Associations of the Healthy Beverage Index (HBI) and the risk of Breast Cancer (BrCa): a case–control study

Abstract

Background

Breast cancer (BrCa) is one of the leading causes of cancer-related deaths. There are several factors for getting BrCa, including some changeable factors related to lifestyle like unhealthy dietary patterns, so modifying them can prevent one third of the complications and deaths caused by BrCa. Therefore, we decided to investigate the relationship between Healthy Beverage Index (HBI) and the risk of BrCa.

Methods

In this hospital-based case–control study, 253 patients with BrCa and 267 non-BrCa controls were enrolled. Food consumption was recorded to calculate the HBI score using a semi-quantitative Food Frequency Questionnaire (FFQ). Additionally, by using binary logistic regression analysis with adjustment for confounders, the relationship between HBI and the risk of BrCa were assessed. HBI was established by Duffey et al. and is used to evaluate the overall quality of beverage intake and identify changes in consumption.

Results

Mean ± SD of age and BMI of the study participants were 47.92 ± 10.33 years and 29.43 ± 5.51 kg/m², respectively. Patients with BrCa considerably outperformed controls in terms of waist circumference (WC), age at first pregnancy, history of abortion, and number of children(Pvalue < 0.05). Compared with those in the lowest quartile of HBI, subjects in the highest quartile had higher intake of energy, carbohydrate, protein, fat, fiber, sodium, potassium, calcium, magnesium, zinc, vitamin C, E, B9, fruits, vegetables, fish and nut as well as lower BMI and WC (Pvalue < 0.05). After adjustment for potential confounders, individuals in highest compared to lowest quartile of HBI had significantly lower risk of BrCa for total population (odds ratio (OR): 0.40; 95% confidence interval (95%CI): 0.21–0.76, Pvalue < 0.05), premenopausal (OR: 0.38; 95% CI: 0.16–0.92, Pvalue = 0.013), and postmenopausal (OR: 0.27; 95% CI: 0.10–0.78, Pvalue = 0.023).

Conclusion

Findings of this study suggested that higher HBI score decreased the risk of BrCa. However, further investigation is needed.

Introduction

Breast cancer (BrCa), with nearly 2.3 million new cases in 2020, was the most often identified disease and the main cause of cancer-related deaths in women globally in 2018 [1, 2]. There are several factors for getting BrCa, including some changeable factors related to lifestyle, the most important of which are physical inactivity [3] and dietary factors [4]. The most important modifiable risk factors for BrCa are unhealthy dietary patterns [5, 6]. For instance, consuming enough and proper vegetables and soft drinks, industrially produced juices, fried meals, and sweets were found to be risk factors for BrCa [7]; as a result, altering dietary habits can avoid one third of the difficulties and fatalities brought on by BrCa [8]. In order to improve healthy beverage choices and effectively evaluate drinking patterns quality in adults, the Healthy Beverage Index (HBI) was developed [9, 10].

The HBI can be utilized to identify the combined effects of numerous drinks as opposed to the marginal impact of a single drink on health outcomes [10]. It comprises fluid intake, eight beverage types, and beverage energy overall [10]. Drinks including milk, coffee, tea, and other unsweetened beverages might affect your overall health in a variety of ways [10]. In general, there are few evidences and studies about the potential effects of various types of drinks on the risk of developing chronic diseases, including breast cancer. In several studies, an increase in the risk of breast cancer and also the death rate was observed with sugar-sweetened beverages and even milk. However, in some evidences, no significant relationship was observed, and even in one study, replacing these types of drinks with tea, coffee, and drinking water did not show a significant effect on disease outcomes [11,12,13].

Due to the confusing nature of the literature and the fact that no study has specifically analyzed HBI in connection to BrCa risk, we set out to conduct the first ever study to look at the relationship between HBI and BrCa risk among Iranian women. Also, in this study, we examined food intake and other factors related to the disease between two groups of women under study, as well as the desired index quartiles.

Material & methods

Case-control research was used in this study. Two hundred and fifty-three persons with BrCa and two hundred and sixty-seven people without the illness were among the 520 participants in this study, who were chosen from Tehran’s Hazrat Rasool and Taleghani Hospitals during the years of 2018 and 2019. The minimal required sample size was calculated using the ability to detect an OR of 2 with a case to control ratio of 1 : 1, power of 90%, and a type I error ($\:\alpha\:$) rate of 5%.

A recent (within six months) diagnosis of BrCa was made by an oncologist in a patient with histologically proven BrCa. Our exclusion criteria included the presence of metastases, the history of previous malignancies, and hormone-related conditions such endometriosis or polycystic ovary syndrome (PCOS). The samples for control group did not have any history of cancer (benign and malignant), hormone-related and inflammatory diseases; they were chosen from other wards of hospitals like ophthalmology, otolaryngology, dermatology, and aesthetics. Also, they have had a regular diet for the past 6 months. The age and BMI matching was done between two groups. Furthermore, to ensure the absence of inflammation, an internal medicine specialist examined each member of this group, and also laboratory tests were done. We used a valid short form of the International Physical Activity Questionnaire (short IPAQ) to determine the level of physical activity of participants. The method of calculating the amount of physical activity in terms of MET-minutes/week was calculated as follows: Walking MET-minutes/week = 3.3 × Walking minutes × Walking days. Moderate MET-minutes/week = 4.0 × Moderate-intensity activity minutes × Moderate days. Vigorous MET-minutes/week = 8.0 × Vigorous-intensity activity minutes × Vigorous-intensity days. Total physical activity MET-minutes/week = Sum of Walking + Moderate + Vigorous METminutes/week scores.

Also, we obtained from all individuals an informed written consent.

Dietary assessment

A reliable semiquantitative food frequency questionnaire with 168 food items was utilized to measure food consumption in comparison to the previous year [14]. The main design element of this FFQ was to replicate normal Iranian cuisine with a standard serving size. For each dish, the participant filled out the FFQ to record how much of a regular quantity they typically ate and how many times they had consumed it. Each meal was consumed in the following ways: never, twice to three times per month, once per week, four to six times per week, and every day. Typical Iranian family measures were used to transform the portion sizes to gram [15]. The daily nutrient intakes for each patient were calculated using the national nutritional databank of the United States Department of Agriculture (USDA) [16]. To determine the consumption of daily nutrient and energy for each individual, Nutritionist IV software was applied.

Assessment of anthropometric variables

Standard techniques were used to acquire anthropometric data. The patients’ weights were determined while they were dressed comfortably and barefoot on a Seca digital scale (manufactured in Germany) with a precision of 100 g. Using a tape meter, standing height was measured with bare feet. The Body Mass Index (BMI) was determined by dividing the weight (kg) by the square of height (m2). The narrowest point of a non-elastic tape was used to measure waist circumference (WC) without applying pressure to the body’s surface.

Calculation of Healthy Beverage Index (HBI)

The Healthy Beverage Index (HBI) was established by Duffey et al. [10] and may be used, like the Healthy Eating Index, to evaluate the overall quality of beverage intake and identify changes in consumption. Health changes are connected to patterns. All beverages that were registered as beverages were split into eight kinds by the beverage guidance system. 100% fruit juice, water, unsweetened coffee and tea, low-fat milk, diet beverages (such as caffeine-free coffee and tea and other artificially sweetened beverages), alcohol (such as beer, wine, and spirits), and full-fat milk are all acceptable drinks. There were eight different types of drinks that people drank, including fruit drinks, sweet coffee and tea, and soft drinks. A higher number denotes better compliance with the drinking norm and a healthy drinking habit, and the final HBI score ranges from 0 to 100 [10]. Because our target group in this study did not use diet drinks (rated between 0 and 5) or alcohol (scoring between 0 and 5), the maximum final HBI score was 90. Because the goal of this study was to investigate adherence to healthy beverage intake recommendations rather than overall fluid intake, fluids ingested as part of a meal (such as soup) were eliminated.

Statistical analyses

We used SPSS software (version 19.0; SPSS Inc, Chicago IL) for all of our statistical analyses. The Shapiro-Wilk tests were used to assess the normality of the variables. For quantitative factors, the baseline characteristics and dietary intakes were presented as mean ± standard deviation (SD), and for qualitative variables, as a number and a percentage. Using independent sample T-Tests or, if necessary, its non-parametric counterpart (Mann-Whitney test) and chi-squared tests for continuous and categorical variables, respectively, we compared the data across two groups. Conditional logistic regression model was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) adjusted for multiple covariates in a different model and in all results, the significance level was determined as P < 0.05.

We adjusted the results in three models using a priori selected potential confounders, which included: model 1- age and BMI, model 2- additional adjustment for waist circumference, physical activity, and energy intake, model 3- the latter model plus first pregnancy age, number of children, history of abortion, hormone replacement therapy, use of NSAIDs drugs, and vitamin D supplementation. In adjusted models, confounders were used from statistical and conceptual approach respectively. In this way, the variables with Pvalue < 0.2 were considered as possible confounders and were entered into the logistic regression and the odds of getting cancer was investigated. Also, in the conceptual approach of adjusting confounders in the model 3, possible confounders were selected based on clinical concepts and based on past articles and added to other confounders.

Results

Specifically, the case groups’ mean (SD) age and BMI were 48.91 (10.46) years and 29.61 (4.55) kg/m², respectively, while the control groups’ mean age and BMI were 47.13 (10.08) years and 29.07 (5.39) kg/m², respectively.

Table 1 displays the demographics, way of life, and medical background of research participants in the case and control groups. Patients with BrCa considerably outperformed controls in terms of waist circumference (WC), age at first pregnancy, history of abortion, and number of children(Pvalue < 0.05). Additionally, the case group’s average use of vitamin D supplements, HBI score, hormone replacement medication, and non-steroidal anti-inflammatory drugs (NSAIDs)drugs was much lower than that of the control group (Pvalue < 0.05). For other traits and factors, however, there were no appreciable distinctions between the case and control patients.

Table 1 Demographic, anthropometric and lifestyle characteristics of participants across case and control groups

Table 2 shows the mean food consumption of the research participants based on the case and control groups. In comparison to controls, subjects with BrCa consumed more macronutrients (energy, carbohydrates, and fat) as well as saturated fatty acids (SFA), cholesterol, carbohydrates, sodium, folate, iron, sugar-sweetened beverages, grains, and starches, while consuming less protein, potassium, phosphorus, calcium, vitamin B12, and micronutrients antioxidants like zinc, magnesium, and vitamins E, C, and D (Pvalue < 0.05). The food intakes, anthropometric measurements, and lifestyle traits of study participants across HBI quartiles are displayed in Table 3. Higher intakes of calories, carbohydrate, protein, fat, fiber, salt, potassium, calcium, magnesium, zinc, vitamin C, E, and B9, as well as reduced BMI and WC, were seen in patients in the highest quartile of HBI compared to those in the lowest quartile(Pvalue < 0.05). In addition, HBI score and physical activity level increased significantly during the index quartiles. Tukey’s test was used to compare food groups and nutrients two by two. The difference between quartiles 1 and 2, 1 and 3, 1 and 4, 2 and 3, 2 and 4, 3 and 4 for energy, potassium, calcium, zinc, magnesium, vitamin C, vitamin B12, and fruit was reported significant. The difference between quartiles 1 and 2, 1 and 3, 1 and 4, 2 and 4, 3 and 4 was significant for fiber, vitamin B9 and vegetables intake. The difference between quartiles 1 and 2, 1 and 3, 1 and 4, 2 and 4 for protein, the difference between quartiles 1 and 3, 1 and 4, 2 and 4 for vitamin D and the difference between quartiles 1 and 4, 2 and 4 for sodium intake were significant.

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

Table 3 Dietary intakes, anthropometric, and lifestyle characteristics of study participants across quartiles of HBI

Table 4 presents the odds ratios (ORs) and 95% confidence intervals (CIs) for BrCa patients based on HBI. In the crude model, highest quartile of HBI scores compare to the lowest quartile, a decrease in the odds of BrCa was observed for the whole population of women (OR: 0.29; 95% CI: 0.17–0.50), premenopausal (OR: 0.30; 95% CI: 0.15–0.63), and postmenopausal women (OR: 0.20; 95% CI: 0.08–0.48). Furthermore, after additional adjusting for potential confounders in the final adjusted model, the reduction in the odds of BrCa remained significant (OR: 0.40; 95% CI: 0.21–0.76 for total population, OR: 0.38; 95% CI: 0.16–0.92 for premenopausal, and OR: 0.27; 95% CI: 0.10–0.78 for postmenopausal women).

Table 4 Odds ratio (OR) and 95% confidence interval (CI) for breast Cancer based on quartiles of Healthy Beverage Index (HBI)

Discussion

Our findings showed a significant and potential relationship between HBI and reducing the odds of BrCa in Iranian women after adjusting for potential confounders. This potential relationship was also observed in women before and after menopause. Although it was slightly lower in pre-menopausal age than after menopause, this decrease in odds was observed.

Rare and contradictory studies on the relationship between several drinks and the risk of cancer, particularly BrCa, exist. For instance, one study found that drinking milk may raise the odds of developing BrCa [13], although no correlation was identified in another research [13]. Also, it was indicated that higher mortality of BrCa was associated with high intake of fruit juice as well as substituting coffee, tea, or water with sugar-sweetened beverages (SSB) was related with a reduced risk of mortality [12]. However, substituting SSB with other low or high calorie liquids, such as fruit juice, skim/low-fat, or whole milk, was not related with noticeably improved survival [12]. Additionally, higher coffee and tea consumption following BrCa diagnosis was associated with reduced all-cause mortality [11]. There are dietary factors related to the risk of BrCa, which are inconclusive. For instance, a diet high in fruits, vegetables, seafood, and olive oil may lessen the hazards associated with BrCa [17]. Juices, fried meals, and sweets made in factories may raise the risk of BrCa [7]. A significant prospective research by Fiolet et al. found that eating ultra-processed foods increased total cancer risk, particularly BrCa risk, among 104 980 individuals who were at least 18 years old [4]. According to population based prospective cohort study by, SSB was related with the risk of total cancer including BrCa [18]. Epidemiological findings regarding BrCa risk associated with beverage consumption are conflicting. According to Boyle et al.‘s meta-analysis, there is no link between drinking sweetened, carbonated beverages and the risk of cancer, especially BrCa [19]. According to Chen et al.‘s research, drinking milk or other dairy products does not significantly increase the odds of developing BrCa [13]. According to Dong et al.‘s meta-analysis of prospective cohort studies, increasing consumption of dairy products in general—not just milk—may be linked to a lower risk of BrCa [20]. In this investigation, we evaluate the relationship between HBI and the risk of BrCa. HBI may be used to identify the cumulative effects of numerous drinks rather than the marginal impact of a single drink on health outcomes since it can assess the quality of adult drinking behaviors [10]. It contains fluid intake, eight beverage types, and beverage energy overall [10]. In comparison to controls, BrCa patients exhibited considerably higher HBI scores, according to our research. Numerous research have been done on the connection between drinking a variety of drinks and the risk of brca, but to the best of our knowledge, this is the first study to assess the association between HBI and BrCa risk in Iranian women.

The findings of our study also showed that some micronutrients, including potassium, phosphorus, calcium, zinc, magnesium and vitamins E, C and D, received less in the case group compared to the control group. Studies suggest that lower intake of these micronutrients, which are usually associated with lower fiber intake, can increase and maintain weight and body fat mass. This accumulation and storage of fat in the body is usually associated with an increase and retention of estrogen in the tissues and can increase the risk of chronic diseases, especially hormone-related cancers such as BrCa [21]. Therefore, differences in these nutrients may be clinically impactful. Special attention should be paid to vitamin D, since it plays a key physiological role in the development and function of the mammary gland [22], although the literature remains conflicting regarding vitamin D status and the risk BrCa. For instance, a meta-analysis of 9 prospective studies suggests a 12% decrease in the risk of BrCa in postmenopausal women for each 5 ng/mL increase in 25(OH) D [23]. However, in a RCT including 36,282 postmenopausal women, a reduction in BrCa (in situ) was found for those patients who underwent 400 IU/d of vitamin D3 combined with 1000 mg/d of elemental calcium carbonate [22]. In our study, the control group reported higher use of vitamin D supplements compared to the case group (24.3% vs. 14.6%, p = 0.005). Nevertheless, due to the nature of our study design and the lack of control over the dosage across vitamin D supplements, we cannot infer that vitamin D supplements are protective for BrCa. Interestingly however, the Vitamin D and Omega-3 Trial (VITAL) represents ongoing research that may be able to elucidate the clinical magnitude of supplementing vitamin D in preventing cancer by addressing the effect of 2000 IU/d vitamin D3 with or without 1 g of omega-3 fatty acids in 25,871 healthy subjects.

The present study’s strong, comprehensive consideration models are one of its advantages. Additionally, since these individuals had just been diagnosed with the condition for a maximum of 6 months, it was far less likely that the sickness would have caused a change in their eating habits. The 168-item food frequency questionnaire that was employed in this study covers the majority of the foods that our study sample ate. Although this study is innovative, there are certain limitations that should be mentioned. Some confounders may not have been taken into account despite the possibility of confounders being taken into account in this study’s analysis. Although we discovered evidence of a link between HBI and BrCa, the retrospective methodology we used in this study prevented us from establishing causality of the observed correlations. Therefore, this finding has to be verified in further prospective investigations and RCTs. Additionally, data were gathered through self-reporting techniques, which are known to be linked to either excessive or inadequate reporting. However, by employing skilled interviewers and technologies with strong validation, we aimed to mitigate this. The statistical methodology was also suitable for reporting at the group level. Another research drawback might be the modest alterations in some dietary products between the time of the interview and before the diagnosis. The precise number of participants who altered their diet was excluded from the research, and we also looked at pre-diagnosis consumption for each food item. Finally, more studies based on solving these limitations and also with follow-up and a higher sample size are suggested.

Conclusion

We found that higher HBI was associated with reduced odds of BrCa in the overall female population as well as in pre- and postmenopausal age. In general, the dietary pattern reflected by this index can serve as a useful guide and recommendation for the prevention of chronic diseases, including breast cancer in women of different menopausal ages, and is of interest to specialists and other relevant experts and consultants.