Relationship of women’s breast nodules with lifestyle, mental health and metabolic characteristics: a large, retrospective, longitudinal cohort study

Abstract

Background

Previous studies specific to breast nodules are relatively scarce and mainly rely on cross-sectional research methods, overlooking the dynamic evolution of breast nodules over time. Thus, the objective is to analyze the impact of lifestyle, mental health, and metabolic characteristics on the occurrence and progression of breast nodules by conducting a longitudinal study.

Methods

This retrospective, longitudinal cohort study conducted from 2012 to 2021 included 114,774 measurements from the Health Management Center at the Third Xiangya Hospital of Central South University in China. All data were collected from the results of anthropometric measurements, laboratory tests, breast ultrasound and online questionnaire surveys. Generalized estimating equations (GEE) models were developed to examine the relationship.

Results

Women who always ate punctual meals or consumed soy products ≥ 5 times per week had lower odds of breast nodule occurrence (adjusted odds ratio [aOR] 0.798, 95% CI 0.761–0.836, p < 0.001; aOR 0.791, 95% CI 0.689–0.908, p = 0.001). In contrast, women who had socializing meals of ≥ 3 times per week had higher odds of breast nodule occurrence (aOR 1.315, 95% CI 1.124–1.538, p = 0.001). Higher serum creatinine and lower serum uric acid were also associated with increased risk of breast nodule occurrence (aOR 1.001, 95% CI 1.000–1.001, p = 0.012; aOR 0.999, 95% CI 0.999–1.000, p < 0.001). Women with fair sleep quality had higher breast nodule occurrence (aOR 1.074, 95% CI 1.027–1.122, p = 0.002) but a lower risk of progression to BI-RADS ≥ 4 A (aOR 0.760, 95% CI 0.607–0.952, p = 0.017).

Conclusions

Our study revealed significant correlations between breast nodules and various lifestyle factors and metabolic characteristics, but the correlations between breast nodules and mental health were not found. These findings may provide insights into potential factors related to breast nodules and guide future research on their prevention and management.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-26180-9.

Introduction

Breast cancer is the most common malignant tumor with the highest incidence rate and seriously affects women’s health worldwide [1]. Breast nodules have been found to be associated with an increased risk of breast cancer [2]. The discovery of breast nodules is often the first sign in the early stages of breast cancer that attracts clinical attention [3]. Therefore, identifying and controlling risk factors for breast nodules is of great significance in the early prevention of breast cancer.

While several studies have explored risk factors for benign breast disease (BBD), research specific to breast nodules is relatively scarce. A cross-sectional study showed a correlation among mental disorders, metabolic factors and breast nodules [4]. Another cross-sectional study with 12,538 female subjects explored the effect of metabolic parameters on the occurrence of breast masses [5]. A prospective study involving 100 patients investigated the relationship among anxiety, depression and BBD [6]. Another prospective study including 9,031 females explored the association between dietary factors and BBD risk [7]. The effect of one factor on breast disease may differ due to confounding factors; what may be positive in one context may be negative in another. Results examining the effect of one factor on breast disease may be biased due to the influence of confounding factors. Therefore, this study comprehensively considered the influence of mental disorders, metabolic factors, and dietary factors when exploring the relevant factors contributing to breast nodules. In addition, the study also explored the effect of other lifestyle factors (e.g. physical intensity of the job and sedentary time) on breast nodules. Understanding how modifiable lifestyle factors contribute to the risk of breast nodules may provide a simple, inexpensive opportunity to reduce the risk of breast nodules and even breast cancer.

In addition, previous research focused on identifying risk factors for breast nodule occurrence and overlooked the risk factors involved in the progression of benign breast lesions to malignant lesions [4]. Therefore, this study not only explored risk factors for the occurrence of breast nodules (from nodule absence to presence) but also delved deeper into the risk factors involved in the progression from benign lesions (Breast Imaging Reporting and Data System (BI-RADS) score of 2 or 3) to malignant lesions (BI-RADS score ≥ 4 A) [8].

Previous studies of breast nodules have primarily utilized cross-sectional research methods, overlooking the dynamic progression of breast nodules over time [4, 5]. To address this limitation, our study employed a longitudinal research design, which, in contrast to cross-sectional approaches, is better positioned to discern and elucidate the pivotal influencing factors throughout this dynamic process [9]. Additionally, our study benefitted from a substantial sample size, allowing for the establishment of a sturdy foundation for a precise and thorough analysis of the key factors influencing breast nodule occurrence and progression.

Therefore, the objective was to comprehensively and accurately analyze the impact of lifestyle, mental health, and metabolic characteristics on the occurrence and progression of breast nodules by conducting a large-sample study with a longitudinal research design.

Materials and methods

Study cohort and sample

This was a retrospective, longitudinal cohort study. We collected data from the Health Management Center at the Third Xiangya Hospital of Central South University in China. From 2012 to 2021, 46,170 women were recruited. By the end of follow-up, 6,396 (13.8%) were lost to follow-up, defined as no further participation after providing consent and completing the first examination and questionnaire. The reasons and corresponding numbers are presented in eFigure 1 of the Supplementary Materials. Ultimately, 39,774 women were included in the present analysis. An annual follow-up design was implemented, with each participant undergoing a medical examination and completing an online questionnaire once per year. For the final analysis, we included participants who completed at least two follow-up assessments (i.e., baseline plus at least one subsequent measurement). The study spanned from 2012 to 2021, during which most participants contributed between 2 and 8 repeated measurements. In total, the dataset comprised 114,774 measurements. Further details are provided in eFigure 1 of the Supplementary Materials.

The inclusion criteria were as follows: (1) women aged 14 years and older; (2) women who agreed to participate in this study; (3) women who underwent a comprehensive medical examination, including anthropometric measurements, breast ultrasound and laboratory tests; and (4) women who completed the online survey. The exclusion criteria were as follows: (1) severe psychiatric disorders or significant cognitive impairment; (2) current lactation; (3) biopsy-confirmed breast cancer (breast ultrasound results classified as BI-RADS category 6); and (4) breast ultrasound results classified as BI-RADS category 0, indicating the need for further diagnostic evaluation.

This research was approved by the Survey and Behavioral Research Ethics Committee of the Third Xiangya Hospital of Central South University (Ref. No. 23455). All participants who underwent a physical examination at the institution were invited to complete the online survey within one week. Participants were informed in the message that the survey was entirely voluntary, and there was no reward for participation. Informed consent was obtained from each participant.

Data collection

All data were obtained from anthropometric measurements, laboratory tests, breast ultrasonography and online questionnaire surveys.

Outcome variable

The primary outcome variable of this study was the status of breast nodules, assessed by ultrasonography performed by trained professionals and defined across two dimensions: occurrence and progression.

BI-RADS is an internationally standardized tool used to categorize breast imaging findings, including ultrasonography and mammography, and to provide clear management recommendations [8]. BI-RADS classifies lesions into seven categories (0–6), where 0 indicates incomplete assessment requiring additional imaging, 1 indicates no abnormal findings, 2 indicates benign findings, 3 is probably benign with a malignancy risk < 2%, 4 is suspicious with increasing malignancy risk (subdivided into 4 A: low, 2–10%; 4B: intermediate, 10–50%; 4 C: high, 50–95%), 5 is highly suggestive of malignancy (> 95%), and 6 is biopsy-proven malignancy [8].

Nodule occurrence was defined as the detection of a lesion with a BI-RADS score ≥ 2, while BI-RADS category 1 was considered normal (no nodule detected).

With reference to relevant clinical guidelines and literature, nodule progression was defined as the transition from a benign state (BI-RADS categories 2 or 3, recommended for short-term follow-up) to a higher-risk state suggestive of malignancy (BI-RADS category ≥ 4 A, biopsy recommended) [8, 10]. This definition captures not only a significant increase in malignancy risk but also a fundamental change in management strategy, shifting from routine follow-up to a high-risk category requiring pathological confirmation. The status of breast nodules was extracted from breast ultrasound reports. In the breast ultrasound report, terms related to solid or mixed masses, breast cysts, breast space-occupying lesions and intraductal masses were all categorized as breast nodules.

Explanatory variables

The explanatory variables of interest included lifestyle habits, mental health status, and metabolic characteristics, which were detailed as follows:

The lifestyle habit information extracted from the results of the online questionnaire survey included dietary habits (variables 5–21 in eTable 1 and eTable 2 of the Supplementary Materials), exercise habits (variables 23–27), sleep habits (variables 32–33), drinking habits (variable 4), and smoking habits (variable 2). Throughout the follow-up period (2012–2021), the content and structure of these core items remained consistent, with no modifications made. Participants reported whether they ate meals punctually (Yes/Mostly yes/No), as well as the weekly frequency (≥ 5, 3–5, 1–2, or < 1 time per week) of consuming soy products and of socializing meals. Socializing meals was defined as eating occasions aimed at establishing or maintaining instrumental social relationships (e.g., work-related interactions or exchanges of benefits), often involving psychological stress and ritualized behaviors to accomplish the underlying social purpose. Sleep quality was assessed using a single-item self-reported measure: “Over the past month, how would you rate your sleep quality?” (Good, Fair, Poor). Duration of sitting outside of work was assessed using a single-item question: “Excluding working and study hours, how many hours per day do you usually spend sitting (e.g., watching television, using the internet, playing mahjong, or playing cards)?” (< 2, 2–4, 4–6, > 6 h). Definitions of additional lifestyle habit variables are detailed in eTable 1 and eTable 2 of the Supplementary Materials and the Supplementary Questionnaire.

The mental health status information of participants extracted from the results of the online questionnaire survey, including depression and anxiety, was measured by the Patient Health Questionnaire-2 (PHQ-2) and Generalized Anxiety Disorder-2 (GAD-2).The presence of depression and anxiety was defined as a score of 2 or higher on the PHQ-2 and GAD-2 [11–13].

Metabolic characteristics were assessed based on the results of laboratory tests. Venous blood samples were collected from the median cubital vein after an overnight fast. The following biochemical indicators were measured: glutamic-pyruvic transaminase (GPT), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), serum creatinine (SCr), serum glucose (SGlu), serum uric acid (SUA), serum urea (SUrea), total bilirubin (TBIL), total cholesterol (TC) and triglyceride (TG). All parameters were determined using standard laboratory procedures with an automated chemistry analyzer (Hitachi 7600 and Hitachi 7170; Hitachi, Japan).

Body mass index (BMI) was calculated as weight (kg) divided by height squared (m²), based on anthropometric measurements. Height and weight were measured by trained staff using a calibrated electronic height–weight scale (SK-L09B, China), which simultaneously recorded both parameters. During the assessment, participants wore light indoor clothing and were barefoot.

Confounding variables

Potential confounders considered in the analysis included sociodemographic characteristics, such as age, marital status, menopausal status, menarche age, fertility status, physical exercise, and measurement year, which were obtained from the online questionnaire survey.

Sample size considerations

As this was a retrospective cohort study, the sample size was not determined a priori based on statistical power calculations. Instead, it was defined by all eligible female individuals with complete records in the Health Management Center database of the Third Xiangya Hospital between 2012 and 2021. A total of 39,774 women were included, contributing 114,774 valid observations.

For the multivariable analysis of breast nodule occurrence, 16 variables were included in the model, with 16,081 positive events. For the multivariable analysis of nodule progression, 8 variables were included, with 478 corresponding progression events. Although the study was based on a convenience sample determined by data availability, the sample size was sufficient according to the commonly accepted rule of thumb that the number of outcome events should be at least 10 times the number of independent variables included in the model [14]. Therefore, the sample size in this study was adequate to support multivariable analyses, ensuring both the stability of model estimation and the reliability of the results.

Statistical analysis

Missing data in this study were mainly due to variations in health check-up items across years, with no systematic association to unobserved outcomes or exposure variables. Therefore, multiple imputation under the missing at random (MAR) assumption was applied. Five imputations (m = 5) were performed using the R package mice, with the following models: Predictive Mean Matching (PMM) for continuous variables, Logistic Regression (LogReg) for binary variables, Polytomous Logistic Regression (PolyReg) for unordered categorical variables, and Proportional Odds Model (POLR) for ordered categorical variables. Missing data proportions and counts were provided in Supplementary Materials eTable 3. Convergence diagnostics showed low relative variability across imputations (see eTable 4), indicating stable and reliable result.

To evaluate potential bias due to loss to follow-up, we compared all baseline variables using the R package CBCgrps —including lifestyle habits, mental health status, metabolic characteristics, and sociodemographic characteristics—between women lost to follow-up and those included in the analysis. Statistical analyses indicated no significant differences between the two groups for any of the variables (all p > 0.05), suggesting a low risk of bias from loss to follow-up (in eTable 5 of the Supplementary Materials).

The descriptive information between the “without breast nodules” group and the “breast nodules” group, as well as between the “BI-RADS score = 2 or 3” group and the “BI-RADS score ≥ 4A” group, was analyzed using the R package CBCgrps. Detailed information is presented in eTable 1 and eTable 2 of the Supplementary Materials. Generalized estimating equation (GEE) models were constructed (rationale for choosing GEE was provided in Supplementary Materials), specifying a first-order autoregressive working correlation structure to account for the approximately annual intervals of repeated measurements. Specifically, “with or without breast nodules” was defined as the dependent variable representing the occurrence of breast nodules, while “BI-RADS score = 2 or 3” versus “BI-RADS score ≥ 4A” was defined as the dependent variable representing the progression of breast nodules. These models were used to analyze the impact of lifestyle, mental health, and metabolic characteristics on both the occurrence and progression of breast nodules via SPSS 18.0 for Windows. Initially, univariate GEE models were employed to screen for variables that may significantly influence the occurrence and progression of breast nodules. Subsequently, based on the literature review and ensuring no clinically important variables were omitted, variables with P < 0.1 in the univariate GEE analysis were incorporated into a multivariate GEE model to explore factors influencing the occurrence and progression of breast nodules. Based on different combinations of independent variables, we constructed multiple GEE models and selected the model with the lowest quasi-likelihood information criterion (QIC) value as the optimal fitting model. Finally, based on the results of the multivariate analysis, estimated marginal means (EM means) plots were generated for the variables of age and years.

Results

Characteristics of the study cohort

This study, conducted from 2012 to 2021, included 39,774 women who contributed a total of 114,774 longitudinal measurements. Among these, breast nodules were detected in 16,081 women (14%). Of these, 15,603 (97%) women had a BI-RADS score = 2 or 3, and 478 (3%) had a BI-RADS score ≥ 4 A (eTable 1 and eTable 2 of the Supplementary Materials). The 41–50 years age group had the highest risk of breast nodule occurrence (Fig. 1), with an aOR of 1.993 (95% CI 1.852–2.146, p < 0.001) compared to the reference group (aged ≤ 30 years) (Table 1). Additionally, beyond the age of 30 years, there was a discernible increasing trend in the risk of breast nodule progression to a BI-RADS score ≥ 4 A with increasing age (Fig. 2). The risk of breast nodule occurrence increased rapidly from 2017 onward (Fig. 1), while the risk of breast nodule progression to a BI-RADS score ≥ 4 A did not exhibit a pronounced increasing trend during the same period (Fig. 2).

Fig. 1.

Estimated marginal means plot for the occurrence of breast nodules

Table 1.

Multifactor GEE analysis for the occurrence of breast nodules

Variables	aOR (95%CI)	P-value
Age (reference: ≤30 years)		< 0.001
≥ 61years	1.185(1.029–1.365)	0.018
51–60 years	1.814(1.633–2.015)	< 0.001
41–50 years	1.993(1.852–2.146)	< 0.001
31–40 years	1.264(1.185–1.349)	< 0.001
BMI (reference: ≤18.4 kg/m2)		< 0.001
≥ 28 kg/m2	0.488(0.425–0.561)	< 0.001
24–27.9 kg/m2	0.680(0.621–0.745)	< 0.001
18.5–23.9 kg/m2	0.861(0.796–0.931)	< 0.001
Eats meals punctually (reference: no)		< 0.001
Always yes	0.798(0.761–0.836)	< 0.001
Mostly yes	0.861(0.817–0.907)	< 0.001
Socializing meals (reference: <1 time peer week)		0.001
≥ 3 times per week	1.315(1.124–1.538)	0.001
1–2 times per week	1.057(0.991–1.127)	0.093
Frequency of consuming soy products (reference: <1 time peer week)		< 0.001
≥ 5 times per week	0.791(0.689–0.908)	0.001
3–5 times per week	0.907(0.831–0.990)	0.029
1–2 times per week	0.966(0.889–1.050)	0.416
Participating in physical exercise (reference: no)		0.069
Yes	1.036(0.997–1.076)	0.069
Physical intensity level of the job (reference: mainly nonmanual work)		< 0.001
Not employed	0.925(0.854–1.002)	0.056
Mainly manual work	0.873(0.833–0.915)	< 0.001
Duration of sitting outside of work (reference: less than 2 h)		0.037
More than 6 h	1.080(1.018–1.145)	0.010
4–6 h	1.025(0.974–1.079)	0.348
2–4 h	1.049(1.006–1.093)	0.027
Fertility status (reference: no)		0.346
Yes	0.970(0.911–1.033)	0.346
Marital status (reference: single)		0.013
Divorced/Widowed	0.854(0.744–0.982)	0.026
Married (including cohabitation)	0.906(0.845–0.972)	0.006
Menopausal status (reference: no)		< 0.001
Yes	0.713(0.660–0.770)	< 0.001
Menarche age (reference: less than 12 years old)		0.465
Forgetful	1.050(0.962–1.146)	0.276
12 years old or older	0.999(0.952–1.048)	0.976
Sleep quality (reference: poor)		0.004
Good	1.071(1.022–1.123)	0.004
Fair	1.074(1.027–1.122)	0.002
Measurement years (reference: year 2021)		< 0.001
Year 2012	0.056(0.044–0.071)	< 0.001
Year 2013	0.061(0.052–0.073)	< 0.001
Year 2014	0.085(0.074–0.096)	< 0.001
Year 2015	0.086(0.077–0.096)	< 0.001
Year 2016	0.104(0.096–0.114)	< 0.001
Year 2017	0.116(0.107–0.126)	< 0.001
Year 2018	0.442(0.419–0.467)	< 0.001
Year 2019	0.688(0.657–0.720)	< 0.001
Year 2020	0.897(0.860–0.936)	< 0.001
Serum uric acid (SUA)	0.999(0.999-1.000)	< 0.001
Serum creatinine (SCr)	1.001(1.000-1.001)	0.012

Quasi-likelihood under the independence model criterion (QIC): 81008.615 aOR Adjusted odds ratio

Fig. 2.

Estimated marginal means plot for the progression of breast nodules

Association between breast nodules and dietary factors

For the final models of breast nodule occurrence and progression, multicollinearity diagnostics revealed that in the occurrence model, variance inflation factors (VIFs) ranged from 1.001 to 3.008 (tolerances 0.324–0.999). In the progression model, VIFs ranged from 1.001 to 2.906 (tolerances 0.344–0.999), indicating no significant multicollinearity in either model. Detailed diagnostic statistics are provided in Supplementary Materials eTables 6 and 7.

In the final multivariate model, after adjusting for age, marital status, menopausal status, menarche age, fertility status, physical exercise, and measurement year, the risk of breast nodule occurrence decreased as the frequency of punctual meals increased, socializing meals decreased, and soy product consumption increased. Compared to individuals who did not eat punctual meals, those who mostly ate punctual meals had an aOR of 0.861 (95% CI 0.817–0.907, p < 0.001), and those who always ate punctual meals had an aOR of 0.798 (95% CI 0.761–0.836, p < 0.001) (Table 1). Compared to the aOR for a frequency of socializing meals of less than once per week, that for a frequency of socializing meals of 1–2 times per week was 1.057 (95% CI 0.991–1.127, p = 0.093), and that for a frequency of socializing meals of ≥ 3 times per week was 1.315 (95% CI 1.124–1.538, p = 0.001) (Table 1). Compared to that for individuals consuming soy products less than once per week, the aOR of breast nodule occurrence for individuals consuming soy products 1–2 times per week was 0.966 (95% CI 0.889–1.050, p = 0.416); that for individuals consuming soy products 3–5 times per week was 0.907 (95% CI 0.831–0.990, p = 0.029); and that for individuals consuming soy products ≥ 5 times per week was 0.791 (95% CI 0.689–0.908, p = 0.001) (Table 1).

Association between breast nodules and physical activity factors

Compared to those with nonmanual work, individuals engaged in manual work had the lowest risk of breast nodule occurrence, with an aOR of 0.873 (95% CI 0.833–0.915, p < 0.001) (Table 1). Compared to that for a duration of sitting outside of work less than 2 h, the aOR for sitting 2–4 h was 1.049 (95% CI 1.006–1.093, p = 0.027), that for sitting 4–6 h was 1.025 (95% CI 0.974–1.079, p = 0.348), and that for sitting more than 6 h was 1.080 (95% CI 1.018–1.145, p = 0.010) (Table 1).

Association between breast nodules and sleep quality factors

Women with both good and fair sleep qualities had a higher risk of breast nodule occurrence, with an aOR of 1.071 (95% CI 1.022–1.123, p = 0.004) for those with good sleep quality and an aOR of 1.074 (95% CI 1.027–1.122, p = 0.002) for those with fair sleep quality (Table 1). However, the risk of breast nodule progression to a BI-RADS score ≥ 4 A was relatively lower in those women, with an aOR of 0.748 (95% CI 0.581–0.963, p = 0.024) for those with good sleep quality and an aOR of 0.760 (95% CI 0.607–0.952, p = 0.017) for those with fair sleep quality (Table 2).

Table 2.

Multifactor GEE analysis for the progression of breast nodules

Variables	OR (95%CI)	P-value
Age (reference: ≥61 years)		0.001
≤ 30years	0.405(0.220–0.745)	0.004
31-40years	0.321(0.187–0.552)	< 0.001
41-50years	0.478(0.289–0.792)	0.004
51-60years	0.563(0.370–0.856)	0.007
Participating in physical exercise (reference: no)		0.313
Yes	1.106(0.909–1.345)	0.313
Fertility status (reference: no)		0.109
Yes	1.349(0.936–1.945)	0.109
Marital status (reference: single)		0.253
Divorced/Widowed	0.496(0.216–1.143)	0.100
Married (including cohabitation)	0.860(0.602–1.228)	0.406
Menopausal status (reference: no)		0.996
Yes	0.999(0.684–1.459)	0.996
Menarche age (reference: less than 12 years old)		0.411
Forgetful	0.736(0.462–1.172)	0.196
12 years old or older	0.900(0.707–1.146)	0.392
Sleep quality (reference: poor)		0.035
Good	0.748(0.581–0.963)	0.024
Fair	0.760(0.607–0.952)	0.017
Measurement years (reference: year 2021)		< 0.001
Year 2012	0.001(0.000-0.001)	< 0.001
Year 2013	0.201(0.028–1.449)	0.111
Year 2014	0.335(0.107–1.052)	0.061
Year 2015	0.809(0.420–1.557)	0.526
Year 2016	0.767(0.463–1.271)	0.304
Year 2017	1.062(0.691–1.632)	0.783
Year 2018	0.659(0.476–0.911)	0.012
Year 2019	1.087(0.846–1.396)	0.514
Year 2020	1.101(0.866-1.400)	0.432

Quasi-likelihood under the independence model criterion (QIC): 4262.920 aOR Adjusted odds ratio

Association between breast nodules and metabolic factors

Higher levels of SCr and lower levels of SUA were associated with an increased risk of breast nodule occurrence, with aORs of 1.001 (95% CI 1.000-1.001, p = 0.012) and 0.999 (95% CI 0.999-1.000, p < 0.001), respectively (Table 1).

Association between breast nodules and mental health

It is noteworthy that we did not observe any correlation between mental health and the occurrence and progression of breast nodules.

Discussion

In this cohort study, we aimed to analyze the impact of lifestyle, mental health, and metabolic characteristics on the occurrence and progression of breast nodules. We examined the relationship by GEE models and identified (1) the association between the risk of breast nodule occurrence and many dietary factors, including the frequency of punctual meals, socializing meals, and soy product consumption, (2) the protective role of physical activity on the occurrence of breast nodules, (3) the risk association between metabolic factors, particularly SCr and SUA, and the occurrence of breast nodules, (4) the association of sleep quality, age and measurement year with both breast nodule occurrence and progression.

These dietary discoveries reinforced the understanding of the potential impact of diet on breast health, as previously explored in research. Ji et al. discovered that meal irregularity was linked to a heightened risk of breast cancer in females [15]. Previous findings have shown that high-level soy consumption exhibits a protective effect against breast cancer [16, 17]. The protective potential attributed to soy isoflavones arises from their structural similarity to 17-β-estradiol. They may modulate the hormonal environment by acting as weak estrogens or blocking endogenous estrogens [16, 18]. This study indicated an association between socializing meals and the occurrence of breast nodules, which may be related to increased intake of high-calorie, high-fat foods and the accompanying psychological stress during such gatherings. A study has indicated that a high-fat, high-calorie diet may establish a microenvironment conducive to the progression of breast tumors [19]. At present, socializing meals have received limited attention in breast health research. However, it is important to note that our study provides preliminary evidence of an association between socializing meals and the occurrence of breast nodules. The dietary variables used in this study were based on non-validated, single-item questions, which do not capture the full complexity of dietary patterns or nutritional quality. As a result, these findings should be interpreted with caution. Furthermore, psychological stress related to social meals was not measured, which may also play a role in the observed association. Future studies using more comprehensive, validated dietary assessments and considering psychological stress factors are necessary to confirm these findings. Prospective studies are needed to verify causality and to explore potential intervention strategies.

Our research suggests a potential protective association between physical activity and the occurrence of breast nodules, diverging from previous studies that primarily emphasized its effects on preventing breast cancer. A previous study indicated an increased risk of breast cancer in individuals with nonmanual occupations [20]. Moreover, Suzanne C et al. found that reducing sedentary time was likely to lower the risk of breast cancer [21]. The mechanisms of breast disease risk reduction through physical activity — including its potential effects on breast nodules — may involve reductions in circulating sex steroids, which are known to increase the risk of breast disease [20]. Our study provides additional evidence that implementing such behavioral changes may be associated with not only a lower risk of breast cancer but also a reduced incidence of breast nodules.

This study provides further insights into the association between renal function-related biomarkers, particularly SCr and SUA, and the occurrence of breast nodules. Our findings supported previous research, which reported that women with breast nodules had slightly lower SUA levels than those without breast nodules [4]. Interestingly, in the same study, SCr levels displayed no significant correlation with the occurrence of breast nodules [4]. A prior study underscored the mediating role of SUA in the relationship between BMI and breast cancer [22]. Notably, despite the observed association between higher levels of SCr and an increased risk of breast nodules in this study, a significant correlation was lacking in other recent studies. Given the large sample size, the statistically significant associations observed for SCr and SUA reflect very small effect sizes, and their clinical relevance remains uncertain. Further prospective studies are needed to clarify the complex relationship between renal function-related biomarkers and breast health and to determine their clinical significance in breast nodule risk assessment.

This study observed a specific association whereby women with both good and fair sleep qualities were more prone to developing breast nodules. However, the risk of progression from lower to higher BI-RADS categories was relatively lower. A previous study found a protective effect of short sleep duration and an adverse effect of long sleep duration on the risk of breast cancer [23]. Research has shown that sleep influences breast health through a complex interplay of potential mechanistic pathways, including oxidative stress, melatonin, inflammation, and metabolic function pathways [24]. We speculate that there may be underlying mechanisms at play, with nodules that develop in women with better sleep quality tending to remain relatively stable and less likely to progress to more malignant states. This study contributes additional observations to the understanding of the relationship between sleep and breast health. Future studies may investigate the potential role of sleep quality in personalized breast health management, especially for women with benign nodules that are less likely to progress to malignant nodules.

Our study found that age was associated with the risk of occurrence and progression of breast nodules. Specifically, the 41–50 years age group exhibited the highest risk of breast nodule occurrence, and this finding was consistent with the results of several other studies [4, 25]. Researchers have explained that this finding could be linked to changes in estrogen levels before and after menopause [4, 25]. Additionally, beyond the age of 30 years, there was a possible increasing trend in the risk of breast nodule progression to a BI-RADS score ≥ 4 A with increasing age. This indicated that beyond the age of 30 years, age was associated with an increased risk for the progression of benign breast nodules to malignant nodules. This finding may provide preliminary information for clinical practitioners to consider when tailoring individualized breast screening plans based on age.

In our study, measurement years were associated with the risk of the occurrence and progression of breast nodules. Our study found an apparent increase in the risk of breast nodule occurrence after 2017. Technological advances in digitized leisure activities have led to increased use of electronic devices and a global rise in sedentary lifestyles beyond work hours [26]. Our study also observed a decreasing trend in the percentage of individuals sitting for 2–4 h outside of work after 2017, alongside an increase in the percentage of individuals sitting for 4–6 h (Table 3). Furthermore, our research indicated that the rise in sedentary time outside of work was associated with the increasing trend in breast nodule occurrence. This may, to some extent, be related to the increased risk of breast nodule occurrence observed after 2017. However, the risk of breast nodules progressing to a BI-RADS score ≥ 4 A after 2017 did not exhibit a pronounced upward trend. This could be linked to the Chinese government’s expanded breast cancer screening efforts, which may have contributed to earlier detection of abnormalities and a lower likelihood of benign lesions advancing to higher-risk categories [27, 28].

Table 3.

Sitting duration outside of work for women from 2012 to 2021

Measurement years	Duration of sitting outside of work
	< 2 h	2–4 h	4–6 h	≥ 6 h
2012	769(29.05)	1136(42.92)	467(17.64)	275(10.39)
2013	1467(28.55)	2190(42.62)	923(17.96)	558(10.86)
2014	1898(27.30)	2967(42.68)	1308(18.81)	779(11.21)
2015	2684(26.20)	4575(44.66)	1938(18.92)	1047(10.22)
2016	4265(26.84)	6987(43.97)	2995(18.85)	1642(10.33)
2017	4284(26.83)	6992(43.78)	2992(18.74)	1701(10.65)
2018	4535(27.49)	6890(41.77)	3133(18.99)	1936(11.74)
2019	4749(27.44)	6956(40.18)	3282(18.96)	2323(13.42)
2020	3653(26.69)	5641(41.21)	2551(18.64)	1842(13.46)
2021	2905(27.82)	4142(39.66)	2024(19.38)	1373(13.15)

Limitations

This study has several limitations that should be considered. First, lifestyle factors such as diet, physical activity, and sleep were assessed using single-item, unvalidated questionnaires, which may not fully capture the complexity of these behaviors. Depression and anxiety were screened using brief tools (PHQ-2 and GAD-2) rather than clinical diagnoses, and no data on medications or biological markers such as cortisol were collected, which limits the interpretation of these results. Second, anthropometric measurements were based solely on BMI, without assessing central obesity or body composition, restricting a more detailed understanding of the role of obesity. Third, the classification of BI-RADS category 5 lesions (without pathological confirmation) as progression may have led to misclassification; however, the number of such events was small (n = 5), and their impact on the overall results is expected to be minimal. Additionally, while several potential confounders were adjusted for, data on family history of breast cancer, hormone therapy use, and detailed socioeconomic factors were unavailable, which could have introduced residual confounding and bias in the estimation of associations. While the number of progression events (n = 478) was adequate for model fitting, further validation in larger samples is necessary. Finally, as an observational study, causal inferences cannot be drawn, and the findings should be interpreted as a longitudinal extension of existing cross-sectional evidence.

Despite these limitations, this study provides valuable longitudinal evidence on the dynamic evolution of breast nodules and established risk factors, offering insights for future prospective research that includes more refined measurements and larger sample sizes.

Conclusions

In conclusion, our study revealed significant correlations between the occurrence of breast nodules and various lifestyle factors and metabolic characteristics. Additionally, we identified that age, sleep quality and measurement years were associated with both the risk of occurrence and progression of breast nodules. These findings suggest that lifestyle factors, particularly dietary habits and physical activity, may be related to the risk of the occurrence of breast nodules in women, which could provide preliminary research direction for the development of simple, practical, and cost-effective interventions for the prevention of breast nodules. When formulating personalized breast health plans, physicians may take age and sleep quality into consideration as part of a comprehensive assessment for the risk of breast nodule occurrence and malignant transformation.

Authors’ contributions

Conceptualization and Methodology: HL, YL, JX. Data curation and Investigation: HL, YD, NQ, GG, ML, YH, PY, YW. Writing - original draft: HL, YL, YD. Writing - review & editing: AC, YX, JX. Formal analysis: HL, YD, NQ, GG, ML. Funding acquisition: JX. Project administration and Supervision: YL, JX.

Funding information

This work was supported by the Wisdom Accumulation and Talent Cultivation Project of the Third Xiangya Hospital of Central South University (NO. BJ202205).

Data availability

The data the support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This research was approved by the Survey and Behavioral Research Ethics Committee of the Third Xiangya Hospital of Center South University (Ref. No. 23455).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.