Prediction of the fertility of stallion frozen-thawed semen using a combination of computer-assisted motility analysis, microscopical observation and flow cytometry.

Abstract: Spermatozoa from some stallions do not maintain an acceptable fertility after freezing and thawing. The selection of frozen ejaculates that would be suitable for insemination is mainly based on post-thaw motility, but the prediction of fertility remains limited. A recent study in our laboratory has enabled the determination of a new protocol for the evaluation of fresh stallion semen, combining microscopical observation, computer-assisted motility analysis and flow cytometry, and providing a high level of fertility prediction. The purpose of the present experiment was to perform similar investigations on frozen semen. A panel of tests evaluating a large number of compartments or functions of the spermatozoa was applied to a population of 42 stallions, 33 of which showing widely differing fertilities (17-67% pregnancy rate per cycle [PRC]). Variability was evaluated by calculating the coefficient of variation (CV=SD/mean) and the intra-class correlation or "repeatability" for each variable. For paired variables, mean within-stallion CV% was significantly lower than between-stallion CV%, which was significantly lower than total CV%. Within-ejaculate repeatability, determined by analysing 6 straws for each of 10 ejaculates, ranged from 0.60 to 0.97. Within-stallion repeatability, determined by analysing at least 5 ejaculates for each of 38 stallions, ranged from 0.12 to 0.95. Principal component regression using a combination of 25 variables, including motility, morphology, viability, oxidation level, acrosome integrity, DNA integrity and hypoosmotic resistance, accounted for 94.5% of the variability regarding fertility, and was used to calculate a prediction of the PRC with a mean standard deviation of 2.2. The difference between the observed PRC and the calculated value ranged from -3.4 to 4.2. The 90% confidence interval (90CI) for the prediction of the PRC for the stallions of unknown fertility ranged from 8 to 30 (mean = 17). The best-fit model using only motility variables, evaluated after 10 min at 36 °C and 2 h at 36 °C or room temperature, accounted for only 74.2% of the variability. The difference between the observed PRC and the calculated value ranged from -7.2 to 14. The 90CI for the prediction of the PRC for the stallions of unknown fertility ranged from 23 to 48 (mean = 33). In conclusion, this study demonstrated that an appropriate combination of computer-assisted motility analysis, microscopical observation and flow cytometry can provide a higher prediction of fertility than motility analysis alone.