Blood microRNAs linked to endometriosis in an Indian population: detailed analysis and confirmation of markers

A comprehensive literature search encompassing publications up to December 16th, 2022, was conducted across electronic databases and the scientific literature platforms, including PubMed and ArrayExpress. The search utilized keywords such as 'biomarker', 'endometriosis', 'microRNA', 'noninvasive diagnosis', 'miRNA', and 'qRT-PCR', both individually and in various combinations. For each study included in the analysis, essential details were accurately extracted and concisely summarized. These details encompassed microRNA expression variances, sample origins, sample sizes, quantification methods employed, and the validation techniques utilized.

Two comprehensive compilations were generated: one included all identified microRNAs from the included studies, while the other comprised only the microRNAs subjected to validation. These inventories offer a comprehensive perspective on the microRNAs examined within the domain of endometriosis. Composite scores were calculated based on the frequency of each miRNA's mention across multiple studies. For each study, a plus sign (+) indicates significantly upregulated miRNAs, a minus sign (−) denotes downregulated miRNAs, and a hash sign (#) represents miRNAs with no significant change. The number preceding each sign indicates the number of studies reporting that observation (Supplementary File S2 (see section ongiven at the end of the article)). Separate cumulative scores were calculated for upregulated, downregulated, and unchanged miRNAs within both the screening and validation study categories, based on the number of studies in which each miRNA was consistently reported. miRNAs consistently reported with the same direction of expression (either up or down) were prioritized. To assess directional consistency, we identified overlaps among replicated differentially expressed miRNAs and flagged those with contradictory expression patterns across studies. miRNAs showing directional conflicts but with high absolute scores (e.g., multiple +1s, −1s, or #1s) were further examined to compare trends between screening and validation datasets. Greater weight was assigned to miRNAs with a consistent expression direction across both study types. In cases of conflicting results, particular emphasis was placed on validated miRNAs demonstrating stable expression trends across diverse populations. Supplementary materials

A higher score denoted heightened representation of a specific microRNA across studies. Employing these composite scores, a conclusive compilation of the foremost differentially regulated microRNAs was precisely curated. This method facilitated a comprehensive investigation directed toward identifying promising microRNAs as potential markers for endometriosis. The integration of validated miRNAs displaying consistent expression patterns across multiple studies aimed to enhance the reproducibility and applicability of the findings.

Initially, participant selection was based on a comprehensive range of criteria, including indications such as suspected endometriosis, pelvic pain with undetermined origin, the presence of adnexal cysts, or infertility issues. Notably, individuals in the postmenopausal phase and pregnant subjects were excluded. Furthermore, participants with a medical history encompassing cancer of any type, hysteromyoma, adenomyosis, endometrial synechiae, polycystic ovary syndrome, hydrosalpinx, or any hormonal or malignant disorders were ineligible for the study. To prevent potential therapy-related effects on miRNA secretion, participants who used any form of hormonal therapy in the last 3 months were excluded from this study. The overall recruitment, screening, and final selection of participants for both patient and control groups are outlined in the study flowchart (Fig. 1).

This schematic representation illustrates the stepwise application of inclusion and exclusion criteria, initial enrollment, diagnostic confirmation by imaging and laparoscopy, and the eventual classification into study and control cohorts. Based on the inclusion and exclusion criteria (Fig. 1), an initial cohort of 30 patients and 15 healthy controls was recruited. Blood samples were collected from all participants during the mid-secretory phase of the menstrual cycle (days 18–24). Plasma samples were assessed for CA-125, estradiol, and progesterone to evaluate the hormonal milieu for all participants. Following clinical evaluation and diagnostic workup – including MRI (performed in five patients before laparoscopy) and laparoscopic visualization with histological confirmation – 12 patients with confirmed endometriosis were included in the final analysis. This group also comprised three individuals with recurrent disease, each of whom had undergone laparoscopic surgery at least 4 years earlier. In addition, eleven age-matched controls were identified, based on specific inclusion criteria: fertile females who had given birth within the last 2 years and who were not pregnant at the time of sample collection. Laparoscopy was considered inappropriate for ethical reasons, and participants needing such procedures for gynecological conditions were excluded to ensure healthy controls. The exclusion criteria for controls were that they lacked any gynecological disorders, were not pregnant or breastfeeding, and had no history of cancer, hysteromyoma, adenomyosis, endometrial synechiae, PCOS, hormonal diseases, or substance abuse. In addition, individuals with chronic medical conditions, psychiatric disorders, or autoimmune disorders, who were receiving hormonal therapy, or who were unwilling to provide consent, were excluded. The control group included participants aged 22–37 years, while the study group included participants aged 24–39 years, and the study adhered to a stringent selection process to ensure the clinical and demographic relevance of the participants.

Five milliliters of whole blood was collected from each participant in EDTA tubes before the surgical intervention. The collected blood specimens were promptly transported to the laboratory within thirty minutes for subsequent processing. The plasma was isolated from the whole blood after centrifugation at 1,900 g for ten minutes at 4°C. Hemolysis in plasma samples was visually identified by observing the color of the supernatant plasma. Furthermore, a quantitative assessment was conducted by measuring the absorbance at 415 nm using a spectrophotometer (Thermo Scientific Nanodrop), and samples with values less than 0.2% were selected for RNA extraction. The separated plasma was then divided into aliquots, each containing 200 microliters, and preserved at a temperature of −80°C for storage.

Two hundred microliters of plasma was subjected to total RNA extraction using the QIAzol Lysis Reagent, following the guidelines provided by the manufacturer (Qiagen, USA). The RNA purity was evaluated by measuring the A260/A280 ratio with a Nanodrop 2000. Moreover, a spiked-in synthetic template miRNA (cel-miR-39) from Qiagen was introduced alongside QIAzol. This addition served to evaluate the efficacy of the RNA isolation, cDNA synthesis, and PCR amplification processes. High-quality RNA samples were used to synthesize complementary DNA (cDNA) using the miRCURY® LNA® miRNA SYBR® Green PCR Kit (Qiagen). The relative expression of miRNAs was determined by SYBR Green Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR) using the miRCURY® LNA® RT Kit (Qiagen). Cel-miR-39 and UniSP6 were used for the extraction and template control processes, respectively. miR-16-5p was used as an endogenous control to determine relative miRNA expression levels. The relative mRNA level was calculated using the formula 2^(−ΔCT).

The data were analyzed using MS Excel and GraphPad Prism v. 8.01. Student's t-test was used to compare clinical characteristics between the endometriosis and control groups, as the data were normally distributed and are expressed as the mean ± standard deviation (SD). The Mann–Whitney U test was used to compare miRNA expression levels, and the results are presented as medians and interquartile ranges (IQR). Post hoc power analysis was performed using the online calculator available at ClinCalc (https://clincalc.com/stats/power.aspx↗, accessed on July 4, 2025), based on observed effect sizes, sample sizes, and a significance level of 0.05. ROC analysis was performed to assess the diagnostic value of each miRNA biomarker, and the area under the ROC curve (AUC) was calculated. A P-value of <0.05 was considered to indicate statistical significance. Furthermore, to account for multiple testing, the false discovery rate (FDR) was controlled using the Benjamini–Hochberg procedure (q = 0.05). Adjusted P-values were calculated with the online tool available at https://tools.carbocation.com/FDR↗. Binary logistic regression was performed using SPSS (Norusis 2008) to assess the predictive value of miR-20a-5p and miR-451a for distinguishing endometriosis cases from controls. The dependent variable was binary (group: 0 = control, 1 = endometriosis), and both miRNAs were entered simultaneously as continuous independent variables using the Enter method. Predicted probabilities were saved and used to generate the ROC curve. Model significance was evaluated using the Omnibus Test of Model Coefficients (χ² test), and classification performance was assessed by sensitivity, specificity, and area under the ROC curve (AUC).

A total of 45 articles discussing circulating miRNAs in blood (plasma/serum) were identified through electronic databases and the scientific literature searches. After a thorough assessment, 23 articles were excluded. These exclusions were due to duplicate titles, secondary research articles, and miRNA studies related to other malignancies in addition to endometriosis. A total of 22 studies that met our inclusion criteria were identified. These investigations, primarily originating from China (n = 6), the USA (n = 5), Iran (n = 3), and other European and Asian countries, were conducted with an almost equal distribution of plasma and serum samples (Table S1). However, it is noteworthy that India, despite having a substantial population and a high prevalence of endometriosis, was not represented in the selected studies.

In the analysis of 22 studies, 102 upregulated miRNAs, 197 downregulated miRNAs, and 118 miRNAs with unaltered or undefined status (NA) were identified (Table S2). Remarkably, only 13 miRNAs were common to both the upregulated and downregulated groups (Fig. 2A). This suggests that most of these miRNAs either consistently trend across multiple studies or are unique to specific studies. The application of the reproducibility criterion, which necessitated confirmation in at least two studies, revealed that most of the identified miRNAs lacked this confirmation and were identified in single studies. Consequently, confirmation was achieved for only 6 upregulated miRNAs, 22 downregulated miRNAs, and 5 miRNAs with unaltered or undefined status (Fig. 2B). This implies that only 28 out of the 286 (10%) significantly differentially expressed miRNAs were detected in at least two studies (Fig. 2A and B). Similarly, in the analysis of miRNAs validated in various laboratories worldwide, only four upregulated and five downregulated miRNAs were identified as being replicated by more than one laboratory, accounting for 18% (9/49) of the miRNAs in the replication category (Table S3, Fig. 2C and D). A combined list of both identified and validated miRNAs has been prepared that replicated in at least two studies (Table S4). Subsequently, we evaluated the expression direction in both the identified and validated (replicated) miRNA groups. A combined score was then computed, with preference given to validated miRNAs in cases where two miRNAs received the same score. This process resulted in the selection of nine miRNAs from the initial pool, making them promising candidates for future investigations in endometriosis (Table 1).

Based on the participant selection criteria outlined above, a total of 23 individuals were included in the final analysis: 12 participants were diagnosed with endometriosis through laparoscopic evaluation, while 11 were classified as healthy controls. Comprehensive diagnostic, clinical, and relevant subject characteristics are detailed in the supplemental file (Table S5).

No statistically significant difference in age between the study groups was observed (P = 0.1663). Among the women in the endometriosis group, seven reported experiencing chronic pelvic pain, three exhibited symptoms of both pelvic pain and infertility, and two described experiencing painful and burning micturition. Furthermore, among the patients who presented with clinical symptoms, nine had been experiencing them for a period ranging from 3 to 7 years, while three had been symptomatic for 2 years. Two patients with a 7-year history of recurrent endometriosis were included. Samples were collected during the presumed luteal phase based on patient-reported menstrual history; however, precise phase determination was limited in some participants due to irregular menstrual cycles. Plasma progesterone and estrogen levels were assessed to approximate luteal phase status. Conversely, all the healthy control subjects included in the study displayed regular menstrual cycles devoid of any discernible gynecological pathology. The mean serum CA-125 level was markedly elevated in the endometriosis group (51.9 ± 9.5 U/mL) compared to controls (16.0 ± 8.0 U/mL), showing a highly significant difference (P < 0.0001). However, serum estradiol (E2) and progesterone levels measured during the presumed mid-secretory phase were comparable between groups. Estradiol averaged 139.7 ± 36.9 pg/mL in the endometriosis group versus 144.1 ± 34.1 pg/mL in controls (P = 0.7689), while progesterone levels were 10.3 ± 2.9 and 10.6 ± 3.2 ng/mL, respectively (P = 0.8557) (Table S5). These findings suggest that while CA-125 is a reliable serum marker to distinguish endometriosis from asymptomatic individuals, circulating levels of estradiol and progesterone may not differ significantly in the mid-luteal phase between groups. In addition, none of the healthy controls had undergone any surgical procedures other than cesarean delivery (4 out of 11 control participants), and these deliveries had occurred at least 2–3 years before sample collection.

In this case-control study, patients with well-characterized advanced-stage endometriosis and matched controls were included. The relative expression levels of nine miRNAs (miR-451a, let-7b, miR-150-5p, miR-17-5p, miR-3613-5p, miR-20a-5p, miR-342-3p, miR-125b-5p, and miR-21-5p) were quantified in plasma samples via quantitative real-time polymerase chain reaction (qRT-PCR). Patients with endometriosis showed lower expression levels of miR-451a and miR-20a-5p than the control group. Post hoc power was estimated at 99.4% for miR-451a (after one outlier adjustment in the control group) and 82.2% for miR-20a-5p. No statistically significant differences in expression were detected for the remaining miRNAs (Fig. 3).

Receiver operating characteristic (ROC) curve analysis was performed to evaluate the discriminatory power of plasma miRNAs that were differentially expressed between endometriosis patients and controls. Our study revealed that miR-451a and miR-20a-5p exhibited a promising diagnostic performance, with an area under the curve (AUC) value of 0.8939 (95% confidence interval (CI) = 0.7508 to 1.000) and 0.8106 (95% confidence interval (CI) = 0.6164–1.000), respectively (Fig. 4). The sensitivity and specificity values of miR-451a (83.33 and 90.91% at a cutoff <2.790) and miR-20a-5p (91.67 and 72.73% at a cutoff <0.5766) collectively indicate their potential as diagnostic biomarkers for endometriosis in the Indian population (Table 2). Among the nine candidate miRNAs tested, miR-451a (P = 0.0007) and miR-20a-5p (P = 0.0106) showed significantly lower expression in endometriosis patients compared to controls. After correction for multiple testing using the Benjamini–Hochberg procedure (q = 0.05), both miR-451a (P_adj = 0.0063) and miR-20a-5p (P_adj = 0.0477) remained statistically significant, confirming their potential as circulating biomarkers (Table 2).

Furthermore, a binary logistic regression was conducted to evaluate whether circulating levels of miR-20a-5p and miR-451a were predictive of endometriosis status. The model was statistically significant, χ²(2) = 20.54, P < 0.001, indicating that the two miRNAs collectively distinguished between patients and controls. The model explained 78.8% of the variance in disease status (Nagelkerke R² = 0.788) and correctly classified 87.0% of cases (81.8% controls, 91.7% patients) (Fig. S2). The Hosmer–Lemeshow test was non-significant (P = 0.995), suggesting good model fit.