A prospective clinical study of the relationship between the computer-assisted assessment of human semen quality and the achievement of pregnancy in vivo.

The objective of this study was to examine the clinical predictive value for the achievement of pregnancy in vivo of the assessment of human semen quality by computer-assisted semen analysis using the Hamilton-Thorn Motility Analyser (HTM-S), alongside classic World Health Organization (WHO) techniques. A prospective follow-up study of 303 couples attending a regional infertility clinic providing tertiary level services to the population of a single geographical region was undertaken. Couples attending the infertility clinic, in whom the female partner was normal on conventional investigations (history, examination, evidence of ovulation, laparoscopy), were studied. After initial assessment, couples were followed for a median of 15 months and treatment-independent pregnancies observed and related to the results of semen assessment. During the period of follow-up, the treatment-independent pregnancy rate was 52%. Several measures of semen quality, determined both manually and by the HTM-S, were found to be predictive of the achievement of pregnancy and were related to the time taken to conceive. Couples who conceived had higher sperm concentrations and motilities, determined both manually and by the HTM-S, as well as differences in sperm head morphometry and sperm velocity determined by the HTM-S. Using multiple logistic regression, the prognostic accuracy of the HTM-S alone was similar to manual techniques, although data from the computer assisted sperm analysis system were preferred. Using proportional hazards regression, several variables were related to the achievement of pregnancy, particularly morphometry and motility. It was concluded that a strong case can be made for the introduction of automated assessment of human semen in routine service andrology laboratories.