic methods used in the investigation of diet-disease associations.