Comparison between cytology and histopathology to evaluate subclinical endometritis in dairy cows

The aim of the present study was to compare endometrial cytology with histopathology to diagnose subclinical endometritis (SCE) in dairy cows. Endometrial cytology samples were collected from Holstein-Friesian cows (n = 32) just before slaughtering. Half of them were obtained by in vivo cytobrush (IV-CB), whereas the other half by in vivo low-volume lavage (IV-LVL). After slaughtering, reproductive tracts were collected, and the endometrium was sampled at eight locations. At each location, both a ex vivo cytobrush sample (EV-CB) and a tissue sample for histopathologic examination were taken. In the histopathology slides, polymorphonuclear (PMN) cell counts were differentiated as PMN cells in direct contact with the epithelial cells of the endometrium (PMN-EP), and PMN cells present in the deeper stratum compactum (PMN-SC). Summation of both countings was referred to as PMN-total. Pearson's correlation and Cohen's kappa coefficient were used to assess the correlation and agreement between both sampling methods (in vivo cytology [IV-CB and IV-LVL] with EV-CB and PMN-total). A Poisson mixed effect model was used to analyze the PMN cells' distribution. The prevalence of SCE was 18.75% (n = 6/32) for in vivo cytology. The SCE prevalence based on EV-CB analyses and on the assessment of PMN-total was determined both at the sample (n = 256) as well as at the cow level (n = 32): EV-CB 25% (n = 64/256) and 35.5% (n = 12/32), and PMN-total 37.11% (n = 95/256) and 59.38% (n = 19/32). Correlation and agreement between IV-CB and EV-CB were r = 0.81 and k = 0.97, whereas between IV-CB and PMN-total r = 0.15 and k = 0.23, respectively. In vivo low-volume lavage correlation and agreement were r = 0.52 and k = 0.66 with EV-CB, and r = 0.45 and k = 0.44 with PMN-total. Moreover, correlation and agreement between EV-CB and PMN-total were r = 0.60 and k = 0.50, respectively. More PMN cells (P < 0.05) were detected in PMN-SC when compared to PMN-EP and EV-CB. A higher SCE prevalence was found using histopathology, rendering the latter as a more sensitive method to diagnose SCE in comparison to in vivo and ex vivo cytology. Although cytology had low and/or moderate sensitivity to diagnose SCE when compared with histopathology, its specificity is 100%, implying that all cows that were indicated to suffer from SCE using in vivo cytology were confirmed to do so by histopathologic examination. There is an uneven distribution of PMN cells throughout the endometrium, generally more PMN cells being found in the deeper stratum compactum than in contact with the superficial layers of the endometrium.