Introduction

Breast cancer is the most commonly diagnosed malignancy and the second leading cause of cancer-associated mortality among women [1]. According to GLOBOCAN estimates, approximately 2 million new breast cancer cases and 626,679 breast cancer deaths occurred worldwide in 2018 [2]. National estimates in Iran have shown that the incidence rate of breast cancer is 33.21 per 100,000 in women [3]. Despite advances in the diagnosis and treatment, the rising prevalence of breast cancer is still a problem [1]. Therefore, it is crucial to identify contributing factors to the occurrence and development of breast cancer.

Several non-modifiable and modifiable risk factors are recognized to contribute to the progression of breast cancer Diet, as a modifiable risk factor, is known to play an important role [4, 5]. Previous studies have demonstrated that diets rich in alcohol and processed meat were associated with a greater risk, while those rich in fruits and vegetables were associated with a lower risk of breast cancer [6,7,8]. Great attention has also been paid to the role of dairy product consumption. Dairy products are rich in protein and calcium, which might have a beneficial role in breast cancer prevention [9]. Consumption of these products was associated with increased plasma insulin-like growth factor Ι (IGF-Ι) concentrations [10], which in turn has a detrimental effect on cancer cell lines [11]. Findings from meta-analyses about the association between dairy product consumption and risk of breast cancer have also been controversial. While some studies have reported an inverse association [12, 13], others reached null findings [14, 15].

Evaluating the association between dairy product consumption and risk of breast cancer is especially relevant for Middle East countries, where dairy product consumption is less than the recommended amounts [16]. In addition, given the dietary culture of people in this region, most dairy products are consumed as whole full-fat non-pasteurized products [17]. Furthermore, most studies on the association between dairy product consumption and risk of breast cancer came from western countries and limited data are available in developing countries. Given the high prevalence of breast cancer in this area along with different dietary culture and other contributing environmental factors, the present study aimed to examine the association between dairy products consumption and risk of breast cancer among Iranian women.

Methods

Study population

This population-based case–control study was conducted on women aged 30 years or older, who were referred to hospitals or private clinics in Isfahan, Iran, between July 2013 and July 2015. To compute the required sample size, we considered the type I error of 5% and study power 80%. Therefore, we used the suggested formula for case–control studies to calculate the needed sample size, in which we assumed the odds ratio as 1.5, meaning that non-healthy dietary intakes raise the risk of breast cancer by 1.5 times. Then, we defined controls-to-cases ratio as 2. Based on these assumptions, we needed almost 350 patients in the case group and 700 controls.

Cases were patients in whom breast cancer had been diagnosed by physical examination, mammography and pathological verification during the maximum of the last 6 months. By using the convenience-sampling method, patients were selected. Given that patients were only requested to respond to some questionnaires, more than 90% of them agreed to take part in the study. Patients were those who had undergone surgical resection, chemotherapy and/or radiotherapy for breast cancer. In the present study, patients with a history of cancer (exception of current breast cancer) and those who were taking hormone replacement therapy were not included. In addition, we did not include those who were on a special diet (e.g., restricted diet for weight reduction and disease-related diet such as diabetic diet).

After recruiting cases, individually age-matched controls (± 5 years) were randomly selected from healthy women, who had no relationship with breast cancer patients. In addition to age, we did our best to match controls in terms of their place of residence as a proxy measure of socioeconomic status (SES) with the cases. Controls met our inclusion criteria (female, Iranian nationally, no history of any malignancy, cysts and medical disorder, having no special diet or hormone replacement therapy) were selected from the general adult population recorded in the Iranian Health Centers data.

All participants provided written informed consent. The study was ethically approved by the Ethical Committee of Isfahan University of Medical Sciences, Isfahan, Iran.

Dietary intake assessment

We used a 106-item semi-quantitative dish-based food frequency questionnaire (FFQ) to assess participant's dietary intakes [18]. The questionnaire was based on Willett format and has been validated in earlier studies. Detailed information about the design and validity of this dish-based FFQ has been reported elsewhere [19]. We completed the questionnaires through face to face interview. The FFQ contained five main categories of foods and dishes: (1) Iranian mixed dishes (cooked or canned, 29 items); (2) grains and carbohydrate-rich foods (including different breads, cakes, biscuits and potato, 10 items); (3) dairy foods (milk, yogurt, butter and cream, 9 items); (4) fruit and vegetables (22 items); and (5) miscellaneous foods and beverages (including sweets, fast foods, nuts, desserts and beverages, 36 items). Participants were requested to report their consumption of foods and dishes based on nine multiple choice frequency response categories which varied from “never or less than once a month” to “12 or more times per day.” All reported daily intakes of foods were converted to grams per day, for which we used a booklet that contained information about the weight of household measures [20]. To obtain grams of foods and dishes, we considered reported frequency along with given portion size as well as the composition of mixed dishes. Total nutrient intake for each person was computed by USDA food composition-based software, Nutritionist IV. However, for local Iranian foods, the database of the software was modified based on an old Iranian food composition dataset.

Dairy products’ definition was done based on the US Food Guide Pyramid [21]. In the present study, low-fat dairy consumption was defined as consumption of pasteurized milk, low-fat cheese, low-fat yogurt, dough (a yogurt-based drink) and low-fat Kashk (a fermented product derived from dairy). High-fat dairy consumption was calculated by summing up the consumption of whole milk, high-fat cheese, high-fat yogurt and high-fat Kashk.

A classic validation study was not performed for this FFQ. However, several publications, which assessed the association between dietary factors derived from this FFQ and different health-related outcomes, indicated that the questionnaire provided valid and reliable measures of the average long-term dietary intakes [19].

Assessment of breast cancer

All patients with BC were females with newly diagnosed stage I–IV breast cancer from Iranian nationality, for whom in situ or invasive status of BC was based on physical examination in combination with mammography and confirmed by pathological assessment. To do mammography, patients were asked to be in a standing position, then breast was compressed for a few seconds between the pages and photography was done.

Assessment of other variables

To assess participants’ information about age, marriages, place of residence, educational status, alcohol consumption, supplement use, smoking, menopausal status, breastfeeding history, family history of breast cancer and history of disease, we applied a pretested questionnaire. Physical activity of participants was assessed using the short form of International Physical Activity Questionnaire (IPAQ) through face to face interviews. All results were expressed as Metabolic Equivalents-hours per week (MET-h/week). A trained nutritionist measured body weight, while the study subjects were without shoes with light clothing, using a weighing scale (Seca, Hamburg, Germany) to the nearest 100 g. Height was measured by a mounted tape, without shoes in a standing position near the wall, using a stadiometer (Seca, Hamburg, Germany) to the nearest 0.5 cm. BMI was calculated by dividing weight in kilograms by height in meters squared.

Statistical methods

To evaluate the relationship between dairy product consumption and risk of breast cancer, individuals were classified based on quartiles of low- and high-fat dairy product consumption. Then, one-way analysis of variance (ANOVA) and Chi-square were applied, respectively, to examine the differences in terms of continuous (such as age, BMI and physical activity) and categorical variables across categories of dairy intake. To assess the relationship between high- and low-fat dairy product and odds of breast cancer, unconditional multivariable logistic regression was applied in different models: (i) a model adjusted for age and energy intake; (ii) a model adjusted for age, energy intake, and other patient backgrounds (such as marriage (single/ married), place of residence (urban/rural), educational status (educated/non-educated), alcohol consumption (yes/no), smoking (non-smoker/smoker), menopausal status (premenopausal/postmenopausal), family history of breast cancer (yes/no), history of disease (yes/no), physical activity (MET-h/week), supplement use (yes/no) and breastfeeding history (yes/no)); (iii) a model adjusted for age, energy intake, other patient’s background variables and food intake (such as fruit, vegetables, meat, nuts, legumes and soy, total grain and dietary total fat); and (iv) a model adjusted for BMI to investigate an obesity-independent association [22]. In addition to considering total dairy intake, we also were interested to examine the associations with individual dairy products including milk, yogurt and cheese. To examine the trend of odds ratios across increasing categories, quartile or tertile categories of intake were treated as an ordinal variable. Furthermore, crude and multivariable-adjusted odds ratios of breast cancer based on additional consumption of milk (50 g/day; equal to one quarter of a glass), yogurt (50 g/day; equal to one quarter of a glass) and cheese (30 g/day) was also evaluated. All analyses were performed by using SPSS software (version 26 Corp, version 19, Chicago, IL, USA). P value less than 0.05 was considered as statistically significant.

Results

The general characteristics of study participants separately by cases and controls are presented in Table 1. Such comparison across quartiles of low- and high-fat dairy product consumption is given in a Supplementary Table 1. Compared to subjects in the control group, cases were more likely to be older, single, postmenopausal, have a family history of breast cancer and had lower mean BMI and were less likely to be educated. Participants in the highest quartile of low-fat dairy product intake had a higher BMI and were more likely to be younger, live in urban areas, university graduated, alcohol consumer and were less likely to be smoker and postmenopausal and have a low socioeconomic background compared with those in the lowest quartile. Looking across categories of high-fat dairy products, we observed that subjects in the top quartile had lower BMI and more likely to be older, have a low socioeconomic background , have a family history of breast cancer, supplement user and postmenopausal, less likely to live in urban areas, university graduated and alcohol user.

Table 1 Characteristics of the study participants across subjects with and without breast cancer
Full size table

Dietary intakes of cases and controls are indicated in Table 2. We have also reported dietary intakes of study participants across quartiles of low and high-fat dairy product consumption in Supplementary Table 2. Patients with breast cancer consumed higher amounts of high-fat dairy products, milk, yogurt, fruit, sugar-sweetened beverages and sweets, energy and fat and lower amounts of low-fat dairy products, cheese, vegetables, nuts, legumes and soy, grains, sodium, carbohydrate, protein, fiber and dietary folic acid, compared with those without breast cancer. Compared to individuals in the lowest quartile, those in the highest quartile of low-fat dairy product consumption had higher intakes of fruit, vegetables, low-fat dairy, milk, yogurt, cheese, nuts, legumes and soy, meats, sodium, energy, carbohydrates, proteins, fats, dietary fiber, dietary folic acid and dietary vitamin E and lower intakes of grains and high-fat dairy. Higher consumption of high-fat dairy products was associated with higher consumption of fruits, high-fat dairy, milk, yogurt, energy, carbohydrate, protein, fat, dietary fiber and vitamin E as well as lower intake of vegetables, low-fat dairy, cheese and grains.

Table 2 Dietary intakes of study participants across case and control groups
Full size table

Crude and multivariable-adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for breast cancer across quartiles of low- and high-fat dairy product consumption are provided in Table 3. In the crude model, individuals in the top quartile of low-fat dairy product intake had lower risk of breast cancer compared to those in the bottom quartile (OR 0.16; 0.95% CI 0.10–0.23). After controlling for age and energy intake, this inverse association did not alter, such that those in the top quartile were less likely to have breast cancer than those in the bottom quartile (OR 0.10; 95% CI 0.06–0.16). When we further took dietary intakes into account, this inverse association remained statistically significant (OR 0.06; 95% CI 0.03–0.11). Additional controlling for BMI did not significantly change the association (OR 0.08; 95% CI 0.05–0.16). In terms of high-fat dairy product consumption, participants in the highest quartile were more likely to have breast cancer than those in the lowest quartile (OR 10.86; 95% CI 7.01–16.83). This positive relationship remained significant even after adjusting for age and energy (OR 7.98; 95% CI 5.07–12.57). In addition, further adjustment for other covariates did not considerably change this relationship (OR 10.17; 95% CI 6.22–16.62). After controlling for dietary variables (OR 7.87; 95% CI 4.55–13.60) and even BMI (OR 8.62; 95% CI 4.78–15.55), this positive association remained significant.

Table 3 Multivariable-adjusted odds ratios (95% CIs) for breast cancer across quartile categories of low- and high-fat dairy consumption
Full size table

Crude and multivariable-adjusted ORs and 95% CIs for breast cancer across tertiles of total milk, yogurt and cheese consumption are shown in Table 4. When we examined the associations by individual dairy products, we observed that individuals in the highest category of total milk intake had a greater risk of breast cancer than those in the lowest category (OR 2.59; 95% CI 1.89–3.55). However, controlling for covariates resulted in a weakened relationship in this analysis, but it was still significant (OR 1.76; 95% CI 1.16–2.65). In terms of total yogurt consumption, although a positive association was seen with breast cancer (OR for the highest vs. lowest tertile: 1.40; 95% CI 1.01–1.95), it disappeared after adjusting for potential confounders (OR 1.02; 95% CI 0.65–1.60). Examining the association with total cheese consumption, we found that compared with those in the bottom tertile, individuals in the top tertile tended to have a reduced chance of having breast cancer (OR 0.69; 95% CI 0.51–0.94). Such a significant inverse association was also observed after controlling for potential confounders (OR 0.62; 95% CI 0.42–0.91). However, when BMI was taken into account, the association was not statistically significant (OR 0.72; 95% CI 0.48–1.07).

Table 4 Multivariable-adjusted odds ratios (95% CIs) for breast cancer across tertile categories of milk, yogurt and cheese consumption
Full size table

Multivariable-adjusted odds ratios and 95% CIs for breast cancer per an additional consumption of a certain amount of dairy products are shown in Table 5. We failed to find any significant association per 50-g additional intake of total milk (OR 0.99; 95% CI 0.90–1.08) or total yogurt (OR 1.01; 95% CI 0.89–1.15) or 30-g cheese (OR 0.65; 95% CI 0.23–1.85) and risk of breast cancer after controlling for potential covariates. However, a significant positive association was observed between each additional 50-g consumption of whole milk and odds of breast cancer (OR 1.18; 95% CI 1.07–1.31). In terms of pasteurized milk, we did not find a considerable association between each 50-g additional intake of pasteurized milk and risk of breast cancer (OR 1.01; 95% CI 0.85–1.20). When the analysis was performed for high-fat and low-fat yogurt, a positive association was seen between each additional 50-g of high-fat yogurt consumption and odds of breast cancer (OR 4.16; 95% CI 3.11–5.56), while additional intake of 50-g low-fat yogurt consumption was associated with reduced risk of breast cancer (OR: OR 0.39; 95% CI 0.31–0.49).

Table 5 Multivariable-adjusted odds ratios and 95% CIs for breast cancer per an additional consumption of certain amount of dairy products
Full size table

Discussion

In this present population-based case–control study, we found a significant inverse association between low-fat dairy consumption and odds of breast cancer, while high-fat dairy intake was associated with elevated odds of this condition. These relationships remained considerable even after adjusting for plausible covariates. Regarding individual dairy product consumption including milk, yogurt and cheese, we observed a positive association between total milk consumption and risk of breast cancer, while no association between cheese or yogurt intake and breast cancer.

Breast cancer remains a major public health concern worldwide [1]. The role of diet in the etiology and progression of breast cancer is well known [5]. Among dietary factors that were examined in relation to odds of breast cancer, dairy product intake has received considerable attention. Our findings revealed an inverse association between low-fat dairy product intake and odds of breast cancer. Similar to our findings, a previous meta-analysis has indicated that low-fat dairy product consumption was inversely associated with risk of breast cancer [12]. However, some studies reached opposite findings. For instance, a prospective cohort study conducted on 35,372 UK women failed to find any significant association between low-fat dairy consumption and breast cancer [23]. Difference in study design and characteristics of study participants might provide some reasons for different findings. For example, Dunnerum et al. had included a high number of vegetarians in their study. Given the controversial findings across published studies and the nature of bias in case–control studies, like ours, it seems that additional prospective studies are required to further examine this issue, in particular in developing nations.

In the present study, a positive association was found between high-fat dairy intake and risk of breast cancer. Similar to our findings, others also reported such a positive association between consumption of high-fat dairy products and risk of breast cancer [24, 25]. In contrast, Farvid et al. found no significant association between high-fat dairy product intake in adolescence and early adulthood and risk of breast cancer in NHS. However, they reported that high-fat dairy products were positively associated with risk of estrogen and progesterone receptor-negative breast cancer and inversely with risk of estrogen and progesterone receptor-positive breast cancer [26]. It must be kept in mind that Farvid et al. used data from remote diet in their investigation, which might be less accurate than the one we used in the current study. However, their investigation was a prospective cohort study that has several advantages over the case–control design we applied in the current study. We did not collect information on the hormone receptor status of patients in this study. It seems that such information is necessary to accurately examine the association between diet and breast cancer. Lack of accurate data about diet–breast cancer links in developing countries, in particular considering the hormone receptor status, highlights the need for additional investigations in these countries in the future.

In the current study, we found that a positive association between total milk intake and breast cancer. Consistent with our findings, a cohort study on 52,795 North American women indicated that higher intake of milk was associated with an increased risk of breast cancer [27]. However, Knekt et al., in a cohort study on 4697 Finnish women, found an inverse association between milk consumption and risk of breast cancer [28]. The positive association between total milk intake and breast cancer might be attributed to high intake of unpasteurized whole milk (rich in trans fatty acids and saturated fatty acids) in Middle Eastern countries [17, 29, 30]. This was also the case in the current study, in which the mean consumption of unpasteurized whole milk was 72.36 g/day, while the consumption of pasteurized milk was 18.34 g/day. Trans fatty acids and SFA in whole milk might increase inflammation which can in turn neutralize the beneficial effects of milk against cancer [31]. Additional well-designed prospective studies, in particular in Middle Eastern nations, are required to shed light on this association.

Based on our findings, yogurt and cheese consumption was not associated with odds of breast cancer. Others also failed to find any significant association between yogurt or cheese consumption and odds of breast cancer [14, 32]. However, in a case–control study among Uruguayan women, high yogurt intake was associated with a reduced risk of breast cancer [33]. Such a protective association was also observed for cheese intake [34]. These conflicting results might be explained by the small sample size in earlier studies as well as different processing used for production of yogurt and cheese in different countries. In addition, lack of adjusting for different covariates as well as different characteristics of the study population might also be some further reasons.

Several mechanisms might explain the association between dairy product intake and odds of breast cancer. Compelling evidence has shown that different bioactive compounds of dairy products might affect this association [35]. The dairy content of conjugated linoleic acid (CLA), vitamin D and calcium might be responsible for its inverse association with breast cancer, while their contents of insulin-like growth factor I (IGF-I) and saturated fats might provide some reasons for their positive associations [9, 30, 36]. CLA manages breast cancer progression by neutralizing the effect of oxidative stress and induction of apoptosis on breast cancer cells [37]. Both vitamin D and calcium affect regulation of cell growth, leading to reduced benign epithelial proliferation [38]. In addition, IGF-I stimulates cancer cell proliferation and transformation through binding to IGF-I receptors in cell membrane and stimulating both MAPK/ERK and PI3K/AKT/mTOR signaling pathways [35, 39]. Saturated fatty acids in dairy products can elevate circulating estrogen levels as a carcinogenic component [40].

This study has several strengths. Adequate sample size, using a validated FFQ to collect information about dietary intake of individuals, taking into account several covariates in the analysis and including participants with newly diagnosed breast cancer are among the strengths of our study. However, the present study has some weaknesses. First, because of the case–control design of the study, our findings are subject to selection and recall bias. Second, given the use of FFQ, misclassification of study participants is unavoidable. Third, despite a comprehensive adjustment, the possible effect of residual confounders cannot be ignored. Fourth, no information about the hormone receptor status (estrogen or progesterone) of breast cancers was collected in the current study. Fifth, the findings of the present study are specific to a particular country and therefore cannot be easily generalized to other populations in the Middle East. Iranians consume high amounts of plant foods based on rice and bread as main sources of energy. High consumption of black tea along with sugar, potatoes, legumes and hydrogenated fats and low consumption of fruits and vegetables are highly prevalent among Iranians. With regard to dairy product consumption, despite a low intake of milk in the country, consumption of other dairy products such as cheese, yogurt and Dough (a yogurt drink) is at a moderate level. Additionally, Iran is an Islamic country where alcohol use is legally prohibited. Although there are some similarities in diet and lifestyle patterns among people living in the Middle East, it seems further studies should be conducted in other Middle Eastern countries to shed light on this issue.

In conclusion, the findings of the current study revealed that high intake of low-fat dairy was associated with a lower odds of breast cancer, while consumption of high-fat dairy products and milk was associated with an elevated odds. No significant association was observed between yogurt or cheese consumption and odds of breast cancer. Further studies with prospective design in Middle Eastern countries are required to confirm these findings.