Scientific Papers

Development and validation of a predictive model of abnormal uterine bleeding associated with ovulatory dysfunction: a case-control study | BMC Women’s Health

A predictive model was developed based on our case–control study, with satisfactory predictive ability and calibration. The predictors involved in the model are age, BMI, SBP, residence, diet, fruits eating, daily sleep duration, parity, and history of ovarian cyst.

Due to the imprecise terminologies and definitions and the lack of a standardized etiologic classification system in the past, the investigation and management of AUB were hampered for a long time [1,2,3,4, 8]. To our knowledge, there are only a few studies focused on the prediction of AUB and no predictive model for AUB-O. Xu et al. proposed the combination of vaginal ultrasonography and bleeding pattern had a good predictive value for AUB. They discovered that BMI, dysmenorrhea, endometrial thickness, diabetes, hypertension, and polycystic ovarian syndrome were related factors of AUB [17]. An analysis of medication-induced heavy bleeding in women with severe mental illnesses conducted in China revealed that some metabolic profiles and antipsychotic therapies were risk factors for heavy bleeding [18]. Four cross-sectional studies [9, 10, 19, 20] conducted in different countries investigated the local prevalence of AUB and associated factors, but with controversial conclusions, suggesting racial differences in AUB prevalence and risk factors might exist.

In our findings, older age was a risk factor of AUB-O. Kazemijaliseh et al. investigated 1393 Iranian women aged 15–45 years to explore the prevalence of AUB and its associated factors, and they observed an increasing prevalence of AUB in older women [10], which agreed with our findings. The Israel researchers carried out an online national questionnaire survey to evaluate menstrual disorders among COVID-19 vaccinated and infected women, and they identified increasing age might be a contributor of AUB after COVID-19 vaccination [13]. However, a Chinese study discovered an inverse relationship between age and HMB in 2356 women aged 18–50 years [20]. The consensus seems to be that ovulatory disorders are more prevalent in females during adolescence and the menopausal transition [7], while more research is still needed to determine the exact mechanisms underlying how menstruation changes as people age.

BMI was another risk factor of AUB-O in our research. Besides aging, the Iranian research also found higher BMI was related to AUB [10]. A Mendelian randomization study [14] investigated 257,193 women of European ancestry in UK Biobank and publicly available genome-wide association studies (GWASs) and revealed that numerous female reproductive disorders are correlated with obesity. Higher BMI was observationally associated with HMB in their study, and leptin and insulin resistance were potential mediators between obesity and female reproductive health [14]. Mena et al. conducted a prospective cohort study based on the Australian Longitudinal Study of Women’s Health (ALSWH) and found that overweight and obese women had a higher risk of irregular periods and HMB, but this effect could be weakened by high levels of physical activity [21]. As estrogen is converted from androstenedione in adipose tissue by aromatase, obese individuals typically have high estrogen levels. In addition, obese women’s sex hormone-binding globulin (SHBG) dropped while their insulin levels rose, promoting the production of androgens. Ovulation and menstrual disorders, including irregular bleeding, oligomenorrhoea, and amenorrhea, are caused by these changes in gonadal steroid concentrations connected to obesity [22].

In Korean female adolescents, sleep duration and irregular menstrual cycles were found to be significantly inversely correlated [23], which was in coincidence with our findings. Hall et al. recruited 11 healthy females to receive a 40-h simultaneous polysomnographic sleep monitoring and luteinizing hormone (LH) measurement, women in the early follicular phase experienced a nocturnal decrease in mean LH and LH pulse frequency due to the inhibited LH and gonadotropin-releasing hormone secretion by sleep [24]. Women’s reproductive function such as folliculogenesis, ovulation, menstruation, hormone synthesis, and secretion can be hampered by sleep deprivation, disruption, dysrhythmia, and disorders, and the complex molecular-genetic and hormonal pathways play a major role in mediating these relationships [25].

Multiparity was found to be a beneficial factor for AUB-O in our study. He et al. reported a lower chance of gravidity and parity in women with oligomenorrhea [26]. Grimes et al. examined anti-mullerian hormone levels in a subset of premenopausal women in the Nurses’ Health Study II. They discovered positive correlations between parity and both the level of anti-mullerian hormone and the timing of menopause, but these relations vanished after further parity adjustment [27]. Meanwhile, a cross-sectional study found a significant correlation between parity and higher levels of ovarian reserve markers [28]. A population-based cohort study included premenopausal participants aged 25 to 42 years at baseline in the Nurses’ Health Study II cohort proposed that parity had an inverse relationship with the risk of early menopause [29]. Similar outcomes were observed in a pooled study carried out by Mishra et al. [30]. Pregnancy inhibits ovulation and may slow the loss of ovarian follicles, which delays menopause [29]. In summary, we speculate that multiparity may have beneficial effects on ovarian function.

Balanced diet is a critical component for the normal functioning of the hypothalamic-ovarian axis, which has considerable promise for improving many chronic gynecologic illnesses and reproductive health [31, 32]. A diet rich in fruits and vegetables has a protective impact that lowers the risk of uterine fibroids and endometriosis [33, 34]. The Mediterranean diet and Dietary Approaches to Stop Hypertension (DASH) diet are two internationally recognized healthy dietary patterns. They have auxiliary therapeutic effects on weight control, prevention and control of cardiovascular diseases, diabetes and many other diseases. Recently, a study reported that these two healthy diets also has the ability to promote the ovarian morphology and function [35]. Compared to the age- and BMI-matched healthy non-vegetarians, women with PCOS who followed an Indian vegetarian diet had greater levels of pro-inflammatory and lower levels of anti-inflammatory markers [36].

Rural Chinese women appeared to be at a higher risk of developing AUB-O than urban residents. There are currently few research focusing on the differences in female reproductive health between urban and rural locations. Rural women, according to Wang et al., have a higher proportion of osteoporosis than women lived in urban areas [37]. A study on Chinese all-cause mortality rate found that the health status of rural residents in China is generally worse than that of urban and suburban residents [38]. In general, the majority of rural Chinese women have lower income and education levels, and poorer health perceptions and medical resources than urban women. The lifestyle disparities between urban and rural women are significant, which may result in physical variations [37].

Despite of the insignificant difference in hypertension between the AUB-O and control groups (Supplemental Table 1), we identified that SBP was a risk factor for AUB-O. In contrast to controls, women with PCOS had significantly higher blood pressure [39]; however, this difference was not seen in women with POI [40]. Estrogen plays a protective effect on the cardiovascular system, ovarian dysfunction could probably interfere with its secretion. Generally, benign ovarian cysts have little effect on ovarian function. However, we found that a history of ovarian cyst was another risk factor of AUB-O. We did not perform any additional pathological classification of ovarian cysts in our study, some of which may impair ovarian ovulation and endocrine function. In the recent Delphi consensus process [7], there was now broad agreement that ovarian tumors, both benign and malignant, may play a role in ovulatory disorders. These findings may prompt more people to consider the impact of an ovarian cyst on ovarian function.

Our study has some limitations. Firstly, recall bias might exist in the data collection process. However, participants completed the questionnaires with the help of trained medical personnel, ensuring the reliability of data. Secondly, limited by the retrospective observational method, our study can only illustrate the association between exposure factors and AUB-O, but cannot prove a causation between them. In addition, the data used to develop and validate the model was completely from China, which might limit the generalizability of the model in other countries. Further research is required in a broader range of ethnicities. Regarding the strength of the study, our predictive model presented a satisfactory prediction and calibration capability with a large sample. Our study proposes the prediction model of AUB-O for the first time, which provides a new tool for public-health workers to assess the risk of AUB-O.

Source link