Comparison of the cytobrush, cottonswab, and low-volume uterine flush techniques to evaluate endometrial cytology for diagnosing endometritis in chronically infertile mares

Endometritis is the most important cause of infertility in barren mares. The quick method of endometrial cytology (EC) has a relatively high reliability in diagnosing endometrial inflammation in the mare. For reliable cytological results, a collection technique that yields many well-preserved cells representative of a large uterine surface area without causing harm to the reproductive tract is required. The aim of the study was to compare three usually employed techniques for collection of endometrial and inflammatory cells (guarded cotton swab, uterine lavage, and cytobrush) in chronically infertile mares. Twenty Standardbred mares were used. In each mare, samples for EC were collected, first by a cotton swab (DGS), then by a cytobrush (CB), and finally by low volume flush (LVF). The slides were stained using the Diff Quick stain. The following parameters were assessed for each tested technique: background content of the slides; quality of the cells harvested; total cellularity; neutrophils; ratio PMN/uterine epithelial cells; inflammatory cells; vaginal epithelium cells. Categorical variables were compared using contingency tables and Pearson Chi-square tests, whereas continuous variables were compared using one-way analysis of variance (ANOVA); P < 0.05 was considered significant. Samplings by DGS and CB resulted easy and quick to perform via a single operator in all cases. LVF was performed easily, but required the presence of 2–3 players and took more time. The background content of the slides prepared by DGS appeared proteinaceous, slides prepared by LVF appeared contaminated by red blood cells or debris, whereas slides prepared by CB appeared clear. All smears showed a good total cellularity. The CB yielded significantly more cells (P < 0.0001) than DGS and LVF. The DGS produced significant more cells than LVF (P < 0.0001). The DGS produced significantly more (P = 0.003) intact cells than CB and LVF. Distorted cells were significantly (P = 0.001) more frequent in smears by LVF. The CB harvested significantly (P = 0.009) more fragmented cells. CB and LVF produced significantly (P < 0.0001; P = 0.02) more PMNs/HPF than DGS. In smears collected by LVF the proportion of PMNs/uterine epithelial cells was significantly (P = 0.0062; P = 0.0023) higher than in smears by CB and DGS. CB collected a significantly higher (P = 0.0011) proportion of PMNs than DGS. Acute endometritis was diagnosed in 50% (10/20) of the mares by DGS cytological samples, 25% (5/20) by CB, and 75% (15/20) by LVF. Inflammatory cells other than PMN (lymphocytes, macrophages, eosinophils) were collected exclusively by CB method. Epithelial cells from the vagina were only detected in LVF slides. The agreement of the diagnosis of endometritis between the three techniques of collection and between the different criteria adopted to evaluate smears obtained with the same technique was poor (k ≤ 0.3). In conclusion, results show that cytobrush and flush specimens were superior in all parameters to cotton swab smears. Even though the cytobrush technique requires specialized equipment, sample collection by this method was easier, more consistent, and quicker than the lavage method, indicating that the brush would be the preferred collection method for use on field in the mare. More studies are needed to establish criteria for interpretation of inflammation in the mare on cytobrush samples.