Hysteroscopy and CD138+ Plasma Cell Density in Chronic Endometritis

Overview

The goal of this observational study is to learn if a hysteroscopy exam alone can find chronic endometritis, a long-lasting inflammation of the lining of the uterus. A hysteroscopy is an exam that looks inside the uterus with a thin camera. Chronic endometritis is usually confirmed with a lab test that counts CD138-positive cells, a type of immune cell, in a small tissue sample from the uterus. About 100 women will take part. All are already having a hysteroscopy and a tissue sample as part of their fertility care, often after repeated failed embryo transfers or repeated miscarriage. The main question it aims to answer is: Do the visual signs seen during hysteroscopy relate closely enough to the CD138 cell count to reliably find chronic endometritis without a lab test? Participants will: Have a hysteroscopy exam and a tissue sample from the uterus, both already part of their routine fertility care Have their hysteroscopy video reviewed by two doctors who do not know the tissue sample results Have their tissue sample checked in the lab for CD138 cells

Chronic endometritis (CE) is a persistent inflammation of the endometrial mucosa associated with recurrent implantation failure (RIF) and recurrent pregnancy loss (RPL). In clinical practice, CE is assessed using two complementary methods: office hysteroscopy, which identifies visual mucosal signs according to the Cicinelli criteria (micropolyps, focal hyperemia, diffuse hyperemia, stromal edema, strawberry pattern, and hemorrhagic spots), and immunohistochemical detection of CD138-positive (syndecan-1) endometrial plasma cells, considered the histopathological reference standard. Published evidence comparing these two methods is inconsistent: individual studies, mostly retrospective, report weak agreement between hysteroscopic findings and plasma cell counts, while pooled diagnostic accuracy estimates suggest good hysteroscopic performance. This discrepancy is plausibly related to heterogeneous and non-comparable plasma cell count thresholds across studies, the use of dichotomized analyses that discard information, and retrospective designs prone to spectrum and verification bias. No prior study has characterized the relationship between hysteroscopic appearance and plasma cell burden as a continuous phenomenon. This is a prospective, cross-sectional, single-center, non-pharmacological, non-device, no-profit observational study correlating an index test (hysteroscopy) with a reference standard (CD138 immunohistochemistry). Both tests are performed in the same session, so the interval between index test and reference standard is zero. The proliferative phase is standardized and confirmed histologically. Hysteroscopic and histopathological assessments are performed with reciprocal blinding, and there is no verification bias, as all participants undergo both the index test and the reference standard regardless of the hysteroscopic result. Objectives and endpoints: The primary objective is to determine whether the number of hysteroscopic signs of CE correlates with the count of CD138-positive endometrial plasma cells, specifically testing whether this correlation is not clinically relevant (equivalence relative to a pre-specified negligible-effect threshold). The primary endpoint is the Spearman correlation coefficient between the CD138-positive plasma cell count (continuous variable) and the number of positive hysteroscopic signs (0-6), evaluated as an equivalence hypothesis against a pre-specified smallest effect size of interest (SESOI) of \|r\| \< 0.25, assessed with two one-sided tests (TOST) and complemented by a Bayes Factor quantifying the evidence in favor of the null hypothesis, together with the posterior probability that \|r\| is below the SESOI. Secondary objectives are to characterize the association of each individual hysteroscopic sign with plasma cell burden; to quantify inter- and intra-observer reproducibility of hysteroscopic scoring and plasma cell counting; to describe the relationship between the indication for examination (RPL, RIF, or other) and both plasma cell burden and hysteroscopic findings; and, descriptively, to report participants' reproductive outcomes. Corresponding secondary endpoints include: the association of each hysteroscopic sign with plasma cell count; inter- and intra-observer agreement (kappa for hysteroscopy, intraclass correlation coefficient for plasma cell counting); diagnostic accuracy of hysteroscopy at conventional CD138 thresholds (≥1 and ≥5 plasma cells) as a secondary sensitivity analysis; the association between clinical indication and findings; and, descriptively, the rate of histological resolution after treatment (test-of-cure) and reproductive outcomes. Population: Eligible participants are women of reproductive age undergoing office hysteroscopy with endometrial biopsy as part of their clinical work-up - including assisted reproduction (PMA) pathways - for any indication (RIF, RPL, or other), with the examination performed during the proliferative phase of a spontaneous cycle not exposed to exogenous progestins, and who provide written informed consent to participation and to data processing. Procedures: Office hysteroscopy is performed vaginoscopically with saline distension, without anesthesia, during the early-to-mid proliferative phase of a spontaneous cycle. The endometrium is systematically inspected and scored for the six Cicinelli criteria, each recorded as present or absent, together with an overall hysteroscopic judgment of CE according to a pre-specified rule. Each procedure is fully video-recorded. Recordings are reviewed offline, independently and in random order, by two experienced hysteroscopists blinded to histology, clinical data, and each other's assessment; a random subsample is re-reviewed by the same observers after a pre-defined interval to estimate intra-observer agreement, with discordances resolved by consensus. In the same session, an endometrial biopsy is obtained. Samples are formalin-fixed, paraffin-embedded, stained with hematoxylin-eosin, and processed for CD138 (syndecan-1) immunohistochemistry with positive and negative controls. CD138-positive stromal plasma cells, identified by unequivocal membranous positivity, are quantified by a pathologist blinded to hysteroscopic findings and all clinical data. Cycle phase is confirmed histologically by endometrial dating; in case of discordance between histological dating and cycle day, the sample is classified according to histological dating. Patients with a histological diagnosis of CE receive first-line doxycycline 100 mg once daily for 15 days per standard clinical practice, followed by a test-of-cure endometrial biopsy after the next menstruation. Persistent CE is treated with second-line metronidazole and clindamycin, followed by a further biopsy. Cure is defined as a post-treatment plasma cell count below the primary diagnostic threshold. These procedures are part of routine clinical care and are not imposed by the study, which only observes and records them. Statistical analysis: The primary endpoint (Spearman correlation) is analyzed under an equivalence hypothesis (SESOI \|r\| \< 0.25) using TOST, complemented by a Bayes Factor. Because a null correlation could reflect measurement error rather than true independence, reproducibility is quantified (Cohen's/Fleiss' kappa for hysteroscopic scoring; intraclass correlation coefficient for plasma cell counting), and an attenuation-corrected correlation is reported; interpretation of an equivalence result is conditional on adequate measurement reliability. Each individual hysteroscopic sign is compared with plasma cell count using the Mann-Whitney test. Plasma cell count is modeled as an overdispersed count variable via negative binomial regression, with hysteroscopic signs as predictors and age, clinical indication, and cycle day as covariates. Analyses stratified by indication, and any sign-by-indication interaction, are pre-specified as exploratory. Diagnostic accuracy of hysteroscopy at CD138 thresholds of ≥1 and ≥5 plasma cells (sensitivity, specificity, predictive values) is reported as a secondary sensitivity analysis. The significance level is set at 0.05 (two-sided). The detailed statistical analysis plan is time-stamped and deposited on the Open Science Framework before the primary analysis is performed. Sample size: Approximately 100 participants will be enrolled, sized to provide adequate precision for the primary equivalence analysis with SESOI \|r\| \< 0.25 (power approximately 80%). With this sample size, an observed correlation near zero yields a 95% confidence interval excluding correlations stronger than approximately 0.20 in either direction. Analyses of individual low-prevalence signs and of the indication interaction are pre-specified as exploratory and hypothesis-generating. Approximately 20 patients already evaluated before study initiation are treated as a separate pilot cohort and are not included in the prospectively registered case series. Ethics and data management: The study is conducted in accordance with the Declaration of Helsinki and applicable good clinical practice standards, and initiation is contingent on a favorable opinion from the competent Ethics Committee. All participants provide written informed consent to participation and, separately, to personal data processing under Regulation (EU) 2016/679. The study does not add invasive procedures beyond routine clinical care: hysteroscopy, biopsy, CD138 immunohistochemistry, and post-treatment control biopsies are part of standard practice; the study adds only standardized scoring, blinded re-reading of video recordings, and systematic data collection, with no additional burden for participants or the health service. Data are collected in pseudonymized form via a study identification code; the code-identity correspondence table is stored separately and securely, accessible only to authorized personnel. Registration and reporting: The study is prospectively registered on ClinicalTrials.gov before enrollment begins. The statistical analysis plan is time-stamped and deposited on the Open Science Framework. Reporting will follow the STARD and STROBE guidelines. Results will be submitted for publication in a peer-reviewed journal regardless of outcome direction. The study is non-profit and entirely funded by Centro A.M.B.R.A., with no external sponsors; investigators declare no relevant conflicts of interest.