Standardization of specimen preparation, staining, and sampling methods improves automated sperm-head morphometry analysis.

Abstract

OBJECTIVES: To evaluate the accuracy and performance of the CellForm-Human automated sperm morphometry instrument (Motion Analysis Corp., Santa Rosa, CA) using different specimen preparation, staining, and analysis techniques. SETTING: Clinical and research andrology and in vitro fertilization laboratories. PATIENTS: Individuals who were undergoing semen evaluation and infertility work-up. RESULTS: The percentage of normal sperm detected by CellForm-Human was not different for washed specimens compared with nonwashed controls. Washing and resuspension to a standard concentration in medium significantly increased spermatozoan density and homogeneity on the slide. Comparison of digitization errors and morphometric measures of sperm stained by the Papanicolaou (PAP) method, the hematoxylin portion of the PAP method, and a new method developed by us, GZIN (pronounced Gee-ZIN), showed that GZIN produced larger measures for length, width, area, and perimeter, but not for width/length, and also produced less variability for most measures than PAP or hematoxylin. The number of digitization errors was significantly less for GZIN than for PAP. Visual inspection revealed that GZIN produced a more consistent stain over the entire sperm head than hematoxylin. The number of sperm analyzed affected the percentage of normal sperm detected. The % normal stabilized only after at least 200 sperm had been analyzed for each man. CONCLUSIONS: Technical variability arising from semen preparation and slide staining methods can be reduced when specimens are washed and resuspended to a standard concentration before smearing. At least 200 sperm should be analyzed to obtain a stable estimate of the percentage of normal sperm. More sperm should be analyzed if the % normal is < 50%. Morphometric measurements are more accurate and precise when sperm are stained with GZIN than when stained with PAP or hematoxylin.