Introduction

Chronic endometritis (CE) is a localized inflammatory disease characterized histologically by stromal edema in the mucosal surface layer, increased stromal density due to leukocyte infiltration into the glands and stroma, and presence of plasma cells within the stroma [1]. The cause of CE has traditionally been attributed to chronic endometrial inflammation resulting from bacterial infection [2]. However, in recent years, non-infectious chronic endometrial inflammation has also been considered a possible cause [3]. CE is found in 55.7% of women with unexplained infertility and 9.3%–60% of women who experience recurrent miscarriages; associations between CE and these conditions have been reported [4–8]. Traditionally, the diagnosis of CE was based on histopathological evidence of plasma cell infiltration into the endometrial stroma [9]. CD138 immunohistochemistry has recently been reported to be superior to classical hematoxylin and eosin (HE) staining in detecting the presence of plasma cells within the endometrial stroma, and it is the current gold standard for CE diagnosis [10,11]. However, there are no internationally accepted fixed diagnostic criteria for CE. In clinical practice, the diagnostic criteria used in a previous study were selected arbitrarily [12].

The use of hysteroscopy for CE diagnosis has been reported [13,14]. However, its diagnostic value for CE has not been established, and the use of hysteroscopy as a substitute for endometrial biopsy is currently not recommended [15–17]. Previous studies have reported endometrial hysteroscopic findings suggestive of CE, such as endometrial congestion, micropolyps, endometrial edema, and endometrial polyps [15–19]. Although some studies have investigated the associations between these findings and the histopathological changes in CE confirmed by biopsy, few studies have compared these hysteroscopic features with those of patients without CE. Therefore, whether these findings are suggestive of CE remains unclear.

In this study, we aimed to identify the true findings suggestive of CE on hysteroscopy by comparing cases diagnosed as CE on endometrial biopsy with those assessed as non-CE on hysteroscopy.

Materials and methods

Hysteroscopy procedure and definition of hysteroscopy results

Hysteroscopy was performed in an outpatient setting without anesthesia using a flexible hysteroscope (Olympus HYF Type V, Tokyo, Japan) during the follicular phase of the menstrual cycle. The uterine lumen was dilated with a saline solution at a pressure of 100 mmHg and observed using the hysteroscope. The following four findings were observed: endometrial congestion, micropolyps, endometrial edema, and endometrial polyps (Fig 1). The mucous membrane of the uterine body was defined as follows: endometrial congestion, if accompanied by a reddened area localized to or scattered in the entire lumen; micropolyps, if accompanied by a mass protruding into the mucous membrane of the uterine body with a diameter of <1 mm; endometrial edema, if accompanied by an irregularly thickened endometrium with multiple microscopic pale bulges on its surface; and endometrial polyps, if accompanied by a mass sized >1 mm protruding from the endometrium. In our institution, six physicians performed hysteroscopy, with each performing 3–4 procedures per month. All six physicians were obstetricians/gynecologists who were specialists in assisted reproductive medicine, with 3–10 years of experience in the field. Chlamydial antibody testing was performed before hysteroscopy, and if a positive result was obtained, the test was repeated after the completion of antimicrobial treatment. Patients were diagnosed as hysteroscopy-positive when one or more of the following features were detected: endometrial congestion, micropolyps, endometrial edema, or endometrial polyps. Patients were diagnosed as hysteroscopy-negative when none of these features were found.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Hysteroscopic findings of the uterine lumen.

(a) Endometrial congestion: Reddened areas localized to the endometrium of the uterine body or scattered throughout the uterine cavity. (b) Micropolyp: A mass sized <1 mm protruding into the endometrium of the uterine body. (c) Endometrial edema: Irregularly thickened endometrium with multiple minute pale bulges on its surface. (a) Endometrial polyp: A mass sized >1 mm protruding from the endometrium of the uterine body.

https://doi.org/10.1371/journal.pone.0303041.g001

Results

Table 1 presents the backgrounds of the eligible patients and the indications for hysteroscopy and/or assisted reproductive technology (ART). Among the 317 eligible cases, 481 hysteroscopies were performed. Of these, 82 procedures were excluded because of indeterminate findings and other reasons. Finally, 399 eligible procedures were included in this study with 200 patients in the CE group and 199 in the non-CE group.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Patient background and indications for hysteroscopy and ART.

https://doi.org/10.1371/journal.pone.0303041.t001

The prevalence of the following hysteroscopic features did not differ significantly between the CE and non-CE groups: micropolyps (36.0% vs. 31.7%, p = 0.36), endometrial edema (9.5% vs. 7.5%, p = 0.48), and endometrial polyps (19.5% vs. 21.6%, p = 0.60). Only the prevalence of endometrial congestions differed significantly between the two groups (26.0% vs. 17.1%, p < 0.05) (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Prevalence of hysteroscopic findings in CE/non-CE groups.

These data contain duplicates. CE, chronic endometritis; N.S., not significant.

https://doi.org/10.1371/journal.pone.0303041.g002

The odds ratio (95% CI) for CE compared with non-endometrial congestion findings was 1.72 (1.05–2.82) for endometrial congestion, 1.16 (0.76–1.77) for micropolyps, 1.41(0.69–2.89) for endometrial edema, and 0.91 (0.56–1.48) for endometrial polyps. Among these hysteroscopic findings, only endometrial congestion was a significant contributing factor to CE diagnosis on endometrial histopathological examination in the multivariate analysis (Fig 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Odds ratio for CE for each hysteroscopic finding.

https://doi.org/10.1371/journal.pone.0303041.g003

The diagnostic concordance rates between each hysteroscopic finding and endometrial histopathology were 60.5%, 53.3%, 55.9%, and 47.6% for endometrial congestion, micropolyps, endometrial edema, and endometrial polyps, respectively. The diagnostic concordance rate for CE in hysteroscopy-positive cases was 53.2%, whereas that for non-CE in hysteroscopy-negative cases was 55.0% (Fig 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Diagnostic concordance rate for CE between each hysteroscopy finding and endometrial histopathology.

CE, chronic endometritis.

https://doi.org/10.1371/journal.pone.0303041.g004

Discussion

The main findings of this study are as follows. First, among the hysteroscopic findings reported to be associated with CE, endometrial congestion was the only contributing factor. Second, the diagnostic concordance rate for CE between hysteroscopic findings and endometrial histopathology was low (53.0%), suggesting that the diagnosis of CE using hysteroscopy is challenging.

In the comparison of the prevalence of each hysteroscopic finding between the two groups, only the prevalence of endometrial congestion was significantly higher in the CE group than in the non-CE group. A systematic review examining the usefulness of hysteroscopy in the diagnosis of CE reported that micropolyps, endometrial edema, and diffuse or focal endometrial congestion are hysteroscopic findings suggestive of CE [18]. However, in this study, among the hysteroscopic findings suggestive of CE, only endometrial congestion had a significantly higher prevalence in the CE group than in the non-CE group, whereas other findings, such as micropolyps, endometrial edema, and endometrial polyps, were not significantly different between the two groups. Furthermore, endometrial congestion was the only contributing factor to CE diagnosis on endometrial histopathological examination in multivariate analysis. The diagnostic concordance rate for CE in patients detected as having endometrial congestion on hysteroscopy was 60.5%, whereas it was approximately 50% for other findings. This suggests that most of the hysteroscopic findings thought to be suggestive of CE may also be present in some non-CE cases. Notably, these findings have been reported in pathophysiological conditions other than CE [20,21] which, in our view, explains the reduced diagnostic rate of CE on hysteroscopy.

The diagnostic concordance rate for CE between the hysteroscopic findings and endometrial histopathology in the present study was low (53.2%). Moreno et al. reported that the diagnostic discordance rate for CE between hysteroscopic findings and endometrial histopathology was high, at 58.5% (2018), which also implies a low diagnostic concordance rate [2]. A low concordance rate between hysteroscopic findings and histological examination was reported more recently, with the authors concluding that hysteroscopic signs are not yet sufficient to make an accurate diagnosis of CE [22]. The view that hysteroscopy alone cannot accurately diagnose CE is corroborated by further studies [15,18,23]. Another reason for the low diagnostic accuracy of hysteroscopy is the dependence of diagnosticians on diagnostic techniques. Additionally, endometrial histopathological examination should always be performed because the sensitivity of hysteroscopic diagnosis depends on the clinician’s subjective evaluation [23]. Overall, the low diagnostic concordance rate for CE between hysteroscopic findings and endometrial histopathology can be attributed to various factors including the lack of uniform diagnostic histopathological criteria [12], low consistency in pathologists’ test results [24], absence of standardized criteria for the timing of endometrial tissue collection, and reliance on the endoscopic surgeon’s subjective and nonspecific observations on hysteroscopy [25]. It is important that these challenges for the diagnosis of CE are addressed in future studies.

The strength of this study is that we compared the hysteroscopic findings of CE cases with those of non-CE cases and showed that the hysteroscopic finding most suggestive of CE was endometrial congestion. However, this study has some limitations. Considering that the degree of endometrial congestion varied from considerably localized to diffuse congestion, known as the “strawberry aspect,” we were not able to examine the difference in the diagnostic rate of according to differences in the degree of hyperemia. Similarly. this limitation extends to the evaluation of variations in the other findings. Another limitation is that because endometrial histopathology was performed using aspiration, the endometrial tissues were not necessarily collected from the same site as those collected using hysteroscopy. However, considering that none of these sites necessarily corresponds to the actual implantation site of the embryo, the meaningfulness of matching these sites is questionable. This study indicates that only endometrial congestion, among the hysteroscopic findings associated with previous CE, contributes to CE. However, because the diagnostic concordance rate for CE between each hysteroscopic finding and endometrial histopathology was low, we believe that hysteroscopy cannot be the first-choice diagnostic tool for CE. It is important to investigate the effect of hysteroscopic findings on clinical outcomes and the relationship between the strength of hysteroscopic findings and endometrial histopathology.

In conclusion, among the hysteroscopic findings suggestive of CE, only endometrial congestion was significantly associated with biopsy-confirmed CE. Making a diagnosis of CE based solely on hysteroscopy presents considerable challenges. Future studies should investigate the correlation between the pathological diagnosis of CE and other factors including the extent of endometrial congestion.

Supporting information