Chronic plasmacytic endometritis (CPE) is considered an infectious or reactive process. Commonly cited causes include transvaginal infection, intrauterine devices (IUDs), submucosal leiomyoma, and endometrial polyp; in other words, almost any cause of chronic irritation to the endometrium may result in a chronic inflammatory reaction (1). The histopathologic diagnosis is characterized by endometrial inflammation rich in plasma cells with or without accompanying acute inflammation and lymphocytes (2). Lymphocytes are a known normal component of the endometrial stroma, whose number fluctuates with the phase of the cycle (3). However, in CPE their numbers are increased (2). Although the diagnostic criteria for significant (eg, clinically significant) chronic endometritis remain controversial, it is generally agreed that the presence of plasma cells within the endometrial stroma is the most useful histologic criterion for diagnosis (2). In addition to plasma cells, a variety of other features can alert the reviewing pathologist to the possible presence of CPE. Disturbances in normal growth and maturation, superficial mucosal stromal edema, stromal breakdown, and characteristic spindle cell alteration of the stroma are other morphologic changes that can be seen with CPE (1,2).
There are few large (100 cases or more) published studies statistically examining both the histopathology and clinical findings of CPE in an ethnically heterogeneous population. Evaluation of CPE in a significant Native American population has not occurred. We hypothesized that thorough examination of the histopathologic findings and clinical history of CPE would further elucidate its clinical significance and major histopathologic features, and would provide information with regard to the most common associated or causative conditions.
MATERIALS AND METHODS
We undertook a retrospective chart and biopsy review that focused on the collection of clinical data and the examination of endometrial tissue samples obtained from patients seen at the University of New Mexico Health Sciences Center (Albuquerque, New Mexico) between 2002 and 2007. A total of 235 endometrial biopsies were culled from the files: 105 women with a diagnosis of chronic endometritis matched to 130 women of similar age and menstrual status without that diagnosis. The search was performed by using the surgical database software program Tamtron PowerPath (IMPAC Medical Systems, Inc. 2310 Corporate Circle, Suite 275 Henderson, NV 89074) for endometrial biopsies. Curettage specimens were not included. Patients were excluded if their biopsies were known to contain hyperplasia or carcinoma.
The original diagnoses were made on the basis of the identification of stromal plasma cells. Subsequently, for this study, 1 senior pathologist assessed biopsies for 11 histologic criteria, whereas 2 other investigators reviewed patient charts for 15 clinical criteria (Table 1); collection of data and slide examination occurred independently. The degree of plasmacytic inflammation was determined as number of plasma cells per 40× high-power field (HPF) using an Olympus BX40 microscope. The pathologist used a grid hand drawn on 1 level of each case so as ensure that all fragments in the tissue section were examined. Three levels were reviewed for each biopsy; only 1 level was reviewed by the grid. Chi-square tests were used to test for differences between groups, and the binomial distribution was used to calculate 95% confidence limits for the undiagnosed CPE proportion. SAS software was used for the analysis.
The mean age of the patients was 46 years, with a range between 21 and 86 years of age. There were no age or ethnic differences (Anglo, Hispanic, Native American, other) related to the diagnosis of CPE versus non-CPE. Overall results are listed in Table 2. The largest difference between the control and CPE groups occurred among White, non-Hispanic women, with 5% fewer such patients in the CPE group (P>0.05). Premenopausal women represented the largest number of individuals (43%) in the overall sample. Premenopausal women composed 34% of CPE cases, whereas menopausal and postmenopausal made up 42% and 19%, respectively. Five percent of the patients had no verifiable menstrual status. The majority (80%) of the women in our study had been pregnant twice or more. Of the entire study population, 24% of the women had used exogenous hormones (oral contraceptive pills, hormone replacement therapy, etc) at the time of biopsy.
There were no significant associations with regard to the presence of neutrophils or eosinophils or the presence or type of reactive epithelial change (squamous metaplasia, tubal metaplasia, eosinophilic syncytium, mixed metaplasia) with the degree of chronic inflammation. Histopathologic review revealed a statistically significant presence of spindled stroma in cases with a high number of plasma cells and a trend toward proportional increases in spindling with higher degrees of inflammation (Fig. 1). Chronic endometritis was identified in a number of cases (51%), with nonaltered endometrium of all phases of the cycle, including secretory (Fig. 2). Macrophages were also shown to be statistically significant in cases of endometritis, with a trend toward higher numbers of macrophages with lower numbers of plasma cells. Lymphocytes increased in number as the plasma cell infiltrate increased (P<0.001). In cases of minimal-to-mild plasmacytic inflammation, there were 26 cases of mild lymphocytic presence, 19 moderate, and 13 florid. Among CPE cases with moderate or greater inflammation, there were 0 cases of no lymphocytic inflammation, 9 cases of mild inflammation, 12 of moderate inflammation, and 21 of florid inflammation. Most cases of CPE had minimal (<10 per HPF) to moderate (20 to 50 per HPF) numbers and fewer had florid (50 to 100 per HPF) or extraflorid amounts (>100 per HPF) (Fig. 2). The degree of inflammation was first assessed by high-power microscopic review, and the case was scored according to the most inflamed area on the slide. All cases contained at least 2 fields with plasma cells present. Plasma cells were most noticeable clustered around stromal vessels (Fig. 3) as opposed to at or immediately below the endometrial surface.
Investigation into associated findings showed that cases involving hyperplasia, polyps, and carcinoma made up less than 5% of cases, and showed no statistically significant association with degree of chronic inflammation. Of the specimens with secretory stroma, 1 case had concurrent tubal metaplasia. The remainder of cases in secretory phase did not have abnormal epithelial changes. Five patients were noted to have histologic evidence of a prior pregnancy; these patients were excluded from the study.
Examination of the control cases revealed 17 cases (16%) of CPE that were not diagnosed as chronic endometritis by the original reviewing pathologist (Fig. 4). Of these, 5% of total control cases were of moderate to extraflorid inflammation. This finding of undiagnosed CPE was statistically significant (P<0.05) when compared with the CPE group, with a confidence interval of 0.02 to 0.12.
Only 3 patients received antibiotics after the diagnosis of chronic endometritis was reported (2.8%). These patients did not have clinical symptoms of fever, pain, or elevated white blood cell counts to distinguish them clinically from nonspecifically treated patients. The few patients treated did not have CPE that was florid or extraflorid in degree.
The duration and type of symptoms (menorrhagia, menometrorrhagia, postmenopausal bleeding, etc) did not differ statistically between CPE patients and the age-matched control group (Fig. 5). Younger patients (<45 y) did not differ in the range of histopathologic features when compared with patients aged 45 years or above.
Our study confirmed that the reported histologic features of CPE diagnosed earlier, including inflammation comprising plasma cells, lymphocytes, and sometimes neutrophils and eosinophils; alterations in proliferative or secretory phase pattern; and stromal breakdown, are commonly observed in CPE (1,2). A direct correlation in lymphocyte numbers with plasma cell numbers is commonly observed, as is an association with lymphoid follicles (1,2). In our study, there was a greater propensity for plasma cells to aggregate around deeper stromal vessels than at the endometrial surface, as has been reported earlier (4,5). Our findings share some similarities with an earlier clinical review by Cadena et al. (6). However, that study did not correlate histologic findings with clinical signs and symptoms. In addition, Cadena et al. found that one third of their patients had pelvic inflammatory disease (PID) and 11% had an IUD. These numbers are much higher than in our study, and it is not clear why there is such a large difference. The difference may be related to a lower rate of PID and IUD placement in our patient population. New methods of IUD placement and new devices may diminish or abrogate an inflammatory response. It is also possible that the retrospective chart review design of our study compromises some accuracy in clinical data when the medical record is incomplete or incorrect. The records abstracted for this study were those of an academic medical institutition with electronic charting, and seemed to be complete. In addition, the Cadena et al. study did not report a possible miss rate of histologic diagnosis of chronic endometritis, as our review does.
A recent study by Cicinelli et al. (7) suggests using a triad of hysteroscopy, standard microbiology culture, and endometrial biopsy to best assess for the presence of CPE. This well-designed, prospective study reported an 89% correlation of hysteroscopy (micropolyps, congestion, and edema) and histology for CPE, coupled with a high microbiology culture rate of 73%. In this study, the histologic criteria used for the diagnosis of CPE are unclear. In addition, it is likely that a high degree of inflammation is necessary for the development of micropolyps. As the presence of micropolyps was the largest factor in predicting a diagnosis of CPE, it is likely that the population of patients with CPE in this study makes up a smaller swath of the overall CPE population, comprising only those patients with more florid inflammation. By contrast, our study consists of the entire histologic spectrum of patients with CPE, including patients with minimal inflammation unlikely to result in hysteroscopically identifiable micropolyps. It would be useful to determine whether patients with CPE of minimal degree correlate histologically with a lower rate of positive microbiology culture.
Although current textbook descriptions state that the endometrium is altered and not datable in the setting of CPE, we found that chronic endometritis was identified in a number of cases with nonaltered endometrium of all phases of the cycle, including secretory (Fig. 6). This finding suggests that the endometrial stroma and cycling is not always affected to the point of disturbing the endometrial phase. Although almost half the cases had stromal breakdown, the finding was not a significant association, though other studies suggest an inflammatory relationship in such cases (5). In fact, few of the histologic findings track in a significant fashion with clinical features. This lack of correlation may be the result of the variety of histologic findings that may be present, of which only 1, the presence of plasma cells, is a key feature. Furthermore, despite a vigorous semiquantitation of the plasma cells in our patients' biopsies, the number of these cells also does not correlate with clinical features. The patients in this study showed the same clinical symptoms and clinical history as the control group. The aforementioned study by Cicinelli et al. (7) described a potential significant relationship with their triad of findings (positive culture, hysteroscopic features of CPE and histologic evidence of CPE) and infertility. None of the patients in our study presented with this complaint, which again, was a notable difference between Cicinelli's patients and the current patient group.
When comparing clinical history with histopathology, we found no association between the degree of inflammation and age, the duration of symptoms and age, or the types of symptoms and age (grouped as <45 y and ≥45 y). These findings suggest that the degree of chronic inflammation may not directly impact symptoms, and are consistent with those of other studies (6,8). The lack of clear association is in accord with the earlier large study by Cadena et al. (6) comprising 152 patients. The lack of any association between distinct histologic features or clinical symptoms and patient ethnicity is similar to the findings in the Cadena et al.'s study.
The 16% calculated underdiagnosis rate is potentially important. When adjusted for greater degrees of inflammation than the minimal criterion of rare plasma cells, the rate decreases to 5%, which remains a relatively high number of undetected CPE cases, given the frequency of endometrial biopsies performed. This rate is consistent with an earlier study showing an underdiagnosis rate when examining apparently normal proliferative tissues with methyl green pyronin of 18% (5). The underdiagnosis rate is most likely attributed to plasma cells being overlooked. Our examination using a grid assured complete, comprehensive review that is not normally achieved in routine pathology practice. It would be unusual for the original diagnosis pathologist to use this technique, and this likely accounts for the majority of underdiagnosed cases.
Interestingly, there was no association between severity of histopathology and clinical treatment, and only a few patients actually received antibiotic therapy. Review of clinical documentation shows that intervention (in terms of antibiotic therapy) was rarely performed in reaction to the CPE diagnosis. Treatment usually comprised surgical ablation, curettage, or simple hysterectomy. Thus, in the current treatment setting, clinicians seem to find it unnecessary to treat CPE patients with antibiotics, despite the classic association of CPE with several bacterial etiologies, including Neisseria gonorrhea, Chlamydia, and Mycoplasma. Cicinelli et al. (7) found that in addition to these bacteria, 70% of CPE cases resulted from nongonococcal, nonchlamydial infections including common bacteria such as Escherichia coli, Streptococci, and Enterococcus faecalis. Our study did not include bacterial cultures of the endometrial biopsies, as it comprises a retrospective review, and attempts to culture the endometrium require a special approach not routinely undertaken at the time of endometrial biopsy.
It is well known that there is an association between PID and its associated organisms (Chlamydia and Neisseria gonorrhea) (1,6,9). In young women with clinical signs and symptoms of PID, clinicians currently reflexively treat with cefotetan and doxycycline (10). As new information suggests that the variety of possible etiological organism is broad, even broad spectrum antibiotics would likely miss many of the possible infectious causes of CPE. Additional studies have shown that curettage and antibiotic treatment decrease the degree of inflammation in cases of CPE (11,12). At our institution, treatment performed on patients with CPE seems to be aimed at addressing symptoms and not at a possible infectious etiology, as only a small minority of clinicians prescribed antibiotics after the diagnosis of CPE was reported to them.
One of the goals of our study was to examine whether a particular suite of histologic features correlated with distinct clinical features. We did not find any such associations. A secondary goal was to determine whether a critical threshold of histologic features (either in number of features or intensity of inflammatory changes) predicted a particular clinical presentation (such as fever, pain, other evidence of PID). In our patient group, no set of histologic features or degree of intensity of inflammation predicted a particular clinical presentation, a response by the clinician to prescribe antibiotics, or outcome. Currently, most cases of CPE are attributed to PID, other common bacteria, polyps, past clinical or surgical procedures, and the presence of IUDs (2,9,10). The retrospective nature of our study identified few cases where cultures were performed coinciding with biopsies. Review of clinical histories did not reveal any cases diagnosed by the clinicians as PID, and only a single patient was noted to have an IUD. Finally, the presence of polyps was rare and not statistically significant in comparison with the control group.
In conclusion, the histologic diagnosis of chronic endometritis comprises a variety of features and varying degrees of intensity of inflammation that do not show any correlation with clinical features such as type of symptoms, duration of symptoms, and severity of symptoms. Premenopausal and postmenopausal women do not differ in the degree or type of inflammatory response that occurs in CPE. The current patient population differs from the last large patient cohort study in that there are no patients with PID in the current patient group and only 1 patient with an IUD. As few clinicians in this study treated their patients with antibiotics, a miss-rate of up to 16% of cases may not be significant, as such patients were treated identically to those patients in whom a diagnosis of chronic endometritis was rendered. Microbiology cultures were not obtained in the majority of the patients in this study; a recent study in which such cultures were obtained (and found a high rate of positivity and a high rate of infertility) most likely included only patients who would qualify as having “florid” or “extraflorid” CPT, according to our criteria. It is unknown whether patients with more minimal histologic examples of chronic endometritis have a similar (90%) rate of positive microbiology culture.
There was no pharmaceutical or industrial support for this project.
1. Greenwood S, Moran J. Chronic endometritis
: morphologic and clinical observations. Obstet Gynecol 1981;58:176–84.
2. Crum CP, Hornstein MD, Stewart EA. Evaluation of cyclic endometrium and benign endometrial disorders. In: Crum CP, Lee KR, eds. Diagnostic Gynecologic and Obstetric Pathology. 1st ed. Philadelphia, PA: Elsevier Saunders; 2006:441–91.
3. Disep B, Innes B, Cochrane H, et al. Immunohistochemical characterization of endometrial leucocytes in endometritis. Histopathology 2004;45:625–32.
4. Bayer-Garner I, Korourian S. Plasma cells
in chronic endometritis
are easily identified when stained with syndecan-1. Mod Pathol 2001;14:877–9.
5. Gilmore H, Fleischhacker D, Hecht J. Diagnosis of chronic endometritis
in biopsies with stromal breakdown. Hum Pathol 2007;38:581–4.
6. Cadena D, Cavanzo FJ, Leone CL, et al. Chronic endometritis
: a omparative clinicopathologic study. Obstet Gynecol 1973;41:733–8.
7. Cicinelli E, DeZiegler D, Nicoletti R, et al. Chronic endometritis
: correlation among hysteroscopic, histologic, and bacteriologic findings in a prospective trial with 2190 consecutive office hysteroscopies. Fertil Steril 2008;89:677–84.
8. Heatley M. The association between clinical and pathological features in histologically identified chronic endometritis
. J Obstet Gynaecol 2004;24:801–3.
9. Ross JDC. What is endometritis, does it require treatment? Sex Transm Inf 2004;80:252–3.
10. Crossman S. The challenge of pelvic inflammatory disease
. Am Fam Physician 2006;73:859–64.
11. Bayer-Garner I, Nickell J, Korourian S. Routine syndecan-1 immunohistochemistry aids in the diagnosis of chronic endometritis
. Arch Pathol Lab Med 2004;128:1000–3.
12. Eckert L, Thwin S, Hillier S, et al. The antimicrobial treatment of subacute endometritis: a proof of concept study. Am J Obstet Gynecol 2004;190:305–13.