RESUMO
For sperm analysis, important inter-laboratory variations have been observed in manual analyses. In this study, a computer-aided sperm analysis (CASA) system was assessed versus manual technique, and specific software modifications were operated to fit the David's classification already used in the laboratory. Four parameters were studied (concentration, motility, vitality and morphology), and at least 30 semen samples from 30 different patients have been tested. Manual and automated analyses were compared using a least-squares regression line analysis, Student's t test, Bland-Altman plots and Passing-Bablok regressions. Repeatability was also assessed, and coefficients of variation (CV) were calculated. Both manual and automated methods gave similar results for sperm concentration (n = 150), motility (n = 30), vitality (n = 90) and morphology (n = 90). Repeatability always showed a decrease in the CV with automated analysis; for example in normal range of sperm values, CV for manual and CASA analyses were, respectively, 9.0% versus 4.4% for sperm concentration, 5.2% versus 4.1% for motility, 7.3% versus 4.2% for vitality and 11.4% versus 4.1% for morphology. All parameters were comparable between automated and manual analysis, and repeatability measures confirm the more reliable values of the SCA compared to those of manual analysis.
