Conjunctival cytology in horses: comparison of two techniques

Aim Conjunctival impression (CI) and cytobrush (CB) are techniques described both in human and veterinary medicine. Information about CI cytology in the horse are scant. Aims of the study were to validate CI technique and to compare feasibility and normal conjunctival cytological pattern of the CI and the CB techniques in healthy horses. Methods Samples were taken by CI from the bulbar conjunctiva of 10 healthy horses to standardize the method following Bolzan???s protocol. In 15 horses (30 eyes) conjunctival samplings were performed both by CI and CB, after the instillation of oxybuprocaine hydrochloride. Impressions were taken applying a 5x7 mm strip of Millipore filter with a pore size of 0.45 µm (Merck-Millipore, Milan, Italy) on the temporal bulbar conjunctiva and then fixed in 95% alcohol. Specimens were stained with Periodic Acid???Schiff (PAS) and with Hematoxilin???Eosin (H&E) stains. Stained samples were immersed in xylene and then mounted on slides cover-slipped. The CB was rolled on the conjunctiva of inferior lid and then rolled gently on the surface of the slide to prevent cell damage. Specimens were stained with H&E stains. Conjunctival and inflammatory cells were differentiated with a light microscope, counting 200 cells for each eye, by scanning each sample in a sinuous and continuous pattern. Cellularity, cell distribution and damage were scored following the criteria of Bauer et al. for both techniques. A score ranging from 0 to 3 was assigned to the feasibility, where 0 was an easy sampling without any restraint and 3 was an impossible to perform sampling. Data were not normally distributed, based on Shapiro-Wilk???s W test results. Mean cellularity was compared between techniques by using a Wilcoxon???s signed rank test, whereas the scores by using a median test. Significance was set at P<0.05. Results Both techniques allowed collecting a sufficient quantity of well-preserved epithelial cells. Goblet cells were observed only in CB smears. Inflammatory cells such as neutrophils and eosinophils were also found for both techniques. Cell debris and naked nuclei were found almost constantly in all samples. PAS staining of the CI was modified, reducing the time of immersion in Schiff???s reagent (3 min) and elongating the immersions in sodium metabisul???te (10 min). CB yield more basal cells (p<0.0001) and less intermediate cells (p