REPEATABILITY OF MEASUREMENTS ON HUMAN, RABBIT, AND BULL SPERM BY COMPUTER-ASSISTED SPERM ANALYSIS WHEN COMPARING INDIVIDUAL FIELDS AND MEANS OF 12 FIELDS

Objective: To compare repeatability of measurements of human, rabbit, and bull sperm on two Hamilton Thorne Integrated Visual Optical System (IVOS) units, software version 10. Design: Semen samples from seven normal human subjects, six rabbits, and eight bulls were obtained at regular intervals. The samples were diluted, two chambers were filled, and 12 fields were recorded, using high resolution recorders. Computer-assisted sperm analysis (CASA) was performed nearly simultaneously with two Hamilton Thorne NOS units. Setting: Reproduction Research Laboratories, Cornell University and Hamilton-Thorne Research, Beverly, Massachusetts. Results: Optimal settings were established for evaluating by CASA sperm from three species. Fifteen variables were analyzed. The correlation coefficients for most variables characterizing sperm motion and concentration, when means of 12 fields were calculated, were 0.95 to 1.00. There were too few hyperactive sperm to obtain a reliable correlation for human sperm (r = 0.63) and repeatability of elongation was lower only for human sperm (r = 0.75). Conclusions: Two units of Hamilton Thorne IVOS, software version 10, were capable of providing nearly identical estimates of many CASA variables of human, rabbit, and bull sperm. Correlations for the paired estimates of many motion characteristics and sperm concentration usually exceeded 0.97.