Low-Fat Dietary Modification and Risk of Ductal Carcinoma In Situ of the Breast in the Women's Health Initiative Dietary Modification Trial

The role of dietary fat intake in the etiology of breast cancer has been the focus of much investigation (1). While observational studies have shown conflicting results (1), randomized controlled trials have shown no statistical association between reduction in fat intake and overall risk of breast cancer (2, 3).

Ductal carcinoma in situ (DCIS) of the breast, characterized by a proliferation of epithelial cells within the basement membrane of the mammary ductal system, is considered a non-obligate precursor of invasive ductal carcinoma of the breast (4), and shares several risk factors with it (5). Whether a reduction in fat intake in women consuming high levels of dietary fat at baseline can lower the risk of DCIS has not been previously evaluated. In this post hoc analysis, we examined the effect of the Women's Health Initiative (WHI) dietary modification (DM) intervention, focused on reducing fat intake, on the risk of DCIS.

The WHI DM randomized controlled trial has been described in detail previously (2, 6). Enrollment in the trial occurred during 1993–1998 and included 48,835 postmenopausal women ages 50–79 years, with no history of breast cancer, and with ≥32% of total energy intake from fat. The DM group (n = 19,541) was randomized to an intervention designed to reduce fat intake to 20% of total energy and to increase consumption of vegetables and fruits to ≥5 servings/day and of grains to ≥6 servings/day. The comparison group (n = 29,294) received written diet-related education material. Some DM trial participants were also enrolled in other WHI trials: 7% in the estrogen only hormone therapy (HT) trial, 9.6% in the estrogen plus progestin HT trial, and 51.6% in the calcium plus vitamin D trial (6). The primary cancer outcomes of the DM trial were breast and colorectal cancers; DCIS was not a prespecified outcome (6). At each participating center, the study was approved by an Institutional Review Board. Written informed consent was obtained from all participants at the time of enrollment. The DM intervention phase ended after a median follow-up time of 8.7 years. Written consent to post-intervention follow-up was obtained for approximately 82% of surviving participants.

At enrollment, participants provided information on demographic, health, and reproductive characteristics, and medication use. Clinical breast exams and mammograms were protocol-mandated biennially for women enrolled in the DM trial and annually for those also participating in an HT trial. Participants completed medical update questionnaires twice a year during the trial. Self-reported breast biopsies were verified through medical record and pathology report. The DCIS diagnoses were centrally adjudicated (7).

Annual incidence rates per 1,000 person-years were calculated over the entire follow-up period and by phases (intervention and post-intervention). Cox proportional hazards regression models stratified by phase (time-dependent), age at enrollment, CaD trial randomization, HT trial randomization, and prior breast biopsy, were used to estimate HRs and 95% confidence intervals for the risk of DCIS in association with the intervention. Follow-up time was calculated from enrollment date to date of diagnosis of DCIS, invasive breast cancer, death, loss to follow-up, or last documented contact before March 1, 2019, whichever came first. Sensitivity analyses entailed adjustment (separately) for mammogram and clinical breast exam frequency, and exclusion of women with <1 year of follow-up or with a prior breast biopsy. Subgroup analyses were conducted to assess effect modification by baseline risk factors such as age, randomization assignment in the HT and CaD trials, years since menopause, ethnicity, body mass index, Gail breast cancer risk score, physical activity, and percentage of energy from total fat at enrollment (8). Statistical analyses were conducted using Stata 16.1 (Stata Corp). P < 0.05 were considered to be statistically significant.

Demographic and health-related characteristics of the participants were similarly distributed in the two trial arms (2). Over a median period of 18.7 years, 817 cases of DCIS were ascertained, of which 56% were identified during the intervention phase. Table 1 shows that there was no significant association between the dietary intervention and risk of DCIS overall or by trial phase. The results of the sensitivity analyses were similar. The subgroup analyses reported in Fig. 1 show that there was no association between the intervention and DCIS risk over the course of the entire follow-up in any of the subgroups; similar results were obtained by trial phase (intervention and post-intervention; Supplementary Table S1).

In this randomized controlled trial designed to reduce fat and increase vegetable, fruit, and grain intake among postmenopausal women, we did not observe an association between the DM and risk of DCIS overall, in the intervention period (median = ∼8.7 years), or in the post-intervention period (median = ∼13.0 years). To the best of our knowledge, this is the first study to evaluate the association of a low-fat DM with risk of DCIS. As reported previously, the WHI DM intervention also did not significantly reduce the risk of breast cancer (2, 3).

Limitations and strengths of the DM trial have been described previously (2). Of particular note here, we used an intention-to treat approach to test the association and the observed results were based on a large number of participating women followed over an extensive period of time. The WHI DM did not quantify adherence to the study intervention, as individual dietary assessments are not sufficiently reliable for this purpose. Many of the women in the intervention arm did not achieve or maintain the trial dietary goals of reducing total fat intake to 20% of total energy and increasing vegetable, fruit, and whole grain intake, despite a significant reduction of fat intake in this group compared with the control group throughout the trial period (2). Frequency of mammogram screening was similar between the two trial arms, reducing the possibility of differential diagnostic rates. Furthermore, screening rates remained high throughout the trial and post-intervention follow-up (3), thereby increasing the likelihood of detecting DCIS cases which might otherwise have remained undiagnosed. The results of this study were based on a secondary analysis in a trial designed to test the association between the DM and breast cancer, and therefore, they should be interpreted with caution. Nevertheless, our results do not support an association between a low-fat dietary intervention and risk of DCIS among postmenopausal women.

R. Chlebowski reports personal fees from Novartis, AstraZeneca, Puma, Immunomedics, and Genentech during the conduct of the study. J.E. Manson reports grants from NIH during the conduct of the study. A. Barac reports personal fees from Takeda outside the submitted work. No disclosures were reported by the other authors.

R. Peila: Conceptualization, formal analysis, investigation, writing–original draft. R. Chlebowski: Writing–review and editing. J.E. Manson: Writing–review and editing. T.E. Crane: Writing–review and editing. D.S. Lane: Writing–review and editing. N. Saquib: Writing–review and editing. A.H. Shadyab: Writing–review and editing. F.K. Tabung: Writing–review and editing. A. Barac: Writing–review and editing. Z. Zhang: Writing–review and editing. K. Pan: Writing–review and editing. S. Wassertheil-Smoller: Writing–review and editing. T.E. Rohan: Conceptualization, supervision, funding acquisition, writing–review and editing.

T.E. Rohan is supported in part by the Breast Cancer Research Foundation (BCRF-20-140).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.