LAUSR

OBJECTIVE To evaluate the role of hysteroscopy in the diagnosis of chronic endometritis (CE). DESIGN Retrospective cohort study. SETTING University teaching hospital. PATIENT(S) A consecutive series of 1,189 cases of diagnostic hysteroscopy. INTERVENTION(S) Endometrial biopsy specimens were obtained after hysteroscopy for routine histology and immunohistochemistry for plasma cells using a CD138 epitope. MAIN OUTCOME MEASURE(S) Observer variability, sensitivity, specificity, and accuracy of the hysteroscopic features in the diagnosis of CE. RESULT(S) Immunohistochemistry of biopsy specimens showed presence of CD138 cells in 322 of 1,189 cases (27.1%). Among cases that tested positive for CD138 cells, the prevalence of hysteroscopic features was as follows: endometrial hyperemia, 169 of 322 (52.5%); endometrial interstitial edema, 27 of 322 (8.4%); and micro-polyps, 11 of 322 (3.4%). The κ value of intraobserver and interobserver agreement on the presence or absence of the hysteroscopic feature of CE was 0.86 and 0.73, respectively. The sensitivity, specificity, positive and negative predictive value, and diagnostic accuracy of the presence of one or more hysteroscopy features were 59.3%, 69.7%, 42.1%, 82.8%, and 66.9%, respectively. CONCLUSION(S) The finding of endometrial hyperemia, micro-polyps or endometrial interstitial edema during hysteroscopy should alert to the diagnosis of CE, but the overall accuracy of hysteroscopic examination with regard to the diagnosis of CE is only 67%, so it should not be used to replace histologic examination as the diagnostic tool of choice. In women in whom a diagnosis of CE is considered likely, endometrial biopsy should be obtained to examine plasma cells by immunohistochemistry, which should remain the preferred method for diagnosis.