Background:

Plant-based diets have been associated with lower risk of various diseases, including type 2 diabetes, cardiovascular disease, and other cardiometabolic risk factors. However, the association between plant-based diet quality and breast cancer remains unclear.

Methods:

We prospectively followed 76,690 women from the Nurses' Health Study (NHS, 1984–2016) and 93,295 women from the NHSII (1991–2017). Adherence to an overall plant-based diet index (PDI), a healthful PDI (hPDI), and an unhealthful PDI (uPDI) was assessed using previously developed indices. Cox proportional hazards models were used to estimate HR and 95% confidence intervals (CI) for incident invasive breast cancer.

Results:

Over 4,841,083 person-years of follow-up, we documented 12,482 incident invasive breast cancer cases. Women with greater adherence to PDI and hPDI were at modestly lower risk of breast cancer [(HRQ5 vs. Q1, 0.89; 95% CI, 0.84–0.95); (HRQ5 vs. Q1, 0.89; 95% CI, 0.83–0.94)]. We observed significant heterogeneity by estrogen receptor (ER) status, with the strongest inverse association between hPDI and breast cancer observed with ER-negative tumors [HRQ5 vs. Q1, 0.77; 95% CI, 0.65–0.90; Ptrend < 0.01]. We also found an inverse association between extreme quintiles of healthy plant foods and ER-negative breast cancer [HR, 0.74; 95% CI, 0.61–0.88; Ptrend < 0.01].

Conclusions:

This study provides evidence that adherence to a healthful plant-based diet may reduce the risk of breast cancer, especially those that are more likely to be aggressive tumors.

Impact:

This is the first prospective study investigating the relation between healthful and unhealthful plant-based dietary indices and risk of total and subtype-specific breast cancer.

Plant-based diets have been associated with lower risk of chronic diseases such as obesity, type 2 diabetes, cardiovascular disease, and some cancers (1–6) and are recommended for both health and environmental benefits (7). Although vegetarian or vegan diets have experienced an increase in popularity, the proportion of people that follow such diets remains limited in Western countries (8, 9) and most people consume a combination of foods from animal and plant sources. Rather than conceptualizing plant-based diets as the whole exclusion of a few or all animal foods, recent studies have evaluated progressive adherence to a plant-based dietary pattern (10, 11). They could have a broader application, because progressive reductions in animal food intake may be easier to adhere to than other recommendations such as the total exclusion of animal foods.

Recently, Satija and colleagues (12) proposed three different approaches of plant-based dietary indices. An overall plant-based diet index (PDI) was created, equivalent to the original provegetarian dietary pattern (10), and two additional scores, a healthful PDI (hPDI), and an unhealthful PDI (uPDI). The hPDI and uPDI overcome the limitations derived from the fact that all plant foods were treated equally in the original provegetarian diet, but the nutritional quality is not equivalent across all plant foods.

The association between plant-based diet quality and breast cancer remains unclear (6, 13). In the NutriNet-Santé study (13), higher intakes of plant-based products along with lower intakes of animal foods was associated with a lower cancer risk, though no association was found for breast cancer. Moreover, in the SUN project (6), a moderate, but not greater, adherence to a provegetarian dietary pattern, was associated with a decreased risk of breast cancer. No further associations were found when the authors distinguished between healthful and unhealthful provegetarian food patterns.

A plant-based dietary pattern may play an important role in breast cancer independent of the estrogen pathway. In fact, epidemiologic evidence supports a beneficial role for a prudent or healthy dietary pattern (characterized by fruit, vegetables, lean meat, fish, and unprocessed grains) with breast cancer risk, particularly estrogen receptor (ER)-positive and/or progesterone receptor (PR)-positive tumors and ER- and/or PR-negative tumors (14). Moreover, a commonality across plant-based dietary patterns is the high consumption of fruits and vegetables. In a large analysis, pooling repeated measures over 30 years of follow-up from the Nurses' Health Studies (NHS), total fruit and vegetable consumption were associated with lower breast cancer incidence, particularly the more aggressive tumors including ER-negative, HER2‐enriched, and basal‐like tumors (15). This is consistent with prior studies where both dietary intake and circulating levels of carotenoids were inversely associated with ER-negative breast cancer. These compounds may be capable of binding and removing free radicals, repairing DNA damage, inhibiting cell proliferation, inducing apoptosis, and stopping angiogenesis (16).

Breast cancer is a heterogeneous disease characterized by multiple tumor types with unique pathologic features and biological behaviors. In this framework, the objective of the current prospective study was to examine the associations between plant-based diet indices and breast cancer incidence characterized by hormone receptor status and molecular subtypes in the NHS and NHSII.

Study population

The NHS is an ongoing prospective cohort study that began in 1976 with enrollment of 121,700 female nurses, and the NHSII began in 1989 with 116,429 female nurses (ages 25–42 years). Every 2 years, participants completed a mailed questionnaire on their medical history and lifestyle factors. Women were followed from 1984 to 2016 in the NHS and from 1991 to 2017 in the NHSII. We excluded women who died prior to baseline, had prevalent cancer, had missing dietary information, or reported implausible total energy intake (<600 or >3,500 kcal/day); leaving 76,690 women from NHS, and 93,295 from NHSII. The study protocol was approved by the Institutional Review Boards of the Brigham and Women's Hospital (Boston, MA) and Harvard T.H. Chan School of Public Health (Boston, MA), and those of participating registries as required.

Dietary assessment and plant-based diet indices

Dietary data were collected using food-frequency questionnaires (FFQ) administered in the NHS in 1980, 1984, 1986 and every 4 years thereafter, and in the NHSII in 1991 and every 4 years thereafter. The number of FFQ food items have changed over time: in NHS, there were 61 items in 1980, 116 items in 1984 and 1986, and >130 items thereafter; in NHSII, the 1991 FFQ had >130 items. Given the complexity and broad array of foods included in our plant-based dietary patterns, we used the 1984 FFQ as baseline questionnaire in the NHS and the 1991 FFQ in the NHSII. FFQs prompted participants to report their food-specific average consumption for a specified serving size over the previous year (nine response categories ranging from “almost never” to “>6/day”). On the basis of the U.S. Department of Agriculture food composition tables, participants' nutrient intakes were calculated by combining frequency information with the nutrient content of a food serving.

As described in a prior publication (12), we derived three versions of a plant-based diet using the FFQ data: an overall PDI, a hPDI, and an uPDI. Briefly, we included 18 food groups on the basis of nutrient and culinary similarities within larger classifications of animal foods and healthy and less healthy plant foods. Of the 18 food groups, seven were healthy plant food groups (whole grains, fruits, vegetables, nuts, legumes, vegetable oils, and tea/coffee), five were unhealthy plant food groups (fruit juices, sugar-sweetened beverages, refined grains, potatoes, and sweets), and six were animal food groups (animal fats, dairy, eggs, fish or seafood, meat, and miscellaneous animal-based foods). Healthy and less healthy plant foods were identified using current knowledge of associations of the foods with type 2 diabetes, cardiovascular disease, some cancers, and intermediate conditions (i.e., obesity, hypertension, or inflammation). Intake of 18 food groups (servings/day) was ranked into quintiles (Q) and given positive or reverse scores. With positive scores, women in the highest quintile of a food group obtained a score of 5, following on through to women in the lowest quintile who obtained a score of 1. With reverse scores, the scoring method was opposite, with a score of 5 for the lowest quintile. For creating the PDI, healthy and less healthy plant food groups received positive scores, while all animal food groups received reverse scores. For the hPDI, we assigned positive scores to healthy plant food groups, and reverse scores to less healthy plant and animal food groups. For the uPDI, the opposite scoring pattern was used. The 18 food group scores were summed to obtain PDI, hPDI, and uPDI, ranging from 18 to 90. Pearson correlation coefficients between hPDI and uPDI were −0.36 and −0.33 for NHS and NHSII, respectively (12). The PDI showed low correlations with hPDI (r = 0.21 in NHS and r = 0.26 in NHSII) and with uPDI (r = −0.11 in NHS and r = −0.21 in NHSII; ref. 12).

Because alcohol has different associations with various health outcomes and the fatty composition of margarine has changed over time from high trans to high unsaturated fats, we did not include these food groups in the indices but rather adjusted for alcohol in the main analysis.

Assessment of breast cancer

We first identified incident breast cancer cases through self-report on the biennial questionnaires. Women who self-reported incident cases of breast cancer were asked permission to review hospital records and pathology reports for diagnosis confirmation and identification of invasive versus in situ, and ER, PR and HER2 status. Given the high confirmation rate of reported breast cancer cases in the NHS and NHSII (>99%), we included both breast cancer cases confirmed via medical record review and self-reported cases confirmed by the nurse but lacking medical record. Fatal events were identified by reports from next of kin, U.S. postal service in response to follow-up questionnaires, or by a search for the National Death Index.

Tissue microarrays, IHC analysis, and subtype classification

We have already reported details of breast cancer tissue block collection and tissue microarray (TMA) construction (17). Briefly, archived formalin-fixed paraffin-embedded breast cancer blocks were collected from women with incident breast cancer diagnosed up through 2006. To classify molecular subtypes, IHC staining information was available for a panel of markers (ER, PR, HER2, cytokeratins 5/6, and EGFR). A subgroup of NHS cases was additionally stained for the proliferative marker Ki-67 (Ki-67 data were not available for NHSII cases). Women with tissue specimens were very similar to all invasive cases with respect to breast cancer risk factors and tumor characteristics.

For a subgroup of cases, classification based on gene expression profiles were used for tumor molecular subtyping (18–23). Histologic grade was used when Ki-67 expression data were not available (NHSII). Thus, luminal A cases were defined as ER-positive and/or PR-positive, HER2-negative, and Ki-67-negative (or histologic grade 1 or 2). Luminal B cases were either (i) ER-positive and/or PR-positive and HER2-positive or (ii) ER-positive and/or PR-positive, HER2-negative, and Ki-67-positive (or histologic grade 3). HER2-enriched tumors were ER-negative, PR-negative, and HER2-positive. Basal-like tumors were negative for ER, PR, and HER2 and positive for CK 5/6 and/or EGFR. ER status was ascertained primarily from TMA slides to evaluate ER-positive versus ER-negative tumors, and if unavailable, secondarily from pathology reports.

Statistical analysis

Data from the NHS and NHSII were pooled. Person-time for each participant was calculated from the date of return of the baseline questionnaire until the date of breast cancer diagnosis, other cancers (excluding non-melanoma skin cancers), death, or the end of follow-up [2016 (NHS) or 2017 (NHSII) for the main analysis and 2006 for molecular subtype analysis], whichever occurred first.

Indices were cumulatively averaged over follow-up to better capture long-term diet. We evaluated the association between PDI, hPDI, and uPDI quintiles and incident breast cancer using multivariable-adjusted time-varying Cox proportional hazards regression models, stratified by age, 2-year time period at risk and cohort. We tested for linear trends by evaluating the quintile median values as a continuous variable. Also, we examined the association between these dietary indices and breast cancer risk by ER status and molecular subtypes. To evaluate whether associations differed by molecular subtype or ER status, we used the Lunn-McNeil approach to derive the Pheterogeneity (24). In ancillary analyses, we fit restricted cubic splines to the fully adjusted model with the plant-based dietary indices entered as continuous variables to examine potential deviation from linearity.

Covariates included race, socioeconomic status, age at menarche, age at menopause, postmenopausal hormone use, oral contraceptive use, parity, age at first birth, breastfeeding history, height, alcohol intake, total caloric intake, physical activity, and body mass index (BMI) at age 18 years. As change in weight from age 18, total carotenoids intake and dietary fiber may be intermediates between diet and breast cancer, we additionally adjusted for them in separate models.

To test whether the indices and breast cancer association differed by menopausal status, current BMI, and physical activity, we added interaction terms and used the Wald test. Mediation analyses (25) were performed to assess the extent to which associations may be potentially mediated by weight gain from age 18 years and estimated the mediation proportion (the proportion of the observed association attributable to a mediator; refs. 26, 27).

To take advantage of repeated diet assessments in these cohorts and evaluate the latency between these indices and breast cancer incidence, we conducted latency analyses, whereby we created different regression models based on dietary data collected at distinct timepoints (28).

Statistical tests were two sided with P < 0.05 indicating statistical significance. All analyses were performed using SAS for UNIX version 9.4. Adjustments were not made for multiple comparisons as our analysis was guided by strong, biologically driven a priori hypotheses and we interpreted the results considering biological plausibility, coherence, and consistency.

Role of the funder/sponsor

The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review or approval of the article; and decision to submit the article for publication.

During 4,841,083 person-years of follow-up, 12,482 participants developed invasive breast cancer (8,220 cases in NHS and 4,262 cases in NHSII). Women with higher PDI were more likely to have lower BMI, lower weight change from age 18, lower prevalence of diabetes, less alcohol intake, higher physical activity and energy intake, lower protein intake (% of energy) and higher carbohydrate intake (% of energy; Table 1).

Table 1.

Age and age-standardized baseline characteristics of women according to quintiles of an overall PDI in the NHS and NHSII.

NHS (1984)NHSII (1991)
Q1 (n = 14,629)Q3 (n = 13,832)Q5 (n = 15,857)Q1 (n = 20,408)Q3 (n = 16,541)Q5 (n = 17,944)
Median PDI 46 54 62 47 55 63 
 Age, yearsa 50.2 (7) 50.7 (7.2) 51.7 (7.3) 36.5 (4.8) 36.6 (4.7) 36.9 (4.5) 
 Body mass index, kg/m2 25.7 (5.1) 25.1 (4.7) 24.5 (4.4) 25.2 (5.7) 24.5 (5.2) 23.7 (4.8) 
 Body mass index at age 18 years, kg/m2 21.7 (3.2) 21.4 (3.0) 21.1 (2.8) 21.6 (3.6) 21.3 (3.3) 20.9 (3.0) 
 Weight change from age 18 years, kg 10.7 (12.3) 9.8 (11.1) 9.0 (10.5) 9.8 (12.4) 8.7 (11.2) 7.7 (10.4) 
 Height, inches 64.5 (2.4) 64.5 (2.4) 64.5 (2.4) 64.9 (2.6) 64.9 (2.6) 64.9 (2.6) 
 Self-reported African heritage (%) 1.5 1.4 1.1 2.0 1.4 1.2 
 Self-reported history of diabetes (%) 3.8 3.3 2.3 1.2 1.0 0.7 
 Family history of breast cancer (%) 7.8 8.1 8.1 5.8 6.1 6.3 
 Personal history of benign breast disease (%) 29.5 30.2 31.0 9.1 9.5 9.6 
 Age at menarche <12 years (%) 23.5 21.6 22.5 24.8 24.0 24.8 
 Oral contraceptives, ever (%) 50.5 49.3 48.3 85.6 84.7 83.4 
 Parous (%) 92.0 92.9 93.3 69.3 75.9 77.5 
 Parity, nb 3.1 (1.5) 3.2 (1.5) 3.2 (1.5) 2.1 (0.9) 2.1 (0.9) 2.2 (0.9) 
 Breastfeeding, ≤6 months (%)b 35.8 36.6 36.1 19.7 16.9 14.1 
 Postmenopausal (%) 48.4 48.6 48.8 3.1 3.2 3.4 
 Postmenopausal hormone use, never (%)c 52.9 52.7 52.8 6.1 6.9 7.8 
 Physical activity, METs-h/week 11.4 (13.2) 11.9 (13.1) 12.9 (13.4) 17.8 (24.4) 20.6 (26.8) 24.9 (31.5) 
 Alcohol intake (g/day) 8.6 (13.6) 6.7 (11) 5.8 (9.3) 3.3 (7.1) 3 (5.8) 3.2 (5.6) 
 Total energy intake (kcal/day) 1,458 (451) 1,723 (488) 2,056 (519) 1,478 (464) 1,792 (504) 2,134 (527) 
 Saturated fat (% of energy) 13.9 (2.9) 12.5 (2.3) 11.2 (2.1) 12.6 (2.5) 11.1 (2.2) 9.8 (2.1) 
 Monounsaturated fat (% of energy) 13.4 (2.6) 12.7 (2.3) 12 (2.2) 12.7 (2.6) 11.9 (2.4) 11.2 (2.3) 
 Polyunsaturated fat (% of energy) 6.5 (1.9) 6.6 (1.7) 6.8 (1.6) 5.7 (1.5) 5.6 (1.4) 5.6 (1.3) 
 Trans fat (% of energy) 1.8 (0.6) 1.9 (0.6) 1.9 (0.6) 1.7 (0.7) 1.6 (0.6) 1.5 (0.5) 
 Protein intake (% of energy) 19.7 (3.9) 17.7 (3.1) 16.2 (2.5) 21 (3.8) 19.2 (3.2) 17.6 (2.9) 
 Carbohydrate intake (% of energy) 40.5 (8) 46.7 (6.8) 51.2 (6.5) 44.8 (7.4) 50.1 (6.5) 54.7 (6.6) 
Healthy plant foods (servings/day) 
 Whole grains 0.7 (0.9) 1.1 (1.1) 1.6 (1.2) 0.9 (0.9) 1.4 (1.1) 2.1 (1.3) 
 Fruits 0.9 (0.9) 1.4 (1) 1.9 (1.1) 0.7 (0.7) 1.2 (0.9) 1.8 (1.1) 
 Vegetables 2.3 (1.4) 2.9 (1.5) 3.7 (1.8) 2.1 (1.4) 2.9 (1.8) 4.1 (2.1) 
 Nuts 0.1 (0.2) 0.2 (0.3) 0.3 (0.4) 0.1 (0.1) 0.2 (0.2) 0.3 (0.3) 
 Legumes 0.3 (0.2) 0.4 (0.3) 0.5 (0.3) 0.2 (0.2) 0.4 (0.3) 0.6 (0.4) 
 Vegetable oil 0.2 (0.3) 0.3 (0.4) 0.4 (0.4) 0.2 (0.3) 0.3 (0.4) 0.5 (0.5) 
 Tea and coffee 2.6 (1.9) 3.1 (1.9) 3.6 (2) 1.8 (1.8) 2.2 (1.9) 2.7 (2) 
Less healthy plant foods (servings/day) 
 Fruit juices 0.4 (0.6) 0.7 (0.7) 1 (0.8) 0.4 (0.6) 0.6 (0.7) 1.1 (1) 
 Refined grains 1.2 (1.1) 1.6 (1.3) 2.2 (1.5) 1.2 (0.9) 1.6 (1) 2 (1.1) 
 Potatoes 0.4 (0.3) 0.5 (0.4) 0.7 (0.4) 0.4 (0.3) 0.5 (0.4) 0.7 (0.4) 
 Sugar-sweetened beverages 0.2 (0.6) 0.3 (0.6) 0.4 (0.6) 0.4 (0.8) 0.5 (0.9) 0.6 (0.9) 
 Sweets and desserts 0.8 (0.9) 1.2 (1.2) 1.8 (1.4) 0.8 (0.9) 1.2 (1.1) 1.6 (1.2) 
Animal foods (servings/day) 
 Animal fat 0.5 (0.9) 0.4 (0.8) 0.3 (0.6) 0.2 (0.5) 0.2 (0.4) 0.1 (0.4) 
 Dairy 2 (1.4) 2 (1.4) 2 (1.3) 2.3 (1.5) 2.3 (1.5) 2.3 (1.4) 
 Eggs 0.4 (0.4) 0.3 (0.3) 0.3 (0.3) 0.2 (0.2) 0.2 (0.2) 0.2 (0.2) 
 Fish and seafood 0.3 (0.3) 0.3 (0.3) 0.3 (0.2) 0.3 (0.2) 0.3 (0.2) 0.3 (0.3) 
 Meat 1.7 (0.8) 1.7 (0.8) 1.7 (0.8) 1.6 (0.8) 1.7 (0.8) 1.6 (0.9) 
 Miscellaneous animal-based foods 0.5 (0.4) 0.4 (0.4) 0.4 (0.4) 0.4 (0.4) 0.4 (0.3) 0.4 (0.3) 
Food groups (servings/day) 
 Healthy plant foods 7.1 (3) 9.3 (3.3) 12.1 (3.7) 6 (2.9) 8.6 (3.5) 12 (4.1) 
 Less healthy plant foods 3 (1.9) 4.4 (2.3) 6.1 (2.7) 3.1 (1.8) 4.4 (2.2) 5.9 (2.4) 
 Animal foods 5.4 (2.3) 5.1 (2.1) 4.9 (1.9) 5 (2.1) 5 (2.1) 4.8 (2) 
NHS (1984)NHSII (1991)
Q1 (n = 14,629)Q3 (n = 13,832)Q5 (n = 15,857)Q1 (n = 20,408)Q3 (n = 16,541)Q5 (n = 17,944)
Median PDI 46 54 62 47 55 63 
 Age, yearsa 50.2 (7) 50.7 (7.2) 51.7 (7.3) 36.5 (4.8) 36.6 (4.7) 36.9 (4.5) 
 Body mass index, kg/m2 25.7 (5.1) 25.1 (4.7) 24.5 (4.4) 25.2 (5.7) 24.5 (5.2) 23.7 (4.8) 
 Body mass index at age 18 years, kg/m2 21.7 (3.2) 21.4 (3.0) 21.1 (2.8) 21.6 (3.6) 21.3 (3.3) 20.9 (3.0) 
 Weight change from age 18 years, kg 10.7 (12.3) 9.8 (11.1) 9.0 (10.5) 9.8 (12.4) 8.7 (11.2) 7.7 (10.4) 
 Height, inches 64.5 (2.4) 64.5 (2.4) 64.5 (2.4) 64.9 (2.6) 64.9 (2.6) 64.9 (2.6) 
 Self-reported African heritage (%) 1.5 1.4 1.1 2.0 1.4 1.2 
 Self-reported history of diabetes (%) 3.8 3.3 2.3 1.2 1.0 0.7 
 Family history of breast cancer (%) 7.8 8.1 8.1 5.8 6.1 6.3 
 Personal history of benign breast disease (%) 29.5 30.2 31.0 9.1 9.5 9.6 
 Age at menarche <12 years (%) 23.5 21.6 22.5 24.8 24.0 24.8 
 Oral contraceptives, ever (%) 50.5 49.3 48.3 85.6 84.7 83.4 
 Parous (%) 92.0 92.9 93.3 69.3 75.9 77.5 
 Parity, nb 3.1 (1.5) 3.2 (1.5) 3.2 (1.5) 2.1 (0.9) 2.1 (0.9) 2.2 (0.9) 
 Breastfeeding, ≤6 months (%)b 35.8 36.6 36.1 19.7 16.9 14.1 
 Postmenopausal (%) 48.4 48.6 48.8 3.1 3.2 3.4 
 Postmenopausal hormone use, never (%)c 52.9 52.7 52.8 6.1 6.9 7.8 
 Physical activity, METs-h/week 11.4 (13.2) 11.9 (13.1) 12.9 (13.4) 17.8 (24.4) 20.6 (26.8) 24.9 (31.5) 
 Alcohol intake (g/day) 8.6 (13.6) 6.7 (11) 5.8 (9.3) 3.3 (7.1) 3 (5.8) 3.2 (5.6) 
 Total energy intake (kcal/day) 1,458 (451) 1,723 (488) 2,056 (519) 1,478 (464) 1,792 (504) 2,134 (527) 
 Saturated fat (% of energy) 13.9 (2.9) 12.5 (2.3) 11.2 (2.1) 12.6 (2.5) 11.1 (2.2) 9.8 (2.1) 
 Monounsaturated fat (% of energy) 13.4 (2.6) 12.7 (2.3) 12 (2.2) 12.7 (2.6) 11.9 (2.4) 11.2 (2.3) 
 Polyunsaturated fat (% of energy) 6.5 (1.9) 6.6 (1.7) 6.8 (1.6) 5.7 (1.5) 5.6 (1.4) 5.6 (1.3) 
 Trans fat (% of energy) 1.8 (0.6) 1.9 (0.6) 1.9 (0.6) 1.7 (0.7) 1.6 (0.6) 1.5 (0.5) 
 Protein intake (% of energy) 19.7 (3.9) 17.7 (3.1) 16.2 (2.5) 21 (3.8) 19.2 (3.2) 17.6 (2.9) 
 Carbohydrate intake (% of energy) 40.5 (8) 46.7 (6.8) 51.2 (6.5) 44.8 (7.4) 50.1 (6.5) 54.7 (6.6) 
Healthy plant foods (servings/day) 
 Whole grains 0.7 (0.9) 1.1 (1.1) 1.6 (1.2) 0.9 (0.9) 1.4 (1.1) 2.1 (1.3) 
 Fruits 0.9 (0.9) 1.4 (1) 1.9 (1.1) 0.7 (0.7) 1.2 (0.9) 1.8 (1.1) 
 Vegetables 2.3 (1.4) 2.9 (1.5) 3.7 (1.8) 2.1 (1.4) 2.9 (1.8) 4.1 (2.1) 
 Nuts 0.1 (0.2) 0.2 (0.3) 0.3 (0.4) 0.1 (0.1) 0.2 (0.2) 0.3 (0.3) 
 Legumes 0.3 (0.2) 0.4 (0.3) 0.5 (0.3) 0.2 (0.2) 0.4 (0.3) 0.6 (0.4) 
 Vegetable oil 0.2 (0.3) 0.3 (0.4) 0.4 (0.4) 0.2 (0.3) 0.3 (0.4) 0.5 (0.5) 
 Tea and coffee 2.6 (1.9) 3.1 (1.9) 3.6 (2) 1.8 (1.8) 2.2 (1.9) 2.7 (2) 
Less healthy plant foods (servings/day) 
 Fruit juices 0.4 (0.6) 0.7 (0.7) 1 (0.8) 0.4 (0.6) 0.6 (0.7) 1.1 (1) 
 Refined grains 1.2 (1.1) 1.6 (1.3) 2.2 (1.5) 1.2 (0.9) 1.6 (1) 2 (1.1) 
 Potatoes 0.4 (0.3) 0.5 (0.4) 0.7 (0.4) 0.4 (0.3) 0.5 (0.4) 0.7 (0.4) 
 Sugar-sweetened beverages 0.2 (0.6) 0.3 (0.6) 0.4 (0.6) 0.4 (0.8) 0.5 (0.9) 0.6 (0.9) 
 Sweets and desserts 0.8 (0.9) 1.2 (1.2) 1.8 (1.4) 0.8 (0.9) 1.2 (1.1) 1.6 (1.2) 
Animal foods (servings/day) 
 Animal fat 0.5 (0.9) 0.4 (0.8) 0.3 (0.6) 0.2 (0.5) 0.2 (0.4) 0.1 (0.4) 
 Dairy 2 (1.4) 2 (1.4) 2 (1.3) 2.3 (1.5) 2.3 (1.5) 2.3 (1.4) 
 Eggs 0.4 (0.4) 0.3 (0.3) 0.3 (0.3) 0.2 (0.2) 0.2 (0.2) 0.2 (0.2) 
 Fish and seafood 0.3 (0.3) 0.3 (0.3) 0.3 (0.2) 0.3 (0.2) 0.3 (0.2) 0.3 (0.3) 
 Meat 1.7 (0.8) 1.7 (0.8) 1.7 (0.8) 1.6 (0.8) 1.7 (0.8) 1.6 (0.9) 
 Miscellaneous animal-based foods 0.5 (0.4) 0.4 (0.4) 0.4 (0.4) 0.4 (0.4) 0.4 (0.3) 0.4 (0.3) 
Food groups (servings/day) 
 Healthy plant foods 7.1 (3) 9.3 (3.3) 12.1 (3.7) 6 (2.9) 8.6 (3.5) 12 (4.1) 
 Less healthy plant foods 3 (1.9) 4.4 (2.3) 6.1 (2.7) 3.1 (1.8) 4.4 (2.2) 5.9 (2.4) 
 Animal foods 5.4 (2.3) 5.1 (2.1) 4.9 (1.9) 5 (2.1) 5 (2.1) 4.8 (2) 

Note: Values are means (SDs) for continuous variables and percentages for categorical variables. All variables except age are standardized to the age distribution of the study population.

Each food's intake expressed as servings/day. The SI equivalent of 1 inch is equivalent to 2.54 cm.

Abbreviations: METs-h/week, metabolic equivalent task hours per week; NHS, Nurses' Health Study; NHSII, Nurses' Health Study II; PDI, plant-based diet index; Q, quintile.

aValue is not age adjusted.

bAmong parous women only.

cAmong postmenopausal women.

In pooled multivariable-adjusted analysis (Table 2), a higher adherence to a PDI was significantly inversely associated with breast cancer (HRQ5 vs. Q1, 0.89; 95% CI, 0.84–0.95; Ptrend < 0.01). Cohort-specific analyses (Supplementary Table S1) showed similar results. Further adjustment for weight change since age 18 slightly attenuated the results (HRQ5 vs. Q1, 0.93; 95% CI, 0.87–0.99; Ptrend = 0.01). When we analyzed hPDI and uPDI separately, we found a modest inverse association between a hPDI and breast cancer incidence (HRQ5 vs. Q1, 0.89; 95% CI, 0.83–0.94; Ptrend < 0.01). Only a modest attenuation with addition of weight change (HRQ5 vs. Q1, 0.91; 95% CI, 0.86–0.97; Ptrend < 0.01) or total carotenoid intake (HRQ5 vs. Q1, 0.91; 95% CI, 0.85–0.98; Ptrend < 0.01) was identified for the hPDI. Further adjustment for dietary fiber, but not weight change or total carotenoid intake, did not further changed the scenario. Moreover, adherence to hPDI during the premenopausal period was inversely associated with both premenopausal and postmenopausal breast cancer. Similarly, adherence to hPDI after menopause was inversely associated with the risk of postmenopausal breast cancer.

Table 2.

Age-adjusted and multivariable-adjusted HRs (95% CIs) for total breast cancer according to quintiles of plant-based diet indices (PDI, hPDI, uPDI) in the NHS and NHSII.

Quintile 1Quintile 2Quintile 3Quintile 4Quintile 5PTrendHR (95% CI) per 10 units
Plant-based dietary index (PDI)  
 Median 47 51 54 58 62   
 Cases/PY 2,540/967,562 2,471/954,550 2,485/979,867 2,500/973,362 2,486/965,742   
 Age-adjusted 1.00 0.98 (0.92–1.03) 0.95 (0.90–1.01) 0.95 (0.90–1.01) 0.94 (0.89–0.99) 0.02 0.96 (0.93–0.99) 
 MV 1.00 0.96 (0.91–1.01) 0.93 (0.87–0.98) 0.92 (0.86–0.97) 0.89 (0.84–0.95) <0.01 0.92 (0.89–0.96) 
Healthful plant-based dietary index (hPDI)  
 Median 46 51 55 58 64   
 Cases/PY 2,413/971,657 2,519/959,433 2,512/969,268 2,540/976,826 2,498/963,900   
 Age-adjusted 1.00 1.02 (0.96–1.07) 0.99 (0.94–1.05) 0.98 (0.93–1.04) 0.94 (0.88–0.99) <0.01 0.97 (0.94–0.99) 
 MV 1.00 0.99 (0.94–1.05) 0.96 (0.90–1.01) 0.94 (0.88–0.99) 0.89 (0.83–0.94) <0.01 0.94 (0.91–0.97) 
Unhealthful plant-based dietary index (uPDI)  
 Median 46 51 55 59 64   
Cases/PY 2,606/971,894 2,526/966,528 2,465/971,276 2,472/962,386 2,413/968,999   
 Age-adjusted 1.00 0.99 (0.94–1.05) 0.98 (0.92–1.03) 0.99 (0.94–1.05) 0.99 (0.94–1.05) 0.80 0.99 (0.97–1.02) 
 MV 1.00 1.00 (0.94–1.05) 0.99 (0.93–1.05) 1.02 (0.96–1.08) 1.04 (0.97–1.10) 0.20 1.02 (0.99–1.05) 
Quintile 1Quintile 2Quintile 3Quintile 4Quintile 5PTrendHR (95% CI) per 10 units
Plant-based dietary index (PDI)  
 Median 47 51 54 58 62   
 Cases/PY 2,540/967,562 2,471/954,550 2,485/979,867 2,500/973,362 2,486/965,742   
 Age-adjusted 1.00 0.98 (0.92–1.03) 0.95 (0.90–1.01) 0.95 (0.90–1.01) 0.94 (0.89–0.99) 0.02 0.96 (0.93–0.99) 
 MV 1.00 0.96 (0.91–1.01) 0.93 (0.87–0.98) 0.92 (0.86–0.97) 0.89 (0.84–0.95) <0.01 0.92 (0.89–0.96) 
Healthful plant-based dietary index (hPDI)  
 Median 46 51 55 58 64   
 Cases/PY 2,413/971,657 2,519/959,433 2,512/969,268 2,540/976,826 2,498/963,900   
 Age-adjusted 1.00 1.02 (0.96–1.07) 0.99 (0.94–1.05) 0.98 (0.93–1.04) 0.94 (0.88–0.99) <0.01 0.97 (0.94–0.99) 
 MV 1.00 0.99 (0.94–1.05) 0.96 (0.90–1.01) 0.94 (0.88–0.99) 0.89 (0.83–0.94) <0.01 0.94 (0.91–0.97) 
Unhealthful plant-based dietary index (uPDI)  
 Median 46 51 55 59 64   
Cases/PY 2,606/971,894 2,526/966,528 2,465/971,276 2,472/962,386 2,413/968,999   
 Age-adjusted 1.00 0.99 (0.94–1.05) 0.98 (0.92–1.03) 0.99 (0.94–1.05) 0.99 (0.94–1.05) 0.80 0.99 (0.97–1.02) 
 MV 1.00 1.00 (0.94–1.05) 0.99 (0.93–1.05) 1.02 (0.96–1.08) 1.04 (0.97–1.10) 0.20 1.02 (0.99–1.05) 

Note: Multivariable model (MV) stratified by age in months, calendar year, and cohort, adjusted for race (non-Hispanic Caucasian, African-American, Asian-American, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45–49, 50–52, 53+), postmenopausal hormone use (never user, past user, current user—estrogen only for <5 years, current user—estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60–1.64, 1.65–1.69, 1.70–1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5–9, 10–14, 15+ g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0–21.9, 22.0–23.9, 24.0–26.9, ≥27.0), and neighborhood-based socioeconomic status indicator (continuous).

We observed a significant heterogeneity by ER status in the hPDI (Pheterogeneity = 0.01) and the uPDI (Pheterogeneity = 0.01; Table 3). We detected an inverse association between a higher hPDI and ER-negative breast cancer (HRQ5 vs. Q1, 0.77; 95% CI, 0.65–0.90; Ptrend < 0.01) and a positive association with uPDI (HRQ5 vs. Q1, 1.28; 95% CI, 1.08–1.51; Ptrend < 0.01). This association was further observed when we performed multivariable spline analysis (Fig. 1). Further adjustment for total carotenoids, dietary fiber, and fruits and vegetables did not substantively alter the effect estimates. For the uPDI, adjustment for dietary fiber or total carotenoid intake attenuated the association and made it not significant. Although we observed no significant heterogeneity by molecular subtypes (Table 3), each 10-unit increase in the hPDI was associated with lower risk of HER2-enriched (HR, 0.80; 95% CI, 0.65–0.99) and basal-like tumors (HR, 0.79; 95% CI, 0.65–0.96).

Table 3.

Multivariable-adjusted HRs (95% CIs) for the association between quintiles (Q) of cumulatively updated PDI, hPDI, and uPDI and breast cancer tumor subtypes in the NHS and NHSII.

Quintile 1Quintile 2Quintile 3Quintile 4Quintile 5PTrendPheterogeneityHR (95% CI) per 10 units
ESTROGEN RECEPTOR STATUS (NHS with follow-up of 1984–2016 and NHSII with follow-up of 1991–2017) 
Plant-based dietary index (PDI) 
ER+ Cases/PY 1,609/968,461 1,599/955,406 1,659/980,656 1,636/974,166 1,632/966,567    
 MV 1.00 0.97 (0.91–1.04) 0.97 (0.90–1.04) 0.94 (0.88–1.02) 0.93 (0.86–1.00) 0.05  0.95 (0.91–0.99) 
ER Cases/PY 368/969,664 358/956,537 334/981,914 345/975,387 345/967,764    
 MV 1.00 0.96 (0.82–1.11) 0.86 (0.74–1.01) 0.87 (0.74–1.02) 0.85 (0.72–1.00) 0.03 0.25a 0.88 (0.80–0.96) 
Healthful plant-based dietary index (hPDI) 
ER+ Cases/PY 1,551/972,520 1,577/960,328 1,671/970,065 1,656/977,670 1,680/964,673    
 MV 1.00 0.96 (0.89–1.03) 0.98 (0.91–1.05) 0.94 (0.87–1.01) 0.91 (0.85–0.99) 0.02  0.96 (0.92–1.00) 
ER Cases/PY 377/973,605 390/961,476 325/971,282 336/978,887 322/966,017    
 MV 1.00 0.99 (0.86–1.15) 0.81 (0.70–0.95) 0.82 (0.70–0.96) 0.77 (0.65–0.90) <0.01 0.01a 0.85 (0.78–0.92) 
Unhealthful plant-based dietary index (uPDI) 
ER+ Cases/PY 1,750/972,733 1,690/967,320 1,577/972,116 1,594/963,230 1,524/969,856    
 MV 1.00 0.99 (0.93–1.06) 0.95 (0.88–1.02) 0.99 (0.92–1.06) 1.00 (0.93–1.08) 0.99  0.99 (0.96–1.03) 
ER Cases/PY 305/974,085 342/968,560 377/973,277 367/964,382 359/970,963    
 MV 1.00 1.17 (1.00–1.36) 1.29 (1.10–1.50) 1.26 (1.07–1.48) 1.28 (1.08–1.51) <0.01 0.01a 1.12 (1.04–1.21) 
MOLECULAR SUBTYPES (NHS with follow-up of 1984–2006 and NHSII with follow-up of 1991–2005) 
Plant-based dietary index (PDI) 
Luminal Ab Cases/PY 571/661,577 573/648,175 611/670,335 592/667,358 572/658,149    
 MV 1.00 0.99 (0.88–1.11) 1.00 (0.89–1.12) 0.95 (0.84–1.08) 0.91 (0.80–1.03) 0.11  0.96 (0.89–1.03) 
Luminal Bb Cases/PY 271/661,825 228/648,509 267/670,634 270/667,639 281/658,411    
 MV 1.00 0.86 (0.72–1.03) 0.98 (0.82–1.17) 1.00 (0.84–1.20) 1.04 (0.86–1.26) 0.39  1.00 (0.89–1.12) 
HER-2b Cases/PY 46/662,033 47/648657 61/670,826 57/667,847 59/658,604    
 MV 1.00 0.97 (0.64–1.47) 1.22 (0.82–1.82) 1.10 (0.73–1.66) 1.10 (0.72–1.69) 0.57  0.99 (0.77–1.26) 
Basal-likeb Cases/PY 65/662,011 73/648,631 53/670,829 54/667,847 54/658,617    
 MV 1.00 1.10 (0.78–1.55) 0.72 (0.50–1.05) 0.70 (0.48–1.03) 0.69 (0.46–1.02) 0.01 0.04a 0.77 (0.62–0.96) 
Healthful plant-based dietary index (hPDI) 
Luminal Ab Cases/PY 561/664,554 523/647,388 626/667,302 604/668,016 605/658,333    
 MV 1.00 0.88 (0.78–1.00) 0.99 (0.88–1.11) 0.93 (0.82–1.05) 0.88 (0.77–1.00) 0.12  0.95 (0.89–1.01) 
Luminal Bb Cases/PY 254/664,818 247/647610 265/667,608 265/668,332 286/658,649    
 MV 1.00 0.94 (0.79–1.13) 0.96 (0.80–1.15) 0.96 (0.80–1.15) 1.00 (0.83–1.20) 0.99  1.01 (0.92–1.11) 
HER-2b Cases/PY 56/664,998 66/647,780 62/667,794 41/668,527 45/658,868    
 MV 1.00 1.10 (0.76–1.58) 1.05 (0.72, 1.53) 0.69 (0.45–1.06) 0.72 (0.47–1.11) 0.04  0.80 (0.65–0.99) 
Basal-likeb Cases/PY 69/664,993 56/647,785 60/667,792 62/668,505 52/658,861    
 MV 1.00 0.82 (0.57–1.17) 0.85 (0.59–1.21) 0.86 (0.60–1.24) 0.70 (0.47–1.04) 0.13 0.19a 0.79 (0.65–0.96) 
Unhealthful plant-based dietary index (uPDI) 
Luminal Ab Cases/PY 621/663,923 611/657,264 576/664,708 575/656,569 536/663,131    
 MV 1.00 1.02 (0.91–1.15) 0.99 (0.88–1.11) 1.02 (0.90–1.15) 0.99 (0.87–1.13) 0.91  0.98 (0.92–1.04) 
Luminal Bb Cases/PY 278/664,221 279/657,567 248/665,003 239/656,848 273/663,379    
 MV 1.00 1.01 (0.85–1.19) 0.90 (0.75–1.07) 0.88 (0.73–1.05) 1.03 (0.86–1.24) 0.81  0.98 (0.89–1.07) 
HER-2b Cases/PY 46/664,428 45/657,782 64/665,164 54/657,032 61/663,561    
 MV 1.00 0.97 (0.64–1.47) 1.43 (0.97–2.12) 1.22 (0.81–1.85) 1.49 (0.98–2.27) 0.04  1.20 (0.98–1.46) 
Basal-likeb Cases/PY 59/664,419 60/657,757 59/665,173 57/657,036 64/663,551    
 MV 1.00 1.04 (0.72–1.49) 1.01 (0.70–1.46) 0.98 (0.67–1.44) 1.13 (0.77–1.66) 0.66 0.21a 1.09 (0.91–1.31) 
Quintile 1Quintile 2Quintile 3Quintile 4Quintile 5PTrendPheterogeneityHR (95% CI) per 10 units
ESTROGEN RECEPTOR STATUS (NHS with follow-up of 1984–2016 and NHSII with follow-up of 1991–2017) 
Plant-based dietary index (PDI) 
ER+ Cases/PY 1,609/968,461 1,599/955,406 1,659/980,656 1,636/974,166 1,632/966,567    
 MV 1.00 0.97 (0.91–1.04) 0.97 (0.90–1.04) 0.94 (0.88–1.02) 0.93 (0.86–1.00) 0.05  0.95 (0.91–0.99) 
ER Cases/PY 368/969,664 358/956,537 334/981,914 345/975,387 345/967,764    
 MV 1.00 0.96 (0.82–1.11) 0.86 (0.74–1.01) 0.87 (0.74–1.02) 0.85 (0.72–1.00) 0.03 0.25a 0.88 (0.80–0.96) 
Healthful plant-based dietary index (hPDI) 
ER+ Cases/PY 1,551/972,520 1,577/960,328 1,671/970,065 1,656/977,670 1,680/964,673    
 MV 1.00 0.96 (0.89–1.03) 0.98 (0.91–1.05) 0.94 (0.87–1.01) 0.91 (0.85–0.99) 0.02  0.96 (0.92–1.00) 
ER Cases/PY 377/973,605 390/961,476 325/971,282 336/978,887 322/966,017    
 MV 1.00 0.99 (0.86–1.15) 0.81 (0.70–0.95) 0.82 (0.70–0.96) 0.77 (0.65–0.90) <0.01 0.01a 0.85 (0.78–0.92) 
Unhealthful plant-based dietary index (uPDI) 
ER+ Cases/PY 1,750/972,733 1,690/967,320 1,577/972,116 1,594/963,230 1,524/969,856    
 MV 1.00 0.99 (0.93–1.06) 0.95 (0.88–1.02) 0.99 (0.92–1.06) 1.00 (0.93–1.08) 0.99  0.99 (0.96–1.03) 
ER Cases/PY 305/974,085 342/968,560 377/973,277 367/964,382 359/970,963    
 MV 1.00 1.17 (1.00–1.36) 1.29 (1.10–1.50) 1.26 (1.07–1.48) 1.28 (1.08–1.51) <0.01 0.01a 1.12 (1.04–1.21) 
MOLECULAR SUBTYPES (NHS with follow-up of 1984–2006 and NHSII with follow-up of 1991–2005) 
Plant-based dietary index (PDI) 
Luminal Ab Cases/PY 571/661,577 573/648,175 611/670,335 592/667,358 572/658,149    
 MV 1.00 0.99 (0.88–1.11) 1.00 (0.89–1.12) 0.95 (0.84–1.08) 0.91 (0.80–1.03) 0.11  0.96 (0.89–1.03) 
Luminal Bb Cases/PY 271/661,825 228/648,509 267/670,634 270/667,639 281/658,411    
 MV 1.00 0.86 (0.72–1.03) 0.98 (0.82–1.17) 1.00 (0.84–1.20) 1.04 (0.86–1.26) 0.39  1.00 (0.89–1.12) 
HER-2b Cases/PY 46/662,033 47/648657 61/670,826 57/667,847 59/658,604    
 MV 1.00 0.97 (0.64–1.47) 1.22 (0.82–1.82) 1.10 (0.73–1.66) 1.10 (0.72–1.69) 0.57  0.99 (0.77–1.26) 
Basal-likeb Cases/PY 65/662,011 73/648,631 53/670,829 54/667,847 54/658,617    
 MV 1.00 1.10 (0.78–1.55) 0.72 (0.50–1.05) 0.70 (0.48–1.03) 0.69 (0.46–1.02) 0.01 0.04a 0.77 (0.62–0.96) 
Healthful plant-based dietary index (hPDI) 
Luminal Ab Cases/PY 561/664,554 523/647,388 626/667,302 604/668,016 605/658,333    
 MV 1.00 0.88 (0.78–1.00) 0.99 (0.88–1.11) 0.93 (0.82–1.05) 0.88 (0.77–1.00) 0.12  0.95 (0.89–1.01) 
Luminal Bb Cases/PY 254/664,818 247/647610 265/667,608 265/668,332 286/658,649    
 MV 1.00 0.94 (0.79–1.13) 0.96 (0.80–1.15) 0.96 (0.80–1.15) 1.00 (0.83–1.20) 0.99  1.01 (0.92–1.11) 
HER-2b Cases/PY 56/664,998 66/647,780 62/667,794 41/668,527 45/658,868    
 MV 1.00 1.10 (0.76–1.58) 1.05 (0.72, 1.53) 0.69 (0.45–1.06) 0.72 (0.47–1.11) 0.04  0.80 (0.65–0.99) 
Basal-likeb Cases/PY 69/664,993 56/647,785 60/667,792 62/668,505 52/658,861    
 MV 1.00 0.82 (0.57–1.17) 0.85 (0.59–1.21) 0.86 (0.60–1.24) 0.70 (0.47–1.04) 0.13 0.19a 0.79 (0.65–0.96) 
Unhealthful plant-based dietary index (uPDI) 
Luminal Ab Cases/PY 621/663,923 611/657,264 576/664,708 575/656,569 536/663,131    
 MV 1.00 1.02 (0.91–1.15) 0.99 (0.88–1.11) 1.02 (0.90–1.15) 0.99 (0.87–1.13) 0.91  0.98 (0.92–1.04) 
Luminal Bb Cases/PY 278/664,221 279/657,567 248/665,003 239/656,848 273/663,379    
 MV 1.00 1.01 (0.85–1.19) 0.90 (0.75–1.07) 0.88 (0.73–1.05) 1.03 (0.86–1.24) 0.81  0.98 (0.89–1.07) 
HER-2b Cases/PY 46/664,428 45/657,782 64/665,164 54/657,032 61/663,561    
 MV 1.00 0.97 (0.64–1.47) 1.43 (0.97–2.12) 1.22 (0.81–1.85) 1.49 (0.98–2.27) 0.04  1.20 (0.98–1.46) 
Basal-likeb Cases/PY 59/664,419 60/657,757 59/665,173 57/657,036 64/663,551    
 MV 1.00 1.04 (0.72–1.49) 1.01 (0.70–1.46) 0.98 (0.67–1.44) 1.13 (0.77–1.66) 0.66 0.21a 1.09 (0.91–1.31) 

Note: Multivariable model (MV) stratified by cohort, age in months, and calendar year, adjusted for race (non-Hispanic Caucasian, African, Asian, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45–49, 50–52, 53+), postmenopausal hormone use (never user, past user, current user—estrogen only for <5 years, current user—estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60–1.64, 1.65–1.69, 1.70–1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5–9, 10–14, 15+ g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0–21.9, 22.0–23.9, 24.0–26.9, ≥27.0), and socioeconomic status (continuous).

Abbreviations: ER, estrogen receptor; HER, human epidermal growth factor receptor 2; PY, person-years.

aFor testing heterogeneity by subtype, we used the Lunn-McNeil approach, for multivariable model (MV).

bBecause of smaller sample sizes in analyses, to ensure that models would run, covariate categorizations were simplified.

Figure 1.

Multivariable spline analysis of the association between adherence to healthful and unhealthful PDIs and risk of incident breast cancer in the NHS (1984–2016) and NHSII (1991–2017). ER, estrogen receptor. Model stratified by cohort, age in months, and 2-year period at risk, adjusted for race (non-Hispanic Caucasian, African, Asian, Hispanic Caucasian), cumulatively updated total caloric intake (kcal/day, quintiles), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45–49, 50–52, 53+), postmenopausal hormone use (never user, past user, current user–estrogen only for <5 years, current user—estrogen only for ≥5 years, current estrogen + progestin user for <5 years, current estrogen + progestin user for ≥5 years, current user of other types), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast disease (yes or no), body mass index at age 18 years (<20.0, 20.0–21.9, 22.0–23.9, 24.0–26.9, ≥27.0), oral contraceptives use (never/ever), height (<1.60, 1.60–1.64, 1.65–1.69, 1.70–1.74, 1.75+ m), cumulatively updated alcohol intake (0, <5, 5–9, 10–14, 15+ g/day), physical activity (linear METs-h/week), and neighborhood-based socioeconomic status indicator (continuous). For ER-negative breast cancer cases, the P values for test of curvature for hPDI = 0.95 and for uPDI = 0.29. For ER-positive cases, the P values for test of curvature for hPDI = 0.35 and uPDI = 0.15.

Figure 1.

Multivariable spline analysis of the association between adherence to healthful and unhealthful PDIs and risk of incident breast cancer in the NHS (1984–2016) and NHSII (1991–2017). ER, estrogen receptor. Model stratified by cohort, age in months, and 2-year period at risk, adjusted for race (non-Hispanic Caucasian, African, Asian, Hispanic Caucasian), cumulatively updated total caloric intake (kcal/day, quintiles), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45–49, 50–52, 53+), postmenopausal hormone use (never user, past user, current user–estrogen only for <5 years, current user—estrogen only for ≥5 years, current estrogen + progestin user for <5 years, current estrogen + progestin user for ≥5 years, current user of other types), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast disease (yes or no), body mass index at age 18 years (<20.0, 20.0–21.9, 22.0–23.9, 24.0–26.9, ≥27.0), oral contraceptives use (never/ever), height (<1.60, 1.60–1.64, 1.65–1.69, 1.70–1.74, 1.75+ m), cumulatively updated alcohol intake (0, <5, 5–9, 10–14, 15+ g/day), physical activity (linear METs-h/week), and neighborhood-based socioeconomic status indicator (continuous). For ER-negative breast cancer cases, the P values for test of curvature for hPDI = 0.95 and for uPDI = 0.29. For ER-positive cases, the P values for test of curvature for hPDI = 0.35 and uPDI = 0.15.

Close modal

In ancillary analyses (Supplementary Table S2), we entered variables for the three food categories together into the fully adjusted model in place of the indices. We found an inverse association between extreme quintiles of healthy plant-based foods and ER-negative breast cancer (HR, 0.74; 95% CI, 0.61–0.88; Ptrend < 0.01).

Stratified analysis showed no significant effect modification by menopausal status, physical activity, or current BMI for the diet indices in relation to ER-negative breast cancer (Fig. 2). Associations were similar by menopausal status [(premenopausal (HR, 0.87; 95% CI, 0.74–1.02) and postmenopausal (HR, 0.86; 95% CI, 0.78–0.95)]. Among postmenopausal women, there was a suggestion of stronger associations among never and past users of postmenopausal hormones (HRQ5 vs. Q1, 0.69; 95% CI, 0.54–0.90; Ptrend < 0.01) than current users. Nonetheless, Pinteraction was nonsignificant.

Figure 2.

Pooled HRs of estrogen receptor negative breast cancer per 10-unit increment in the three dietary indices (PDI, hPDI, and uPDI) across subgroups (physical activity, current BMI, and menopausal status). The HRs and P values for women were obtained after combining all two cohorts (NHS; NHSII). Stratified by age in months, cohort and calendar year, adjusted for race (non-Hispanic Caucasian, African, Asian, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45–49, 50–52, 53+), postmenopausal hormone use (never user, past user, current user—estrogen only for <5 years, current user—estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60–1.64, 1.65–1.69, 1.70–1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5–9, 10–14, 15+ g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0–21.9, 22.0–23.9, 24.0–26.9, ≥27.0), and socioeconomic status (continuous). For the analysis of BMI, we further adjusted for current body mass index (linear, kg/m2). To test whether the PDI, hPDI, uPDI, and breast cancer association differed by current BMI, physical activity, or menopausal status, we added interaction terms and used the Wald test. BMI, body mass index; CI, confidence interval; MET, metabolic equivalent task.

Figure 2.

Pooled HRs of estrogen receptor negative breast cancer per 10-unit increment in the three dietary indices (PDI, hPDI, and uPDI) across subgroups (physical activity, current BMI, and menopausal status). The HRs and P values for women were obtained after combining all two cohorts (NHS; NHSII). Stratified by age in months, cohort and calendar year, adjusted for race (non-Hispanic Caucasian, African, Asian, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45–49, 50–52, 53+), postmenopausal hormone use (never user, past user, current user—estrogen only for <5 years, current user—estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60–1.64, 1.65–1.69, 1.70–1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5–9, 10–14, 15+ g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0–21.9, 22.0–23.9, 24.0–26.9, ≥27.0), and socioeconomic status (continuous). For the analysis of BMI, we further adjusted for current body mass index (linear, kg/m2). To test whether the PDI, hPDI, uPDI, and breast cancer association differed by current BMI, physical activity, or menopausal status, we added interaction terms and used the Wald test. BMI, body mass index; CI, confidence interval; MET, metabolic equivalent task.

Close modal

We evaluated the extent to which the inverse association with higher PDI and hPDI may be mediated by less weight gain from age 18. The calculated mediation proportion was 11.0% (95% CI = 3.4–30.2; P < 0.01), indicating that less weight gain could statistically explain 11.0% of the inverse association with PDI. Moreover, the proportion of hPDI effect mediated by weight change since age 18 was 7.3% (95% CI = 3.0–16.5; P < 0.01). Results from latency analysis (Supplementary Table S3) indicated the highest versus the lowest adherence to a hPDI 0–4 and 4–8 years before diagnosis was associated with lower risk of ER-negative breast cancer [(HRQ5 vs. Q1, 0.85; 95% CI, 0.72–1.01; Ptrend = 0.01); (HRQ5 vs. Q1, 0.80; 95% CI, 0.66–0.96; Ptrend = 0.01)].

In two large prospective cohort studies in the United States, we found that a higher hPDI score, a measure of adherence to a high-quality plant-based diet, was associated with lower risk of total breast cancer, independent of weight change, total carotenoid intake and dietary fiber. The association was most evident in relation to ER-negative tumors for which non-hormonal exposures may be most important (29–31). Less weight gain could statistically explain 7.3% of the inverse association with hPDI.

Only few prospective studies have assessed the association between hPDI and uPDI defined by Satija and colleagues, (12, 32–37). Because of the impracticality of assessment of a pure vegetarian diet in these cohorts and the fact that it may not easily be embraced by many individuals, consuming preferentially plant-derived foods would be a more comprehensible message. In this context, these scores are designed to understand common dietary patterns that incorporates a range of progressively increasing proportions of plant foods and accompanying reductions in animal-foods.

Previous studies that examined associations between these plant-based diets and breast cancer are not consistent in the literature (6, 13). In the SUN project, a moderate, but not greater, adherence to a provegetarian dietary pattern, was associated with a decreased risk of breast cancer, otherwise no particular associations were found when assessing healthful and unhealthful provegetarian dietary patterns separately (6). Furthermore, a higher pro plant-based dietary score was associated with decreased risks of overall cancer though no association was found for breast cancer (13) which may partly be due to the restricted number of cases (n = 487 breast cancer cases, only 13.6% representing ER‐negative/PR‐negative and ER-negative/PR-positive).

In 2007, the World Cancer Research Fund supported that there was insufficient evidence to make a judgment about the association between dietary patterns and the risk of breast cancer (38). Subsequently, in 2010, a systematic review and meta-analysis showed that a Prudent dietary pattern characterized by high intakes of fruit, vegetables, whole grains, low-fat dairy products, fish, and poultry, was linked to a 11% breast cancer risk reduction (39). In a recent systematic review and meta-analysis of 32 observational studies (14), a Western and a Prudent diet patterns were associated with a 14% increased and a 18% reduced risk of breast cancer, respectively.

Hormone receptor status in an important diagnostic and prognostic characteristic of breast tumor and, therefore, deserves consideration. In fact, estrogen exposure is one of the strongest risk factors for breast cancer, but it may have less impact on ER-negative tumors than ER-positive tumors. In fact, in ER-positive tumors, any potential influence of dietary factors may be hard to detect due to the strong influence of hormonal factors. Contrarily, in ER-negative tumors, risk factors such as diet may have a relatively greater influence and be more easily detected. In previous analyses within the NHS, the Prudent dietary pattern, the Dietary Approaches to Stop Hypertension (40, 41), the Mediterranean Diet (42), and a diabetes risk reduction diet (43) have been significantly associated with a decreased risk of ER-negative breast tumors. Thus, results are in close agreement with previous findings, which could be expected, as most healthy plant foods positively weighted in the hPDI [e.g., vegetables, fruits (15, 44, 45), whole grains (46), olive oil (47), or coffee (43, 48)] have been associated with lower risk of breast cancer in prospective cohort studies, including our own. Moreover, higher intakes of dietary fiber (49, 50) and carotenoids, particularly α-carotene, β-carotene, and lutein/zeaxanthin as shown in a pooled analysis of 18 prospective cohort studies, were inversely associated with risk of ER-negative breast cancer (51). A healthful plant-based diet may reduce the development of ER-negative breast cancer through several mechanisms. The abundance of antioxidants, vitamins, and dietary fiber may have antioxidant, anti-inflammatory, and antiproliferative activity (52, 53), and may neutralize free radicals, and prevent DNA damage (54). In addition, plant-based dietary patterns have been shown to improve insulin resistance and glycemic control (55) which are more strongly associated with ER-negative than with ER-positive breast cancer (54). However, adjustment for carotenoid intake, dietary fiber or fruits and vegetables slightly attenuated the observed associations with regard to hPDI adherence and ER-negative breast tumors, indicating that other constituents in plant-based foods also account for the observed findings. In this sense, other phytochemicals present in fruits, vegetables, whole grains, and legumes, such as flavonoids or other phenolic compounds, or an interaction among several phytochemicals, could be responsible for the observed association (56). Future studies should explore the associations between additional phytochemical components and breast cancer risk. On the other hand, less healthy plant-based foods positively weighted in the uPDI [refined grains, pastries (57), sugary drinks (58–60) or processed foods (61)] have been associated with higher risk of breast cancer in other studies.

In our analyses, the consistent opposite associations of hPDI and uPDI also strengthened the value of considering plant-based dietary quality and increasing intake of healthy plant foods while lessening consumption of less healthy plant foods. This has important public health implications as future nutrition policies and public health efforts to lower breast cancer risk (and other chronic diseases) should take the quality of plant foods into consideration. A healthful version of a PDI represents many advantages such as the representation of a healthy plant-based diet without complete exclusion of animal foods. In a previous analysis from the NHS and the Health Professionals Follow-Up Study (37) where changes in plant-based diet quality and mortality were evaluated, a 10-point increase in hPDI could be reached by increasing healthy plant foods (i.e., fruits, vegetables, and whole grains) by about 3 servings/day and decreasing less healthy plant foods (i.e., refined grains and sugary beverages) and some animal foods (i.e., processed meat) by approximately 2 servings/day. Such an approach is preferable because it is flexible and allows individuals to make gentle changes to their diets. Besides, excluding all animal foods might not be convenient for all populations as there is limited suggestive evidence that moderate intakes of some animal foods such as fish and dairy may be associated with lower breast cancer risk (62).

Strengths and limitations of study

The strengths of this study include the two large prospective cohorts, the large sample size, the long follow-up period, and low attrition. Detailed collection of updated dietary, lifestyle, and medical information (such as tissue information for the determination of molecular subtypes) over several decades permitted the evaluation of quality of plant-based diets and the adjustment for a widely recognized confounders of these associations. Moreover, the plant‐based diet indices we used in this study are easily reproducible in other cohort studies. However, several limitations should be acknowledged. First, generalizability may be limited because participants in our study were all health professionals and were predominantly white. However, the high educational status of our participants should be considered as a strength because it allowed us to gather detailed and accurate information on diet, lifestyle, and other health variables and minimize confounding by socioeconomic status. Second, while we controlled for a wide variety of lifestyle factors and excluded participants with cancer, or implausible energy intakes, the possibility of residual confounding cannot be excluded because of the observational nature of the study. Third, our dietary assessment was based on self-reported questionnaires, which inevitably produce measurement errors; however, these would likely be nondifferential in relation to risk of breast cancer and therefore, have caused underestimation of associations. Nonetheless, the FFQs used in the current study were extensively validated against diet records and biomarkers (63, 64). Fourth, one challenge in examining ER-negative breast cancer in epidemiologic studies is that it only accounts for 15% to 20% of breast cancer (65). Thus, further analyses exploring molecular subtypes beyond ER status are warranted given our limited power.

Conclusions

A healthful plant-based diet was significantly associated with lower risk of total breast cancer, independently of total carotenoid intake, dietary fiber, and weight change, and specifically for ER-negative tumors. Our results support dietary guidelines that emphasize increasing intake of healthy plant-based foods for breast cancer prevention. While there is some mechanistic support for the associations with breast cancer subtypes, further confirmatory studies are warranted.

A. Romanos-Nanclares reports grants from Spanish Association Against Cancer during the conduct of the study. B.A. Rosner reports grants from NIH during the conduct of the study. A.H. Eliassen reports grants from NIH during the conduct of the study. No disclosures were reported by the other authors.

A. Romanos-Nanclares: Data curation, software, formal analysis, investigation, methodology, writing–original draft, writing–review and editing. W.C. Willett: Conceptualization, resources, software, supervision, funding acquisition, validation, methodology, writing–review and editing. B.A. Rosner: Conceptualization, resources, data curation, software, formal analysis, supervision, investigation, methodology, writing–review and editing. L.C. Collins: Supervision, investigation, writing–review and editing. F.B. Hu: Conceptualization, resources, data curation, supervision, investigation, methodology, writing–review and editing. E. Toledo: Supervision, investigation, writing–review and editing. A.H. Eliassen: Conceptualization, resources, data curation, software, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–review and editing.

We would like to thank the participants of the NHS and NHSII and the following state cancer registries for their help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, WY. The authors assume full responsibility for analyses and interpretation of these data.

This study was supported by grants UM1 CA186107, U01 CA176726, P01 CA87969, and R01 CA50385 from the NIH, and the Breast Cancer Research Foundation. A. Romanos-Nanclares was supported by a fellowship from the Spanish Association Against Cancer Scientific Foundation (FC AECC).

The funding sources did not participate in the design or conduct of the study; collection, management, analysis or interpretation of the data; or preparation, review, or approval of the article.

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.

1.
Hemler
EC
,
Hu
FB
. 
Plant-based diets for personal, population, and planetary health
.
Adv Nutr
2019
;
10
:
S275
83
.
2.
Huang
T
,
Yang
B
,
Zheng
J
,
Li
G
,
Wahlqvist
ML
,
Li
D
. 
Cardiovascular disease mortality and cancer incidence in vegetarians: a meta-analysis and systematic review
.
Ann Nutr Metab
2012
;
60
:
233
40
.
3.
Kwok
CS
,
Umar
S
,
Myint
PK
,
Mamas
MA
,
Loke
YK
. 
Vegetarian diet, Seventh Day Adventists and risk of cardiovascular mortality: a systematic review and meta-analysis
.
Int J Cardiol
2014
;
176
:
680
6
.
4.
Tonstad
S
,
Stewart
K
,
Oda
K
,
Batech
M
,
Herring
RP
,
Fraser
GE
. 
Vegetarian diets and incidence of diabetes in the Adventist Health Study-2
.
Nutr Metab Cardiovasc Dis
2013
;
23
:
292
9
.
5.
Yokoyama
Y
,
Nishimura
K
,
Barnard
ND
,
Takegami
M
,
Watanabe
M
,
Sekikawa
A
, et al
Vegetarian diets and blood pressure ameta-analysis
.
JAMA Intern Med
2014
;
174
:
577
87
.
6.
Romanos-Nanclares
A
,
Toledo
E
,
Sánchez-Bayona
R
,
Sánchez-Quesada
C
,
Martínez-González
,
Gea
A
. 
Healthful and unhealthful provegetarian food patterns and the incidence of breast cancer: results from a Mediterranean cohort
.
Nutrition
2020
;
70–80
:
110884
.
7.
Willett
W
,
Rockström
J
,
Loken
B
,
Springmann
M
,
Lang
T
,
Vermeulen
S
, et al
Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from sustainable food systems
.
Lancet
2019
;
393
:
447
92
.
8.
Leitzmann
C
. 
Vegetarian nutrition: past, present, future
.
Am J Clin Nutr
2014
;
100
:
496S
502S
.
9.
Allès
B
,
Baudry
J
,
Méjean
C
,
Touvier
M
,
Péneau
S
,
Hercberg
S
, et al
Comparison of sociodemographic and nutritional characteristics between self-reported vegetarians, vegans, and meat-eaters from the nutrinet-santé study
.
Nutrients
2017
;
9
:
1023
.
10.
Martínez-González
MA
,
Sánchez-Tainta
A
,
Corella
D
,
Salas-Salvadó
J
,
Ros
E
,
Arós
F
, et al
A provegetarian food pattern and reduction in total mortality in the Prevención con Dieta Mediterránea (PREDIMED) study
.
Am J Clin Nutr
2014
;
100
:
320S
8S
.
11.
Satija
A
,
Hu
FB
. 
Plant-based diets and cardiovascular health
.
Trends Cardiovasc Med
2018
;
28
:
437
41
.
12.
Satija
A
,
Bhupathiraju
SN
,
Rimm
EB
,
Spiegelman
D
,
Chiuve
SE
,
Borgi
L
, et al
Plant-based dietary patterns and incidence of type 2 diabetes in US men and women: results from Three Prospective Cohort Studies
.
PLoS Med
2016
;
13
:
e1002039
.
13.
Kane-Diallo
A
,
Srour
B
,
Sellem
L
,
Deschasaux
M
,
Latino-Martel
P
,
Hercberg
S
, et al
Association between a pro plant-based dietary score and cancer risk in the prospective NutriNet-santé cohort
.
Int J Cancer
2018
;
143
:
2168
76
.
14.
Xiao
Y
,
Xia
J
,
Li
L
,
Ke
Y
,
Cheng
J
,
Xie
Y
, et al
Associations between dietary patterns and the risk of breast cancer: a systematic review and meta-analysis of observational studies
.
Breast Cancer Res
2019
;
21
:
16
.
15.
Farvid
MS
,
Chen
WY
,
Rosner
BA
,
Tamimi
RM
,
Willett
WC
,
Eliassen
AH
. 
Fruit and vegetable consumption and breast cancer incidence: repeated measures over 30 years of follow-up
.
Int J Cancer
2019
;
133
:
1496
510
.
16.
Buja
A
,
Pierbon
M
,
Lago
L
,
Grotto
G
,
Baldo
V
. 
Breast cancer primary prevention and diet: an umbrella review
.
Int J Environ Res Public Health
2020
;
17
:
4731
.
17.
Tamimi
RM
,
Baer
HJ
,
Marotti
J
,
Galan
M
,
Galaburda
L
,
Fu
Y
, et al
Comparison of molecular phenotypes of ductal carcinoma in situ and invasive breast cancer
.
Breast Cancer Res
2008
;
10
:
R67
.
18.
Nielsen
TO
,
Hsu
FD
,
Jensen
K
,
Cheang
M
,
Karaca
G
,
Hu
Z
, et al
Immunohistochemical and clinical characterization of the basal like subtype of invasive breast carcinoma
.
Clin Cancer Res
2004
;
10
:
5367
74
.
19.
Cheang
MC
,
Voduc
D
,
Bajdik
C
,
Leung
S
,
McKinney
S
,
Chia
SK
, et al
Basal-like breast cancer defined by five biomarkers has superior prognostic value than triple-negative phenotype
.
Clin Cancer Res
2008
;
14
:
1368
76
.
20.
Cheang
MC
,
Chia
SK
,
Voduc
D
,
Gao
D
,
Leung
S
,
Snider
J
, et al
Ki67 index, HER2 status, and prognosis of patients with luminal B breast cancer
.
J Natl Cancer Inst
2009
;
101
:
736
50
.
21.
Van de Rijn
M
,
Perou
CM
,
Tibshirani
R
,
Haas
P
,
Kallioniemi
O
,
Kononen
J
, et al
Expression of cytokeratins 17 and 5 identifies a group of breast carcinomas with poor clinical outcome
.
Am J Pathol
2002
;
161
:
1991
6
.
22.
Brenton
JD
,
Carey
LA
,
Ahmed
AA
,
Caldas
C
. 
Molecular classification and molecular forecasting of breast cancer: ready for clinical application?
J Clin Oncol
2005
;
23
:
7350
60
.
23.
Abd El-Rehim
DM
,
Pinder
SE
,
Paish
CE
,
Bell
J
,
Blamey
RW
,
Robertson
JF
, et al
Expression of luminal and basal cytokeratins in human breast carcinoma
.
J Pathol
2004
;
203
:
661
71
.
24.
Lunn
M
,
McNeil
D
. 
Applying cox regression to competing risks
.
Biometrics
1995
;
51
:
524
32
.
25.
Lin
DY
,
Fleming
TR
,
De Gruttola
V
. 
Estimating the proportion of treatment effect explained by a surrogate marker
.
Stat Med
1997
;
16
:
1515
27
.
26.
Stryker
WS
,
Kaplan
LA
,
Stein
EA
,
Stampfer
MJ
,
Sober
A
,
Willett
WC
. 
The relation of diet, cigarette smoking, and alcohol consumption to plasma beta-carotene and alpha-tocopherol levels
.
Am J Epidemiol
1988
;
127
:
283
96
.
27.
Jun
H-J
,
Austin
SB
,
Wylie
SA
,
Corliss
HL
,
Jackson
B
,
Spiegelman
D
, et al
The mediating effect of childhood abuse in sexual orientation disparities in tobacco and alcohol use during adolescence: results from the Nurses' Health Study II
.
Cancer Causes Control
2010
;
21
:
1817
28
.
28.
Lee
JE
,
Willett
WC
,
Fuchs
CS
,
Smith-Warner
SA
,
Wu
K
,
Ma
J
, et al
Folate intake and risk of colorectal cancer and adenoma: modification by time
.
Am J Clin Nutr
2011
;
93
:
817
25
.
29.
Sisti
JS
,
Collins
LC
,
Beck
AH
,
Tamimi
RM
,
Rosner
BA
,
Eliassen
AH
. 
Reproductive risk factors in relation to molecular subtypes of breast cancer: results from the nurses' health studies
.
Int J Cancer
2016
;
138
:
2346
56
.
30.
Barnard
ME
,
Boeke
CE
,
Tamimi
RM
. 
Established breast cancer risk factors and risk of intrinsic tumor subtypes
.
Biochim Biophys Acta Rev Cancer
2015
;
1856
:
73
85
.
31.
Martínez
ME
,
Cruz
GI
,
Brewster
AM
,
Bondy
ML
,
Thompson
PA
. 
What can we learn about disease etiology from case-case analyses? Lessons from breast cancer
.
Cancer Epidemiol Biomarkers Prev
2010
;
19
:
2710
4
.
32.
Satija
A
,
Bhupathiraju
SN
,
Spiegelman
D
,
Chiuve
SE
,
Manson
JE
,
Willett
W
, et al
Healthful and unhealthful plant-based diets and the risk of coronary heart disease in U.S. adults
.
J Am Coll Cardiol
2017
;
70
:
411
22
.
33.
Kim
H
,
Caulfield
LE
,
Garcia-Larsen
V
,
Steffen
LM
,
Grams
ME
,
Coresh
J
, et al
Plant-based diets and incident CKD and kidney function
.
Clin J Am Soc Nephrol
2019
;
14
:
682
91
.
34.
Gómez-Donoso
C
,
Martínez-González
,
Martínez
JA
,
Gea
A
,
Sanz-Serrano
J
,
Perez-Cueto
FJA
, et al
A provegetarian food pattern emphasizing preference for healthy plant-derived foods reduces the risk of overweight/obesity in the SUN cohort
.
Nutrients
2019
;
11
:
1553
.
35.
Baden
MY
,
Satija
A
,
Hu
FB
,
Huang
T
. 
Change in plant-based diet quality is associated with changes in plasma adiposity-associated biomarker concentrations in women
.
J Nutr
2019
;
149
:
676
86
.
36.
Kim
H
,
Caulfield
LE
,
Rebholz
CM
. 
Healthy plant-based diets are associated with lower risk of all-cause mortality in US Adults
.
J Nutr
2018
;
148
:
624
31
.
37.
Baden
MY
,
Liu
G
,
Satija
A
,
Li
Y
,
Sun
Q
,
Fung
TT
, et al
Changes in plant-based diet quality and total and cause-specific mortality
.
Circulation
2019
;
140
:
979
91
.
38.
Wiseman
M
. 
The Second World Cancer Research Fund/American Institute for Cancer Research expert report. Food, nutrition, physical activity, and the prevention of cancer: a global perspective
.
Proc Nutr Soc
2008
;
67
:
253
6
.
39.
Brennan
SF
,
Cantwell
MM
,
Cardwell
CR
,
Velentzis
LS
,
Woodside
JV
. 
Dietary patterns and breast cancer risk: a systematic review and meta-analysis
.
Am J Clin Nutr
2010
;
91
:
1294
302
.
40.
Fung
TT
,
Hu
FB
,
Hankinson
SE
,
Willett
WC
,
Holmes
MD
. 
Low-carbohydrate diets, dietary approaches to stop hypertension-style diets, and the risk of postmenopausal breast cancer
.
Am J Epidemiol
2011
;
174
:
652
60
.
41.
Hirko
KA
,
Willett
WC
,
Hankinson
SE
,
Rosner
BA
,
Beck
AH
,
Tamimi
RM
, et al
Healthy dietary patterns and risk of breast cancer by molecular subtype
.
Breast Cancer Res Treat
2016
;
155
:
579
88
.
42.
Fung
TT
,
Hu
FB
,
McCullough
ML
,
Newby
PK
,
Willett
WC
,
Holmes
MD
. 
Diet quality is associated with the risk of estrogen receptor–negative breast cancer in postmenopausal women
.
J Nutr
2006
;
136
:
466
72
.
43.
Kang
JH
,
Peng
C
,
Rhee
JJ
,
Farvid
MS
,
Willett
WC
,
Hu
FB
, et al
Prospective study of a diabetes risk reduction diet and the risk of breast cancer
.
Am J Clin Nutr
2020
;
112
:
1492
503
.
44.
Jung
S
,
Spiegelman
D
,
Baglietto
L
,
Bernstein
L
,
Boggs
DA
,
van den Brandt
PA
, et al
Fruit and vegetable intake and risk of breast cancer by hormone receptor status
.
J Natl Cancer Inst
2013
;
105
:
219
36
.
45.
Fung
TT
,
Chiuve
SE
,
Willett
WC
,
Hankinson
SE
,
Hu
FB
,
Holmes
MD
. 
Intake of specific fruits and vegetables in relation to risk of estrogen receptor-negative breast cancer among postmenopausal women
.
Breast Cancer Res Treat
2013
;
138
:
925
30
.
46.
Farvid
MS
,
Cho
E
,
Eliassen
AH
,
Chen
WY
,
Willett
WC
. 
Lifetime grain consumption and breast cancer risk
.
Breast Cancer Res Treat
2016
;
159
:
335
45
.
47.
Toledo
E
,
Salas-Salvadó
J
,
Donat-Vargas
C
,
Buil-Cosiales
P
,
Estruch
R
,
Ros
E
, et al
Mediterranean diet and invasive breast cancer risk among women at high cardiovascular risk in the PREDIMED trial: a randomized clinical trial
.
JAMA Intern Med
2015
;
175
:
1752
60
.
48.
Sánchez-Quesada
C
,
Romanos-Nanclares
A
,
Navarro
AM
,
Gea
A
,
Cervantes
S
,
Martínez-González
, et al
Coffee consumption and breast cancer risk in the SUN project
.
Eur J Nutr
2020
;
59
:
3461
71
.
49.
Chen
S
,
Chen
Y
,
Ma
S
,
Zheng
R
,
Zhao
P
,
Zhang
L
, et al
Dietary fibre intake and risk of breast cancer: a systematic review and meta-analysis of epidemiological studies
.
Oncotarget
2016
;
7
:
80980
9
.
50.
Reynolds
A
,
Mann
J
,
Cummings
J
,
Winter
N
,
Mete
E
,
Te Morenga
L
. 
Carbohydrate quality and human health: a series of systematic reviews and meta-analyses
.
Lancet
2019
;
393
:
434
45
.
51.
Zhang
X
,
Spiegelman
D
,
Baglietto
L
,
Bernstein
L
,
Boggs
DA
,
van den Brandt
PA
, et al
Carotenoid intakes and risk of breast cancer defined by estrogen receptor and progesterone receptor status: a pooled analysis of 18 prospective cohort studies
.
Am J Clin Nutr
2012
;
95
:
713
25
.
52.
Bertram
JS
. 
Dietary carotenoids, connexins and cancer: what is the connection?
Biochem Soc Trans
2004
;
32
:
985
9
.
53.
Maru
GB
. 
Understanding the molecular mechanisms of cancer prevention by dietary phytochemicals: from experimental models to clinical trials
.
World J Biol Chem
2016
;
7
:
88
9
.
54.
Du
M
,
Liu
SH
,
Mitchell
C
,
Fung
TT
. 
Associations between diet quality scores and risk of postmenopausal estrogen receptor-negative breast cancer: a systematic review
.
J Nutr
2018
;
148
:
100
8
.
55.
Qian
F
,
Liu
G
,
Hu
FB
,
Bhupathiraju
SN
,
Sun
Q
. 
Association between plant-based dietary patterns and risk of type 2 diabetes: a systematic review and meta-analysis
.
JAMA Intern Med
2019
;
179
:
1335
44
.
56.
Liu
RH
. 
Health benefits of fruit and vegetables are from additive and synergistic combinations of phytochemicals
.
Am J Clin Nutr
2003
;
78
:
517S
20S
.
57.
Chandran
U
,
McCann
SE
,
Zirpoli
G
,
Gong
Z
,
Lin
Y
,
Hong
C-C
, et al
Intake of energy-dense foods, fast foods, sugary drinks, and breast cancer risk in African American and European American women
.
Nutr Cancer
2014
;
66
:
1187
99
.
58.
Romanos-Nanclares
A
,
Toledo
E
,
Gardeazabal
I
,
Jiménez-Moleón
JJ
,
Martínez-González
MA
,
Gea
A
. 
Sugar-sweetened beverage consumption and incidence of breast cancer: the Seguimiento Universidad de Navarra (SUN) Project
.
Eur J Nutr
2018
;
58
:
2875
86
.
59.
Chazelas
E
,
Srour
B
,
Desmetz
E
,
Kesse-Guyot
E
,
Julia
C
,
Deschamps
V
, et al
Sugary drink consumption and risk of cancer: results from NutriNet-Santé prospective cohort
.
BMJ
2019
;
366
:
l2408
.
60.
Hodge
AM
,
Bassett
JK
,
Milne
RL
,
English
DR
,
Giles
GG
. 
Consumption of sugar-sweetened and artificially sweetened soft drinks and risk of obesity-related cancers
.
Public Health Nutr
2018
;
21
:
1618
26
.
61.
Fiolet
T
,
Srour
B
,
Sellem
L
,
Kesse-Guyot
E
,
Allès
B
,
Méjean
C
, et al
Consumption of ultra-processed foods and cancer risk: results from NutriNet-Santé prospective cohort
.
BMJ
2018
;
360
:
k322
.
62.
World Cancer Research Fund
,
American Institute for Cancer Research
. 
Diet, nutrition, physical activity and breast cancer
; 
2017
. Available from: https://www.wcrf.org/wp-content/uploads/2021/02/Breast-cancer-report.pdf.
63.
Salvini
S
,
Hunter
DJ
,
Sampson
L
,
Stampfer
MJ
,
Colditz
GA
,
Rosner
B
, et al
Food-based validation of a dietary questionnaire: the effects of week-to-week variation in food consumption
.
Int J Epidemiol
1989
;
18
:
858
67
.
64.
Willett
WC
,
Sampson
L
,
Stampfer
MJ
,
Rosner
B
,
Bain
C
,
Witschi
J
, et al
Reproducibility and validity of a semiquantitative food frequency questionnaire
.
Am J Epidemiol
1985
;
122
:
51
65
.
65.
Ries
LAG
,
Young
JL
,
Keel
GE
,
Eisner
MP
,
Lin
YD
,
Horner
M-J
, editors.
SEER Survival Monograph: Cancer Survival Among Adults: U.S. SEER Program, 1988–2001, Patient and Tumor Characteristics
.
Bethesda, MD
:
NCI
; 
2007
.