Density adjustment of software settings minimizes bias in automated sperm motility estimation

To minimize overestimation of motility, it is recommended that fresh semen be diluted with seminal plasma prior to automated analysis. However, for glycerolated or cryopreserved semen this is impractical, and alternative methods are needed to minimize automated motility bias. In the present study, the proportion of motile spermatozoa was determined in fresh, diluted and cryopreserved semen (n = 25 ejaculates) using visual and automated methods. The effect of software settings on motility was investigated by assessing samples at a range of modified settings. At standard settings, automated motility was biased in fresh semen (+7.2%) after dilution with cryopreservative (-2.9%) and after cryopreservation (-7.8%) (P < 0.0001 versus visual). Automated motility was inversely related to the minimum number of frames for motility sampling (P < 0.0001), with mean estimates of 41.0, 46.1, 52.0 and 58.2% generated at settings of 8, 4, 2 and 1 frame(s) respectively (n = 15 fresh, diluted and cryopreserved samples). Based on an arbitrary ordinal scale, a method was developed whereby motility sampling was adjusted prior to analysis according to sperm density. Analysis of an independent set of semen samples with density-adjusted software settings reduced bias in automated estimates (n = 30) before and after freezing (P < 0.0001). In addition, bias was no longer related to sperm density. In conclusion, modification of software settings is an effective alternative to dilution to minimize bias in automated motility estimates in fresh, diluted and cryopreserved human semen.