The second semen analysis: does it make any sense?

Abstract

OBJECTIVE: To evaluate whether taking the better of two routinely performed semen analyses added value to the prediction of natural conception, compared to only one semen analysis. DESIGN: Prospective multicenter cohort study. MATERIALS AND METHODS: We calculated the associations between the results of the male partner's semen analysis and natural conception within one year, based on a prospective cohort study of consecutive couples presenting for subfertility in the period 2002 to 2004. For each man, two semen samples were routinely collected and analysed according to the WHO criteria. Primary outcome was spontaneous ongoing pregnancy. We constructed two prognostic models for the prediction of pregnancy based on four semen parameters each by using Cox' multivariable regression analyses. Model I was based on the first semen analysis, and model II on the better result of two semen analyses. RESULTS: In 706 men the semen samples were performed twice routinely. Follow-up data were available for 692 couples (98%), of whom 107 (15%) had a pregnancy by natural conception. The calculated probabilities of an ongoing pregnancy in model I ranged from 3% to 27%, for model II the calculated ranged from 4% to 32%. In 226 couples (33%) both models calculated a poor probability of an ongoing pregnancy without treatment within one year (<25%). In 5 couples (1%) both models calculated a moderate to good probability (>=25%). In 422 couples (61%), the second model calculated a moderate to good probability (>=25%), whereas the first model calculated a poor probability (<25%). In this group, the actually observed pregnancy rate within one year was 16%, indicating that model I was better in predicting pregnancy, whereas model II, using the results of the better of two semen samples, overestimated the probability of an ongoing pregnancy without treatment. CONCLUSIONS: One semen sample seems enough to predict the probability of a spontaneous ongoing pregnancy. Routinely repeating of a semen analysis does not seem to have additional prognostic value.