Introduction: Breast cancer (BC) is the most common cancer among women globally. Vitamin D has been considered a protective factor; however, its relationship with any aspect of the disease remains controversial. Methods: A cross-sectional, single-center clinical study was conducted between 2015 and 2018, including 141 women diagnosed with BC and 239 women in the control group, with mean ages of 43.1 and 41.7 years, respectively (p = 0.103). Serum levels of vitamin D and lipid profile were measured. Clinical and nutritional data were obtained through interviews and medical records. Results: The vitamin D dosage presented an average value of 25.5 ng/mL and 31.0 ng/mL in the case and control groups, respectively (p < 0.001). The vitamin D cut-off point for discriminating the presence of BC was 27.45 ng/mL. Additionally, low-density lipoprotein cholesterol levels were higher in the case group (121.4 mg/dL) compared to the control group (110.7 mg/dL) (p = 0.002), whereas high-density lipoprotein cholesterol levels were lower in the case group (47.6 mg/dL) compared to the control group (53.3 mg/dL) (p = 0.001). Alcohol consumption was significantly higher in the case group than in the control group (2.7 vs. 5.3 doses/day; p < 0.001). Conclusion: The results indicate a significant association between lower vitamin D levels and BC, persisting after multivariate analysis (p < 0.001). These findings could inform prevention strategies, highlighting the importance of maintaining adequate vitamin D levels and potentially identifying a risk group.

Breast cancer (BC) remains a pressing global health concern, particularly affecting women, and surpasses the incidence of nonmelanoma skin cancers. Approximately one-sixth of all cancer-related mortalities among women are attributed to BC, making it the primary cause of death in 110 countries [1].

Despite early detection and treatment progress, new BC cases have shown little change over the past 2 decades [2]. This unchanging trend highlights the urgency of identifying effective strategies to mitigate risk factors. The factors contributing to BC can be categorized into unmodifiable and modifiable factors, the latter often related to lifestyle [3]. In this context, diet and vitamin intake are modifiable factors with complex implications for BC risk. High consumption of fruits and vegetables can reduce the risk of BC development. Conversely, increased fatty food intake and vitamin D deficiency may lead to a higher risk [4, 5].

The active form of vitamin D interacts with the vitamin D receptor (VDR), which is involved in various physiological functions. The VDR is present in normal and cancerous tissues, influencing antiproliferative and immunomodulatory effects. The presence of VDR in BC tissue and the potential antitumor effects of vitamin D indicate its role in disease development [6].

BC presents a challenge due to incomplete knowledge of its pathogenesis and the significant heterogeneity of the disease. The relationship between circulating 25-hydroxyvitamin D (25(OH)D3) levels and BC incidence varies across prospective population studies, fueling ongoing debates [7, 8]. This study aimed to investigate the correlation between BC occurrence, vitamin D levels, and associated risk factors, in a reference cancer center in Curitiba, Brazil.

The study was approved by the Ethics Research Committee of Erasto Gaertner Hospital (EGH; protocol number: CAEE: 51367415.2.0000.0098) in December 15, 2015. All participants received information about the study’s objectives, risks, and potential benefits and provided their consent by signing the informed consent form in duplicate.

Study Design and Patients

This cross-sectional study was conducted between March 2016 and August 2017 encompassing women aged 18–69 years diagnosed with invasive breast carcinoma. All participants with BC were originated from the primary care health system after suspicious lump for pathological diagnosis, and in case of malignancy, for curative oncological treatment at EGH, Curitiba, in Paraná state, in the south of Brazil. Only after histopathological analysis confirmed the disease, participants were invited for this study.

Cases were included in the study within 1 month from the diagnostic tumor biopsy and before having initiated any local (surgery resection or adjuvant radiotherapy) or systemic therapy (hormone therapy, adjuvant or neoadjuvant chemotherapy, or monoclonal antibody). For the control group, healthy female volunteers from the blood bank of EGH without any suspected or previous diagnosis of cancer or any precursor breast benign condition were recruited under invitation, after being selected under pairing characteristics with cases. They were matched to the case group based on age ±5 years, self-reported ethnicity, and weight ±10 kg, in a 2:1 ratio. The season of the year was also considered in the matching process, with controls included within 15 days of each case’s inclusion, aiming for a maximum of 10 cases and 20 controls/month. Exclusion criteria included incomplete medical record data, inflammatory, endocrine, autoimmune diseases, osteoporosis, and the use of vitamin D supplements or artificial tanning in the last 6 months.

From the initial records, a consecutive series of 390 women were recruited. However, 10 participants were later excluded: nine due to ongoing laboratory-monitored vitamin D use and one due to ductal carcinoma in situ diagnosis. Consequently, the final sample consisted of 380 women, comprising 141 with BC (case group) and 239 health controls (control group). The study design is depicted in Figure 1.

Fig. 1.

Flowchart of the study population.

Fig. 1.

Flowchart of the study population.

Close modal

Data Collection

During the recruitment phase, study participants were interviewed and completed a comprehensive self-reported questionnaire under supervision of biomedicine undergraduate students for any doubt. This questionnaire included questions about age, educational attainment, prior medical history, family cancer history, alcohol consumption, and extent of sun exposure.

To determine plasma biochemical parameters, 4 mL of whole blood was collected through peripheral venipuncture into an ethylenediaminetetraacetic acid tube. The quantitative determination of vitamin D [25(OH)D] was performed using a competitive immunoassay technique with Vitros Immunoassay Nutritional Assessment 25-hydroxyvitamin D (Ortho-Clinical Diagnostics, NJ, USA). Total cholesterol, high-density lipoprotein (HDL)-cholesterol, and triglyceride levels were measured using the dry slide technique with Vitros CHOL Slides, Vitros Direct HDL Slides, and Vitros TRIG Slides (Ortho-Clinical Diagnostics, NJ, USA). The calculation of low-density lipoprotein (LDL)-cholesterol followed Martin's formula [9].

All nutritional data were collected by the EGH oncological nutritionist for both cases and controls. The anthropometric evaluation measured weight, body mass composition with four-point bioimpedance test and height. To assess nutritional status, the body mass index (BMI) was calculated using the formula BMI = weight (kg)/height (m2), and classification was done according to the cut-off points recommended by the World Health Organization (WHO) in 2000. The food record questionnaire accessed dietary intake over the past 3 days. The macro and micronutrient content was determined based on the Nutritional Composition Table of foods consumed in Brazil [10].

Histopathological reports of breast tumor tissues were obtained through patients’ medical charts and the Pathology Department records of EGH. The expression status of estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 (HER) were determined by immunohistochemistry, and used for the molecular classification of tumors. For estrogen receptor and progesterone receptor status, expression was reported in percentages, and for HER, in scores from 0 to +++, following recommendation from the American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP) guidelines. In case of “indeterminate” range of HER, those specimens were re-classified in “not-amplified” or “amplified” after fluorescence in situ hybridization or chromogenic in situ hybridization.

Statistical Analysis

In the descriptive analysis, quantitative variables were presented with the mean, standard deviation, median, minimum, and maximum values. Categorical variables were described using absolute frequency and percentage. The normality condition of quantitative variables was assessed using the Kolmogorov-Smirnov test. The independent samples t test or the nonparametric Mann-Whitney test was employed to compare two groups concerning quantitative variables. The one-factor analysis of variance (ANOVA) model was used when dealing with more than two groups.

Categorical variables were analyzed using Fisher’s exact or χ2 tests. Correlation analysis between quantitative variables was performed by estimating the Pearson or Spearman correlation coefficients. Logistic regression models were adjusted to investigate the association of factors with BC, and the Wald test was used to evaluate the significance of variables [11]. The multivariate model initially included variables that demonstrated statistical significance in the univariate analysis.

The final model was derived using a stepwise backward variable selection approach, with an entry probability of 0.05 and a removal probability of 0.10. The estimated measure of association was the odds ratio, accompanied by 95% confidence intervals (CIs). Statistical significance was set at p < 0.05. Data were analyzed using the IBM SPSS Statistics software version 28.0 (IBM Corp., Armonk, NY, USA).

Participant Characteristics

The study population included 239 women in the control group and 141 in the BC case group. Table 1 presents a comparison of clinical and demographic variables between the groups. The mean age was 41.7 (median 40) and 43.1 years (median 45) in the control and case groups, respectively. The majority of participants self-reported themselves as white, representing 76% and 86% in the control and case groups, respectively. The case group had a lower level of education, with 48.2% having less than a high school education, compared to 22.6% in the control group (p < 0.001).

Table 1.

Sociodemographic and clinical characteristics of participants in the control and case groups

ControlCasep value*
Quantitative variables, mean±SD; median [range] 
 Age, years 41.7±7.8 43.1±7.9 0.103 
40 [24–65] 45 [21–61] 
 Total cholesterol, mg/dL 191.2±35.5 198.5±38.1 0.068 
 LDL-C Martin, mg/dL 110.7±29.5 121.4±32.8 0.002 
 HDL-C, mg/dL 53.3±14.1 47.6±10.6 <0.001 
 Triglyceride, mg/dL 161.8±75.9 176.2±96.3 0.122 
Categorical variables, n (%) 
 Race   0.019 
  White 164 (76.3) 122 (86.5)  
  Non-white 51 (23.7) 19 (13.5)  
 Education   <0.001 
  < High school 49 (22.6) 68 (48.2)  
  High school 88 (40.7) 57 (40.4)  
  > High school 79 (36.6) 16 (11.3)  
 Childbirth   0.268 
  0 39 (18.1) 15 (12.1)  
  1–2 137 (63.4) 67 (54)  
  ≥3 40 (18.5) 42 (33.9)  
 Breastfeeding   0.009 
  No 47 (22.2) 11 (10)  
  Yes 165 (77.8) 99 (90)  
 Oral contraceptive   0.296 
  No 40 (18.5) 26 (23.4)  
  Yes 176 (81.5) 85 (76.6)  
 Oophorectomy   0.020 
  No 214 (99.1) 105 (94.6)  
  Yes 2 (0.9) 6 (5.4)  
 Postmenopausal    
  No 172 (79.6) 77 (60.6)  
  Yes 44 (20.4) 50 (39.4) <0.001 
 HRT   0.127 
  No 201 (93.1) 117 (97.5)  
  Yes 15 (6.9) 3 (2.5)  
 BMI   0.137 
  Underweight/normal 75 (31.4) 53 (37.6)  
  Overweight 95 (39.7) 46 (32.6)  
  Obese 69 (28.9) 42 (29.8)  
 Diabetes   0.293 
  No 212 (98.1) 117 (95.9)  
  Yes 4 (1.9) 5 (4.1)  
 SAH   <0.001 
  No 198 (92.1) 96 (78.0)  
  Yes 17 (7.9) 27 (22.0)  
 Tabagism   0.786 
  No 153 (71.2) 83 (69.7)  
  Yes 62 (28.8) 36 (30.3)  
 Alcohol intake   0.002 
  No 162 (75.3) 52 (57.8)  
  Yes 53 (24.7) 38 (42.2)  
ControlCasep value*
Quantitative variables, mean±SD; median [range] 
 Age, years 41.7±7.8 43.1±7.9 0.103 
40 [24–65] 45 [21–61] 
 Total cholesterol, mg/dL 191.2±35.5 198.5±38.1 0.068 
 LDL-C Martin, mg/dL 110.7±29.5 121.4±32.8 0.002 
 HDL-C, mg/dL 53.3±14.1 47.6±10.6 <0.001 
 Triglyceride, mg/dL 161.8±75.9 176.2±96.3 0.122 
Categorical variables, n (%) 
 Race   0.019 
  White 164 (76.3) 122 (86.5)  
  Non-white 51 (23.7) 19 (13.5)  
 Education   <0.001 
  < High school 49 (22.6) 68 (48.2)  
  High school 88 (40.7) 57 (40.4)  
  > High school 79 (36.6) 16 (11.3)  
 Childbirth   0.268 
  0 39 (18.1) 15 (12.1)  
  1–2 137 (63.4) 67 (54)  
  ≥3 40 (18.5) 42 (33.9)  
 Breastfeeding   0.009 
  No 47 (22.2) 11 (10)  
  Yes 165 (77.8) 99 (90)  
 Oral contraceptive   0.296 
  No 40 (18.5) 26 (23.4)  
  Yes 176 (81.5) 85 (76.6)  
 Oophorectomy   0.020 
  No 214 (99.1) 105 (94.6)  
  Yes 2 (0.9) 6 (5.4)  
 Postmenopausal    
  No 172 (79.6) 77 (60.6)  
  Yes 44 (20.4) 50 (39.4) <0.001 
 HRT   0.127 
  No 201 (93.1) 117 (97.5)  
  Yes 15 (6.9) 3 (2.5)  
 BMI   0.137 
  Underweight/normal 75 (31.4) 53 (37.6)  
  Overweight 95 (39.7) 46 (32.6)  
  Obese 69 (28.9) 42 (29.8)  
 Diabetes   0.293 
  No 212 (98.1) 117 (95.9)  
  Yes 4 (1.9) 5 (4.1)  
 SAH   <0.001 
  No 198 (92.1) 96 (78.0)  
  Yes 17 (7.9) 27 (22.0)  
 Tabagism   0.786 
  No 153 (71.2) 83 (69.7)  
  Yes 62 (28.8) 36 (30.3)  
 Alcohol intake   0.002 
  No 162 (75.3) 52 (57.8)  
  Yes 53 (24.7) 38 (42.2)  

HRT, hormone replacement therapy; BMI, body mass index; SAH, systemic arterial hypertension; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol.

*Logistic regression model and Wald’s test, p < 0.05.

The gynecological history revealed differences in postmenopausal status, with a significant disparity between the groups: 39.4% of the case group were postmenopausal compared to 20.4% in the control group (p < 0.001). Regarding the number of biological children, most women in the case group (33.9%) had three or more children, compared to only 18.5% in the control group (p = 0.008). More than half of the participants practiced breastfeeding, which was significantly more common in the case group than in the control group (p = 0.009). The use of hormonal oral contraceptives (HOC), oophorectomy, and hormone replacement therapy was also assessed and showed no significant differences (Table 1).

Regarding clinical characteristics, 22% of women with BC reported the presence of systemic arterial hypertension compared to 7.9% in the control group (p < 0.001). Other clinical characteristics, such as diabetes mellitus and BMI, showed no significant differences between the groups. The lipid profiles also differed; LDL-C levels were higher (p = 0.001), and HDL-C levels were lower (p < 0.001) in the case group compared to the control group. Total cholesterol and triglyceride levels were similar between the groups. Among the participants’ habits, alcohol consumption was more frequent among women in the case group (42.2% vs. 24.7%; p = 0.002), whereas smoking prevalence did not differ between groups.

Several known risk factors for BC were compared between the case and control groups (Table 2). The case group exhibited a higher lifetime usage of HOC (127.7 vs. 70.4 months; p = 0.038) and a greater alcohol intake (5.3 vs. 2.7 doses per week; p < 0.001). Age in menarche, menopause, first childbirth, duration of lactation, and BMI were not statistically different between groups (p > 0.05).

Table 2.

BC risk factors analysis of participants in the control and case groups

Elements of RiskMean±SD; median [range]p value**
control (n = 237)case (n = 141)
Menarche (age) 12.7±1.6; 13 [9–18] 12.6±1.5; 13 [9–7] 0.326 
Menopause (age) 43.6±7.6; 47 [27–56] 45.0±5.8; 45.5 [27–56] 0.338 
First childbirth (age) 22.6±5.7; 21 [14–42] 22.0±4.7; 20 [15–34] 0.433 
Lactation, months 20.2±24.3; 12.5 [1–252] 26.6±26.2; 18 [1–132] 0.069 
HOC, months 70.4±58.0; 60 [1–180] 127.7±98.1; 120 [4–360] 0.038 
Alcohol intake* 2.7±2.5; 2 [0.5–10] 5.3±4.2; 3.5 [1–13] <0.001 
BMI, kg/m2 27.7±4.9; 27.1 [18.8–46.4] 27.4±5.3; 26.9 [17.6–46.4] 0.557 
Elements of RiskMean±SD; median [range]p value**
control (n = 237)case (n = 141)
Menarche (age) 12.7±1.6; 13 [9–18] 12.6±1.5; 13 [9–7] 0.326 
Menopause (age) 43.6±7.6; 47 [27–56] 45.0±5.8; 45.5 [27–56] 0.338 
First childbirth (age) 22.6±5.7; 21 [14–42] 22.0±4.7; 20 [15–34] 0.433 
Lactation, months 20.2±24.3; 12.5 [1–252] 26.6±26.2; 18 [1–132] 0.069 
HOC, months 70.4±58.0; 60 [1–180] 127.7±98.1; 120 [4–360] 0.038 
Alcohol intake* 2.7±2.5; 2 [0.5–10] 5.3±4.2; 3.5 [1–13] <0.001 
BMI, kg/m2 27.7±4.9; 27.1 [18.8–46.4] 27.4±5.3; 26.9 [17.6–46.4] 0.557 

HOC, hormonal oral contraceptives; BMI, body mass index.

*Dose per week.

**Student’s t test for independent samples or Mann-Whitney nonparametric test, p < 0.05.

The serum vitamin D levels differed significantly between the case and control groups across all samples, with medians of 25.5 ng/mL and 31.0 ng/mL, respectively (p < 0.001) (Fig. 2a). This disparity persisted when examining vitamin D levels throughout the year and by season (p < 0.0001) (Fig. 2c), despite both groups have spended an equivalent amount of time of sun exposure during the day (Fig. 2b).

Fig. 2.

Evaluation of serum vitamin D in the case and control groups, and the interference of sun exposure in vitamin D levels along the year. a Comparison of vitamin D levels consolidated on each group. b Frequency of sun exposure (hours a day). c Vitamin D levels throughout the year. d Sunshine hours for Curitiba, Southern Brazil. Statistical analysis: (a) t test; (c) data were analyzed by fitting a mixed model for differences between months and groups.

Fig. 2.

Evaluation of serum vitamin D in the case and control groups, and the interference of sun exposure in vitamin D levels along the year. a Comparison of vitamin D levels consolidated on each group. b Frequency of sun exposure (hours a day). c Vitamin D levels throughout the year. d Sunshine hours for Curitiba, Southern Brazil. Statistical analysis: (a) t test; (c) data were analyzed by fitting a mixed model for differences between months and groups.

Close modal

Multivariate analysis was conducted to assess the impact of other significant variables on the association between vitamin D levels and BC. The results indicated that vitamin D was significantly associated with BC, regardless of ethnicity, education, hypertension, alcohol consumption, HDL, LDL, and menopause. For each unit improvement of vitamin D in the blood, the likelihood of belong into the BC group decreased by 7% (1 − 0.93 = 0.07). This inversely associating effect of vitamin D was independent of HDL-C, alcohol consumption, education, and race (Fig. 3).

Fig. 3.

Forest plot of significant variables for association with BC of the case and control groups. HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SAH, systemic arterial hypertension. Alcohol intake: yes × no; breastfeeding: yes × no; education: > high school complete × < high school; white × non-white race.

Fig. 3.

Forest plot of significant variables for association with BC of the case and control groups. HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SAH, systemic arterial hypertension. Alcohol intake: yes × no; breastfeeding: yes × no; education: > high school complete × < high school; white × non-white race.

Close modal

To ensure that level of vitamin D was minimally protective against BC, a receiver operating characteristic (ROC) curve was fitted to determine the discriminating cut-off point between cases and controls. The area under the curve was 0.65, indicating statistical significance (p < 0.001), with a sensitivity of 61.2% and a specificity of 64.0%. A cut-off point of 27.45 was established to discriminate between cases and controls with 65% accuracy; values bellow 27.45 were linked to BC, whereas values above indicated its absence or protection. Considering that this cut-off point is remarkably close to the one used clinically (28 ng/mL), the following analyses used values below and above 28 as dichotomous categories for vitamin D. The relationship between dichotomous vitamin D and diagnostic data in the case group was assessed. Patients with vitamin D ≥28 were predominantly clinically staged II, whereas those with vitamin D <28 were mostly clinically staged III (p = 0.01). Regarding molecular subtypes, lower levels <28 were enriched in the luminal B, luminal B HER2, and triple-negative groups, but significant only the last (p = 0.041). As observed, 18 out of 23 cases (78.2% of triple-negative subtypes were found with vitamin D <28 (p = 0.041) (Table 3). The most common histological subtype was the non-special type (invasive ductal carcinoma); however, the limited number of cases for other histological subtypes precluded statistical analysis.

Table 3.

Clinical and anatomopathological characteristics of women with BC and association with vitamin D levels

CharacteristicsClassificationVitamin Dp value*
<28 (n = 87)≥28 (n = 52)
n (%)n (%)
Histological type Invasive ductal carcinoma 79 (91) 42 (84)  
Invasive lobular 5 (6) 5 (10)  
NST and invasive lobular carcinoma 2 (2) 1 (2)  
Mucinous 1 (1) 0 (0)  
Occult 0 (0) 1 (2)  
Microinvasor 0 (0) 1 (2) 
Degree 10 (14) 5 (12)  
42 (61) 18 (44)  
17 (25) 18 (44) 0.108 
Stage IA e IB 10 (12) 10 (19)  
IIA e IIB 30 (34) 29 (56)  
IIIA, B, C 32 (37) 10 (19)  
IV 15 (17) 3 (6) 0.010 
Molecular subtype Luminal A 6 (7) 5 (10)  
Luminal B 46 (53) 30 (58)  
Luminal B HER2 10 (11) 1 (2)  
HER2 7 (8) 10 (19)  
Triple-negative 18 (21) 5 (10) 0.041 
CharacteristicsClassificationVitamin Dp value*
<28 (n = 87)≥28 (n = 52)
n (%)n (%)
Histological type Invasive ductal carcinoma 79 (91) 42 (84)  
Invasive lobular 5 (6) 5 (10)  
NST and invasive lobular carcinoma 2 (2) 1 (2)  
Mucinous 1 (1) 0 (0)  
Occult 0 (0) 1 (2)  
Microinvasor 0 (0) 1 (2) 
Degree 10 (14) 5 (12)  
42 (61) 18 (44)  
17 (25) 18 (44) 0.108 
Stage IA e IB 10 (12) 10 (19)  
IIA e IIB 30 (34) 29 (56)  
IIIA, B, C 32 (37) 10 (19)  
IV 15 (17) 3 (6) 0.010 
Molecular subtype Luminal A 6 (7) 5 (10)  
Luminal B 46 (53) 30 (58)  
Luminal B HER2 10 (11) 1 (2)  
HER2 7 (8) 10 (19)  
Triple-negative 18 (21) 5 (10) 0.041 

HER2, human epidermal growth factor receptor 2.

*Fisher’s exact test or χ2 test, p < 0.05.

A 3-day dietary recall was collected from cases and controls to assess the influence of diet on vitamin D levels, including the intake of nutrients that might interfere with vitamin D absorption. The nutritional data evaluated in the case group, including calorie intake (VET-total energy value and kcal/kg) and macro and micronutrient assessment, when compared according to the vitamin D cut-off point, revealed no significant differences between the analyzed groups (Table 4). These small variations suggest similar nutrient intake regardless of vitamin D levels, indicating other factors may influence vitamin D status.

Table 4.

Evaluation of nutritional data in relation to the cut-off point of vitamin D in the case group

VariableVitamin DnAverage±SDMedian (min-max)p value*
TEV <28 86 1,889.3±220.4 1,958.5 (1,274–2,163)  
≥28 52 1,857.6±228 1,939 (1,425–2,202) 0.425 
kcal/kg <28 86 29.5±7 29.7 (14.7–50)  
≥28 52 28.1±6.1 27.7 (14.7–39.5) 0.214 
Carbohydrate, g <28 86 293±28.5 302 (157–341)  
≥28 52 291.1±32.3 292 (198–341) 0.582 
Consumption protein/weight, g/kg <28 86 1.3±0.4 1.3 (0.5–2.2)  
≥28 52 1.3±0.4 1.2 (0.5–2.3) 0.198 
Protein, g <28 86 89.8±21 91.5 (36–140)  
≥28 52 85.9±21 85 (50–140) 0.139 
Lipid consumption/weight, g/kg <28 86 0.7±0.3 0.6 (0.3–1.6)  
≥28 52 0.6±0.2 0.6 (0.3–1.2) 0.267 
Lipids, g <28 86 46.1±15.2 43 (24–74)  
≥28 52 43.8±11.7 44.5 (25–74) 0.817 
Dietary fiber, g <28 86 18.5±3 19.3 (10.9–24.1)  
≥28 52 18±3 17.4 (13.7–24.1) 0.373 
Calcium, mg <28 86 1,038.6±143.1 1,048 (524–1,258)  
≥28 52 1,045.5±158.4 1,028 (789–1,265) 0.797 
Iron, mg <28 86 6.5±1.3 6.8 (3.9–8.5)  
≥28 52 6.4±1.3 6.4 (4.3–8.5) 0.605 
Sodium, mg <28 86 1,725±198.8 1,734 (1,355–2,132)  
≥28 52 1,695.8±216.1 1,625 (1,355–2,140) 0.430 
Zinc, mg <28 86 9.8±1.8 9.8 (6.6–13.4)  
≥28 52 9.4±1.7 9.3 (6.6–12.1) 0.210 
Potassium, mg <28 86 4.8±0.9 4.8 (2.9–6.1)  
≥28 52 4.7±0.9 4.8 (2.9–6.2) 0.685 
VariableVitamin DnAverage±SDMedian (min-max)p value*
TEV <28 86 1,889.3±220.4 1,958.5 (1,274–2,163)  
≥28 52 1,857.6±228 1,939 (1,425–2,202) 0.425 
kcal/kg <28 86 29.5±7 29.7 (14.7–50)  
≥28 52 28.1±6.1 27.7 (14.7–39.5) 0.214 
Carbohydrate, g <28 86 293±28.5 302 (157–341)  
≥28 52 291.1±32.3 292 (198–341) 0.582 
Consumption protein/weight, g/kg <28 86 1.3±0.4 1.3 (0.5–2.2)  
≥28 52 1.3±0.4 1.2 (0.5–2.3) 0.198 
Protein, g <28 86 89.8±21 91.5 (36–140)  
≥28 52 85.9±21 85 (50–140) 0.139 
Lipid consumption/weight, g/kg <28 86 0.7±0.3 0.6 (0.3–1.6)  
≥28 52 0.6±0.2 0.6 (0.3–1.2) 0.267 
Lipids, g <28 86 46.1±15.2 43 (24–74)  
≥28 52 43.8±11.7 44.5 (25–74) 0.817 
Dietary fiber, g <28 86 18.5±3 19.3 (10.9–24.1)  
≥28 52 18±3 17.4 (13.7–24.1) 0.373 
Calcium, mg <28 86 1,038.6±143.1 1,048 (524–1,258)  
≥28 52 1,045.5±158.4 1,028 (789–1,265) 0.797 
Iron, mg <28 86 6.5±1.3 6.8 (3.9–8.5)  
≥28 52 6.4±1.3 6.4 (4.3–8.5) 0.605 
Sodium, mg <28 86 1,725±198.8 1,734 (1,355–2,132)  
≥28 52 1,695.8±216.1 1,625 (1,355–2,140) 0.430 
Zinc, mg <28 86 9.8±1.8 9.8 (6.6–13.4)  
≥28 52 9.4±1.7 9.3 (6.6–12.1) 0.210 
Potassium, mg <28 86 4.8±0.9 4.8 (2.9–6.1)  
≥28 52 4.7±0.9 4.8 (2.9–6.2) 0.685 

TEV, total energy value.

*Student’s t test for independent samples or Mann-Whitney nonparametric test, p < 0.05.

Experimental studies in vitro and using in vivo models have demonstrated a relationship between vitamin D and BC development. However, epidemiological evidence remains limited [12]. Concurrently, although studies have identified and defined risk factors related to BC development, prevention strategies based on these findings have not sufficiently reduced disease incidence [1, 2]. In this study, women with vitamin D levels below 27.45 ng/mL were more frequently associated with the BC group than those with levels above this value. Here, we conducted a cross-sectional study with BC case and health control groups in Curitiba, known as the coldest capital in Brazil and among the lowest sunshine hours per month.

The findings of the present study closely align with those from a systematic review and meta-analysis that reported average vitamin D levels of 26.88 ng/mL (95% CI: 22.8–30.96 ng/mL) in patients with BC and 31.41 ng/mL (95% CI: 19.31–43.5 ng/mL) in the control group [13]. However, in a previous meta-analysis, the relationship between vitamin D and BC, when stratified by the menopausal state, demonstrated that the protective association of the vitamin was only significant in the premenopausal group [12]. In line with that, Dimitrakopoulou et al. [14] found scant evidence of a linear causal relationship between circulating vitamin D levels and the risk of various cancers, including BC, suggesting that vitamin D deficiency screening and supplementation should not be recommended as a primary cancer prevention strategy.

Vitamin D deficiency and insufficiency constitute a global health issue, affecting over one billion children and adults worldwide [15]. Typically, deficiency results from inadequate skin exposure to sunlight [16]. Skin synthesis depends on several factors, including the amount of exposed skin, ethnicity, age, and duration of sun exposure, which are influenced by latitude, seasonality, and the use of sun protection, sunscreen and clothing. Ultraviolet B radiation is more intense at higher altitudes and in sunnier regions. However, our study site, Curitiba in southern Brazil, experiences few sunny days throughout the year, impairing the production of vitamin D [17]. The monthly levels of vitamin D changed in the opposite direction of sunshine along the year, increasing to highest levels in November to February (summer: December–February) and decreasing from June to August (winter: June–August), in agreement with other studies [18‒20]. Several studies investigating the impact of vitamin D supplementation on disease risk have reported a marginal protective effect.

The clinical aspects and other risk factors for BC assessed in the present study did not significantly affect the association between vitamin D levels and BC, suggesting that vitamin D levels may be an independent factor. However, clinical factors such as low serum HDL-C levels were more prevalent with the BC group, and this seems not to be related with diet or BMI. Physical activity and exercise was not assessed in our study. Higher cholesterol levels are considered essential for cancer cell proliferation and tumor progression [21]. Consequently, a role for HDL-C in the disease’s pathogenesis has been proposed, although current knowledge still remains inconclusive about it [22].

Lifestyle and habits have not been shown to influence the relationship between vitamin D and BC. However, some habits, such as alcohol consumption, are independently associated with cancer. The World Cancer Research Fund (WCRF) has identified alcohol consumption as a causal factor in postmenopausal BC and as having probable evidence in premenopausal cases, with an additional 9% risk for every 10 g/day increase in alcohol intake. Among BC patients, those with more advanced tumor stages (III and IV) showed a significant association with vitamin D levels <28 ng/mL. Yao et al. [23] described a similar relationship in a prospective cohort study that evaluated serum vitamin D levels during diagnosis. They found a significant association between the mean serum level of 25-hydroxyvitamin D and tumor stage, ranging from 21.5 ng/mL in stage I to 18.5 ng/mL in stage IV.

The relatively young average age observed in our study population, 41.7 years for the control group and 43.1 years for the case group, can be explained partly by the study design. This case-control study used female blood donors as the control group, which may introduce a selection bias for the higher donation rates among younger women. Despite this, the relatively young average age observed in our study aligns with trends noted in several populations. Evidence shows a rising incidence of BC in younger women, particularly in the 25–39 age group, with a forecasted 11–12% increase by 2030 [24]. The incidence has been increasing since 2000 and is reflected globally, including in Brazil [25, 26]. This trend is linked to more aggressive pathology and advanced disease at diagnosis [27] with disparities in outcomes for younger patients further highlighting the need for revisiting screening guidelines and preventive measures for younger women.

Our study had limitations. First, despite adjusting for potential confounding factors, the cross-sectional nature of this study precluded the establishment of a causal relationship between serum vitamin D levels and BC. Second, selection bias was present, as the control group was derived from a blood donation service with an age limit of 69 years, thus restricting the age range of the case group. Lastly, we had no power for addressing race and prototype as co-variables since Curitiba’s population is constituted by European descendants. A notable strength of this study was the inclusion of participants throughout the year considering sunshine hours, with cases and controls being matched within the same month, allowing for the observation of seasonal variations in vitamin D levels.

The findings presented here underscore the potential role of vitamin D in reducing BC risk. Moreover, this study has shown that establishing a threshold for serum vitamin D levels could be advantageous in identifying populations at risk for BC, informing prevention strategies beyond mere supplementation.

The study was approved by the Ethics Research Committee of Erasto Gaertner Hospital, Approval No. CAEE: 51367415.2.0000.0098) in December 15, 2015. All participants received information about the study’s objectives, risks, and potential benefits and provided their consent by signing the informed consent form in duplicate.

The authors have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

Conceptualization: LCCR; methodology: LCCR, C.B.P., and O.C.B.; data analysis: M.O. and T.S.S.P.; writing – original draft preparation: T.S.S.P. and S.S.A.M.; writing, review and editing, supervision, and project administration: S.E.-E. and LCCR.

The data in this study were obtained from Erasto Gaertner Hospital database where restrictions may apply as information could compromise the privacy of research participants. Datasets may be requested from the corresponding author (S.E.-E.).

1.
Sung
H
,
Ferlay
J
,
Siegel
RL
,
Laversanne
M
,
Soerjomataram
I
,
Jemal
A
, et al
.
Global cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries
.
CA Cancer J Clin
.
2021
;
71
(
3
):
209
49
.
2.
Siegel
RL
,
Miller
KD
,
Jemal
A
.
Cancer Statistics, 2017
.
CA Cancer J Clin
.
2017
;
67
(
1
):
7
30
.
3.
Youn
HJ
,
Han
W
.
A review of the epidemiology of breast cancer in asia: focus on risk factors
.
Asian Pac J Cancer Prev
.
2020
;
21
(
4
):
867
80
.
4.
De Cicco
P
,
Catani
MV
,
Gasperi
V
,
Sibilano
M
,
Quaglietta
M
,
Savini
I
.
Nutrition and breast cancer: a literature review on prevention, treatment and recurrence
.
Nutrients
.
2019
;
11
(
7
):
1514
.
5.
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
(
13
):
4731
.
6.
Vanhevel
J
,
Verlinden
L
,
Doms
S
,
Wildiers
H
,
Verstuyf
A
.
The role of vitamin D in breast cancer risk and progression
.
Endocr Relat Cancer
.
2022
;
29
(
2
):
R33
55
.
7.
Scragg
R
,
Khaw
KT
,
Toop
L
,
Sluyter
J
,
Lawes
CMM
,
Waayer
D
, et al
.
Monthly high-dose vitamin D supplementation and cancer risk: a post hoc analysis of the vitamin D assessment randomized clinical trial
.
JAMA Oncol
.
2018
;
4
(
11
):
e182178
.
8.
Manson
JE
,
Cook
NR
,
Lee
IM
,
Christen
W
,
Bassuk
SS
,
Mora
S
, et al
.
Vitamin D supplements and prevention of cancer and cardiovascular disease
.
N Engl J Med
.
2019
;
380
(
1
):
33
44
.
9.
Martin
SS
,
Blaha
MJ
,
Elshazly
MB
,
Toth
PP
,
Kwiterovich
PO
,
Blumenthal
RS
, et al
.
Comparison of a novel method vs the Friedewald equation for estimating low-density lipoprotein cholesterol levels from the standard lipid profile
.
JAMA
.
2013
;
310
(
19
):
2061
8
.
10.
IBGE
Pesquisa de orçamentos familiares 2008-2009 : tabelas de composição nutricional dos alimentos consumidos no Brasil
;
2011
.
11.
Meysami
M
,
Kumar
V
,
Pugh
M
,
Lowery
ST
,
Sur
S
,
Mondal
S
, et al
.
Utilizing logistic regression to compare risk factors in disease modeling with imbalanced data: a case study in vitamin D and cancer incidence
.
Front Oncol
.
2023
;
13
:
1227842
.
12.
Estébanez
N
,
Gómez-Acebo
I
,
Palazuelos
C
,
Llorca
J
,
Dierssen-Sotos
T
.
Vitamin D exposure and risk of breast cancer: a meta-analysis
.
Sci Rep
.
2018
;
8
(
1
):
9039
.
13.
Voutsadakis
IA
.
Vitamin D baseline levels at diagnosis of breast cancer: a systematic review and meta-analysis
.
Hematol Oncol Stem Cell Ther
.
2021
;
14
(
1
):
16
26
.
14.
Dimitrakopoulou
VI
,
Tsilidis
KK
,
Haycock
PC
,
Dimou
NL
,
Al-Dabhani
K
,
Martin
RM
, et al
.
Circulating vitamin D concentration and risk of seven cancers: mendelian randomisation study
.
BMJ
.
2017
;
359
:
j4761
.
15.
Holick
MF
.
The vitamin D deficiency pandemic: approaches for diagnosis, treatment and prevention
.
Rev Endocr Metab Disord
.
2017
;
18
(
2
):
153
65
.
16.
Reid
IR
,
Bolland
MJ
.
Controversies in medicine: the role of calcium and vitamin D supplements in adults
.
Med J Aust
.
2019
;
211
(
10
):
468
73
.
17.
Chang
SW
,
Lee
HC
.
Vitamin D and health: the missing vitamin in humans
.
Pediatr Neonatol
.
2019
;
60
(
3
):
237
44
.
18.
Klingberg
E
,
Oleröd
G
,
Konar
J
,
Petzold
M
,
Hammarsten
O
.
Seasonal variations in serum 25-hydroxy vitamin D levels in a Swedish cohort
.
Endocrine
.
2015
;
49
(
3
):
800
8
.
19.
Fontanive
TO
,
Dick
NRM
,
Valente
MCS
,
Laranjeira
VDS
,
Antunes
MV
,
Corrêa
MP
, et al
.
Seasonal variation of vitamin D among healthy adult men in a subtropical region
.
Rev Assoc Med Bras
.
2020
;
66
(
10
):
1431
6
.
20.
Santana
KVS
,
Oliver
SL
,
Mendes
MM
,
Lanham-New
S
,
Charlton
KE
,
Ribeiro
H
.
Association between vitamin D status and lifestyle factors in Brazilian women: implications of sun exposure levels, diet, and health
.
EClinicalMedicine
.
2022
;
47
:
101400
.
21.
Kuzu
OF
,
Noory
MA
,
Robertson
GP
.
The role of cholesterol in cancer
.
Cancer Res
.
2016
;
76
(
8
):
2063
70
.
22.
Mazzuferi
G
,
Bacchetti
T
,
Islam
MO
,
Ferretti
G
.
High density lipoproteins and oxidative stress in breast cancer
.
Lipids Health Dis
.
2021
;
20
(
1
):
143
.
23.
Yao
S
,
Kwan
ML
,
Ergas
IJ
,
Roh
JM
,
Cheng
TYD
,
Hong
CC
, et al
.
Association of serum level of vitamin D at diagnosis with breast cancer survival: a case-cohort analysis in the pathways study
.
JAMA Oncol
.
2017
;
3
(
3
):
351
7
.
24.
Kehm
RD
,
Yang
W
,
Tehranifar
P
,
Terry
MB
.
40 Years of change in age- and stage-specific cancer incidence rates in US women and men
.
JNCI Cancer Spectr
.
2019
;
3
(
3
):
pkz038
.
25.
Huang
J
,
Chan
PS
,
Lok
V
,
Chen
X
,
Ding
H
,
Jin
Y
, et al
.
Global incidence and mortality of breast cancer: a trend analysis
.
Aging
.
2021
;
13
(
4
):
5748
803
.
26.
Bonadio
RC
,
Moreira
OA
,
Testa
L
.
Breast cancer trends in women younger than 40 years in Brazil
.
Cancer Epidemiol
.
2022
;
78
:
102139
.
27.
Fabiano
V
,
Mando
P
,
Rizzo
M
,
Ponce
C
,
Colo
F
,
Loza
M
, et al
.
Breast cancer in young women presents with more aggressive pathologic characteristics: retrospective analysis from an Argentine national database
.
JCO Glob Oncol
.
2020
;
6
:
639
46
.