Cancer Prev Res 2008 Rohan 275 84

Cancer Prevention Research

Low-Fat Dietary Pattern and Risk of Benign Proliferative Breast Disease:A Randomized, Controlled Dietary Modification Trial

Thomas E. Rohan,1 Abdissa Negassa,1 Bette Caan,2 Rowan T. Chlebowski,3 J. David Curb,4 Mindy Ginsberg,1

Dorothy S. Lane,5 Marian L. Neuhouser,6 James M. Shikany,7 Sylvia Wassertheil-Smoller1 and David L. Page8

Abstract Modifiable factors, including diet, might alter breast cancer risk. We used the Women'sHealth Initiative Dietary Modification trial to test the effect of the intervention on risk of be-nign proliferative breast disease, a condition associated with increased risk of, and consid-ered to be on the pathway to, invasive breast cancer. The Women's Health Initiative DietaryModification trial was a randomized, controlled, primary prevention trial conducted in 40 U.S.clinical centers from 1993 to 2005. A total of 48,835 postmenopausal women, ages 50 to79 years, without prior breast cancer, were enrolled. Participants were randomly assigned tothe dietary modification intervention group or to the comparison group. The intervention wasdesigned to reduce total dietary fat intake to 20% of total energy intake, and to increase fruitand vegetable intake to ≥5 servings/d and intake of grain products to ≥6 servings/d, butresulted in smaller, albeit significant, changes in practice. Participants had biennial mammo-grams and regular clinical breast exams. We identified women who reported breast biopsiesfree of cancer, obtained the histologic sections, and subjected them to standardized centralreview. During follow-up (average, 7.7 years), 570 incident cases of benign proliferativebreast disease were ascertained in the intervention group and 793 in the comparison group.The hazard ratio for the association between dietary modification and benign proliferativebreast disease was 1.09 (95% confidence interval, 0.98-1.23). Risk varied by levels of base-line total vitamin D intake but it varied little by levels of other baseline variables. These resultssuggest that a modest reduction in fat intake and increase in fruit, vegetable, and grainintake do not alter the risk of benign proliferative breast disease.

Breast cancer incidence rates vary 4- to 5-fold internationallyand have been observed to change on migration from low-riskto high-risk countries (1). These observations have led to thehypothesis that modifiable factors, including diet, might influ-ence breast cancer risk. Most of the focus in epidemiologicstudies of the association between diet and breast cancer hasbeen on the roles of dietary fat, fiber, and fruit and vegetable

intake. In general, cohort studies have not supported an asso-ciation between dietary fat intake and breast cancer risk (2–5),although several recent cohort studies have reported smallincreases in risk in postmenopausal women (6–8). Addition-ally, there is little recent evidence from cohort studies thatfruit and vegetable consumption (9, 10) and dietary fiber in-take (11) are associated with altered breast cancer risk,although earlier studies provided some support for inverseassociations (5).

The Women's Health Initiative (WHI) randomized con-trolled Dietary Modification trial was the first large-scale ran-domized trial to test the effect of adoption of a low-fat dietarypattern (decrease in total fat intake, increase in fruit, vegetable,and grain intake) on breast cancer risk (12). After ∼8 years offollow-up, risk in the intervention group was 9% lower thanthat in the comparison group, but the effect was not statisti-cally significant. Among the possible explanations for this lackof effect is that the follow-up period was too short. If that isthe case, it does not preclude the possibility of an effect of theintervention on earlier stages in the natural history of breastcancer in the short term, with a subsequent (and consequent)reduction in breast cancer incidence. Therefore, we used theWHI Dietary Modification trial to test the effect of adoptionof a low-fat dietary pattern on risk of benign proliferativebreast disease, a condition which is associated with increased

Authors' Affiliations: 1Department of Epidemiology and Population Health,Albert Einstein College of Medicine, Bronx, New York; 2Division ofResearch, Kaiser Permanente Northern California, Oakland, California;3Harbor-University of California at Los Angeles Medical Center, Torrance,California; 4Department of Geriatric Medicine, University of Hawaii, Honolulu,Hawaii; 5Department of Preventive Medicine, School of Medicine, StateUniversity of New York at Stony Brook, Stony Brook, New York; 6Public HealthSciences Division, Fred Hutchinson Cancer Research Center, Seattle,Washington; 7Division of Preventive Medicine, School of Medicine, Universityof Alabama, Birmingham, Alabama; and 8Department of Pathology,Vanderbilt University Medical School, Nashville, TennesseeReceived 01/04/2008; revised 02/25/2008; accepted 03/13/2008.

Grant support: NIH grant RO1 CA077290-07.Requests for reprints: Thomas E. Rohan, Department of Epidemiology

and Population Health, Albert Einstein College of Medicine, 1300 Morris ParkAvenue, Bronx, NY 10461. Phone: 718-430-3355; Fax: 718-430-8653; E-mail:[email protected].

©2008 American Association for Cancer Research.doi:10.1158/1940-6207.CAPR-08-0003

275 Cancer Prev Res 2008;1(4) September 2008www.aacrjournals.org

for Cancer Research. on July 15, 2015. © 2008 American Associationcancerpreventionresearch.aacrjournals.org Downloaded from

Published OnlineFirst July 9, 2008; DOI: 10.1158/1940-6207.CAPR-08-0003

http://cancerpreventionresearch.aacrjournals.org/

risk of, and is considered to be on the pathway to, invasivebreast cancer (13–15).

Materials and Methods

Study populationThe WHI Dietary Modification trial has been described in detail

elsewhere (12, 16, 17). In brief, 48,835 postmenopausal women, ages50 to 79 y at initial screening, who were likely to reside in the areafor 3 y, and who provided written informed consent, were enrolledbetween 1993 and 1998 at 40 clinical centers throughout the UnitedStates. Major reasons for ineligibility for the trial included a historyof breast or colorectal cancer, a history of other cancer except nonme-lanoma skin cancer in the last 10 y, medical conditions with predictedsurvival of <3 y, adherence and retention considerations (e.g., alcohol-ism, dementia), or a diet at baseline with fat intake of <32% of totalenergy. All participants had a baseline mammogram and clinicalbreast examination; abnormal findings suggestive of breast cancer re-quired clearance before study entry. Eligible women were randomizedeither to the dietary modification intervention group (40%; n = 19,541)or to the comparison group (60%; n = 29,294) using a permuted blockalgorithm with stratification by clinic and age. Dietary modificationparticipants were also eligible to be in the WHI hormone therapyand calcium/vitamin D supplementation trials (16). The WHI andthe ancillary study reported here (in which all 40 WHI clinical centersparticipated) were approved by institutional review boards at allparticipating institutions.

Dietary intervention and retentionThe dietary intervention was designed to promote dietary change

with the goals of reducing intake of total dietary fat to 20% of totalenergy intake (as a consequence of which, saturated fat intake wasexpected to be reduced to ∼7% of energy intake), increasing fruitand vegetable intake to ≥5 servings/d, and increasing intake of grainproducts to at least 6 servings/d (12). Women in the comparisongroup were not offered a nutrition intervention program becausethe general strategy for this group was minimum interference withcustomary diets while collecting nutritional data considered appropri-ate for comparison with the intervention group. However, they wereprovided with a copy of the dietary guidelines for Americans. Neitherthe intervention nor the comparison group was asked to modifyhealth-related behaviors, including use of dietary supplements.

Dietary change was encouraged by providing social support andpositive interactions. To this end, women randomized to the interven-tion arm were assigned to a permanent group of 8 to 15 members ledby a designated nutritionist. During the first year of follow-up, thegroup met weekly for 6 wk, biweekly for 6 wk, and monthly for9 mo; subsequently, it met quarterly. Each participant had an indivi-dual counseling session with her group nutritionist between 12 and16 wk from the beginning of the intervention sessions.

Baseline data collection, follow-up, and assessment ofadherenceComprehensive information on breast cancer risk factors was

obtained at baseline by interview (for lifetime hormone use) and byself-report (other covariates) using standardized questionnaires (16).

Study participants were contacted every 6 mo for collection of in-formation on outcomes. Clinic visits were required annually, at whichtime height and weight were measured using standardized proce-dures. Mammograms were required every 2 y, and regular clinicalbreast exams were done as well.

Dietary intake for all participants was monitored using the WHIfood frequency questionnaire (12, 18). The food frequency question-naire was administered to all participants at baseline and 1 y afterrandomization. Thereafter, about one third of participants completedthe food frequency questionnaire annually in a rotating sample.

Furthermore, 4-d food records were provided by all women beforerandomization (12), a 4.6% sample provided further 4-d food recordsat 1 y after randomization, as well as 24-h dietary recalls at 3 and 6 ypostrandomization, and 1% samples of women provided a 24-hdietary recall annually. The dietary data presented here were derivedfrom the baseline food frequency questionnaire.

Ascertainment of outcomesThe outcome of interest for the present study was histologically

confirmed incident benign proliferative breast disease with or withoutatypia (see “Histology”). Clinical events including breast cancers andbreast biopsies for noncancerous lesions were initially identified fromself-administered questionnaires completed every 6 mo. Breast cancerswere confirmed by local and central adjudicators who reviewed med-ical records and pathology reports and who were blinded both totreatment assignment and to symptoms due to study interventions.For the present study, women who reported breast biopsies that werefree of cancer were identified, and clinical centers were sent lists ofpotentially eligible subjects quarterly. Clinic staff contacted partici-pants to obtain written informed consent to solicit the histologicsections resulting from the biopsies. To investigate the possibility thatbreast biopsies were missed by using this approach, the charts of100 randomly selected participants who did not report a breastbiopsy were reviewed at one center and none was found to haveunreported biopsies.

HistologyH&E-stained histologic sections were reviewed by the study pathol-

ogist (D.L.P.), who was blinded to the randomization assignment. Thebenign lesions were classified using well-established criteria as non-proliferative lesions, proliferative lesions without atypia (classifiedfurther according to whether they were mild, moderate, or florid inextent), or atypical (ductal/lobular) hyperplasia (19–21).

Statistical analysisIncidence rates of benign proliferative breast disease in the dietary

modification and comparison groups were compared based on theintention-to-treat principle using time-to-event analyses. The primaryanalysis used a weighted (two-sided) log-rank test with weightincreasing linearly from zero at randomization to a maximum of 1at 10 y and constant thereafter to enhance statistical power underthe design assumptions (12). The time to benign proliferative breastdisease was defined as the number of days from the date of randomi-zation to the date of the first postrandomization diagnostic biopsy.Follow-up time was censored at the date of last documented contact,diagnosis of breast cancer, mastectomy, death, or trial close-out (be-tween October 2004 and March 2005), whichever came first. Womenwho developed a nonproliferative benign breast lesion continued to befollowed up because they remained at risk of developing a subsequentproliferative lesion. Event rates over time were summarized usingcumulative hazard plots. The intervention effect was summarizedusing hazard ratios (HR) and 95% confidence intervals (95% CI) esti-mated from Cox proportional hazards models (22), with stratificationby age, prior breast biopsies, and randomization to the WHI hormonetherapy and calcium/vitamin D supplementation trials. Stratificationwas time dependent in the case of the calcium/vitamin D supplemen-tation trial (16). The possibility that the intervention effect differed bylevels of other characteristics of the study population was investigatedby including product terms between treatment assignment and indi-cator variables for the subsets of interest in Cox proportional hazardsmodels stratified by age, prior breast biopsies, and randomization tothe hormone therapy and calcium/vitamin D trials (16), and was as-sessed by testing the equality of the product-term coefficients. Giventhat a total of 22 interactions were tested, approximately one signifi-cant test at an α level of 0.05 was expected by chance alone. The pro-portional hazards assumption, which was tested both by fittingmodels containing a product term between the intervention and


276Cancer Prev Res 2008;1(4) September 2008 www.aacrjournals.org




follow-up time and assessing the coefficient of the product term forstatistical significance, and by fitting piecewise constant interventioneffects on nonoverlapping time intervals and testing for equality ofsuch terms with the intervals chosen in advance, was shown not tobe violated. Annualized event rates were calculated for comparisonsof absolute disease rates. Results were considered statistically signif-icant when two-sided P ≤ 0.05.

Results

The study groups differed little at baseline with respect toage, ethnicity, breast cancer risk factors, participation in otherWHI trials, or intake of energy or selected nutrients and vita-mins (Table 1).

Data on follow-up and adherence, as well as change innutrient and food intake and blood biomarker levels, werereported elsewhere (12). In brief, 4.7% of women in the inter-vention group and 4.0% of those in the comparison groupwithdrew from the study or were lost to follow-up. Estimatedadherence rates were 87% and 75% in the comparison groupand 57% and 31% in the intervention group at years 3 and 6,respectively. Based on estimates from the food frequencyquestionnaires, the intervention and comparison groups wereessentially identical at baseline with respect to intake of fat,fruit and vegetables, and grains (as well as other dietaryitems). One year after randomization, statistically significant,albeit modest, between-group differences in intake of thesedietary items were evident: For the intervention and compar-ison group, the mean (SD) percent energy from fat, servings offruit and vegetables per day, and servings of grains per daywere 24.3 (7.5) versus 35.1 (6.9), 5.1 (2.3) versus 3.9 (2.0), and5.1 (2.7) versus 4.2 (2.3), respectively. By year 6, the mean (SD)difference in change (change in intervention group minuschange in comparison group) in percent energy from fatbetween the intervention and comparison groups was –8.1(7.8), whereas the mean differences in change for servingsper day of fruit and vegetables and of grains were 1.1 (2.1)and 0.4 (2.6), respectively; all of these differences werestatistically significant. Compared with blood levels in thecomparison group, the intervention group showed a greaterreduction in blood levels of γ-tocopherol, a greater increasein β-cryptoxanthin, and smaller decreases in α- and β-carotene.Estradiol levels decreased more and sex hormone-bindinglevels increased more in the intervention group than inthe control group.

During follow-up (average duration, 7.7 years), we identi-fied 3,383 potentially eligible biopsies that had been donefor benign breast disease. The eligibility of 65 biopsies couldnot be determined due to lack of consent, hospital refusal, andother reasons. Of the 3,318 biopsies confirmed to be eligible,consent was provided for review of 3,314, and 3,254 histologicsections were obtained. Of the sections reviewed, 172 werefrom biopsies that occured outside the trial period and 217had no breast tissue. The remaining 2,865 sections were from2,610 women. Of these women, 19 were censored (so that thecorresponding section was excluded from consideration), 7had no pathological diagnosis, 1,221 had a nonproliferative le-sion (and therefore continued to be eligible to develop a pro-liferative lesion), and 1,363 had a proliferative lesion.

Overall, 570 cases of benign proliferative breast diseasewere ascertained in the intervention group and 793 wereascertained in the comparison group. The estimated HR for

the association between dietary modification and benign pro-liferative breast disease was 1.09 (95% CI, 0.98-1.23; Table 2).The intervention was associated with a slight, statistically non-significant increase in risk of benign proliferative breast dis-ease without atypia (HR, 1.10; 95% CI, 0.97-1.25) and with astatistically significant increase in risk for either atypical hy-perplasia or moderately extensive or florid proliferative dis-ease without atypia (HR, 1.16; 95% CI, 1.02-1.33). Risk ofatypical hyperplasia was essentially unaltered by the inter-vention (HR, 1.05; 95% CI, 0.79-1.39). The Kaplan-Meier es-timate of the cumulative hazard of benign proliferativebreast disease (all types combined) revealed that there waslittle difference between the intervention and comparisongroups for the first 4 years of follow-up, after which the ha-zard for the dietary modification group exceeded slightlythat for the comparison group (Fig. 1).

The overall dietary modification effect was largely un-changed by exclusion of the first year of follow-up (HR,1.09; 95% CI, 0.97-1.23), exclusion of women with a breastbiopsy before the commencement of the trial (HR, 1.07; 95%CI, 0.93-1.23), or adjustment for annual measures of heightand weight (HR, 1.09; 95% CI, 0.98-1.23) or use of prior hor-mone therapy (estrogen alone or estrogen plus progestin; HR,1.09; 95% CI, 0.97-1.22).

Risk of benign proliferative breast disease in associationwith dietary modification varied by levels of baseline totalvitamin D intake (from diet and supplements), albeit not sub-stantially and not monotonically. Risk did not vary by levels ofother dietary variables (Table 3), and it varied little by levelsof most of the other baseline variables examined (Table 4).Although risk varied significantly by levels of estrogen plusprogestin use as reported at baseline, the magnitude of thedifference in risk between categories of this variable was notlarge.

To address the possible effect of nonadherence on the studyresults, we used an inverse probability weighting scheme toestimate a “full adherence” relative risk function for the inter-vention as described earlier (12). This yielded a HR of 1.37(95% CI, 0.95-1.99) for all proliferative breast diseases com-bined.

The WHI study protocol mandated that participants hadbiennial mammograms and regular clinical breast exams.Compliance with these exams was high, and there was essen-tially no difference between the intervention and comparisongroups with respect to the frequency with which these weredone (12). Neither adjustment for the frequency of mammo-grams (HR, 1.10; 95% CI, 0.98-1.23) nor adjustment for thefrequency of clinical breast exams (HR, 1.09; 95% CI, 0.98-1.23) had much impact on the estimated dietary modificationeffect.

Discussion

The results of this study suggest that the WHI low-fat diet-ary intervention, which yielded a statistically significant, al-beit modest, decrease in fat intake and increase in intake offruit, vegetables, and grains, was not associated with alteredrisk of benign proliferative breast disease after almost 8 yearsof follow-up. Although there was a small increase in overallrisk (9%) that was even larger (37%) when the extent of adher-ence to the dietary intervention was taken into account,

Diet Modification and Risk of Benign Breast Disease





Table 1. Baseline characteristics of participants in the WHI Dietary Modification trial

No. participants (%)

Dietary modification (n = 19,541) Comparison (n = 29,294)

Age, y* 62.26 (6.87) 62.26 (6.86)Race/ethnicity, %

White 15,871 (81.22) 23,891 (81.56)Black 2,135 (10.93) 3,127 (10.67)Hispanic 751 (3.84) 1,094 (3.73)American Indian 88 (0.45) 114 (0.39)Asian/Pacific Islander 431 (2.21) 674 (2.30)Other 221 (1.13) 339 (1.16)Unknown 44 (0.23) 55 (0.19)

Family history of breast cancer, %† 3,396 (17.38) 4,929 (16.83)Gail model 5-y risk ≥1.75, % 6,510 (33.31) 9,696 (33.10)Body mass index, kg/m2*,‡ 29.16 (6.05) 29.17 (6.46)Prior breast disease

No 13,812 (70.68) 20,559 (70.18)1 biopsy 2,553 (13.06) 3,914 (13.36)≥2 biopsies 1,014 (5.19) 1,553 (5.30)Unknown 2,162 (11.06) 3,268 (11.16)

Age at menarche, y≤10 1,311 (6.71) 1,907 (6.51)11-14 16,338 (83.61) 24,520 (83.70)≥15 1,834 (9.39) 2,771 (9.46)Unknown 58 (0.30) 96 (0.33)

Age at first full-term pregnancy, yNever had term pregnancy 496 (2.54) 732 (2.50)<20 2,709 (13.86) 4,183 (14.28)20-29 11,577 (59.24) 17,220 (58.78)≥30 1,372 (7.02) 2,008 (6.85)Unknown 3,387 (17.33) 5,151 (17.58)

ParityNever 2,114 (10.82) 3,214 (10.97)1 1,682 (8.61) 2,463 (8.41)2 4,766 (24.39) 7,002 (23.90)3 4,714 (24.12) 7,183 (24.52)4+ 6,159 (31.52) 9,294 (31.73)Unknown 106 (0.54) 138 (0.47)

Age at natural menopause, y* 48.01 (6.47) 47.91 (6.51)Oral contraceptive use

Ever used, % 8,751 (44.78) 12,992 (44.35)Duration of use, y* 5.32 (5.14) 5.27 (5.24)

Postmenopausal hormone useEstrogen alone

Ever used, % 7,279 (37.25) 10,842 (37.01)Duration of use, y* 9.96 (8.9) 10.02 (8.86)

Estrogen plus progestinEver used, % 5,345 (27.35) 7,995 (27.29)Duration of use, y* 5.78 (5.18) 5.72 (5.06)

Mammography screeningwithin 2 y, %

15,729 (80.49) 23,708 (80.93)

NOTE: Percentages may not sum to 100% because of rounding error.*Mean (SD).†First-degree female relative.‡Calculated as weight in kilograms divided by the square of height in meters.






neither of these effect estimates was statistically significant.Furthermore, although the intervention was associated witha statistically significant increase in the risk of the combinedoutcome category consisting of either atypical hyperplasia ormoderately extensive or florid proliferative disease withoutatypia, the magnitude of this effect was small (16% increasein risk). Similarly, although there were statistically significantinteractions between the intervention and baseline total vita-min D intake, postmenopausal hormone use before entry intothe trial, and enrollment in the WHI estrogen plus progestintrial, the magnitude of the variations in risk by categories ofthese variables was not large.

It seems that there have not been any previous randomizedtrials of the effect of dietary modification on the risk of benignproliferative breast disease. However, the role of diet in theetiology of benign proliferative breast disease has been exam-ined in several case-control (23–27) and cohort studies (28–31).Of the case-control studies, two (24, 25) showed positive asso-ciations between saturated fat intake (or indices thereof) andrisk of atypical (24) or proliferative forms (25) of benign breastdisease, whereas the other studies (23, 26, 27) provided littlesupport for a role for dietary fat. With respect to other nutri-ents, one case-control study provided some evidence forinverse associations between retinol and β-carotene intakeand risk (26) and also showed strong inverse associations fordietary fiber and its components (soluble and insoluble non-starch polysaccharides and cellulose; ref. 32). These findingswere supported to some extent by those of another study(23) in which risk of benign epithelial hyperplasia was re-duced in association with consumption of fruit and leafyand orange-red vegetables, both of which contain fiber andmicronutrients such as β-carotene. However, in another study(33), carotene and retinol intakes were not associated with riskof atypical or nonatypical forms of benign proliferative breastdisease. Findings from cohort studies have been less suppor-tive of associations. Of the four cohort studies to date (28–31),

three (29–31) measured diet in adulthood and showed thatintakes of dietary fat, carotenoid, fiber (29, 30), and folate(31) were not related to risk of subsequent benign proliferativebreast disease, whereas one study (28) showed that vitamin Eand fiber intakes during adolescence had inverse associationswith risk that were of borderline statistical significance.

Although the dietary intervention was not associated withaltered risk of benign proliferative breast disease overall, itwas associated with an increase in risk in some subgroup ana-lyses. It is conceivable that the latter represent chance findingsdue to the many subgroup analyses that were done. However,the dietary intervention was complex, and although it seemsunlikely that detrimental changes might have been introducedas a result of the intervention, an alteration in the balancebetween nutrients that mitigate the risk of cancer and thosethat increase risk might have occurred.

The mammary gland is particularly susceptible to environ-mental influences early in life (34, 35). Indeed, early life eventsare considered to play a role in the etiology of breast cancer, asevidenced by the positive associations of breast cancer riskwith birthweight and height (a marker of the adequacy ofnutritional status in childhood; ref. 36). Hence, it is conceiva-ble that the essentially null results of this trial (and those of theparent trial; ref. 12) reflect the fact that the intervention wasadministered well after the critical exposure period.

Limitations of the parent trial were discussed elsewhere(12). In brief, recruitment took longer than expected so thatthe average follow-up duration was less than had beenplanned and perhaps insufficiently long to show an effect(in this regard, follow-up of the Dietary Modification trial par-ticipants is continuing and an updated report is expectedsoon); the between-group (intervention versus comparison)difference in percent energy from fat was only ∼70% of thetarget; relatively few women managed to reduce their fatintake to 20% of energy; self-report methods were used to as-sess between-group differences in dietary intake, although the

Table 1. Baseline characteristics of participants in the WHI Dietary Modification trial (Cont'd)

No. participants (%)


Enrollment in WHI estrogen alone trial, %No 16,359 (83.72) 24,426 (83.38)Active 615 (3.15) 1,039 (3.55)Control 670 (3.43) 1,068 (3.65)

Enrollment in WHI estrogen plus progestin trial, %No 16,359 (83.72) 24,426 (83.38)Active 972 (4.97) 1,457 (4.97)Control 925 (4.73) 1,304 (4.45)

Enrollment in WHI calcium plus vitamin D supplementation trial, %No 9,896 (50.64) 13,729 (46.87)Intervention 4,767 (24.39) 7,827 (26.72)Control 4,878 (24.96) 7,738 (26.41)

Total daily energy intake, kcal* 1,790 (710) 1,789 (703)Total daily fat intake, g* 75.7 (34.1) 75.7 (33.6)Total daily calcium intake (supplements plus diet), mg* 1,123.7 (686.8) 1,117.5 (662.9)Total daily vitamin D intake (supplements plus diet), IU* 349.0 (262.9) 348.5 (265.1)






observed relative changes in blood levels of γ-tocopherol andcarotenoids are consistent with reported differences betweenthe randomization groups with respect to consumption of fats,oils, and fruits and vegetables; and the complexity of the inter-vention (reduction in fat intake and increase in fruit, vegeta-ble, and grain intake) limits the ability to attribute effects toany specific component of the dietary changes. With respectto the present study, there are two additional potential limita-

tions. First, differential ascertainment of the outcome in thetwo randomization groups may have occurred, although thisseems unlikely given that compliance with the biennial mam-mograms and regular clinical breast exams was high and es-sentially the same in both groups, and that adjustment for thefrequency of these exams had minimal effect on the estimateof the dietary modification effect (underascertainment in bothgroups is a possibility given that breast biopsies were not

Table 2. Risk of benign proliferative breast disease in association with dietary modification, overall and by thepresence/absence of atypia

No. cases (annualized %) HR (95% CI)* P

Dietary modification(n = 19,541)

Comparison(n = 29,294)

Unweighted* Weighted†

Benign proliferative breast disease, all 570 (0.38) 793 (0.35) 1.09 (0.98-1.23) 0.13 0.18Benign proliferative breast disease without atypia 477 (0.32) 660 (0.29) 1.10 (0.97-1.25) 0.12 0.22Benign proliferative breast disease without atypia

(moderately extensive or florid) or with atypia429 (0.29) 565 (0.25) 1.16 (1.02-1.33) 0.03 0.05

Atypical hyperplasia 93 (0.06) 133 (0.06) 1.05 (0.79-1.39) 0.76 0.59

*Proportional hazards model stratified by age, prior breast disease, and treatment assignment in the hormone therapy and calcium plusvitamin D supplementation trials.†Weighted log-rank test stratified by age, prior breast disease, and treatment assignment in the hormone therapy trial and adjusted forcalcium plus vitamin D supplementation trial randomization as a time-dependent covariate. Weights increase linearly from zero at rando-mization to a maximum of 1 at 10 y.

Fig. 1. Kaplan-Meier estimates of the cumulative hazard of benign proliferative breast disease in the dietary modification and comparison groups.






done on all study subjects). Second, it is possible that there wassome misclassification of the outcome, although this is likely tohave been nondifferential, with consequent biasing of the effectestimate for dietary intervention toward the null (37).

In conclusion, the results of the trial reported here suggest thata modest reduction in fat intake and a modest increase in fruit,vegetable, and grain intake in postmenopausal women do notalter their risk of subsequent benign proliferative breast disease.

Table 3. Risk of benign proliferative breast disease in association with dietary modification, by intake ofselected dietary variables at baseline

No. cases of BPBD (annualized %)* HR (95% CI) Pinteraction†


Energy intake, kcal<1,392 169 (0.37) 232 (0.33) 1.05 (0.85-1.29) 0.911,392-<1,664 112 (0.40) 153 (0.38) 1.08 (0.83-1.39)1,664-<1,959 97 (0.38) 135 (0.35) 1.19 (0.90-1.57)≥1,959 190 (0.38) 272 (0.36) 1.09 (0.90-1.32)

Percentage of energy from fat<32.3 76 (0.42) 104 (0.40) 1.14 (0.83,1.56) 0.5632.3-<36.8 248 (0.44) 322 (0.38) 1.16 (0.97-1.38)≥36.8 244 (0.32) 366 (0.32) 1.02 (0.86-1.21)

Total fat intake, g/d<46.2 100 (0.40) 126 (0.34) 1.10 (0.83-1.45) 0.8746.2-<59.8 112 (0.38) 152 (0.35) 1.15 (0.89-1.49)59.8-<76.0 137 (0.40) 207 (0.40) 1.01 (0.80-1.26)≥76.0 219 (0.36) 307 (0.34) 1.12 (0.93-1.34)

Vegetable and fruit, servings/d<2.3 130 (0.34) 183 (0.32) 1.07 (0.84-1.36) 0.902.3-<3.3 146 (0.40) 204 (0.37) 1.05 (0.84-1.32)3.3-<4.6 141 (0.38) 187 (0.34) 1.18 (0.94-1.48)≥4.6 151 (0.40) 218 (0.38) 1.07 (0.86-1.33)

Grains, servings/d<3 133 (0.36) 187 (0.33) 1.05 (0.83-1.33) 0.503-<4.3 161 (0.41) 188 (0.33) 1.26 (1.00-1.57)4.3-<5.9 136 (0.37) 210 (0.39) 0.99 (0.79-1.24)≥5.9 140 (0.38) 208 (0.36) 1.09 (0.87,1.37)

Total daily calcium intake (supplements plus diet), mg<635.6 135 (0.37) 164 (0.29) 1.28 (1.00-1.63) 0.33635.6-<979.4 152 (0.41) 216 (0.38) 1.06 (0.86-1.32)979.4-<1,459.6 134 (0.36) 215 (0.38) 0.94 (0.75-1.18)≥1,459.6 147 (0.39) 197 (0.36) 1.14 (0.91-1.42)

Total daily vitamin D intake (supplements plus diet), IU<133.2 125 (0.34) 169 (0.30) 1.12 (0.87-1.43) 0.02133.2-<262.6 174 (0.46) 192 (0.33) 1.41 (1.13-1.75)262.6-<533.8 145 (0.39) 210 (0.38) 1.04 (0.83-1.30)≥533.8 124 (0.33) 221 (0.39) 0.85 (0.68-1.08)

Caffeine intake, mg/d0-<76.8 153 (0.41) 200 (0.36) 1.12 (0.93-1.34) 0.9576.8-<177.4 138 (0.39) 188 (0.33) 1.12 (0.93-1.36)177.4-<196.7 139 (0.37) 208 (0.37) 1.06 (0.87-1.28)≥196.7 138 (0.36) 196 (0.34) 1.07 (0.88-1.29)

Alcohol intake, g/d0 236 (0.38) 315 (0.34) 1.07 (0.92-1.23) 0.650-<5 191 (0.39) 263 (0.36) 1.14 (0.97-1.35)5-≤15 99 (0.39) 141 (0.37) 1.12 (0.90-1.40)>15 42 (0.31) 73 (0.36) 0.92 (0.67-1.27)

Abbreviation: BPBD, benign proliferative breast disease.*For some variables, the number of events does not equal the total number shown in Table 2 due to missing values.†Unweighted proportional hazards model stratified by age, prior disease, and randomization group.






Table 4. Risk of benign proliferative breast disease in association with dietary modification, by selected baselinecharacteristics



Age, y50-59 249 (0.43) 311 (0.36) 1.24 (1.04-1.48) 0.1360-69 257 (0.38) 371 (0.36) 1.04 (0.88,1.23)70-79 64 (0.27) 111 (0.31) 0.88 (0.63-1.21)

Race/ethnicityWhite 474 (0.39) 688 (0.37) 1.03 (0.91-1.17) 0.06Black 52 (0.32) 51 (0.21) 1.65 (1.10-2.48)Hispanic 18 (0.32) 18 (0.22) 1.92 (0.96-3.85)Other 25 (0.46) 33 (0.39) 1.20 (0.71-2.03)

Family history of breast cancer in first-degree relativeNo 434 (0.37) 591 (0.33) 1.13 (0.99-1.29) 0.23Yes 105 (0.40) 161 (0.43) 0.95 (0.74-1.23)

Gail model 5-y risk, %<1.25 190 (0.35) 231 (0.29) 1.24 (1.01-1.53) 0.311.25-1.74 174 (0.36) 256 (0.35) 1.09 (0.89-1.33)≥1.75 206 (0.42) 306 (0.42) 1.00 (0.83-1.20)

Body mass index, kg/m2‡

<25 171 (0.43) 234 (0.39) 1.08 (0.88-1.33) 0.9925-29 206 (0.39) 292 (0.36) 1.09 (0.90-1.31)30-34 116 (0.34) 163 (0.32) 1.09 (0.85-1.41)≥35 77 (0.34) 104 (0.31) 1.13 (0.83-1.54)

Age at menarche, y≤10 32 (0.32) 46 (0.31) 0.94 (0.58-1.52) 0.7911-14 483 (0.38) 669 (0.35) 1.11 (0.98-1.26)≥15 54 (0.39) 74 (0.34) 1.07 (0.74-1.56)

Age at first full-term pregnancy, yNever had term 15 (0.40) 23 (0.41) 0.91 (0.46-1.83) 0.19<20 94 (0.46) 105 (0.33) 1.44 (1.07-1.94)20-29 345 (0.39) 496 (0.37) 1.03 (0.89-1.20)≥30 47 (0.45) 54 (0.35) 1.27 (0.84-1.92)

Parity0 54 (0.33) 79 (0.32) 1.13 (0.79-1.62) 0.551 56 (0.44) 61 (0.32) 1.46 (1.00-2.14)2 154 (0.42) 218 (0.40) 0.99 (0.79-1.24)3 136 (0.38) 192 (0.35) 1.07 (0.85-1.35)≥4 169 (0.36) 242 (0.34) 1.10 (0.89-1.35)

Age at natural menopause, y<48 236 (0.43) 312 (0.37) 1.11 (0.93-1.32) 0.9948-<50 55 (0.40) 70 (0.35) 1.16 (0.80-1.67)50-<53 154 (0.39) 221 (0.38) 1.09 (0.88-1.35)≥53 104 (0.34) 149 (0.33) 1.10 (0.85-1.44)

Oral contraceptive use, yNone 271 (0.33) 396 (0.32) 1.04 (0.88-1.22) 0.32<5 151 (0.41) 225 (0.40) 1.04 (0.84-1.30)≥5 148 (0.47) 172 (0.38) 1.28 (1.01-1.61)

Baseline postmenopausal hormone use, yEstrogen alone, y

None 329 (0.35) 420 (0.29) 1.23 (1.06-1.43) 0.08

*For some variables, the number of events does not equal the total number shown in Table 2 due to missing values.†Unweighted proportional hazards model stratified by age, prior disease, and randomization group.‡Calculated as weight in kilograms divided by the square of height in meters.






Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank the WHI investigators and staff for their outstanding dedication andcommitment. A list of key investigators involved in this research follows. A fulllisting of WHI investigators can be found at the following website: http://www.whi.org.

Program Office: National Heart, Lung, and Blood Institute (Bethesda, MD)—Elizabeth Nabel, Jacques Rossouw, Shari Ludlam, Linda Pottern, Joan McGo-wan, Leslie Ford, and Nancy Geller.

Clinical Coordinating Centers: Fred Hutchinson Cancer Research Center(Seattle, WA)—Ross Prentice, Garnet Anderson, Andrea LaCroix, Charles L.Kooperberg, Ruth E. Patterson, and Anne McTiernan; Wake Forest UniversitySchool of Medicine (Winston-Salem, NC)—Sally Shumaker; Medical ResearchLabs (Highland Heights, KY)—Evan Stein; University of California at San Fran-cisco (San Francisco, CA)—Steven Cummings.

Clinical Centers: Albert Einstein College of Medicine (Bronx, NY)—SylviaWassertheil-Smoller; Baylor College of Medicine (Houston, TX)—Jennifer Hays;Brigham and Women's Hospital, Harvard Medical School (Boston, MA)—JoAnnManson; Brown University (Providence, RI)—Annlouise R. Assaf; Emory Univer-sity (Atlanta, GA)—Lawrence Phillips; Fred Hutchinson Cancer Research Center(Seattle, WA)—Shirley Beresford; George Washington University Medical Cen-ter (Washington, DC)—Judith Hsia; Los Angeles Biomedical Research Instituteat Harbor-University of California at Los Angeles Medical Center (Torrance, CA)—Rowan Chlebowski; Kaiser Permanente Center for Health Research (Portland,OR)—Evelyn Whitlock; Kaiser Permanente Division of Research (Oakland, CA)

—Bette Caan; Medical College of Wisconsin (Milwaukee, WI)—Jane MorleyKotchen; MedStar Research Institute/Howard University (Washington, DC)—Barbara V. Howard; Northwestern University (Chicago/Evanston, IL)—LindaVan Horn; Rush Medical Center (Chicago, IL)—Henry Black; Stanford Preven-tion Research Center (Stanford, CA)—Marcia L. Stefanick; State University ofNew York at Stony Brook (Stony Brook, NY)—Dorothy Lane; The Ohio StateUniversity (Columbus, OH)—Rebecca Jackson; University of Alabama at Bir-mingham (Birmingham, AL)—Cora E. Lewis; University of Arizona (Tucson/Phoenix, AZ)—Tamsen Bassford; University at Buffalo (Buffalo, NY)—JeanWactawski-Wende; University of California at Davis (Sacramento, CA)—JohnRobbins; University of California at Irvine (Irvine, CA)—F. Allan Hubbell; Univer-sity of California at Los Angeles (Los Angeles, CA)—Howard Judd; University ofCalifornia at San Diego (LaJolla/Chula Vista, CA)—Robert D. Langer; Universityof Cincinnati (Cincinnati, OH)—Margery Gass; University of Florida (Gainesville/Jacksonville, FL)—Marian Limacher; University of Hawaii (Honolulu, HI)—DavidCurb; University of Iowa (Iowa City/Davenport, IA)—Robert Wallace; Universityof Massachusetts/Fallon Clinic (Worcester, MA)—Judith Ockene; University ofMedicine and Dentistry of New Jersey (Newark, NJ)—Norman Lasser; Universityof Miami (Miami, FL)—Mary Jo O'Sullivan; University of Minnesota (Minneapolis,MN)—Karen Margolis; University of Nevada (Reno, NV)—Robert Brunner; Uni-versity of North Carolina (Chapel Hill, NC)—Gerardo Heiss; University of Pitts-burgh (Pittsburgh, PA)—Lewis Kuller; University of Tennessee (Memphis, TN)—Karen C. Johnson; University of Texas Health Science Center (San Antonio,TX)—Robert Brzyski; University of Wisconsin (Madison, WI)—Gloria E. Sarto;Wake Forest University School of Medicine (Winston-Salem, NC)–DeniseBonds; Wayne State University School of Medicine/Hutzel Hospital (Detroit,MI)—Susan Hendrix.

The WHI program is funded by the National Heart, Lung and Blood Institute,U.S. Department of Health and Human Services.

References1. Hankinson SE, Hunter DJ. Breast Cancer. In: Adami

H-O, Hunter DJ, Trichopoulos D, editors. Textbookof cancer epidemiology. New York: Oxford Univer-sity Press; 2002. p.301–39.

2. Duncan AM. The role of nutrition in the prevention ofbreast cancer. AACN Clin Issues 2004;15:119–35.

3. Hunter DJ, Spiegelman D, Adami HO, et al. Co-hort studies of fat intake and the risk of breast

cancer—a pooled analysis. N Engl J Med 1996;334:356–61.

4. Lee MM, Lin SS. Dietary fat and breast cancer.Annu Rev Nutr 2000;20:221–48.

Table 4. Risk of benign proliferative breast disease in association with dietary modification, by selected baselinecharacteristics (Cont'd)



<5 83 (0.39) 117 (0.38) 0.96 (0.71-1.30)≥5 158 (0.46) 256 (0.49) 0.93 (0.76-1.15)

Estrogen plus progestin, yNone 341 (0.31) 531 (0.32) 0.97 (0.84-1.12) 0.02<5 97 (0.45) 117 (0.36) 1.30 (0.97-1.73)≥5 132 (0.69) 145 (0.50) 1.38 (1.08-1.76)

Enrollment in WHI hormone trials§

Estrogen aloneActive 19 (0.40) 32 (0.40) 0.81 (0.43-1.51) 0.35Placebo 13 (0.26) 18 (0.22) 1.26 (0.63-2.53)

Estrogen plus progestinActive 20 (0.27) 49 (0.44) 0.60 (0.34-1.04) 0.90Placebo 11 (0.15) 28 (0.28) 0.63 (0.30-1.32)

Enrollment in WHI calcium plus vitamin D trial§

Intervention 149 (0.40) 239 (0.39) 1.09 (0.88-1.34) 0.48Placebo 166 (0.44) 214 (0.35) 1.21 (0.98-1.49)

§Restricted to those who were randomized.






5. Glade MJ. Food, nutrition, and the prevention ofcancer: a global perspective. American Institutefor Cancer Research/World Cancer Research Fund,American Institute for Cancer Research, 1997. Nu-trition 1999;15:523–6.

6. Wirfalt E, Mattisson I, Gullberg B, Johansson U,Olsson H, Berglund G. Postmenopausal breast can-cer is associated with high intakes of ω6 fatty acids(Sweden). Cancer Causes Control 2002;13:883–93.

7. Freedman LS, Potischman N, Kipnis V, et al. Acomparison of two dietary instruments for evaluat-ing the fat-breast cancer relationship. Int J Epide-miol 2006;35:1011–21.

8. Thiebaut AC, Kipnis V, Chang SC, et al. Dietary fatand postmenopausal invasive breast cancer in theNational Institutes of Health-AARP Diet and HealthStudy cohort. J Natl Cancer Inst 2007;99:451–62.

9. Smith-Warner SA, Spiegelman D, Yaun SS, et al.Intake of fruits and vegetables and risk of breastcancer: a pooled analysis of cohort studies. JAMA2001;285:769–76.

10. van Gils CH, Peeters PH, Bueno-de-MesquitaHB, et al. Consumption of vegetables and fruitsand risk of breast cancer. JAMA 2005;293:183–93.

11. Cho E, Spiegelman D, Hunter DJ, Chen WY,Colditz GA, Willett WC. Premenopausal dietary car-bohydrate, glycemic index, glycemic load, and fiberin relation to risk of breast cancer. Cancer Epide-miol Biomarkers Prev 2003;12:1153–8.

12. Prentice RL, Caan B, Chlebowski RT, et al. Low-fatdietary pattern and risk of invasive breast cancer: theWomen's Health Initiative Randomized ControlledDietary Modification Trial. JAMA 2006;295:629–42.

13. Lakhani SR. The transition from hyperplasia to in-vasive carcinoma of the breast. J Pathol 1999;187:272–8.

14. Rohan TE, Kandel RA. Breast In: Franco EL,Rohan TE, editors. Cancer precursors: epide-miology, detection, and prevention. New York:Springer; 2002. p.232–48.

15. Santen RJ, Mansel R. Benign breast disorders. NEngl J Med 2005;353:275–85.

16. Hays J, Hunt JR, Hubbell FA, et al. The Women'sHealth Initiative recruitment methods and results.Ann Epidemiol 2003;13:S18–S77.

17. Ritenbaugh C, Patterson RE, Chlebowski RT,et al. The Women's Health Initiative Dietary Modifi-cation trial: overview and baseline characteristics ofparticipants. Ann Epidemiol 2003;13:S87–S97.

18. Patterson RE, Kristal AR, Tinker LF, Carter RA,Bolton MP, Agurs-Collins T. Measurement char-acteristics of the Women's Health Initiative foodfrequency questionnaire. Ann Epidemiol 1999;9:178–87.

19. Hartmann LC, Sellers TA, Frost MH, et al. Benignbreast disease and the risk of breast cancer. N EnglJ Med 2005;353:229–37.

20. Page DL, Rogers LW. Combined histologic andcytologic criteria for the diagnosis of mammary aty-pical ductal hyperplasia. Hum Pathol 1992;23:1095–7.

21. Page DL, Schuyler PA, Dupont WD, Jensen RA,Plummer WD, Jr., Simpson JF. Atypical lobular hy-perplasia as a unilateral predictor of breast cancerrisk: a retrospective cohort study. Lancet 2003;361:125–9.

22. Cox DR. Regression models and life tables (withdiscussion). J R Stat Soc B 1972;34:187–220.

23. Ingram DM, Nottage E, Roberts T. The role of dietin the development of breast cancer: a case-controlstudy of patients with breast cancer, benign epithe-lial hyperplasia and fibrocystic disease of thebreast. Br J Cancer 1991;64:187–91.

24. Lubin F, Wax Y, Ron E, et al. Nutritional factorsassociated with benign breast disease etiology: acase-control study. Am J Clin Nutr 1989;50:551–6.

25. Hislop TG, Band PR, Deschamps M, et al. Dietand histologic types of benign breast disease de-fined by subsequent risk of breast cancer. Am JEpidemiol 1990;131:263–70.

26. Rohan TE, Cook MG, Potter JD, McMichael AJ. Acase-control study of diet and benign proliferativeepithelial disorders of the breast. Cancer Res1990;50:3176–81.

27. London SJ, Sacks FM, Stampfer MJ, et al. Fattyacid composition of the subcutaneous adiposetissue and risk of proliferative benign breast dis-ease and breast cancer. J Natl Cancer Inst 1993;85:785–93.

28. Baer HJ, Schnitt SJ, Connolly JL, et al. Adoles-cent diet and incidence of proliferative benignbreast disease. Cancer Epidemiol Biomarkers Prev2003;12:1159–67.

29. Rohan TE, Jain M, Miller AB. A case-cohort studyof diet and risk of benign proliferative epithelial dis-orders of the breast (Canada). Cancer Causes Con-trol 1998;9:19–27.

30. Webb PM, Byrne C, Schnitt SJ, et al. A pro-spective study of diet and benign breast disease.Cancer Epidemiol Biomarkers Prev 2004;13:1106–13.

31. Cui Y, Page DL, Chlebowski RT, et al. Alcoholand folate consumption and risk of benign prolifera-tive epithelial disorders of the breast. Int J Cancer2007;121:1346–51.

32. Baghurst PA, Rohan TE. Dietary fiber and risk ofbenign proliferative epithelial disorders of thebreast. Int J Cancer 1995;63:481–5.

33. London SJ, Stein EA, Henderson IC, et al. Caro-tenoids, retinol, and vitamin E and risk of prolifera-tive benign breast disease and breast cancer.Cancer Causes Control 1992;3:503–12.

34. Colditz GA, Frazier AL. Models of breast cancershow that risk is set by events of early life: preven-tion efforts must shift focus. Cancer Epidemiol Bio-markers Prev 1995;4:567–71.

35. Michels KB, Mohllajee AP, Roset-Bahmanyar E,Beehler GP, Moysich KB. Diet and breast cancer: areview of the prospective observational studies.Cancer 2007;109:2712–49.

36. Hankinson SE, Colditz GA, Willett WC. Towardsan integrated model for breast cancer etiology: thelifelong interplay of genes, lifestyle, and hormones.Breast Cancer Res 2004;6:213–8.

37. Rothman KJ. Modern epidemiology 1986. Little,Brown & Company: Boston106.






2008;1:275-284. Published OnlineFirst July 9, 2008.Cancer Prev Res Thomas E. Rohan, Abdissa Negassa, Bette Caan, et al. Modification TrialBreast Disease: A Randomized, Controlled Dietary Low-Fat Dietary Pattern and Risk of Benign Proliferative

Updated version

10.1158/1940-6207.CAPR-08-0003doi:

Access the most recent version of this article at:

Cited articles

http://cancerpreventionresearch.aacrjournals.org/content/1/4/275.full.html#ref-list-1

This article cites 34 articles, 10 of which you can access for free at:

Citing articles

http://cancerpreventionresearch.aacrjournals.org/content/1/4/275.full.html#related-urls

This article has been cited by 1 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

[email protected] at

To request permission to re-use all or part of this article, contact the AACR Publications



http://cancerpreventionresearch.aacrjournals.org/lookup/doi/10.1158/1940-6207.CAPR-08-0003

http://cancerpreventionresearch.aacrjournals.org/content/1/4/275.full.html#ref-list-1

http://cancerpreventionresearch.aacrjournals.org/content/1/4/275.full.html#related-urls

http://cancerpreventionresearch.aacrjournals.org/cgi/alerts

mailto:[email protected]

mailto:[email protected]


Documents

Cancer Prev Res 2008 Rohan 275 84