Associations of Essential and Non-Essential Trace Elements' Levels in the Blood, Serum, and Urine in Women with Premature Ovarian Insufficiency.

The study was conducted at the University Medical Centre Ljubljana, Slovenia, from January 2021 to September 2023. A total of 81 participants, including 40 POI patients and 41 healthy controls, were enrolled in this cross-sectional case–control study. To ensure sufficient statistical power, we calculated the required sample size using the reported standard deviations (SDs) of TEs from previous studies on POI (33–41). However, significant inter-element variability in SDs prevented the determination of a reasonable unified sample size. As a result, we opted for a sample size that aligns with those used in prior studies. The POI patients were recruited from the gynecological outpatient clinic of the University Medical Centre Ljubljana. The inclusion criteria for POI were based on the ESHRE diagnostic criteria (6), which include: (I) age below 40 years at the time of diagnosis, (II) oligo/amenorrhea for at least 4 months, and (III) elevated FSH levels (> 25 IU/L) on two occasions at least 4 weeks apart.Patients with known causes of POI, such as a diagnosed abnormal karyotype or Fragile X syndrome / FMR1 1 gene premutation, as well as those with additional endocrine or chronic diseases, including polycystic ovary syndrome (PCOS), endometriosis, diabetes, Crohn's disease, celiac disease, cardiovascular diseases, liver diseases, kidney diseases, asthma, chronic obstructive pulmonary disease, musculoskeletal diseases, neurological diseases, cancer, thyroid disorders, and hyperprolactinemia, were excluded from the study. The control group consisted of healthy women with regular menstrual cycles, recruited from routine health checkups at the gynecological outpatient clinic in the Community Health Center Ljubljana. Both POI and control group participants met the following common inclusion criteria: a body mass index (BMI) between 18.5 and 29.9 kg/m 2 and an age range of 20 to 49 years. Common exclusion criteria for both groups included a history of reproductive, endocrine, or other chronic systemic disorders, as well as the use of any form of hormonal therapy. All study participants had resided permanently in Slovenia for at least 5 years prior to the date of study inclusion and denied any occupational exposure to analyzed TEs. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Republic of Slovenia National Medical Ethics Committee (No. 0120–158/2022/9). All participants gave their informed written consent to participate in the study. Fasting morning blood and spot urine samples were collected concurrently at the Clinical Institute of Clinical Chemistry and Biochemistry (KIKKB) of the University Medical Centre Ljubljana, Slovenia, where a portion of each fresh sample was analyzed for basic biochemistry parameters (hemogram, urinalysis, liver, and kidney function) using standard enzymatic and colorimetric methods [ 37 – 40 ]. Samples from POI patients were collected at the time of recruitment, while samples from the control group were collected during the early follicular phase (from the second to the fifth day of the menstrual cycle), which reflects the baseline status of ovarian function. The glomerular filtration rate (GFR) was estimated using the Modification of Diet in Renal Disease (MDRD) [ 41 , 42 ], and Chronic Kidney Disease Epidemiology Collaboration (CDK-EPI) formula [ 43 ], and liver function was evaluated using the Fibrosis-4 (FIB-4) index and the aspartate aminotransferase to platelet ratio index (APRI) [ 44 ]. The remaining aliquoted serum, whole blood, and urine samples were frozen at −80 °C and subsequently shipped on dry ice to the Institute Jožef Stefan, Department of Environmental Sciences in Ljubljana, Slovenia, where all analyses of TEs were conducted. TEs analysis was performed in whole blood, serum, and urine samples by validated analytical methods [ 45 , 46 ]. About 0.3 g of whole blood/ serum sample or 0.5 mL of urine sample was weighed into pre-cleaned Teflon tubes. 0,5 mL of 65% HNO3 (suprapure) was added, and samples were subjected to a closed vessel microwave digestion (ULTRAWAVE, Single Reaction Chamber Microwave Digestion System, MILESTONE Srl, Sorisole, Italy). Digested solutions were transferred into measuring tubes and diluted to 5 mL with MilliQ water. The same procedure was applied for blank samples and reference materials. Determination of selected elements (Hg, Pb, Cd, As, Zn, Se, Cu and Mo) was performed using an Agilent 8800 triple quadrupole inductively coupled plasma mass spectrometer (ICP-QQQ, Agilent Technologies, Tokyo, Japan). Isotopes monitored were 55 Mn, 63 Cu, 66 Zn, 75 As, 78 Se, 95,98 Mo, 111,114 Cd, 202 Hg, 206−208 Pb. For the elimination of interferences collision, a reaction cell was used with helium, hydrogen, or oxygen gas. Internal standards (Y, Rh, Sc, and Gd) were added online. The concentrations were determined based on external calibration. Blank samples, control samples, and reference materials were tested together with the samples daily. The limits of detection (LOD) for all nine TEs in each biological sample, calculated as three times the standard deviation of the blank sample, are provided in Supplement 2 . The quality of the results was checked via the regular use of the reference materials Seronorm Trace Elements Whole Blood Level 1 and Level 2, Seronorm Trace Elements Serum Level 1 and Level 2, Seronorm Trace element Urine Level 2 (SERO, Billingstad, Norway) and ClinChek® Urine Control, lyophil., for Trace Elements, Level I, (Receipe, Chemicals,&Instruments GmbH, Germany) and also by participation in the German External Quality Assessment Scheme (G-EQUAS, Erlangen -Nuremberg, Germany). The results of the analysis of the reference materials are combined in the supplement (Supplement 2 ). The results obtained were in good agreement with the reference values. Element concentrations in urine were adjusted for variations in diuresis among samples using creatinine concentration and specific gravity (SG). Since creatinine levels can vary based on factors such as age, body mass index, fat-free mass, and race/ethnicity [ 47 ], and due to systematic variation in urinary flow rates, there is a risk of underestimating exposure due to creatinine over-compensation, particularly for Cd and As [ 46 , 48 ]. Therefore, we adjusted the measurements using SG, as it appears to be a more reliable alternative in the context of environmental exposures [ 46 , 47 , 49 , 50 ]. For accurate exposure and nutritional status comparisons across individuals and within the study population, we present urine concentrations expressed per volume (µg/L) and normalized by creatinine [ 51 ] and SG [ 52 ]. Upon recruitment, participants completed a comprehensive questionnaire that included socio-demographic information, medical, gynecological, and reproductive history, current health status, as well as lifestyle and nutritional habits. The questionnaire was adapted from a previously validated questionnaire used in a national biomonitoring study of TEs in the Slovenian population [ 46 ]. Descriptive statistics included the determination of mean and median values, standard deviation, and 95% confidence interval (CI) of the mean. The normality of the distribution of the variables was tested with the Shapiro–Wilk test. Measurements below the LOD were imputed using the SPSS multiple imputation procedure on natural logarithm-transformed data [ 53 ]. Subsequently, random values were imputed from the estimated distribution within the defined limits (0-corresponding LOD value). To compare levels of TEs between POI and the control group, we performed an independent samples t-test for normally distributed data and the Mann–Whitney test for non-normally distributed data. The Fisher's exact test was employed to compare categorical data between the study groups. To account for multiple comparisons and reduce the risk of false positives, a Bonferroni correction was applied to the bivariate analyses, adjusting the significance threshold to p = 0.000115. Kendall's tau correlation analyses were performed to evaluate correlations between TEs concentrations and statistically significant biochemical and demographic variables. This method was chosen due to the non-normal distribution of the data and the relatively small sample size. The associations between POI (primary outcome) and TE levels (primary exposure) were evaluated using multiple logistic regression models. TE levels were the primary exposure variable, and POI status was the binary outcome variable. Due to the small sample size, only statistically significant ( p  < 0.05) known POI risk factors with correlations below 0.8 (Ƭ < 0.8) were included as covariates. Age and BMI, established POI risk factors [ 54 ], were included in all models as confounders. Hematocrit was adjusted for in models analyzing whole blood TE concentrations to account for erythrocyte-bound TE normalization [ 55 , 56 ]. In total, 27 logistic regression models were performed, each evaluating the association between a specific TE and POI. Further details of the models are provided in Supplement 8 . All tests were two-sided, and a level of p  < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS software (version 29, IBM, Chicago, IL, USA).

In this cross-sectional-case–control study, demographic, biochemical, and environmental parameters, as well as levels of TE in urine, serum, and whole blood, were compared between POI patients and the control group. The demographic and biochemical characteristics of the POI and control groups are summarized in Table  1 , Table  2 , and Supplement 3 . While POI patients were significantly older at enrolment compared to the control group, there was no significant difference between the age at diagnosis in the POI group and the age at enrolment in the control group ( p  = 0.606, Table  1 ). The median age at POI diagnosis was 32 years, ranging from 16 to 38 years. At the time of recruitment, the median (min–max) levels of follicle-stimulating hormone (FSH), luteinizing hormone (LH), and anti-Müllerian hormone (AMH) in the POI group were 59.2 IU/L (24–153 IU/L), 36 IU/L (18.9–87.3 IU/L), and 0.13 ng/L (0.01–1.85 ng/L), respectively. Hormonal levels were not analyzed in the control group, which consisted of healthy women with regular menstrual cycles. No significant differences were observed between the two groups regarding the age at menarche, parity, number of siblings, or liver and kidney function, as assessed by the FIB-4, APRI, MDRD, and CKD-EPI formulas, respectively. The POI group had a significantly higher BMI, with a 1.10-fold increase compared to the control group ( p  = 0.005), as well as elevated red blood cell (RBC) indices, including hemoglobin (1.04-fold, p  = 0.007), hematocrit (1.03-fold, p  = 0.019), and red cell distribution width (RDW) levels (1.08-fold, p  < 0.001). However, after Bonferroni correction, statistical significance was retained only for age at enrolment and RDW levels. Table 1 Distributions of statistically significant demographic and biochemical numerical variables Non-normally distributed variables Median (min–max) POI group ( n  = 40) Control group ( n  = 41) p value Age at enrolment (years) 36 (21 – 47) 31 (20—39)  < 0.001 2 Age at diagnosis (years) 32 (16 – 38) / 0.606 2 Time from diagnosis (years) 4.5 (0 – 12) / / Haemoglobin (g/l) 136 (103 – 150) 131 (105 – 147) 0.007 2 RDW (%) 13.60 (12.5 – 18.7) 12.6 (4.8 – 17.0)  < 0.001 2 Frequency of weekly pasta intake (x times/week) 1 (0 – 7) 2 (0.5 – 5) 0.003 2 Daily time spent driving car (hours) 1 (0 – 5) 1 (0 – 2.5) 0.030 2 Normally distributed variables Mean ± SD POI group (n = 40) Control group (n = 41) p value Body mass index 23.3 ± 4.2 21.1 ± 1.8 0.005 1 Haematocrit 0.405 ± 0.022 0.393 ± 0.024 0.019 1 POI premature ovarian insufficiency, SD standard deviation, RDW red cell distribution width, 1 Student T test for independent samples, 2 Mann Whitney test Table 2 Distributions of statistically significant categorical variables Categorical variables POI group ( n  = 40) Control group ( n  = 41) p value 3 Medication use – nonhormonal (%) Yes 42.9 15.2 0.023 No 57.1 84.8 Mineral supplementation (%) Yes 39.3 78.8 0.003 No 60.7 21.2 Region of residence (%) Osrednjeslovenska 45.5 80.5 0.006 Gorenjska 6.1 2.4 Goriška 3.0 2.4 Jugovzhodna Slovenija 18.2 2.4 Obalno-Kraška 6.1 2.4 Podravska 3.0 2.4 Savinjska 9.1 2.4 Zasavska 9.1 0 Primorsko-Notranjska 0 2.4 Pomurska 0 2.4 Posavska 0 2.4 Type of residence (%) one family house 59.4 26.8 0.013 multi-family house 18.8 24.4 apartment building 18.8 48.8 other 3.1 0 Location of residence (%) city center 25.0 51.2 0.029 city suburb 31.3 36.6 industrial-craft zone 3.1 0 countryside, village 37.5 12.2 countryside, remote area 3.1 0 Home surroundings agriculture (%) Yes 68.8 34.2 0.008 No 31.3 65.8 Home surroundings vineyards (%) Yes 21.9 2.4 0.011 No 78.1 97.6 Home surroundings fruit farming (%) Yes 34.4 13.2 0.034 No 65.6 86.8 Is waste being burned in the neighborhood (%) Yes 46.7 10.0  < 0.001 No 53.3 90.0 Employment status (%) Employed 90.9 68.3 0.019 Student 6.1 29.3 Other 3.0 2.4 Alcohol consumption (%) Never 34.4 2.5  < 0.001 Sometimes 56.3 87.5 Regularly in small amounts 9.4 10.0 Last remembered rice intake (%) not in the last week 44.8 17.9 0.004 in the last week 31.0 71.8 in the last 24 h 24.1 10.3 Last remembered still mineral water intake (%) not in the last week 92.6 60.5 0.009 in the last week 7.4 23.7 in the last 24 h 0 15.8 Last remembered fruit tea intake (%) Not in the last week 53.6 26.3 0.034 In the last 24 h 32.1 34.2 In the last week 14.3 39.5 POI premature ovarian insufficiency, 3 Fisher exact test Distributions of statistically significant demographic and biochemical numerical variables Non-normally distributed variables Median (min–max) Normally distributed variables Mean ± SD POI premature ovarian insufficiency, SD standard deviation, RDW red cell distribution width, 1 Student T test for independent samples, 2 Mann Whitney test Distributions of statistically significant categorical variables POI premature ovarian insufficiency, 3 Fisher exact test Based on the questionnaire responses, no significant differences were observed between the two groups regarding the level of education, current workplace, smoking status, coffee consumption, number of amalgam fillings, or intake of fruits, seafood, vegetables, cereals, dairy products, and meat. In the group of former smokers, the number of responses regarding the duration since smoking cessation (POI n  = 3; HC n  = 4) was insufficient for further statistical analysis. However, the control group reported significantly higher levels of alcohol consumption, showing a 1.55-fold increase compared to the POI group ( p  < 0.001) and more frequent pasta consumption (median: 2 times/week HC vs. 1 time/week POI, p  = 0.003). The control group also had shorter time intervals since their last intake of rice (71.8% HC vs. 31.0% POI within the last week, p  = 0.004), fruit tea (39.5% HC vs. 14.3% POI within the last week, p  = 0.034), and still mineral water (23.7% HC vs. 7.4% POI within the last week, p  = 0.009). Additionally, the control group showed a 2.01-fold higher frequency of mineral supplement use ( p  = 0.003) and reported non-hormonal medication use at 15.2%, compared to 42.9% in the POI group ( p  = 0.023). The distribution of residential settings differed significantly between the groups. The POI group more frequently resided in agricultural areas, with 68.8% living near agricultural areas compared to 34.2% of the control group ( p  = 0.008), including vineyards (21.9% POI vs. 2.4% HC, p  = 0.011) and fruit farms (34.4% POI vs. 13.2% HC, p  = 0.034). The control group predominantly resided in the Osrednjeslovenska urban area, where their frequency was 76.9% higher than the POI group (80.5% HC vs. 45.5% POI, p  = 0.006). Furthermore, the POI group reported significantly higher levels of waste burning in their neighborhoods (4.67-fold increase, p  < 0.001) and spent more time driving daily (median: 1 h/day [0–5] POI vs. 1 h/day [0–2.5] HC, p  = 0.030). After Bonferroni correction, statistical significance was maintained only for differences in alcohol consumption and the occurrence of waste burning in the vicinity of residences. Kendall's tau correlation analyses evaluated the relationships among nine TEs and the statistically significant demographic and biochemical variables. The correlation coefficients and corresponding p-values for the elements in each biofluid are presented in Supplement 4 . The concentrations of TEs measured in this study were generally within the established reference intervals [ 38 ] for spot urine, whole blood, and serum specimens (Supplement 5 ) and were generally consistent with levels reported in women from various environmental studies (Supplement 6 ) [ 46 , 57 – 75 ]. In both groups, most elements showed the highest concentrations in whole blood, except for Cu, Mo, and As, which were highest in serum and urine, respectively (Supplement 7 ). The urinary concentrations (unadjusted and adjusted for creatinine and SG) and detection rates for healthy controls and POI patients are summarized in Supplement 5 . Except for Mn, the TEs detection rates were higher than 80% in both groups. The detection rate for Mn was only 15% (POI group: 7.5%, control group: 22%); therefore, Mn was excluded from further analyses. After adjustment for SG, the median urinary levels of Cu, Zn, Se, Mo, Cd, Hg, and Pb were significantly higher in the control group than in the POI group. When adjusting for age and BMI, the difference in Pb levels was no longer significant, but the differences in Cu, Zn, Se, Mo, Cd, and Hg levels remained significant. However, the median As concentrations did not differ significantly between the two groups in either the unadjusted or adjusted models (Supplement 8 ) Figs. 1 , 2 , and 3 . Fig. 1 Scatter box-and-whisker plots illustrating the distributions of essential TEs Se (Figures a , b , c ) and Mo (Figures d , e , f ) across different biofluids in the POI group and healthy controls (HC), along with their corresponding adjusted p -values Fig. 2 Scatter box-and-whisker plots illustrating the distributions of essential TEs Mn (Figures a , b , c ), Cu (Figures d , e , f ), and Zn (Figures g , h , i ) across different biofluids in the POI group and healthy controls (HC), along with their corresponding adjusted p -values Fig. 3 Scatter box-and-whisker plots illustrating the distributions of non-essential TEs Hg (Figures a , b , c ), As (Figures d , e , f ), Pb (Figures g , h , i ), and Cd (Figures j , k , l ) across different biofluids in the POI group and healthy controls (HC), along with their corresponding adjusted p -values Scatter box-and-whisker plots illustrating the distributions of essential TEs Se (Figures a , b , c ) and Mo (Figures d , e , f ) across different biofluids in the POI group and healthy controls (HC), along with their corresponding adjusted p -values Scatter box-and-whisker plots illustrating the distributions of essential TEs Mn (Figures a , b , c ), Cu (Figures d , e , f ), and Zn (Figures g , h , i ) across different biofluids in the POI group and healthy controls (HC), along with their corresponding adjusted p -values Scatter box-and-whisker plots illustrating the distributions of non-essential TEs Hg (Figures a , b , c ), As (Figures d , e , f ), Pb (Figures g , h , i ), and Cd (Figures j , k , l ) across different biofluids in the POI group and healthy controls (HC), along with their corresponding adjusted p -values The whole blood concentrations and detection rates for healthy controls and POI patients are summarized in Supplement 5 . The detection rate for all elements was 100%. Median blood levels of Cu, Zn, and Pb were significantly higher in the POI group compared to the control group. In contrast, the median blood level of Mo was significantly lower in the POI group. However, these differences were no longer statistically significant after adjusting for age, BMI, and hematocrit (for Zn, Cd, Pb, Hg, Mo). In the bivariate analysis, Mn levels did not differ significantly between the POI and control groups. However, after adjusting for age, BMI, and hematocrit (Ht), Mn levels were significantly higher in the POI group. Subsequent analysis showed that Mn levels remained significant when age was included as a covariate, but this significance was lost when the model included either BMI or Ht individually or when both BMI and Ht were included together in the model (Supplement 8 ). The serum concentrations and detection rates for healthy controls and POI patients are summarized in Supplement 5 . In unadjusted models, the median serum levels of Mn and Cu were significantly higher, while the median serum levels of Cd, Hg, and Pb were significantly lower in the POI group compared to the control group. In adjusted models (accounting for age and BMI), the differences in element levels remained significant except for Mn, which was no longer significant (Supplement 8 ). Kendall's tau correlation analyses were conducted to evaluate the relationships among nine TEs across different fluids, with the results detailed in Supplement 4 . The analysis revealed several notable correlations. In both the overall study cohort and subgroup analyses, strong inter-biofluid correlations were observed between As, Mo, and Hg across different biofluids, with Kendall's tau coefficients ranging from 0.6 to 0.8. Additionally, Cu and Se showed moderate correlations between their serum and whole blood concentrations, with coefficients between 0.4 and 0.6. In urine, within the overall study cohort, a strong correlation was found between Cu and Se, with Kendall's tau coefficient of 0.625 ( p  < 0.01). Moderate correlations were observed between Cu and other elements such as Zn (Ƭ = 0.447, p  < 0.01), Mo ( Ƭ  = 0.453, p  < 0.01), and Cd ( Ƭ  = 0.476, p < 0.01). Selenium also showed moderate correlations with Zn (Ƭ = 0.433, p  < 0.01) and Mo ( Ƭ  = 0.469, p  < 0.01) in urine. In the subgroup correlation analysis, a moderate correlation was observed between urinary Cu and Zn levels ( Ƭ  = 0.497, p  < 0.01), as well as between Cu and Cd levels ( Ƭ  = 0.472, p  < 0.01) in the POI group. In contrast, the control group showed generally weak correlations among urinary TEs ( Ƭ  < 0.4), with the exception of a moderate correlation between Cu and Se ( Ƭ  = 0.436, p  < 0.01). In both subgroup and overall analyses, moderate correlations ( Ƭ  = 0.4—0.6) were also observed between whole blood Hg and As levels across urine, serum, and whole blood. Only weak, statistically significant Kendall's tau correlations ( Ƭ  < 0.4) were observed between the TEs and significant demographic and biochemical parameters (Tables 1 and 2 ). These correlations were inconsistent, appearing in the overall analyses but not consistently across subgroups, as detailed in Supplement 4 . For instance, mineral supplementation was weakly negatively correlated with urinary Cu ( Ƭ  = −0.311, p  < 0.01) and Se ( Ƭ  = −0.330, p  < 0.01) levels in the overall analysis. However, this correlation was insignificant in the control group, which had significantly higher urinary Cu and Se levels and a higher reported intake of mineral supplements. In the POI group, mineral supplementation showed a weak negative correlation with whole blood As ( Ƭ  = −0.350, p  < 0.01), a TE that did not differ significantly between groups. Urinary Hg levels were weakly correlated with agricultural home surroundings in the overall analysis ( Ƭ  = 0.301, p  < 0.01) and in the control group ( Ƭ  = 0.366, p  < 0.01), but this association was not significant in the POI group, which had a higher proportion of individuals from rural, agricultural areas. Medication use was weakly negatively correlated with serum Se in the control group ( Ƭ  = −0.312, p  < 0.01) but was not significant in the overall or POI group analyses. Employment status, though not significant in the overall analysis, was weakly correlated with urinary Cu ( Ƭ  = −0.369, p  < 0.01) and whole blood Cu ( Ƭ  = −0.339, p  < 0.01) in the control group, as well as with urinary As ( Ƭ  = −0.354, p  < 0.01), whole blood As ( Ƭ  = −0.363, p  < 0.01), whole blood Pb ( Ƭ  = 0.338, p  < 0.01), and urinary Mo ( Ƭ  = −0.327, p  < 0.01) in the POI group.

Premature ovarian insufficiency (POI) is a multifaceted medical condition of young women that can significantly affect their overall quality of life. It refers to a cessation of ovarian function that occurs in women under the age of 40 years and may be associated with intermittent resumption of ovarian activity in over 25% of patients [ 1 ]. POI is characterized by hypergonadotropic hypoestrogenic oligo/amenorrhea, which arises from dysfunction or early depletion of the ovarian follicle pool [ 2 – 4 ]. POI is uncommon but not rare, with a generally estimated prevalence between 0.3 and 3.7% in women under 40 years of age [ 3 , 5 ]. According to the European Society of Human Reproduction and Embryology (ESHRE), the diagnostic criteria for POI are primary or secondary oligo/amenorrhea of more than 4 months duration, associated with an increase in serum levels of follicle-stimulating hormone (FSH) (> 25 IU/L), confirmed by a second measurement one month later, in women before the age of 40 [ 6 ]. Women with POI exhibit symptoms similar to those of natural menopause; however, these symptoms are accompanied by an earlier loss of fertility, which can be particularly distressing for young women [ 4 ]. Additionally, POI is associated with well-recognized long-term health sequelae, including osteoporosis, cardiovascular disease, dementia, and reduced life expectancy [ 1 , 5 ]. These health issues may lead to increased stress levels, impairments in mental health, and diminished social functioning, all of which result in a significant decrease in quality of life [ 5 , 7 ]. Etiologies of POI are still poorly understood, with only 25% of cases having an identified cause. Possible causes include genetic factors (related to the X chromosome and autosomal), autoimmune disorders, infectious, metabolic disorders, toxicant-related, and iatrogenic factors (such as chemotherapy, radiation, or surgery) [ 1 , 3 , 8 ]. In the majority of cases, the etiology remains unknown, leading to classification as idiopathic POI [ 9 ] highlighting the need to identify potential pathogenic factors, including environmental influences, that remain to be elucidated. Almost three decades ago, a well-characterized cluster of POI in female workers exposed to environmental pollutant 2-bromopropane illustrated the plausibility of an environmental chemical contribution to this condition [ 3 , 10 ]. Further studies suggested that environmental factors seem to be a significant determinant of ovarian reserve and are therefore suspected of contributing to the onset of POI [ 8 , 11 – 14 ]. Trace elements (TEs) are defined as elements that are present at low concentrations (mg kg −1 or less) in most soils, plants, and living organisms [ 15 ]. Manganese (Mn), copper (Cu), zinc (Zn), selenium (Se), and molybdenum (Mo) are essential TEs, as they are vital for numerous biological processes at normal levels but can exert toxic effects at higher concentrations. Conversely, non-essential TEs arsenic (As), cadmium (Cd), mercury (Hg), and lead (Pb) are not considered to have any essential or beneficial effects on organisms and may be toxic at any concentration [ 16 ]. The ovary contains various essential TEs such as Zn, Cu, Se, Mn and Mo, which are vital for oogenesis, oocyte maturation, ovulation as well as normal oocyte function [ 17 , 18 ]. Their involvement in a broad range of biological processes, particularly cellular metabolism and antioxidant defenses, makes them especially relevant for the development of oocyte competence [ 17 , 19 ]. Conversely, the accumulation of nonessential TEs and alterations in the ovarian content of essential TEs can damage the reserve of ovarian follicles, interfere with folliculogenesis, and disrupt ovarian steroidogenesis [ 16 , 20 ]. An overview of the suggested pathological mechanisms of TEs reproductive toxicity, which may coexist, includes disruption of the endocrine and immune systems and the induction of oxidative stress [ 8 ]. Most studies investigating the impact of TEs exposure on human reproductive health have focused on infertile couples undergoing in vitro fertilization (IVF) cycles [ 16 , 21 – 24 ]. Cumulative epidemiological and experimental data from human and animal studies suggest that TEs can accumulate in the ovaries [ 16 ], exerting detrimental effects on ovarian function and female fertility, even at relatively low exposure levels [ 3 , 8 , 14 , 16 , 20 – 27 ]. In our literature search, we identified nine studies that have analyzed levels of TEs in patients with POI to date, reporting inconclusive results [ 28 – 36 ] (Supplement 1 ). Despite limited and incomplete investigations to date, associations between analyzed TEs and POI have been suggested. Previous studies have primarily focused on assessing Zn, Cu, and Se in a single biological sample. To our knowledge, this is the first study to evaluate differences in the concentrations of nine common TEs (Mn, Cu, Zn, As, Se, Mo, Hg, Cd, Pb) across urine, serum, and whole blood in patients with POI compared to a control group. The overall objective of this study was to investigate the concentrations and distribution of TEs in various biofluids (urine, whole blood, and serum) of women with POI compared to healthy controls (HC), and to explore the potential associations between these TEs and POI.

The development and diagnosis of POI in young women have profound physical and emotional consequences. POI significantly shortens a woman's fertile lifespan and causes distressing vasomotor and genitourinary symptoms associated with menopause. Additionally, it has widespread effects on general health, psychological and sexual well-being, and long-term bone, cardiovascular, and cognitive health [ 1 ]. Therefore, ovarian health should be viewed as an early indicator of premature somatic aging and reduced longevity, analogous to the 'canary in the coal mine' as a predictor of broader health issues [ 159 ]. Although the environmental etiology of POI is well established, the precise mechanisms involved are still not fully understood. Our study is the first to comprehensively analyze multiple essential and non-essential TEs across urine, whole blood, and serum in POI patients compared to healthy controls. These findings provide a foundation for future research into the mechanisms of POI and potential strategies for its management. Larger longitudinal studies are needed to confirm these results and investigate the therapeutic potential of TEs modulation in POI.

The results of this study suggest that levels of certain TEs differ between POI patients and healthy controls. POI patients had significantly lower urinary levels of Cu, Zn, Se, Mo, Cd, Hg, and Pb compared to controls. In whole blood, Mn levels were significantly higher in POI patients after adjustment for age, hematocrit, and BMI. Serum Cu levels were significantly higher in POI patients, while Pb, Cd, and Hg levels were consistently lower. Only As levels showed no significant differences in any biological samples. The TEs concentrations measured in our study align with those reported in women from various environmental health studies [ 46 , 57 – 72 ], indicating a degree of consistency across different research settings. However, comprehensive comparisons of all analyzed elements in POI patients are limited. We identified nine studies comparing TE levels in POI patients to healthy controls [ 28 – 31 , 35 , 36 ] (Supplement 1 ). Our findings reveal both similarities and differences with these studies, with some TEs showing consistent patterns and others varying significantly. This variability highlights the complexity of evaluating TEs in POI and suggests influences such as geographical differences, dietary habits, and methodological variations. Acknowledging the potential for exposure misclassification, we collected data on demographic, lifestyle, and dietary factors that may help to explain differences between the POI and control groups. However, our analyses showed only weak and inconsistent correlations between significant questionnaire variables (Table  1 , Table  2 ) and measured TEs levels (Supplement 4 ). Additionally, kidney and liver function did not differ significantly between groups, excluding organ function as a confounding factor for TEs levels. Our analysis of essential TEs showed the most notable differences in urinary concentrations, with Cu, Zn, Se, and Mo all being significantly lower in the POI group compared to healthy controls. Particularly, Cu emerged as a key element displaying significant differences in both urine and serum. Furthermore, Cu exhibited correlations with other TEs (Se, Zn, Mo, Cd) in urine, highlighting its central role in TEs metabolism and its potential involvement in the pathophysiology of POI. Three previous studies have investigated Cu status in POI patients. Consistent with our findings, Kebapcilar et al. (2013) [ 28 ] reported significantly higher Cu levels in the serum of POI patients. Conversely, Verma et al. 2018 [ 30 ] and Li et al. (2021) [ 34 ] reported significantly lower serum and higher urinary levels of Cu in POI patients, respectively. Studies on women undergoing in vitro fertilization highlight the essential role of Cu in ovarian function [ 17 , 76 , 77 ]. Disruptions in Cu homeostasis, as shown in animal studies, may accelerate ovarian aging through DNA damage, oxidative stress, and impaired steroidogenesis [ 18 , 78 – 80 ], reducing follicle numbers across all developmental stages [ 79 ] hence affecting ovarian reserve. Studies indicate that women generally have higher blood Cu levels than men [ 81 – 83 ], which was primarily attributed to the influence of estradiol (E2) [ 84 ]. In healthy, non-pregnant women not using hormonal therapies, serum Cu shows an inverse relationship with E2 levels [ 84 , 85 ]. Furthermore, Ferdous et al. (2019) reported higher serum Cu levels in postmenopausal women compared to premenopausal women [ 86 ]. Cellular studies suggest that E2 increases Cu uptake by upregulating Cu transporters [ 87 , 88 ], meeting higher metabolic demands during the reproductive cycle [ 87 , 89 ]. This may lead to increased cellular Cu uptake, reduced serum Cu, and elevated urinary Cu excretion, explaining the lower serum and higher urinary Cu in controls compared to the opposite pattern in hypoestrogenic POI patients. Beyond Cu, our analysis revealed that three additional essential TEs; Zn, Se, and Mo were also significantly lower in the urine of POI patients compared to controls, suggesting a broader pattern of altered TE metabolism and excretion in POI. Urinary Zn levels were lower in the POI group compared to controls. However, Li et al. (2021) found no significant difference in urinary Zn levels but reported a negative correlation between urinary Zn and serum FSH in POI patients [ 34 ]. In contrast, two other studies reported lower serum Zn levels in POI patients compared to healthy controls [ 28 , 30 ]. Additionally, a large epidemiological study found significantly lower urinary Zn levels in postmenopausal women [ 90 ], suggesting altered Zn homeostasis under hypoestrogenic conditions. This finding supports evidence linking hormonal regulation to Zn homeostasis via Zn-specific transporter gene expression [ 91 , 92 ]. Zn homeostasis is crucial in the female reproductive system, influencing oocyte maturation, quality, and functionality [ 93 , 94 ]. Reduced Zn levels in POI patients have been associated with increased oxidative stress [ 30 ]. Since balanced levels of reactive oxygen species (ROS) and reactive nitrogen species (RNS) are essential for follicular growth, angiogenesis, and sex hormone synthesis in ovarian tissue, an imbalance between Zn and ROS may contribute to the decline in oocyte quantity and quality [ 95 ]. However, further research is needed to elucidate the mechanisms underlying Zn metabolism in POI. POI patients had significantly lower urinary Se levels compared to controls, contrasting with Li et al. (2021), who reported higher urinary Se levels in POI patients [ 45 ] (Supplement 1 ). Two previous studies reported decreased serum Se levels in POI patients [ 30 , 31 ]. Se is essential for follicle development, steroid hormone synthesis, and embryonic development [ 18 ] and is found in granulosa cells of healthy follicles but is absent in atretic ones [ 22 ]. The lower urinary Se levels in POI patients may indicate increased Se utilization to counteract ROS associated with POI due to its antioxidant properties [ 96 ]. However, data on ceruloplasmin (Cu-binding glycoprotein) and Cu, Zn superoxide dismutase (SOD) activity are needed for a relevant conclusion. Urinary Mo levels were significantly lower in POI patients compared to controls, contrasting with a previous study reporting higher serum Mo levels in POI patients linked to markers of vascular endothelial injury [ 35 ]. Mo is conserved at low intakes or higher metabolic demands but excreted rapidly at high levels. The effects of Mo on human reproductive health have been primarily studied in relation to male infertility, while its impact on female reproduction remains unclear [ 17 ]. Studies in women undergoing IVF have shown conflicting results: one found a positive correlation with oocyte retrieval [ 76 ], while another linked elevated urinary Mo to reduced implantation and pregnancy rates [ 17 ]. Animal studies suggest Mo may affect oocyte quality by modulating oxidative stress, vascular permeability, and androgen and estrogen receptor activity in a dose-dependent manner [ 35 , 97 – 99 ]. Whole blood Mn levels did not significantly differ between the POI and control group in unadjusted analyses; however, after adjusting for age, BMI, and hematocrit, levels were significantly higher in the POI group. This significance was maintained with age as a covariate but lost when only BMI and hematocrit were included, suggesting age moderates the relationship between Mn levels and POI. Previous study found no link between urinary Mn levels and POI risk [ 34 ], and population studies show Mn levels stabilize in adulthood, with no direct age-related variability [ 100 – 103 ]. However, premenopausal women have higher Mn levels than postmenopausal women [ 102 , 104 , 105 ], possibly due to estrogen's role in upregulating intestinal Mn absorption [ 102 , 106 , 107 ]. In the ovary, Mn is essential for steroidogenesis, oocyte maturation, and corpus luteum function [ 18 , 108 ], but both excess and deficiency have been linked to ovarian aging and impaired oocyte development in animal studies [ 18 , 108 – 111 ]. Further research is needed to clarify Mn's role in hypoestrogenic conditions like POI and natural menopause [ 102 ] The levels of nonessential TEs Cd, Hg, and Pb were significantly lower in the serum and urine of patients with POI compared to controls. Three previous studies have examined nonessential TEs in POI patients, with one study reporting no significant changes in serum levels of As, Hg, Cd, and Pb [ 29 ]. In contrast, Pan et al. (2020, 2021) found significantly higher urinary As and Cd levels in POI patients compared to healthy women [ 32 , 33 ]. As summarized in Supplement 1 , exposure levels in POI patients varied considerably between studies. For example, Pan et al. reported median As levels 4.7 times higher and geometric mean (GM) Cd levels 2.3 times higher in POI patients than those in our study. Although our study also found lower As and Cd levels in healthy controls, the difference was less pronounced. Overall, the levels of non-essential TEs observed in both POI patients and controls in our study were relatively low, suggesting a low level of exposure. In environmental toxicology, it's generally accepted that the likelihood of an agent causing harm increases with higher doses or exposure levels, including for non-essential TEs [ 65 ]. While high doses of non-essential TEs are known to be harmful, low-dose exposure poses a subtle but significant threat [ 112 , 113 ]. Animal and in vitro studies indicate that the toxic effects of Cd, Pb, and Hg on the ovaries are remarkably similar when each is administered individually. These elements can accumulate in the ovaries [ 23 , 114 – 116 ], causing decreased follicular growth, an increased number of atretic follicles, degeneration of the corpus luteum, and prolonged and/or irregular cycles [ 115 , 117 – 127 ]. These effects are linked to alterations in hormone levels, specifically FSH, LH, E2, P, AMH, and oxidative stress indices, such as increased levels of malondialdehyde (MDA) and decreased levels of antioxidants [ 128 – 134 ]. Even though detecting significant interactions or adverse effects on reproductive health may be challenging at low exposure levels, animal studies have demonstrated that even low doses of Cd, Pb, and Hg can accumulate in the ovaries [ 116 , 122 , 126 , 134 – 140 ] leading to increased follicular apoptosis [ 116 , 122 , 135 ], the presence of atretic oocytes at various developmental stages [ 140 ] and disrupted steroidogenesis [ 116 ]. Low Cd and Pb exposure in healthy premenopausal women was linked to hormonal variations, with Pb increasing mean FSH levels while reducing its amplitude, and both Cd and Pb correlating with higher mean E2 levels [ 141 ], patterns suggesting early ovulatory resistance, resembling perimenopausal transitions where elevated FSH and E2 levels are needed for ovulation [ 141 , 142 ]. These findings, combined with the high sensitivity of the applied methodology and the low rate of undetectable TEs values, suggest that the differences observed between the POI group and healthy controls are important, even at low exposure levels. TEs concentrations in blood, serum, and urine primarily reflect exposure levels and may not accurately represent TEs accumulation in specific organs like the ovaries [ 143 ]. For example, Cd has been shown to accumulate in ovarian tissue over time, with concentrations increasing in women aged 30 to 65 years [ 144 ]. Higher TEs levels in healthy controls found in the present study may, besides differing exposure patterns, reflect lower cellular accumulation with more efficient detoxification and excretion processes. Further research is needed to elucidate the effects of low-level exposures and their potential interactions in the context of POI. TE accumulation in the body is influenced not only by dietary intake and environmental exposure but also by genetic variations [ 143 ]. Genome-wide association studies have linked genetic variants in metal transporter genes to levels of Cu, Se, Zn, Mn, As, Cd, Pb, and Hg [ 143 , 145 – 149 ]. Such genetic differences may increase susceptibility to TE accumulation and toxicity, affecting excretion or retention in hair, blood, erythrocytes, or urine [ 112 ]. However, little information is available on gene-TEs interactions in POI patients. A recent study by Mirinzehad et al. (2024) investigated the association between some POI-related genotypes and serum levels of Cu, Zn, calcium, phosphate, magnesium and vitamin D. The study reported that the C and G alleles of the rs4806660 polymorphism in the TMEM150B 2 gene and the rs244715 polymorphism in the ZNF346 3 gene, respectively, are independently associated with serum Cu levels in women with POI. In POI patients, the C allele of rs4806660 was linked to lower Cu levels compared to the TT genotype, while the GG genotype of rs244715 was associated with higher Cu levels than the AA genotype. Both TMEM150B and ZNF346 encode proteins potentially involved in cell growth and survival [ 36 ]. According to the epidemiological studies carried out so far, a genetic predisposition to altered TEs homeostasis or vulnerability to non-essential TEs may help explain varying health impacts from environmental TEs exposure. However, further research is needed to clarify the genetic basis of susceptibility to TEs toxicity in POI pathogenesis [ 112 ]. Our study has several limitations that require cautious interpretation. The primary limitation is the small sample size, which reduces statistical power. However, participants were rigorously selected according to strict, internationally recognized criteria for POI [ 6 ], with the exclusion of individuals with known causes of POI, such as those with abnormal karyotypes or Fragile X syndrome/ FMR1 4 gene premutation. Relevant biological samples were used for specific TEs. In clinical practice, the most commonly used biological specimens to assess exposure to non-essential TEs and the nutritional status of essential TEs are serum for Cu [ 57 ], Se [ 150 ], Zn [ 151 ], and Mo [ 152 ]; whole blood for Pb [ 153 ], Mn [ 154 ], and organic Hg species [ 155 ]; and urine for As [ 156 ], Cd [ 157 ], and inorganic Hg species [ 158 ]. This study thus comprehensively evaluates TE exposure and nutritional status by including all major biofluids. The cross-sectional design limits our ability to establish temporal relationships or causality between the analyzed elements and POI. Additionally, reliance on single time-point TE measurements may misclassify associations, as this approach does not capture relevant exposure windows, particularly since our study included patients with established ovarian failure. For a more accurate assessment of cumulative TEs exposure, alternative matrices like hair or nails, which better reflect long-term exposure, may be preferable. Finally, although we controlled for selected known confounders, residual confounding cannot be entirely ruled out due to the observational nature of the study. Addressing critical factors such as exposure timing, dosage, and susceptibility in relation to POI requires longitudinal studies. Such studies, with larger sample sizes and diverse populations, are necessary to clarify the role of TEs in POI and establish the directionality of these associations.

Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 29 KB) Supplementary file2 (DOCX 26 KB) Supplementary file3 (DOCX 34 KB) Supplementary file4 (DOCX 71 KB) Supplementary file5 (DOCX 25 KB) Supplementary file6 (DOCX 30 KB) Supplementary file7 (DOCX 40 KB) Supplementary file8 (DOCX 147 KB) Supplementary file9 (DOCX 978 KB) Supplementary file1 (DOCX 29 KB) Supplementary file2 (DOCX 26 KB) Supplementary file3 (DOCX 34 KB) Supplementary file4 (DOCX 71 KB) Supplementary file5 (DOCX 25 KB) Supplementary file6 (DOCX 30 KB) Supplementary file7 (DOCX 40 KB) Supplementary file8 (DOCX 147 KB) Supplementary file9 (DOCX 978 KB)