The association between dietary diversity and patterns in Chinese university students with abnormal uterine bleeding: a cross-sectional study.

The sample comprised 698 students, with 434 (62.2%) having experienced AUB. Among these, 87 (20.05%) reported abnormal frequency, 250 (57.60%) reported irregular menstrual cycle, 41 (9.45%) reported prolonged duration, and 268 (61.75%) reported abnormal flow volume. Freshmen and sophomores comprised 468 students (67.05%), while 609 (87.25%) were nonsmokers and abstained from alcohol consumption. Menarche occurred primarily between the ages of 11 and 13 years for 514 students (73.64%). Dysmenorrhea was reported by 439 students (62.89%). The median PSST score was 19.00 (range: 0.00–57.00). Light physical activity was reported by 331 (47.97%) students. Moderate-to-severe sleep disorders affected 434 students (62.18%), and 568 (81.38%) reported moderate perceived stress. In terms of various factors, including age, height, weight, BMI, grade, major, smoking and drinking habits, mother's menstrual disorder history, family history of hypertension/diabetes, menarche age, VAS score, premenstrual symptoms, physical activity, sleep status, and perceived stress, no significant differences were observed in the statistical data between students with AUB and those without AUB. (Table  1 ). Table 1 Baseline characteristics of the participants (N = 698) Total Normal AUB Variables n = 698 (n = 264; 37.8%) (n = 434; 62.2%) p value Frequncy, n (%) –   Normal 611 (87.54) 264 (100.00) 347 (79.95)   Abnormal 87 (12.46) 0 (0.00) 87 (20.05) Regularity, n (%) –   Normal 448 (64.18) 264 (100.00) 184 (42.40)   Irregular 250 (35.82) 0 (0.00) 250 (57.60) Duration, n (%) –   Normal 657 (94.13) 264 (100.00) 393 (90.55)   Prolonged 41 (5.87) 0 (0.00) 41 (9.45) Volume, n (%) –   Normal 430 (61.60) 264 (100.00) 166 (38.25)   Abnormal 268 (38.40) 0 (0.00) 268 (61.75)   Age(year),mean (SD) 19.93 (1.41) 20.03 (1.41) 19.87 (1.41) 0.145   Height(cm), mean (SD) 162.27 (5.49) 162.37 (5.84) 162.21 (5.27) 0.699   Weight(kg), mean (SD) 52.77 (8.04) 53.02 (8.59) 52.62 (7.69) 0.527   BMI(kg/m 2 ), mean (SD) 20.01 (2.67) 20.07 (2.85) 19.97 (2.55) 0.627 Grade, n (%) 0.184   Freshman and sophomore 468 (67.05) 169 (64.02) 299 (68.89)   Junior, senior and above 230 (32.95) 95 (35.98) 135 (31.11)   Major, n (%) 0.593   Science 55 (7.88) 20 (7.58) 35 (8.06)   Engineering 90 (12.89) 32 (12.12) 58 (13.36)   Medical Science 190 (27.22) 66 (25.00) 124 (28.57)   Humanities 363 (52.01) 146 (55.30) 217 (50.00) Habit, n (%) 0.817   None 609 (87.25) 233 (88.26) 376 (86.64)   Smoking or drinking 74 (10.60) 26(9.85) 48 (11.06)   Smoking and drinking 15 (2.15) 5(1.89) 10(2.30) Mother’s menstrual disorder history, n (%) 0.612   Yes 123 (17.62) 49 (18.56) 74 (17.05)   No or unclear 575 (82.38) 215 (81.44) 360 (82.95) Family history of Hypertension/Diabetes, n (%) 0.992   Yes 177 (25.36) 67 (25.38) 110 (25.35)   No or unclear 521 (74.64) 197 (74.62) 324 (74.65) Menarche age(year), n (%) 0.214   9–10 45 (6.45) 22 (8.33) 23 (5.30)   11–13 514 (73.64) 189 (71.59) 325 (74.88)   14–16 136 (19.48) 53 (20.08) 83 (19.12)    > 16 3 (0.43) 0 (0.00) 3 (0.69) VAS, n (%) 0.170   0 259 (37.11) 87 (32.95) 172 (39.63)   Mild pain 182 (26.07) 71 (26.89) 111 (25.58)   Moderate pain 114 (16.33) 52 (19.70) 62 (14.29)   Severe pain 143 (20.49) 54 (20.45) 89 (20.51)   PSST total, Median (Min–Max) 19.00 (0.00–57.00) 19.00 (0.00–56.00) 19.00 (0.00–57.00) 0.237   PSST symptoms, Median (Min–Max) 15.00 (0.00–42.00) 15.00 (0.00–41.00) 15.00 (0.00–42.00) 0.259   PSST functional,Median (Min–Max) 3.00 (0.00–15.00) 3.00 (0.00–15.00) 3.00 (0.00–15.00) 0.260 IPAQ, n (%) 0.518   Light physical activity 331 (47.97) 119 (45.59) 212 (49.42)   Moderate physical activity 105 (15.22) 39 (14.94) 66 (15.38)   Vigorous physical activity 254 (36.81) 103 (39.46) 151 (35.20) SRSS, n (%) 0.490   Good sleep status 13 (1.86) 7 (2.65) 6 (1.38)   Fair sleep status 36 (5.16) 12 (4.55) 24 (5.53)   Mild sleep disorders 215 (30.80) 86 (32.58) 129 (29.72)   Moderate to severe sleep disorders 434 (62.18) 159 (60.23) 275 (63.36) PSS-10, n (%) 0.778   Low levels 88 (12.61) 35 (13.26) 53 (12.21)   Moderate levels 568 (81.38) 215 (81.44) 353 (81.34)   High levels 42 (6.02) 14 (5.30) 28 (6.45) AUB abnormal uterine bleeding, VAS visual analog scale, PSST premenstrual symptoms screening tool, IPAQ International Physical Activity Questionnaire, SRSS Self-Rating Scale of Sleep, PSS-10 Perceived Stress Scale-10, Continuous data are shown as the mean ± standard deviation or median (min–max), and categorical data are shown as n (%) Baseline characteristics of the participants (N = 698) AUB abnormal uterine bleeding, VAS visual analog scale, PSST premenstrual symptoms screening tool, IPAQ International Physical Activity Questionnaire, SRSS Self-Rating Scale of Sleep, PSS-10 Perceived Stress Scale-10, Continuous data are shown as the mean ± standard deviation or median (min–max), and categorical data are shown as n (%) The results of the factor analysis adaptability test were as follows: KMO = 0.828; Bartlett’s sphericity test: χ 2  = 5215.87, p  < 0.001. The scree plots and factor analysis revealed eigenvalues for the first five principal components of 4.126, 3.167, 2.538, 2.364, and 2.115, accounting for 16.504%, 12.670%, 10.152%, 9.457%, and 8.462% of the total variance postvarimax rotation, respectively. Thus, the cumulative variance explained by the first five components was 57.245%. An adequate representation of food items or categories on a principal component was considered if they had an absolute factor loading of 0.5 or higher. The first group, high in porridge, flours, desserts, fried food, stuffing, coarse grains, and potatoes, was termed the “carb-rich” dietary pattern owing to its carbohydrate-rich foods. The second group, which included seafood, shrimp, beer, yellow wine, and white spirit, was labeled the “aqua-alcohol” dietary pattern. The third group, with high loadings of dairy, processed meat, soybean products, and nuts, was named the “mixed protein-based” dietary pattern. The fourth food group included dark vegetables, light-colored vegetables, mushrooms, and fruits, corresponding to the “low fat plant-based” pattern. The fifth group comprised red meats and poultry, aligned with the “red and white meat-based” pattern. Thus, five types of dietary patterns were identified: “carb-rich”, “aqua-alcohol”, “mixed protein-based”, “low fat plant-based”, and “red and white meat-based” (Table  2 ). Table 2 Distribution of factor loadings for each food or food group across the five dietary patterns of university students Foods or food groups Correlation coefficients (n = 426) Carb-rich Aqua-alcohol Mixed protein-based Low fat plant-based Red and white meat-based Rice Porridge 0.575 Flours 0.554 Desserts 0.740 Fried food 0.724 Stuffing 0.799 Coarse grains 0.814 Potatoes 0.750 Dairy 0.577 Egg Red meats 0.885 Poultry 0.863 Processed meat 0.682 Soybean products 0.554 Seafood 0.514 Shrimp 0.540 Nuts 0.504 Dark vegtables 0.836 Light-colored vegetables 0.873 Mushroom 0.574 Fruits 0.502 Sweetened beverages Beer 0.732 Yellow wine 0.908 White spirit 0.904 The absolute value of the factor loadings < 0.5 is excluded Distribution of factor loadings for each food or food group across the five dietary patterns of university students The absolute value of the factor loadings < 0.5 is excluded The factor analysis resulted in the extraction of five distinct dietary patterns, each represented by the corresponding factor scores. Preference for a specific dietary pattern was determined via the score assigned to that dietary pattern. A higher score indicates a greater preference [ 45 ]. We examined each individual's factor scores across the five dietary patterns and identified and labeled the predominant pattern on the basis of the highest score. The highest factor score was identified as the primary pattern score for each participant. The dietary patterns were categorized as follows: FAC1 represented “carb-rich”, FAC2 represented “aqua-alcohol”, FAC3 represented “mixed protein-based”, FAC4 represented “low fat plant-based”, and FAC5 represented “red and white meat-based”. Among the 426 university students, 67 (15.73%) were “Carb-Rich”, 75 (17.61%) were “aqua-alcohol”, 102 (23.94%) were “mixed protein-based”, 84 (19.72%) were “low fat plant-based”, and 98 (23.00%) were “red and white meat-based”. Univariate binary logistic regression revealed that AUB risk decreased by 20% ( p  = 0.001) per unit increase in the DDS (Table  3 ). Table 3 Univariate analysis for AUB Variables Statistics OR (95%CI) p Value Age(year),mean (SD) 19.93 ± 1.41 0.92 (0.83, 1.03) 0.145 Height(cm),mean (SD) 162.27 ± 5.49 0.99 (0.97, 1.02) 0.698 Weight(kg),mean (SD) 52.77 ± 8.04 0.99 (0.98, 1.01) 0.527 BMI(kg/m 2 ),mean (SD) 20.01 ± 2.67 0.99 (0.93, 1.04) 0.627 IPAQ, n (%)   Light physical activity 331 (47.97) Reference   Moderate physical activity 105 (15.22) 0.95 (0.60, 1.50) 0.825   Vigorous physical activity 254 (36.81) 0.82 (0.59, 1.15) 0.256 SRSS, n (%)   Good sleep status 13 (1.86) Reference   Fair sleep status 36 (5.16) 2.33 (0.64, 8.49) 0.199   Mild sleep disorders 215 (30.80) 1.75 (0.57, 5.39) 0.329   Moderate to severe sleep disorders 434 (62.18) 2.02 (0.67, 6.11) 0.214 PSS-10, n (%)   Low levels 88 (12.61) Reference   Moderate levels 568 (81.38) 1.08 (0.68, 1.72) 0.730   High levels 42 (6.02) 1.32 (0.61, 2.85) 0.479   DDS as continuous variable, mean (SD) 6.74 ± 1.60 0.80 (0.70, 0.91) 0.001 Factor score for each dietary pattern   FAC1, Median (Min–Max) 0.00 (− 0.07–14.27) 0.82 (0.59, 1.13) 0.226   FAC2, Median (Min–Max) 0.00 (− 0.19–10.24) 1.29 (0.87, 1.92) 0.200   FAC3, Median (Min–Max) 0.00 (− 0.12–10.55) 0.92 (0.71, 1.19) 0.510   FAC4, Median (Min–Max) 0.00 (− 0.09–4.84) 0.82 (0.64, 1.05) 0.121   FAC5,Median (Min–Max) 0.00 (− 0.04–3.60) 1.17 (0.87, 1.59) 0.298 Dietary Pattern   Carb-Rich, n (%) 67 (15.73) 1.0   Aqua-Alcohol, n (%) 75 (17.61) 1.53 (0.77, 3.03) 0.228   Mixed Protein-Based, n (%) 102 (23.94) 1.11 (0.59, 2.08) 0.738   Low Fat Plant-Based, n (%) 84 (19.72) 0.79 (0.41, 1.51) 0.475   Red and White Meat-Based, n (%) 98 (23.00) 1.55 (0.81, 2.96) 0.182 CI confidence interval, OR odds ratio, IPAQ International Physical Activity Questionnaire, SRSS Self-Rating Scale of Sleep, PSS-10 Perceived Stress Scale-10, DDS dietary diversity score The factor score for each dietary pattern was categorized as follows: FAC1 corresponded to “carb-rich”, FAC2 corresponded to “aqua-alcohol”, FAC3 corresponded to “mixed protein-based”, FAC4 corresponded to “low fat plant-based”, and FAC5 corresponded to “red and white meat-based”. Continuous data are shown as the mean ± standard deviation, and categorical data are shown as n (%) Univariate analysis for AUB CI confidence interval, OR odds ratio, IPAQ International Physical Activity Questionnaire, SRSS Self-Rating Scale of Sleep, PSS-10 Perceived Stress Scale-10, DDS dietary diversity score The factor score for each dietary pattern was categorized as follows: FAC1 corresponded to “carb-rich”, FAC2 corresponded to “aqua-alcohol”, FAC3 corresponded to “mixed protein-based”, FAC4 corresponded to “low fat plant-based”, and FAC5 corresponded to “red and white meat-based”. Continuous data are shown as the mean ± standard deviation, and categorical data are shown as n (%) Regardless of the adjustment for confounding variables, the inverse relationship between the DDS and AUB risk was significant in all the multivariate logistic regression models. Specifically, Model 1 yielded an OR (odds ratio) of 0.80 (95% CI: 0.70, 0.91), whereas Model 2 produced an OR of 0.80 (95% CI: 0.70, 0.91). Notably, after adjusting for potential confounding variables in Model 3, a 1-unit increase in the DDS was linked to a 22% lower risk of AUB (odds ratio 0.78, 95% CI: 0.67, 0.90). A high DDS (DDS ≥ 6) was consistently associated with a lower risk of AUB across all models, with an OR of 0.50 (95% CI: 0.30, 0.84) in Model 1, an OR of 0.50 (95% CI: 0.30, 0.84) in Model 2, and an OR of 0.52 (95% CI: 0.31, 0.89) in Model 3 (Table  4 ). Table 4 Associations between the DDS and AUB in different models Characteristics OR(95% CI), p Value Model 1 Model 2 Model 3 DDS as continuous variable 0.80 (0.70, 0.91) 0.0007* 0.80 (0.70, 0.91) 0.0006* 0.78 (0.67, 0.90) 0.0008* DDS as categorical variable (ref. = Low DDS) High DDS 0.50 (0.30, 0.84) 0.0081* 0.50 (0.30, 0.84) 0.0083* 0.52 (0.31, 0.89) 0.0169* Model 1: no adjustment; Model 2: adjusted for age and BMI; Model 3: adjusted for age, BMI, grade, habit, mother’s menstrual disorder history, family history of hypertension/diabetes, IPAQ score, SRSS score, PSS-10 score, PSST total score, VAS score, FAC1, FAC2, FAC3, FAC4, and FAC5 Abbreviations: CI confidence interval, OR odds ratio, DDS , dietary diversity score, High DDS: DDS ≥ 6; low DDS: DDS < 6 * p < 0.05 Associations between the DDS and AUB in different models Model 1: no adjustment; Model 2: adjusted for age and BMI; Model 3: adjusted for age, BMI, grade, habit, mother’s menstrual disorder history, family history of hypertension/diabetes, IPAQ score, SRSS score, PSS-10 score, PSST total score, VAS score, FAC1, FAC2, FAC3, FAC4, and FAC5 Abbreviations: CI confidence interval, OR odds ratio, DDS , dietary diversity score, High DDS: DDS ≥ 6; low DDS: DDS < 6 * p < 0.05 Our findings revealed a nonlinear connection between the "Low Fat Plant-Based" dietary pattern factor score (FAC4) and AUB, as shown by the GAM and smooth curve fitting in Fig.  2 . Fig. 2 Curve fitting analysis of FAC4 and AUB. FAC4, “low fat plant-based” dietary pattern factor score Curve fitting analysis of FAC4 and AUB. FAC4, “low fat plant-based” dietary pattern factor score Table 5 presents the results of the two-piecewise linear regression model and recursive algorithm. The log-likelihood ratio test, which compares the linear and two-piecewise linear regression models, yielded a p value of 0.02, indicating that the latter is a better fit. In the fully adjusted model, a nonlinear reverse L-shaped association was observed between FAC4 and AUB, with an inflection point at 1.45. Before this point, there was a significant negative correlation between FAC4 and AUB (OR: 0.42, 95% CI: 0.21, 0.84). After the inflection point, the association was not significant (OR: 1.73, 95% CI: 0.93, 3.22). These results indicate that a "low fat plant-based" dietary pattern positively affects AUB prevention within a certain range. Table 5 Threshold effect analysis of FAC4 on AUB via a two-piecewise linear regression model Outcome: AUB Adjust OR (95%CI) p Value Fitting by linear regression model 0.90 (0.68, 1.19) 0.46 Fitting by two-piecewise linear regression model Inflection point 1.45   1.45 1.73 (0.93, 3.22) 0.09 p for log likelihood ratio test 0.02* Age, BMI, grade, habit, mother’s menstrual disorder history, family history of hypertension/diabetes, IPAQ score, SRSS score, PSS-10 score, PSST total score, VAS score, FAC1, FAC2, FAC3, and FAC5 were adjusted * p  < 0.05 Threshold effect analysis of FAC4 on AUB via a two-piecewise linear regression model Age, BMI, grade, habit, mother’s menstrual disorder history, family history of hypertension/diabetes, IPAQ score, SRSS score, PSS-10 score, PSST total score, VAS score, FAC1, FAC2, FAC3, and FAC5 were adjusted * p  < 0.05

From April to May 2024, cross-sectional research was conducted among university students in Shenzhen, Guangdong, South China, via Questionnaire Star, a platform for creating and collecting data on various devices. The study utilized a self-administered online questionnaire, allowing participants to complete surveys independently, ensuring cost-effective and accessible data collection. To participate, individuals were required to click “Agree” on the informed consent form, and those who did not comply were disqualified. Completing the questionnaires was deemed consenting. The participants were freshman- to fifth-year undergraduates aged at least 17 years who voluntarily participated in this study. The following criteria were used to exclude individuals from the study: (1) pregnancy and abortion within three months, (2) use of ovulation-affecting drugs or anticoagulants during the past three months, and (3) significant organ diseases, including but not limited to brain, heart, lung, liver, and kidney. Examples include: stroke, myocardial infarction, type I diabetes, secondary dyslipidemia, renal dysfunction, epilepsy, hemophilia, and infectious diseases. One study indicated that 18.4% of Koreans aged 19–29 years experience AUB [ 19 ]. In this cross-sectional study, the allowable error δ was set to 3%. α was set at 0.05, u α/2  = 1.96, the degree of confidence was set at 0.8, and the required sample size was 640. Considering a 10% nonresponse rate and factoring in constraints such as human, material, and financial limitations, 728 questionnaires were distributed. Among the 724 university students who participated, 698 valid responses were obtained (effective recovery rate: 95.8%), with 426 completing the dietary questionnaire (Fig.  1 ). Fig. 1 Participant inclusion flow chart Participant inclusion flow chart Information on demographics: age, height, weight, grade, and major. Smoking or drinking habits, mother’s menstrual disorder history, family history of hypertension or diabetes, and age at menarche were recorded. According to the FIGO AUB System 1 [ 34 ], the criteria for diagnosing AUB include menstrual flow characteristics, such as frequency, regularity, duration, and amount outside the scope of normal ranges. Normal frequency: 24–38 days. The regularity was defined by variations in the cycle length. Irregular menstrual cycles were defined as variations in cycle length of ≥ 8–10 days, extending from the shortest to the longest cycle. Prolonged flow was defined as a flow lasting > 8 days [ 35 ]. The categorization of flow volume was determined on the basis of the number of pads or sanitary towels that changed per day. The following figures were obtained: light (less than 4 pads), moderate (5–7 pads), and heavy (> 8 pads). Dysmenorrhea was assessed by measuring pain intensity and menstrual symptoms [ 36 ]. The visual analog scale (VAS) was used to assess pain intensity. A 10-cm line was marked with the label from “no pain” to “most extreme pain”. In our research, scores from 1 to 4 were mild, scores from 5 to 6 were moderate, and scores from 7 to 10 were severe [ 37 ]. The premenstrual symptom screening tool (PSST)[ 38 ], which contains 14 measures of symptoms and 5 measures of functional impairment due to premenstrual symptom (PMS), was utilized to evaluate the severity of PMS. Each item is rated on a scale from 0 to 3: 0 (not at all), 1 (mild), 2 (moderate), and 3 (severe) [ 39 ]. The International Physical Activity Questionnaire (IPAQ) was used to assess students' physical activity levels in the previous week. The IPAQ short form (IPAQ-SF) [ 40 ] comprises seven questions, summarizing data by intensity (metabolic equivalent of task (METs)), duration (hours/week), and average energy expenditure. It uses standardized methods, assigning 1.0, 4.0, 8.0, and 3.3 METs for sedentary activity, moderate activity, vigorous activity, transportation or walking, respectively. These scoring criteria for short scales assess university students' physical activity into three distinct categories: light, moderate, and vigorous. The Self-Rating Scale of Sleep (SRSS) was modified by psychologists for use in the Chinese population and has been widely employed in many studies [ 41 ]. The SRSS comprises ten items, yielding total scores between 10 and 50. Sleep quality was categorized into four levels: scores of 10–19, 20–21, 22–25, and 26–50 denoted good sleep, fair sleep, mild sleep disorder, and moderate–to-severe sleep disorders, respectively. The Perceived Stress Scale (PSS-10) is widely employed for the assessment of long-term stress as perceived [ 42 ]. It comprises ten items that explore the emotions and thoughts associated with personal difficulties and behaviors encountered during the preceding month. Each item is scored from 0 to 4, with scores ranging from 0 to 40. The categorization of the scores for perceived chronic stress levels was as follows: low (≤ 13), moderate (14–26), and high (27–40). The Food Frequency Questionnaire (FFQ25) collects data on food consumption and frequency over the past six months. A modified food frequency questionnaire (FFQ) was validated among nonpregnant young women [ 43 ]. The survey created a 25-item FFQ25 food list to identify common foods among university students. Dietary patterns were evaluated and classified via factor analysis. Principal factors with eigenvalues exceeding one were extracted via the Kaiser criterion, and varimax orthogonal rotation ensured a meaningful factor structure. Each food or food group's frequency was documented as weekly consumption, with solid foods in grams and liquid foods in milliliters per serving and with solid foods in grams and liquid foods in milliliters per serving. Trained researchers have assessed the dietary diversity score (DDS) via the FFQ25 [ 44 ], classifying university students' diets into nine groups. A score of 1 was allocated to each food group on a weekly basis, with the DDS varying from 0 to 9. The DDS score measures dietary diversity, with a score below 6 indicating a low level of diversity. Categorical variables were presented as frequencies and percentages, and continuous variables were reported as the mean ± standard deviation (SD) or median (min, max). Student’s t test, the χ 2 test or the Mann‒Whitney U test was used for differences between the variables. To investigate the association between the DDS and AUB, we developed models using univariate and multivariate binary logistic regression. There were three types of models: nonadjusted, minimally adjusted, and fully adjusted. We explored the nonlinear associations between the factor scores for each dietary pattern (FAC) and AUB via a generalized additive model (GAM) with smooth curve fitting. Upon identifying nonlinearity, a recursive algorithm was employed to detect significant inflection points in the FAC-AUB relationships, followed by threshold effect analysis, which compared the logistic regression model with the two-part logistic regression model. R software (version 4.2.0, http://www.R-project.org ) and EmpowerStats ( http://www.empowerstats.com ) were used for all analyses. The statistical significance level was set at p  < 0.05 (two-sided).

Two evaluation and categorization systems for AUB have been developed by FIGO. The FIGO System 2 [ 35 ] employs the PALM-COEIN acronym: "PALM" for structural causes (Polyp, Adenomyosis, Leiomyoma, Malignancy) and "COEI" for nonstructural causes (Coagulopathy, Ovulatory dysfunction, Endometrial, Iatrogenic). "N" stands for unclassified entities. FIGO System 1 precedes System 2 in evaluating AUB [ 8 ]. The new FIGO classification system's application to assess AUB prevalence among Chinese college students is under-researched. Globally, self-reported menstrual patterns help identify AUB. We used a self-assessment survey to classify bleeding as abnormal, based on FIGO's four parameters. Our study revealed a 62.2% AUB prevalence among Shenzhen university students, surpassing the 32.6% and 18.4% reported in Ethiopian [ 46 ] and Korean studies [ 19 ], potentially due to regional, genetic, socio-economic factors. A Chinese study involving gynecology outpatients aged 18–57 years reported a 60.8% prevalence via self-assessment [ 47 ]. The higher prevalence rate in our study may stem from college students' increased symptom awareness and reporting ability. Ovulatory dysfunction-associated AUB (AUB-O) is the principal cause of AUB, with an approximately 50% prevalence among Chinese women [ 48 – 50 ]. Anovulatory cycles can lead to irregular menstrual bleeding. Our study revealed that 250 students (57.60%) had an irregular menstrual frequency, suggesting possible AUB-O. Previous research revealed that AUB proportions increased during the early and late reproductive years, indicating that age and BMI were significantly associated with AUB [ 2 ]. AUB is also more prevalent among athletes involved in anaerobic and resistance sports or training continuously for more than 60 min/day [ 51 ]. However, our study revealed no correlations between AUB and age, BMI, or physical activity. A previous study revealed a connection between menstrual regularity and the severity of dysmenorrhea [ 52 ]. In contrast, our research did not reveal any substantial relationship between AUB and the intensity of dysmenorrhea, age at menarche, sleep status, or stressful life events, possibly due to different study populations and designs. Nutritionists have long regarded dietary diversity as crucial for enhancing dietary quality [ 53 ]. Enhancing dietary diversity is essential for improving nutritional status and fostering overall health [ 54 ]. The DDS assesses the quantity of food items or categories ingested within a specific timeframe [ 55 ] and functions as a sign of sufficient micronutrient intake [ 56 ]. Our analysis demonstrated that a 1-unit increase in the DDS was linked to a 22% lower risk of AUB (OR: 0.78, 95% CI: 0.67, 0.90). A high DDS (DDS ≥ 6) consistently indicated a reduced risk of AUB across all the models. Dietary diversity reduces the risk of AUB through various biological processes. A low DDS is associated with an undernutrition risk [ 57 ]. Diets that lack nutrients, vitamins, and minerals adversely affect women's health, increasing the likelihood of illness and menstrual disorders [ 58 , 59 ]. A diverse diet ensures better nutritional status and provides essential vitamins and minerals necessary for blood clotting and reproductive function [ 60 ]. Polycystic ovary syndrome (PCOS), a common cause of AUB-O, affects 5%−13% of reproductive-aged women presenting with AUB, infertility, hirsutism, acne, or obesity [ 61 , 62 ]. Research has indicated a connection between PCOS and gut microbiota dysbiosis. A varied diet supports a healthy gut microbiome, producing anti-inflammatory short-chain fatty acids that affect hormonal balance [ 63 , 64 ]. Future PCOS studies could benefit from personalized gut microbiota manipulation [ 65 ]. A high DDS alleviates oxidative stress and inflammation [ 66 ]. One study found that women on an anti-inflammatory diet had fewer menstrual distress symptoms [ 67 ]. It modulates inflammation through nutrients and bioactive compounds that regulate inflammatory pathways [ 68 ]. Dietary diversity enhances well-being and reduces stress, which can disrupt the hypothalamic-pituitary-ovarian axis and cause hormonal imbalances, leading to AUB [ 69 ]. These pathways emphasize the importance of a varied diet in supporting reproductive health and preventing AUB. In our study, we identified the predominant dietary patterns for each individual on the basis of factor scores from the factor analysis, similar to the analyses employed in other studies [ 70 , 71 ]. We conducted a more in-depth analysis to determine whether certain dietary patterns are associated with AUB. The results revealed that a “low fat plant-based” dietary pattern had a positive effect on the prevention of AUB within a specific range. A previous study corroborated these results, indicating that a diet rich in fruits and vegetables can mitigate the risk of uterine fibroids, a cause of AUB [ 72 ]. This finding might be attributed to the fact that a strict low-fat, high-fiber diet in healthy premenopausal women reduces estradiol and estrone levels without influencing ovulation [ 16 ]. The Mediterranean diet, which includes fruits, whole grains, vegetables, and legumes, positively impacts ovarian morphology and function [ 73 , 74 ]. Plant-based nutrients are linked to reduced systemic inflammation, possibly due to antioxidants such as carotene and vitamin C [ 58 , 75 ]. An anti-inflammatory diet aids in weight reduction and improves glucose tolerance, thereby enhancing reproductive parameters and menstrual regulation in women by lowering total testosterone and the free androgen index [ 76 , 77 ]. Furthermore, a low-glycemic diet improves insulin sensitivity and menstrual regularity [ 76 – 78 ]. A previous study linked the consumption of yogurt and cream to anovulation [ 79 ]. Higher meat intake generally impairs ovulation [ 80 , 81 ], as it alters gonadotropin hormone release and follicular maturation, extending the menstrual cycle [ 82 ]. Recent studies have focused on advanced glycation end products (AGEs). Diets rich in fat and protein, such as those comprising cheese, meat, and egg yolks, tend to increase AGE levels [ 83 , 84 ]. In contrast, low-fat milk, grains, legumes, fruits, and vegetables generally have the lowest AGE contents [ 83 ]. High-AGE diets lead to abnormal gene expression associated with steroidogenesis and folliculogenesis and increased macrophage infiltration in ovarian tissue, causing ovarian dysfunction [ 85 – 88 ]. Numerous studies indicate that AGEs are associated with PCOS and its metabolic consequences [ 89 – 91 ], possibly worsening the reproductive dysfunction linked to PCOS [ 92 ]. The observation that there is no significant connection between FAC4 and AUB beyond a certain point is intriguing because it suggests that a "low fat plant-based" diet is beneficial for preventing AUB only within a specific range. Strict adherence to certain diets did not appear to reduce the risk of AUB. A previous study corroborated the present findings, showing that Indian vegetarian diets increase both anti-inflammatory and proinflammatory marker levels in women with PCOS [ 93 ]. The reasons for this remain unclear; however, adherence to a single dietary pattern may reduce dietary diversity, which, as previously discussed, can increase the risk of AUB. The limited associations observed between other dietary patterns and AUB, with the exception of the" low fat plant-based" pattern, may be attributed to the complex nature of dietary patterns, which potentially obscures their individual health effects due to interrelated components [ 94 ]. Methodological limitations in assessing dietary patterns, such as the utilization of FFQs, may not accurately capture the nuances of dietary intake, potentially resulting in unsuitable classification of participants [ 95 ]. The underlying biological mechanisms linking diet to AUB remain incompletely elucidated, and while certain dietary components in a low fat plant-based diet may confer protective effects, other dietary patterns may not exhibit comparable benefits [ 96 ]. Inter-individual variability in response to dietary patterns, influenced by genetic factors, pre-existing health conditions, and lifestyle choices, further complicates the analysis [ 97 ]. In conclusion, the absence of similar associations for other patterns may be attributed to dietary complexity, methodological constraints, unclear biological mechanisms, and individual variability, necessitating further investigation to elucidate these relationships and explore potential dietary interventions to mitigate AUB risk. To the best of our knowledge and on the basis of the literature that is available, this research seems to be the initial effort to investigate the links among dietary diversity, dietary patterns and AUB. Previous studies [ 31 , 32 , 67 , 69 ] have explored the link between menstrual disorders and dietary conditions; however, inconsistent terminology and a lack of standardized etiological classification have led to variability in study populations and potentially influenced results. Our study addressed this by diagnosing menstrual conditions via the term of AUB, as defined by FIGO system 1 [ 1 ]. Notably, we utilized the DDS to quantify the relationship between dietary diversity and AUB. Furthermore, we identified a nonlinear, reverse L-shaped association between the "low fat plant-based" dietary pattern (FAC4) and AUB in a fully adjusted model. These findings indicate the positive effect of this dietary pattern in preventing AUB within a specific range. However, recognizing the methodological constraints of our study is crucial, with several limitations that need to be considered. First, the cross-sectional design precludes drawing causal conclusions; therefore, future longitudinal research is necessary to establish causal relationships. Second, the AUB diagnosis relied on self-reported questionnaires, which may contain inaccuracies. Comprehensive assessments involving gynecological examinations, laboratory tests, and imaging could better exclude other causes of abnormal bleeding and identify organic lesions. However, the lack of data on dietary supplement use and detailed gynecological history may also have influenced the outcomes. Third, the study's focus on a Chinese sample limits its generalizability to other ethnic groups given the variability in dietary habits and AUB prevalence. Selection bias is possible because university student participants may not represent the broader population owing to their educational and socioeconomic status. Additionally, the use of an online survey and self-reported data introduces potential biases, relying on participants' accurate recall of height, weight, dietary habits, and related information, which may be compromised by social desirability and acquiescence biases, leading to underreporting of sensitive information or overreporting of socially acceptable behaviors [ 98 ], which may impact the study's reliability and validity. College students who are exposed to scientific knowledge may report menstrual indices closer to normal, potentially reducing the reported incidence of AUB, and may also tend to report healthier diets, potentially increasing the reported rate of low fat plant-based diets.

The prevention of AUB through lifestyle changes, especially dietary interventions, remains underexplored. Our study underscores the importance of a varied diet as a preventive measure for AUB, utilizing a DDS to quantify the relationship between dietary diversity and AUB. In particular, a “low fat plant-based” dietary pattern had a positive effect on AUB prevention within a certain range. Our study may guide government bodies in implementing policies on the reproductive health of Chinese university students. Future follow-up studies are necessary to establish causality.

Abnormal uterine bleeding (AUB) is a gynecological disorder characterized by irregularities in menstrual volume, frequency, duration, and regularity among women of reproductive age [ 1 ]. Globally, researchers have reported that the prevalence of AUB ranges between 5 and 60% [ 2 – 9 ]. The chronic complications associated with AUB include anemia, infertility, and endometrial cancer. If left untreated, acute AUB can result in severe anemia, low blood pressure, shock, and even death [ 9 ]. Furthermore, it can negatively affect work productivity, quality of life and education [ 10 ]. Factors contributing to AUB include hormonal imbalances and reproductive system abnormalities. In addition to pathological factors, certain lifestyle trends and psychological stressors, such as age, body mass index (BMI) [ 11 ], menarche age [ 12 ], smoking habits [ 13 – 19 ], physical activity, sleep status [ 20 ], depression, and stress [ 21 – 24 ], have been reported to influence menstrual patterns. Dietary diversity, which involves consuming a wide range of foods within a specific timeframe, is closely associated with a decreased risk of chronic diseases [ 25 ]. Various dietary patterns have been linked to the development of persistent health issues such as cardiovascular disorders, malignancies, and diabetes [ 26 ]. A healthy diet is essential for efficient operation of the hypothalamic‒ovarian axis and offers significant promise for enhancing a range of chronic gynecological issues and reproductive health results [ 27 , 28 ]. Omitting breakfast can have detrimental effects on menstrual irregularities in young female college students [ 29 ]. Limiting food consumption for aesthetic purposes during adolescence may have detrimental and long-lasting consequences for the reproductive capabilities of young women [ 30 ]. Irregular menstruation is correlated with a high intake of carbohydrates, protein, fat and calories [ 31 ]. The consumption of junk food directly affects the menstrual cycle [ 32 ]. Spanish students with low adherence to the Mediterranean diet exhibited longer menstrual durations and shorter menstrual periods [ 33 ]. Research on how dietary diversity and patterns influence AUB in Chinese university students is limited, with some studies not clearly defining irregular menstruation. This research investigated the incidence of AUB according to the International Federation of Gynecology and Obstetrics (FIGO) criteria and sought to identify preventative factors.

Additional file 1. Additional file 1.