Association Between Plant-Based Diet Index and Breast Cancer Risk Stratified by Menopausal and Hormone Receptor Status: A Case-Control Study Among Korean Women

Abstract

Background

This study aims to examine the association between a plant-based diet index (PDI) and the risk of breast cancer in Korean women. Additionally, we aim to assess whether the associations differ by menopausal status and in estrogen receptor (ER) and progesterone receptor (PR) status.

Methods

In this case-control study, we included 395 breast cancer patients and 395 age-matched controls from the National Cancer Center Hospital. The overall PDI, healthful plant-based diet index (hPDI), and unhealthful plant-based diet index (uPDI) were calculated based on dietary intake assessed using a 106-item semi-quantitative food frequency questionnaire. Logistic regression was used to estimate odds ratios and 95% confidence intervals for the association between plant-based diets and breast cancer, adjusting for potential confounders.

Results

We found that PDI was inversely associated with breast cancer, while uPDI was positively associated. However, hPDI was not significantly associated with breast cancer after multivariable adjustment. In analyses stratified by menopausal status, PDI showed an inverse association with breast cancer in premenopausal women, while uPDI was positively associated with breast cancer in both pre- and postmenopausal women. Additionally, when stratified by ER and PR status, we observed that both PDI and uPDI showed stronger associations in the ER+/PR+ group, whereas hPDI remained non-significant in both the ER+/PR+ and ER−/PR− groups.

Conclusion

A higher PDI was associated with lower odds of breast cancer, supporting the recommendation to increase plant-based food intake and reduce animal food intake as part of dietary guidance for breast cancer prevention.

Graphical Abstract

INTRODUCTION

Breast cancer cases are increasing worldwide; in 2020, 2.3 million women were diagnosed with breast cancer, and there were 685,000 deaths globally.1 In South Korea, breast cancer is also increasing, with no exception to the global trend. According to cancer statistics in Korea, the age-standardized incidence rate for breast cancer was 68.6 per 100,000 in 2021.2 Experts predict that the upward trend in breast cancer cases is likely to continue in the future. Studies on the prevention and treatment of breast cancer are ongoing, and nutritional research has focused on the intake of specific nutrients and their associations with breast cancer risk. Numerous previous studies have specifically explored the association between dietary fat and breast cancer risk, confirming that higher dietary fat intake is associated with an increased risk of breast cancer and may influence the prognosis of breast cancer patients.3 A cohort study on vegetable intake and breast cancer survival confirmed that increased consumption of vegetables, particularly green leafy and cruciferous vegetables is associated with improved survival rates in breast cancer patients.4 Also, previous studies have shown that increased consumption of fruits, vegetables, legumes, and nuts is associated with a reduced risk of breast cancer.5,6 Recently, studies have expanded beyond examining specific nutrients to investigate overall dietary patterns. For instance, the associations between the Western dietary pattern, the Mediterranean dietary pattern, and the risk of breast cancer have been explored.7

While studies on the association between dietary patterns and breast cancer are increasing, such as those focusing on the Western diet, Mediterranean diet, and plant-based foods, research using the plant-based diet index (PDI) remains particularly limited.7 A plant-based diet refers to a dietary pattern that primarily consists of plant-derived foods and is gaining attention due to its health and environmental benefits. In recent studies on plant-based diets, unlike earlier research on vegetarianism that partially or completely excludes animal foods, the PDI is being used to distinguish between plant and animal foods by assigning positive and negative scores.8 The PDI has been shown to be associated with health outcomes. A higher PDI is inversely associated with the risk of obesity, which in turn increases the risk of chronic diseases and overall health deterioration.9 Also, a higher PDI has been associated with a lower risk of cardiovascular disease, type 2 diabetes, and prostate cancer.8,10,11 There are numerous studies examining the association between plant-based diets and chronic diseases, including cancer, however, few studies have investigated the association between the PDI and breast cancer.

Previous studies examining the relationship between plant-based diets and breast cancer risk have reported inconsistent findings across different populations.12,13,14,15 These discrepancies may stem from variations in the compositions of plant-based diets such as type of plant foods as well as difference in cultural dietary practices. Additionally, methodological differences in how plant-based diet indices were constructed, differences in study designs, and variations in population characteristics may further contribute to the inconsistencies observed in the literature. By focusing on the Korean population, our study aims to address this gap and provide region-specific insights into the association between plant-based dietary patterns and breast cancer risk. To our knowledge, this is the first study conducted in Northeast Asia to examine the association between the PDI and breast cancer. Additionally, we aimed to assess differences in the results when stratified by menopausal status and by estrogen receptor (ER) and progesterone receptor (PR) status, both of which are important prognostic factors for breast cancer.

METHODS

Study population

Eligible breast cancer patients were those newly diagnosed between July 2007 and September 2008 at the Center for Breast Cancer of the National Cancer Center, Korea. Of the 415 patients who agreed to participate, 20 were excluded due to a previous history of cancer, inability to complete the interview, or implausible energy intakes. Therefore, 395 patients were included in the final analysis. Controls were enrolled between October 2007 and December 2008 from individuals visiting the Center for Cancer Prevention and Detection at the same hospital for regular health check-ups. Of the 1,687 subjects who agreed to participate, 601 were excluded due to a history of malignant neoplasm or benign breast diseases, or implausible energy intakes. From the remaining 1,086 eligible controls, 395 were selected by frequency matching to the cases by 5-year age groups. Thus, the final analysis included 395 cases and 395 controls (Fig. 1).

Fig. 1. Flow chart of study population selection.

NCC = National Cancer Center, FFQ = food frequency questionnaire.

Data collation

A trained dietitian conducted face-to-face interviews using a structured questionnaire that included demographics and lifestyle information. Dietary intake was assessed using a 106-item semiquantitative food frequency questionnaire (SQFFQ), which had been previously validated for reliability and accuracy.16 The SQFFQ included the portion sizes (small, medium, large) and frequency of consumption (never or rarely, 1 time per month, 2–3 times per month, 1–2 times per week, 3–4 times per week, 5–6 times per week, 1 time per day, 2 times per day, 3 times per day) for each food item consumed during the previous year. ER and PR status were analyzed by immunohistochemistry on tissue sections cut from formalin-fixed, paraffin-embedded breast tumors (Ventana Medical Systems, Tucson, AZ, USA). More detailed information regarding the immunohistochemistry procedure is provided elsewhere.17

Assessment of the plant-based diet indices

We created three versions of a PDI for each food frequency questionnaire: an overall PDI, a healthful plant-based diet index (hPDI), and an unhealthful plant-based diet index (uPDI). All food items were categorized into 17 food groups based on nutrient content and culinary similarities. These included healthy plant foods (whole grains, fruits, vegetables, nuts, legumes, tea and coffee), less healthy plant foods (refined grains, potatoes, sugar-sweetened beverages, sweets and desserts, salty foods), and animal foods (animal fat, dairy, eggs, fish or seafood, meat, miscellaneous animal foods).11 The classification of mixed-composition foods was based on the predominant. The detailed list of food items corresponding to each food group used to calculate the PDI components is summarized in Supplementary Table 1. For each food group, we adjusted for total energy intake using the residual method and ranked participants into quintiles, which were then assigned either positive or reverse scores. For positive scoring, participants in the highest quintile of a food group received a score of 4, and those in the lowest quintile received a score of 0. For reverse scoring, this pattern was inverted. In the PDI, plant-based food groups (healthy plant food groups and less healthy plant groups) received positive scores, while animal-based food groups received reverse scores. In the hPDI, healthy plant food groups received positive scores, and less healthy plant and animal food groups received reverse scores. In contrast, for the uPDI, less healthy plant food groups received positive scores, and healthy plant and animal food groups received reverse scores. The scores for all food groups were summed to calculate the PDI, hPDI, and uPDI values, which ranged from 0 to 68. A higher index score indicates a more plant-based and less animal-based dietary pattern.

Statistical analyses

Differences between cases and controls were examined using Student’s t-test for continuous variables and χ² test for categorical variables. A linear residual regression method was used to adjust food intake in the plant-based diet indices for total energy intake. The plant-based diet indices were categorized into tertiles based on their distribution among the controls. To evaluate the association between plant-based diets and breast cancer risk, logistic regression analysis was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs). The lowest tertile group served as the reference category. Multivariate models were adjusted for age, body mass index (BMI), first-degree family history of breast cancer, education level, occupation, alcohol consumption, and total energy intake. Subgroup analyses were conducted to assess dietary patterns and breast cancer risk by menopausal status, with additional adjustment for postmenopausal hormone use in the analysis of postmenopausal women. Stratified analyses were also performed to evaluate the association between plant-based diets and combined ER/PR status using polytomous logistic regression. A test for linear trend across increasing tertiles of the plant-based diet indices was conducted by modeling the median value of each tertile as a continuous variable. As a sensitivity analysis, we additionally evaluated the association between PDI and breast cancer risk by modelling PDI as a continuous variable in the logistic regression model. All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA), and a two-sided P value < 0.05 was considered statistically significant.

Ethics statement

The study protocols and consent forms were approved by the Institutional Review Board (IRB) of the National Cancer Center Hospital (IRB protocol number NCCNCS 07-083), and all subjects provided informed consent.

RESULTS

Table 1 summarizes the characteristics of the 790 participants. Women with breast cancer had a higher BMI (P = 0.009), a family history of breast cancer (P = 0.044), lower education levels (P < 0.001), and lower alcohol consumption (P < 0.001). When stratified by menopausal status, the case group had lower education levels and alcohol consumption in both premenopausal and postmenopausal groups. The premenopausal case group also reported lower physical activity (P = 0.022), while the postmenopausal case group had a higher BMI (P = 0.012), later age at menarche (P = 0.019), and earlier age at menopause, along with differences in menopause type and hormone use. The overall PDI was lower in cases than in controls (49.79 vs. 50.94), whereas the hPDI (51.68 vs. 50.16) and uPDI (53.29 vs. 51.07) were higher in cases. Stratified analyses showed that premenopausal cases had lower PDI and higher uPDI, while postmenopausal cases had higher energy intake, hPDI, and uPDI.

Table 1. General characteristics of study subjects.

Characteristics		Total (N = 790)			Premenopausal women (n = 495)			Postmenopausal women (n = 295)
		Control (n = 395)	Case (n = 395)	P valuea	Control (n = 250)	Case (n = 245)	P valuea	Control (n = 145)	Case (n = 150)	P valuea
Age, yr		47.4 ± 8.4	46.9 ± 9.3	0.437	42.6 ± 4.9	42.1 ± 5.9	0.305	55.7 ± 6.5	54.7 ± 8.7	0.305
Body mass index, kg/m2		22.9 ± 2.8	23.4 ± 3.2	0.009	22.5 ± 2.8	22.9 ± 3.2	0.156	23.6 ± 2.5	24.3 ± 2.9	0.012
family history of breast cancer, yes		10 (2.5)	21 (5.3)	0.044	7 (2.8)	11 (4.5)	0.315	3 (2.1)	10 (6.7)	0.054
Marital status				1.000			0.638			0.568
	Married	320 (81.0)	320 (81.0)		214 (85.6)	206 (84.1)		106 (73.1)	114 (76.0)
	Unmarried, Divorced, widowed	75 (19.0)	75 (19.0)		36 (14.4)	39 (15.9)		39 (26.9)	36 (24.0)
Education, yr				< 0.001			< 0.001			< 0.001
	≤ 9	47 (11.9)	104 (26.3)		10 (4.0)	33 (13.5)		37 (25.5)	71 (47.3)
	9–12	144 (36.5)	182 (46.1)		86 (34.4)	119 (48.6)		58 (40.0)	63 (42.0)
	> 12	204 (51.7)	109 (27.6)		154 (61.6)	93 (38.0)		50 (34.5)	16 (10.7)
Occupation				0.016			0.161			0.099
	Professionals, office worker	93 (23.5)	66 (16.7)		74 (29.6)	56 (22.9)		19 (13.1)	10 (6.7)
	Sales, service positions	66 (16.7)	51 (12.9)		46 (18.4)	37 (15.1)		20 (13.8)	14 (9.3)
	Agriculture, laborer	12 (3.0)	18 (4.6)		8 (3.2)	10 (4.1)		4 (2.8)	8 (5.3)
	Housewife, unemployment	224 (56.7)	260 (65.8)		122 (48.8)	142 (58.0)		102 (70.3)	118 (78.7)
Smoking status				0.637			0.824			0.590
	Never	357 (90.4)	353 (89.4)		223 (89.2)	217 (88.6)		134 (92.4)	136 (90.7)
	Ever	38 (9.6)	42 (10.6)		27 (10.8)	28 (11.4)		11 (7.6)	14 (9.3)
Alcohol consumption				< 0.001			0.004			0.001
	Never	145 (36.7)	205 (51.9)		84 (33.6)	113 (46.1)		61 (42.1)	92 (61.3)
	Ever	250 (63.3)	190 (48.1)		166 (66.4)	132 (53.9)		84 (57.9)	58 (38.7)
Physical activity, Met-min/wk		2,019.6 ± 2,375.0	1,741.8 ± 1,918.9	0.072	2,008.7 ± 2,540.8	1,559.3 ± 1,707.0	0.022	2,038.6 ± 2,064.4	2,040.0 ± 2,195.6	0.995
Age at menarche, yr		14.5 ± 1.6	14.7 ± 1.9	0.071	14.2 ± 1.5	14.2 ± 1.7	0.755	14.9 ± 1.7	15.4 ± 2.0	0.019
Parity, yes		358 (90.6)	347 (87.9)	0.332	220 (88.0)	206 (84.1)	0.296	138 (95.2)	141 (94.0)	0.833
Menopausal status				0.713
	Premenopause	250 (63.3)	245 (62.0)
	Postmenopause	145 (36.7)	150 (38.0)
Age at menopause,b yr								49.4 ± 4.3	48.0 ± 5.4	0.013
Type of menopauseb										0.009
	Natural							115 (79.3)	98 (65.3)
	Surgery, other							30 (20.7)	51 (34.0)
Postmenopausal hormone useb										0.010
	Never							98 (67.6)	118 (78.7)
	Ever							47 (32.4)	28 (18.7)
Energy intake, kcal/day		1,622.3 ± 539.6	1,734.6 ± 465.1	0.002	1,716.9 ± 566.5	1,724.0 ± 432.6	0.874	1,459.3 ± 446.4	1,752.0 ± 514.7	< 0.001
PDI		50.94 ± 5.12	49.79 ± 5.83	0.003	50.40 ± 4.95	48.92 ± 5.70	0.002	51.86 ± 5.30	51.20 ± 5.79	0.312
hPDI		50.16 ± 7.65	51.68 ± 8.34	0.008	50.37 ± 7.87	50.43 ± 8.41	0.934	49.80 ± 7.25	53.72 ± 7.83	< 0.001
uPDI		51.07 ± 6.19	53.29 ± 6.49	< 0.001	51.03 ± 6.14	52.47 ± 6.29	0.010	51.14 ± 6.30	54.63 ± 6.61	< 0.001

Values are expressed as the mean ± standard deviation or number (%). Missing data are included in the total %.

^aCategorical variables were analyzed using the χ² test; continuous variables were analyzed using the Student’s t-test. ^bIn postmenopausal women.

Table 2 shows the intake of PDI components by food item. The case group consumed fewer hPDI foods including fruits, vegetables, nuts, legumes as well as fewer animal foods such as dairy, fish, meat, but had higher intake of refined grains, and lower intake of potatoes (an uPDI foods). When stratified by menopausal status, premenopausal cases had higher intake of tea and coffee (hPDI foods) but lower intake of meat (animal foods), while postmenopausal cases had lower intake of vegetables, nuts, and legumes (hPDI foods), as well as lower intake of potatoes and sugar-sweetened beverages (uPDI foods), and animal foods.

Table 2. Comparison of plant-based diet index food items intake of the study subjects.

Food groups,a g/day	Total (N = 790)			Premenopausal women (n = 495)			Postmenopausal women (n = 295)
	Control (n = 395)	Case (n = 395)	P valueb	Control (n = 250)	Case (n = 245)	P valueb	Control (n = 145)	Case (n = 150)	P valueb
Whole grains	372.5 ± 264.7	432.7 ± 347.5	0.006	381.9 ± 283.5	400.6 ± 339.3	0.507	356.1 ± 228.8	485.1 ± 355.4	< 0.001
Fruits	268.2 ± 196.6	231.6 ± 183.8	0.007	268.3 ± 196.4	227.8 ± 183.2	0.018	268.0 ± 197.5	237.6 ± 185.3	0.175
Vegetables	178.9 ± 126.7	114.6 ± 81.0	< 0.001	166.3 ± 117.6	109.6 ± 77.5	< 0.001	200.6 ± 138.8	122.9 ± 86.1	< 0.001
Nuts	2.2 ± 6.8	1.0 ± 3.1	0.003	1.4 ± 4.5	1.2 ± 3.6	0.549	3.4 ± 9.5	0.8 ± 2.1	0.001
Legumes	39.7 ± 44.1	32.3 ± 44.0	0.017	34.5 ± 37.0	33.6 ± 42.9	0.814	48.8 ± 53.1	30.0 ± 45.9	0.001
Tea and coffee	55.2 ± 90.5	83.4 ± 149.6	0.002	56.8 ± 89.2	93.7 ± 163.7	0.002	52.5 ± 93.0	66.4 ± 121.8	0.270
Refined grains	208.8 ± 292.6	256.1 ± 341.8	0.037	222.6 ± 301.0	265.2 ± 327.8	0.133	185.0 ± 276.7	241.3 ± 364.2	0.136
Potatoes	97.2 ± 423.8	26.9 ± 113.7	0.002	53.6 ± 270.1	19.8 ± 40.6	0.052	172.4 ± 596.9	38.3 ± 176.8	0.010
Sugar sweetened beverages	244.3 ± 1,640.7	89.9 ± 641.7	0.082	232.0 ± 1926.6	115.8 ± 791.3	0.379	265.5 ± 973.1	47.6 ± 245.9	0.010
Sweets and desserts	42.4 ± 165.8	25.6 ± 43.7	0.053	33.5 ± 78.3	26.7 ± 41.4	0.225	57.6 ± 253.4	23.8 ± 47.3	0.116
Salty food group	186.2 ± 113.6	181.3 ± 91.9	0.507	173.9 ± 104.2	179.3 ± 95.4	0.543	207.5 ± 125.6	184.6 ± 86.2	0.070
Animal fat	6.7 ± 17.6	7.4 ± 12.6	0.518	6.4 ± 10.0	7.9 ± 11.0	0.115	7.1 ± 26.0	6.5 ± 14.9	0.809
Dairy	442.2 ± 1,314.8	218.5 ± 738.6	0.003	371.1 ± 1,167.1	223.0 ± 772.8	0.096	564.7 ± 1,533.2	211.2 ± 681.6	0.012
Eggs	13.7 ± 14.6	12.1 ± 12.5	0.111	13.6 ± 13.8	13.2 ± 12.1	0.703	13.8 ± 16.1	10.4 ± 12.9	0.050
Fish	38.0 ± 26.4	33.2 ± 26.4	0.011	35.4 ± 25.1	34.2 ± 26.6	0.598	42.4 ± 28.0	31.5 ± 26.2	0.001
Meat	57.7 ± 43.3	41.1 ± 34.9	< 0.001	57.4 ± 39.5	46.4 ± 37.0	0.002	58.3 ± 49.3	32.5 ± 29.2	< 0.001
Miscellaneous animal foods	28.0 ± 106.2	15.0 ± 50.4	0.029	21.5 ± 66.0	19.0 ± 59.9	0.655	39.1 ± 152.1	8.6 ± 27.8	0.019

Values are presented as mean ± standard deviation.

^aAdjusted for total energy intake using the residual method. ^bVariables were analyzed using the Student’s t-test.

Table 3 presents age- and multivariable-adjusted ORs for breast cancer by tertiles of the PDI, hPDI, and uPDI. Higher PDI was associated with lower odds of breast cancer (T2 vs. T1: OR, 0.51; 95% CI, 0.35–0.76; T3 vs. T1: OR, 0.60; 95% CI, 0.42–0.87; P = 0.008). In contrast, higher uPDI was associated with increased odds of breast cancer (T2 vs. T1: OR, 1.51; 95% CI, 1.02–2.24; T3 vs. T1: OR, 2.08; 95% CI, 1.41–3.05; P < 0.001). The hPDI was significantly associated with higher odds of breast cancer in the age-adjusted model; however, this association was not observed in the multivariable adjusted model. When stratified by menopausal status, higher PDI was associated with a lower odds of breast cancer among premenopausal women (T2 vs. T1: OR, 0.49; 95% CI, 0.31–0.78; T3 vs. T1: OR, 0.57; 95% CI, 0.36–0.91; P = 0.011). Higher uPDI was associated with increased odds of breast cancer in both premenopausal (T3 vs. T1: OR, 1.64; 95% CI, 1.03–2.62; P = 0.040) and postmenopausal women (T3 vs. T1: OR, 3.36; 95% CI, 1.57–7.17; P = 0.002). PDI and hPDI were not significantly associated with breast cancer among postmenopausal women. Consistent results were observed in the sensitivity analysis when PDI was treated as a continuous variable. A 1-score increase in PDI and uPDI was associated with lower odds (OR, 0.95; 95% CI, 0.93–0.98) and higher odds (OR, 1.04; 95% CI, 1.02–1.07) of breast cancer, respectively in the total population. Among premenopausal women, a 1-score increase in PDI was associated with lower odds of breast cancer (OR, 0.94; 95% CI, 0.91–0.98), whereas among postmenopausal women, a 1-score increase in uPDI was associated with higher odds of breast cancer (OR, 1.08; 95% CI, 1.03–1.14).

Table 3. Age and multivariable-adjusted ORs and CIs for breast cancer across tertiles of PDI, hPDI, and uPDI.

Tertile of plant-based diet indicesa			Median	No. of controls	No. of cases	OR (95% CI)
						Age adjusted	Multivariate adjustedb
Total women (N = 790)
	PDI
		< 49	46	122	161	1.00	1.00
		49–52	51	129	103	0.61 (0.43–0.87)	0.51 (0.35–0.76)
		≥ 53	56	144	131	0.70 (0.50–0.98)	0.60 (0.42–0.87)
		P for trendc				0.036	0.008
		Continuous PDI				0.96 (0.94–0.99)	0.95 (0.93–0.98)
	hPDI
		< 47	43	127	114	1.00	1.00
		47–53	50	129	109	0.96 (0.67–1.37)	0.72 (0.48–1.07)
		≥ 54	57	139	172	1.43 (1.01–2.01)	0.94 (0.61–1.43)
		P for trend				0.033	0.866
		Continuous hPDI				1.03 (1.01–1.04)	1.01 (0.98–1.03)
	uPDI
		< 49	45	145	85	1.00	1.00
		49–53	51	127	126	1.71 (1.19–2.46)	1.51 (1.02–2.24)
		≥ 54	57	123	184	2.59 (1.82–3.68)	2.08 (1.41–3.05)
		P for trend				< 0.001	< 0.001
		Continuous uPDI				1.06 (1.03–1.08)	1.04 (1.02–1.07)
Premenopausal women (n = 495)
	PDI
		< 49	46	87	117	1.00	1.00
		49–52	51	82	61	0.56 (0.36–0.86)	0.49 (0.31–0.78)
		≥ 53	56	81	67	0.63 (0.41–0.97)	0.57 (0.36–0.91)
		P for trendc				0.023	0.011
		Continuous PDI				0.95 (0.92–0.98)	0.94 (0.91–0.98)
	hPDI
		< 47	43	82	85	1.00	1.00
		47–53	50	82	73	0.87 (0.56–1.35)	0.77 (0.47–1.26)
		≥ 54	58	86	87	1.01 (0.65–1.55)	0.85 (0.51–1.43)
		P for trend				0.954	0.583
		Continuous hPDI				1.00 (0.98–1.03)	1.00 (0.97–1.02)
	uPDI
		< 49	46	93	59	1.00	1.00
		49–53	51	77	76	1.56 (0.99–2.46)	1.35 (0.83–2.19)
		≥ 54	57	80	110	2.18 (1.41–3.37)	1.64 (1.03–2.62)
		P for trend				0.001	0.04
		Continuous uPDI				1.04 (1.01–1.07)	1.02 (0.99–1.05)
Postmenopausal women (n = 295)
	PDI
		< 50	47	48	54	1.00	1.00
		50–54	52	55	56	0.92 (0.54–1.59)	0.85 (0.44–1.62)
		≥ 55	58	42	40	0.84 (0.47–1.50)	0.87 (0.43–1.77)
		P for trendc				0.553	0.695
		Continuous PDI				0.98 (0.94–1.02)	0.98 (0.93–1.02)
	hPDI
		< 47	42	45	29	1.00	1.00
		47–53	50	47	36	1.24 (0.65–2.36)	0.77 (0.35–1.67)
		≥ 54	56	53	85	2.65 (1.47–4.78)	1.05 (0.48–2.32)
		P for trend				0.001	0.86
		Continuous hPDI				1.07 (1.04–1.11)	1.03 (0.94–1.03)
	uPDI
		< 49	46	52	26	1.00	1.00
		49–53	51	50	50	2.04 (1.10–3.77)	1.92 (0.91–4.06)
		≥ 54	57	43	74	3.61 (1.96–6.64)	3.36 (1.57–7.17)
		P for trend				< 0.001	0.002
		Continuous uPDI				1.09 (1.05–1.13)	1.08 (1.03–1.14)

OR = odds ratio, CI = confidence interval, PDI = plant-based diet index, hPDI = healthful plant-based diet index, uPDI = unhealthful plant-based diet index.

^aPlant-based diet indices were categorized into tertile according to the distribution of the control groups. The tertile do not have equal sample size because many participants received the same scores.

^bTotal women was adjusted adjusted for age, body mass index, family history of breast cancer, education level, occupation and alcohol consumption, and total energy intake; Premenopausal women were adjusted for age, body mass index, education level, alcohol consumption, physical activity, and total energy intake; Postmenopausal women were adjusted for age, body mass index, family history of breast cancer, education level, alcohol consumption, age at menarche, age at menopause, type of menopause, postmenopausal hormone use, and total energy intake.

^cTest for trends calculated with the median intake for each category of nutrient intake as a continuous variable.

Table 4 shows age- and multivariable-adjusted ORs for breast cancer risk by ER/PR status and tertiles of the PDI, hPDI, and uPDI. In the ER+/PR+ group, higher PDI was associated with lower odds of breast cancer (T2 vs. T1: OR, 0.55; 95% CI, 0.36–0.83; T3 vs. T1: OR, 0.56; 95% CI, 0.37–0.84; P = 0.005), while higher uPDI was associated with increased odds of breast cancer (T2 vs. T1: OR, 1.64; 95% CI, 1.06–2.54; T3 vs. T1: OR, 2.15; 95% CI, 1.40–3.29; P = 0.005). In the ER-/PR- group, higher uPDI was also associated with increased odds of breast cancer (T3 vs. T1: OR, 2.20; 95% CI, 1.17–4.12; P = 0.010). Among premenopausal women in the ER+/PR+ group, higher PDI was associated with lower odds of breast cancer (T2/T3 vs. T1: OR, 0.54; 95% CI, 0.33-–0.90; P = 0.013), while higher uPDI was associated with increased odds of breast cancer (T3 vs. T1: OR, 1.81; 95% CI, 1.10–3.00; P = 0.020). No significant associations were found in the ER−/PR− group for premenopausal women. Among postmenopausal women, higher uPDI was associated with increased odds of breast cancer in both the ER+/PR+ group (T3 vs. T1: OR, 2.79; 95% CI, 1.10–7.09; P = 0.045) and the ER−/PR− group (T3 vs. T1: OR, 2.90; 95% CI, 1.07–7.90; P = 0.021).

Table 4. ORs and 95% CIs of breast cancer risk characterized by ER and PR status.

Tertile of plant-based diet indicesa			No. of controls	ER+/PR+			ER−/PR−
				No. of cases	Age adjusted OR (95% CI)	Multivariate adjusted OR (95% CI)b	No. of case	Age adjusted OR (95% CI)	Multivariate adjusted OR (95% CI)
Total women (N = 790)
	PDI
		< 49	122	108	1.00	1.00	36	1.00	1.00
		49–52	129	71	0.64 (0.43–0.94)	0.55 (0.36–0.83)	24	0.61 (0.34–1.09)	0.49 (0.26–0.90)
		≥ 53	144	76	0.62 (0.42–0.91)	0.56 (0.37–0.84)	33	0.74 (0.43–1.27)	0.63 (0.35–1.12)
		P for trendc			0.014	0.005		0.292	0.126
		Continuous PDI			0.95 (0.92–0.98)	0.94 (0.91–0.97)		0.97 (0.93–1.01)	0.96 (0.92–1.00)
	hPDI
		< 47	127	80	1.00	1.00	22	1.00	1.00
		47–53	129	68	0.85 (0.57–1.28)	0.67 (0.43–1.05)	28	1.24 (0.68–2.29)	0.85 (0.44–1.66)
		≥ 54	139	107	1.30 (0.89–1.91)	0.90 (0.57–1.42)	43	1.75 (0.98–3.10)	1.03 (0.52–2.04)
		P for trend			0.152	0.728		0.055	0.902
		Continuous hPDI			1.02 (1.00–1.04)	1.00 (0.98–1.03)		1.04 (1.01–1.07)	1.01 (0.97–1.05)
	uPDI
		< 49	145	54	1.00	1.00	19	1.00	1.00
		49–53	127	84	1.81 (1.19–2.75)	1.64 (1.06–2.54)	28	1.67 (0.89–3.13)	1.41 (0.73–2.76)
		≥ 54	123	117	2.62 (1.75–3.92)	2.15 (1.40–3.29)	46	2.82 (1.57–5.08)	2.20 (1.17–4.12)
		P for trend			< 0.001	0.001		< 0.001	0.010
		Continuous uPDI			1.05 (1.02–1.08)	1.04 (1.01–1.07)		1.08 (1.04–1.12)	1.07 (1.03–1.11)
Premenopausal women (n = 495)
	PDI
		< 49	87	84	1.00	1.00	20	1.00	1.00
		49–52	82	49	0.61 (0.39–0.98)	0.54 (0.33–0.90)	12	0.67 (0.31–1.47)	0.56 (0.25–1.28)
		≥ 53	81	47	0.59 (0.37–0.95)	0.54 (0.33–0.90)	12	0.72 (0.33–1.58)	0.60 (0.26–1.37)
		P for trend			0.023	0.013		0.382	0.205
		Continuous PDI			0.94 (0.91–0.98)	0.94 (0.90–0.97)		0.97 (0.91–1.03)	0.95 (0.89–1.02)
	hPDI
		< 47	82	64	1.00	1.00	13	1.00	1.00
		47–53	82	53	0.83 (0.51–1.33)	0.74 (0.44–1.25)	15	1.22 (0.54–2.73)	0.96 (0.40–2.31)
		≥ 54	86	63	0.93 (0.59–1.49)	0.79 (0.46–1.37)	16	1.31 (0.59–2.92)	0.87 (0.35–2.18)
		P for trend			0.793	0.425		0.520	0.748
		Continuous hPDI			1.00 (0.98–1.02)	0.99 (0.96–1.02)		1.02 (0.98–1.06)	1.00 (0.95–1.05)
	uPDI
		< 49	93	42	1.00	1.00	10	1.00	1.00
		49–53	77	54	1.55 (0.94–2.57)	1.38 (0.82–2.34)	17	2.05 (0.88–4.76)	1.83 0.76–4.43)
		≥ 54	80	84	2.32 (1.44–3.74)	1.81 (1.10–3.00)	17	2.01 (0.87–4.67)	1.59 (0.66–3.85)
		P for trend			0.001	0.020		0.122	0.369
		Continuous uPDI			1.04 (1.01–1.08)	1.02 (0.99–1.06)		1.04 (0.98–1.09)	1.01 (0.96–1.08)
Postmenopausal women (n = 295)
	PDI
		< 50	48	29	1.00	1.00	16	1.00	1.00
		50–54	55	28	0.85 (0.45–1.64)	0.76 (0.35–1.64)	12	1.09 (0.51–2.30)	0.99 (0.42–2.34)
		≥ 55	42	18	0.70 (0.34–1.45)	0.77 (0.33–1.80)	21	0.74 (0.31–1.75)	0.65 (0.24–1.78)
		P for trend			0.335	0.529		0.502	0.429
		Continuous PDI			0.96 (0.91–1.01)	0.96 (0.91–1.02)		0.96 (0.91–1.02)	0.96 (0.89–1.02)
	hPDI
		< 47	45	16	1.00	1.00	9	1.00	1.00
		47–53	47	15	0.94 (0.42–2.14)	0.55 (0.21–1.44)	13	1.43 (0.55–3.68)	1.05 (0.34–3.27)
		≥ 54	53	44	2.55 (1.25–5.20)	0.89 (0.34–2.30)	27	2.70 (1.14–6.42)	1.29 (0.41–4.11)
		P for trend			0.007	0.877		0.023	0.700
		Continuous hPDI			1.08 (1.03–1.12)	1.02 (0.97–1.07)		1.07 (1.02–1.12)	1.04 (0.98–1.10)
	uPDI
		< 49	52	12	1.00	1.00	9	1.00	1.00
		49–53	50	30	2.66 (1.22–5.77)	2.52 (1.01–6.30)	11	1.29 (0.49–3.38)	1.10 (0.37–3.27)
		≥ 54	43	33	3.49 (1.60–7.61)	2.79 (1.10–7.09)	29	4.03 (1.71–9.47)	2.90 (1.07–7.90)
		P for trend			0.002	0.045		0.001	0.021
		Continuous uPDI			1.07 (1.02–1.12)	1.05 (0.99–1.10)		1.13 (1.07–1.19)	1.09 (1.03–1.16)

OR = odds ratio, CI = confidence interval, ER = estrogen receptor, PR = progesterone receptor, PDI = plant-based diet index, hPDI = healthful plant-based diet index, uPDI = unhealthful plant-based diet index.

^aPlant-based diet indices were categorized into tertile according to the distribution of the control groups. The tertile do not have equal sample size because many participants received the same scores.

^bTotal women was adjusted adjusted for age, body mass index, family history of breast cancer, education level, occupation and alcohol consumption, and total energy intake; Premenopausal women were adjusted for age, body mass index, education level, alcohol consumption, physical activity, and total energy intake; Postmenopausal women were adjusted for age, body mass index, family history of breast cancer, education level, alcohol consumption, age at menarche, age at menopause, type of menopause, and postmenopausal hormone use.

^cTest for trends calculated with the median intake for each category of nutrient intake as a continuous variable.

When PDI was treated as a continuous variable, a 1-score increase in PDI was associated with lower odds of breast cancer among ER+/PR+ women (OR, 0.94; 95% CI, 0.91–0.97). A 1-score increase in uPDI was associated with higher odds of breast cancer in both ER+/PR+ (OR, 1.04; 95% CI, 1.01–1.07) and ER−/PR− (OR, 1.07; 95% CI, 1.03–1.11) groups. Among premenopausal women with ER+/PR+ tumors, a 1-score increase in PDI was associated with lower odds of breast cancer (OR, 0.94; 95% CI, 0.90–0.97), whereas among postmenopausal women with ER−/PR− tumors, a 1-score increase in uPDI was associated with higher odds of breast cancer (OR, 1.09; 95% CI, 1.03–1.16). These results were consistent with those observed when PDI was analyzed as a categorical variable (Table 4).

DISCUSSION

In this case-control study of 790 Korean participants, we confirmed that PDI was inversely associated and uPDI positively associated with breast cancer risk, while hPDI showed no significant association after multivariable adjustment. When stratified by menopausal status, higher PDI was associated with lower breast cancer risk among premenopausal women, and higher uPDI was associated with increased risk in both premenopausal and postmenopausal women. These associations were stronger in the ER+/PR+ group for both PDI and uPDI, whereas hPDI showed no significant association in either the ER+/PR+ or ER−/PR− groups.

In this study, increased uPDI, comprising refined grains, potatoes, sugary beverages, sweets, desserts, and salty foods, was associated with a higher risk of breast cancer after adjustment. This aligns with findings from a previous case-control study.13 In contrast, a case-control study found no significant association between uPDI and breast cancer risk, which differs from our findings.15 Increased PDI was associated with lower odds of breast cancer, consistent with previous studies reporting similar findings.13,15 Dietary patterns rich in whole grains, fruits, and vegetables may be associated with a reduced risk of breast cancer.18 A few other studies found that plant-based diets were associated with reduced risk of overall, digestive, and lung cancers but not breast cancer, findings that are inconsistent with our results.14,19 Although hPDI intake was lower in the breast cancer group after adjustment for total energy intake, no significant association with breast cancer was observed in the multivariable analysis. In the age-adjusted analysis, higher hPDI was associated with an increased risk of breast cancer, however, this association was not significant after further adjustment, differing from previous findings that suggested a protective effect.12,15 A previous study found that low intake of hPDI foods, such as vegetables, soy, and nuts was associated with a higher risk of breast cancer, which contrasts with our findings.20 Further research is needed to explore the associations between specific hPDI components and breast cancer risk. Stratified analysis showed that higher uPDI was associated with increased breast cancer risk in both pre- and postmenopausal women, while higher PDI was associated with reduced risk in premenopausal women but not in postmenopausal women, findings consistent with previous studies showing stronger protective associations of prudent diets in premenopausal women.21 The hPDI was significantly associated with increased breast cancer risk in the age-adjusted model for the total population and postmenopausal women; however, this association was not observed in the multivariable model for either group. The attenuation of the association between hPDI and breast cancer in the fully adjusted model may be due to collinearity between dietary patterns and lifestyle-related covariates, limited statistical power in subgroup analyses, or the varying effects of individual food items classified as “healthful” within the hPDI. These findings suggest that the association between plant-based diets and breast cancer risk varies by menopausal status, highlighting the importance of considering dietary differences across menopausal groups.

Differences in breast cancer risk were observed based on ER and PR status. PDI was strongly associated with a reduced risk of breast cancer in the ER+/PR+ group, particularly among the total population and premenopausal women, but not in the ER−/PR− group. Increased uPDI was associated with higher breast cancer risk among the total population as well as in pre- and postmenopausal women in the ER+/PR+ group, while in the ER−/PR− group, the association was significant only in the total and postmenopausal groups. hPDI showed no significant association with breast cancer risk regardless of ER and PR status. These findings are consistent with previous studies that reported a significant association between a Western dietary pattern and breast cancer risk in ER+/PR+ cases, but not in ER−/PR− cases.21 Previous studies have shown that higher consumption of vegetables, legumes, and certain fruits is inversely associated with breast cancer risk, while increased intake of animal foods is linked to higher risk. However, no significant differences were observed between ER+/PR+ and ER−/PR− cases in those studies, which contrasts with our findings.22

More studies have focused on vegetable-based diets and cancer than on plant-based diets. Previous research has shown that vegetarian diets significantly reduce overall cancer risk, and that postmenopausal women have a lower risk of breast cancer when following a vegetable-based diet.23 Caution is needed with vegetarian diets, as they may lack essential nutrients such as protein, omega-3 fatty acids, vitamin D, B12, calcium, and iron, especially when the diet is overly restrictive.24 Some studies suggest that poorly managed vegetarian diets may increase fracture risk and require additional nutritional support.24 The PDI includes a variety of nutrients from both plant and animal foods, which may help prevent nutritional deficiencies. This study distinguishes between hPDI (whole grains, fruits, vegetables, nuts, legumes, tea, and coffee) and uPDI (refined grains, potatoes, sugary drinks, sweets, and salty foods), unlike previous studies that treated all plant foods equally.19

PDI's positive association with vegetables and fruits, rich in antioxidants and anti-inflammatory compounds like flavonoids, may help regulate oxidative stress, influencing breast cancer risk and progression.25 Isoflavones in beans, part of hPDI, have anti-proliferative effects and regulate estrogen-receptor signaling pathways, reducing breast cancer risk.26 Isoflavones also have protective effects against breast cancer, including anti-angiogenic, anti-oxidative, pro-apoptotic, and anti-inflammatory properties, regulated by mechanisms beyond the estrogen pathway.27 It may also be influenced by Insulin-like Growth Factor 1 (IGF-1), a hormone that stimulates cell growth. Previous studies have linked higher IGF-1 levels to an increased breast cancer risk in premenopausal women.28 Total IGF-1 levels were significantly lower in vegans compared to meat-eaters and vegetarians. Increased milk consumption and saturated fat from animal foods may raise serum IGF-1 levels.29,30 Increased PDI, characterized by more fruits and vegetables and less animal food, may lower breast cancer risk by reducing IGF-1 levels. PDI helps reduce inflammation, maintain cell health, and regulate hormones.

This study is significant for confirming the role of dietary patterns, particularly plant-based diets, and specific foods in breast cancer risk. Notably, the PDI includes a variety of nutrients without excluding animal foods, distinguishing it from vegetarian diets. The study also highlights the influence of menopausal status and hormonal factors. With a sample size of 790 participants, it provides valuable insights into various risk factors of breast cancer. Additionally, it distinguishes between hPDI (whole grains, fruits, vegetables, nuts, legumes, tea, coffee) and uPDI (refined grains, potatoes, sweets, salty foods).

One of the key limitations of this study lies in the retrospective nature of dietary data collection and the potential issues related to temporality between dietary exposure and breast cancer diagnosis. Although dietary intake was assessed for the year preceding diagnosis (for cases) and health check-up (for controls) using a validated SQFFQ, the possibility of recall bias or changes in dietary behavior among cases cannot be entirely excluded. Some breast cancer patients, upon experiencing symptoms or after receiving a diagnosis, may have modified their dietary habits, which could affect the accuracy of reported intake and lead to reverse causality. To minimize such bias, dietary assessments for cases were conducted shortly after diagnosis and before the initiation of treatment, with instructions to reflect on their habitual intake during the year prior to diagnosis. Nevertheless, these efforts may not completely eliminate bias inherent to retrospective recall.

In addition, the selection of controls from individuals participating in routine health check-ups may introduce a degree of selection bias. These individuals might differ from the general population in terms of health awareness and dietary behavior, potentially resulting in a healthier diet profile. In South Korea, however, the national health screening program provided free of charge through the National Health Insurance Service is mandatory for employees and workplace subscribers, and strongly recommended for the entire population. This significantly reduces the potential for selection bias and provide a practical and relevant comparison group for epidemiologic studies. Nevertheless, the strength of observed associations may still be underestimated due to differential health behaviors among participants.

Given these concerns, the case-control design limits the ability to establish a clear temporal relationship between diet and breast cancer risk. Moreover, the observed associations between plant-based dietary patterns and breast cancer should not be interpreted as causal. Future prospective cohort studies with pre-diagnostic dietary data and long-term follow-up would be better suited to address issues of temporality and reduce the likelihood of reverse causation or selection bias. Additionally, we were unable to account for certain potential confounding factors such as reproductive history, genetic predisposition, stress, and sleep, which may have influenced the observed associations. As dietary patterns in Korea have changed since the time of data collection (2007–2008), future studies using more recent cohorts are needed to confirm whether the observed associations persist in the context of evolving dietary behaviors and to further investigate biological mechanisms, including hormonal and receptor-related differences, that may explain subgroup variation.

In conclusion, our study shows that a higher PDI is associated with a lower risk of breast cancer, while a higher uPDI is linked to an increased risk. These findings highlight the importance of overall dietary patterns and quality, rather than focusing solely on specific foods like vegetables and fruits. Our results support the recommendation to increase plant food intake in dietary guidelines for breast cancer patients. This study may help inform future dietary guidelines for breast cancer prevention, with an emphasis on maintaining nutrient balance. Future cohort studies are needed to investigate the long-term associations of diet on breast cancer incidence and recurrence, including among patients undergoing treatment.

SUPPLEMENTARY MATERIAL

Supplementary Table 1

Scoring system and classification of food items