CHRONIC ENDOMETRITIS: WHAT IS THE PREVALENCE AMONG HEALTHY CONTROLS?

Abstract

To evaluate the prevalence of chronic endometritis (CE) in women with unexplained recurrent pregnancy loss (RPL) and healthy controls. Cohort Study. IRB approval was obtained. The RPL group included women with ≥ two pregnancy losses, endometrial biopsy (EMB) between 1/2016 and 12/2018, TSH values < 4 mU/L, negative antiphospholipid antibodies and normal uterine anatomy. The control group included volunteers without a history of RPL, infertility, uterine fibroids, polyps, pelvic inflammatory disease or retained pregnancy tissue recruited between 5/2019 and 2/2020. Women 18-50 years of age were included. H&E and CD138 immunohistochemical staining were performed on all endometrial biopsies. A single pathologist blinded to patient history recorded the number of plasma cells per 10 high power fields (HPF) and the presence of endometrial stromal changes (spindling, edema, breakdown, pigment deposition, areas of hypercellularity, and presence of inflammatory cells). Assuming a 3% prevalence of CE in controls and 30% in RPL, 25 controls and 50 women with RPL were needed to detect a difference in the rate of CE with 80% power and alpha of 0.05. 76 women were included, 50 with RPL and 26 controls. The cohort had a mean age of 34.6 (SD 4.1) years and mean BMI of 26.0 (SD 6.2) kg/m2. Women with RPL had a mean of 3.1 (SD 0.9) prior pregnancy losses, while controls had no prior pregnancy losses. When the presence of plasma cells alone was used to define CE, the prevalence among women with RPL was 56% (28/50) with ≥ 1 plasma cell, 44% (22/50) with ≥ 2 plasma cells and 26% (13/50) with ≥ 5 plasma cells by CD138 staining. Using these same criteria, the prevalence of CE among controls would be 31% (8/26) with ≥ 1 plasma cell, 27% (7/26) with ≥ 2 plasma cells and 8% (2/26) with ≥ 5 plasma cells. However, when both endometrial stromal changes and plasma cells were required for CE diagnosis, there were no cases of CE in the control group. Using this definition, the prevalence in the RPL group was 30% (15/50) with ≥1 plasma cells, 28% (14/50) with ≥2 plasma cells and 16% (8/50) with ≥5 plasma cells by CD138 staining, and was statistically significantly higher in women with RPL compared to controls (P=0.0015, 0.0016, and 0.04, respectively). Establishing specific diagnostic criteria for CE is necessary for both research and evidence-based treatment guidelines. We demonstrate that while rare plasma cells were found in EMB samples from healthy controls, the presence of plasma cells and endometrial stromal changes was isolated to the RPL cohort. Given that no controls were found to have both plasma cells and endometrial stromal changes, we propose that the definition of CE include 1 or more plasma cells in addition to the presence of endometrial stromal changes. Cases in which endometrial stromal changes are visualized, but no plasma cells are identified by H&E, would benefit from CD138 staining to identify plasma cells. We demonstrate that when using strict criteria to define CE, women with RPL have a significantly higher rate of CE compared to controls. An EMB should be considered as part of the evaluation for RPL.