Comparison of Clinical Signs, Endometrial Culture, Endometrial Cytology, Uterine Low-Volume Lavage, and Uterine Biopsy and Combinations in the Diagnosis of Equine Endometritis

Abstract

Endometritis is the most common cause of infertility in mares; however, many mares fail to be diagnosed despite availability of many diagnostic tests. Our objective was to compare different diagnostic methods and establish a cutoff value for the number of polymorphonuclear cells in cytology samples. Fifty-four mares were classified positive for endometritis based on endometrial biopsy (“gold standard”), and in a later analysis, the mare was reclassified as positive for endometritis if two or more of the following five criteria on a checklist were present (“new gold standard” [NGS]): (1) abnormal clinical findings, (2) abnormal gross character of low-volume lavage (LVL) fluid, (3) positive endometrial cytology, (4) bacterial growth on culture of the LVL pellet, and (5) histologic evidence of inflammation on endometrial biopsy. Kappa (κ) coefficient and percentages were calculated for sensitivity and positive predictive value (PPV) using SAS 9.3 software. Endometritis was diagnosed in 35/44 (79.5%) mares by biopsy. Based on the endometritis checklist, 33/51 (64.7%) mares were positive for endometritis. The character of LVL was 45% sensitive, whereas culture was 22% sensitive, when compared with endometrial biopsy. One percent of neutrophil-to-epithelial cell ratio was the most sensitive cutoff value (93.3%) when using a guarded swab (κ = 0.64). Endometrial biopsy was the most sensitive diagnostic method when compared against the NGS (sensitivity = 86%). Abnormal clinical findings and positive cytology showed moderate agreement with the NGS (κ = 0.41 and κ = 0.38, respectively). These studies demonstrate the importance of combining clinical findings and laboratory data when evaluating mares for endometritis.