QUESTION Do two semen analyses predict natural concep on be er than a single semen analysis and will adding mul ple semen parameters to the Hunault predic on model for natural concep on improve predic ons ? SUMMARY

STUDY QUESTION Do two semen analyses predict natural concep on be er than a single semen analysis and will adding mul ple semen parameters to the Hunault predic on model for natural concep on improve predic ons? SUMMARY ANSWER A second semen analysis does not add helpful informa on for predic ng natural concep on compared to using the results of a single semen analysis. It did not so for rou nely performed second semen analyses, nor for a second semen analysis in selected only. One semen analysis is all that is needed for the predic on of natural concep on in the basic fer lity workup. WHAT IS KNOWN ALREADY A major problem with semen analyses is the large variability of semen parameters within an individual. High quality evidence on how many semen analyses need to be performed during the fer lity workup is lacking. STUDY DESIGN, SIZE, DURATION Prospec ve cohort study of 897 consecu ve couples presen ng with subfer lity in two university hospitals in the period 2002 to 2004 in the Netherlands. PARTICIPANTS/MATERIALS, SETTING, METHODS We constructed models for three strategies for the predic on of natural concep on, using univariable and mul variable cox hazard regression analyses. We evaluated the performance of the three strategies by comparing goodness-of-fi t, discrimina on and calibra on. First, we analysed the semen parameters only. Second, we analysed the semen parameters in addi on to the mul variable Hunault predic on model. MAIN RESULTS AND THE ROLE OF CHANCE Of the 897 couples 132 (15%) achieved a pregnancy by natural concep on. Using the results of a single semen analysis only, the calculated probabili es of natural concep on within 12 months ranged from 0.12 to 0.38, with a median of 0.16 [IQR: 0.16 to 0.17]. Using the results of two semen analyses did not lead to a be er goodness-of-fi t. Discrimina ve capacity was rather poor, with area under the ROC curve (AUC) of ranging from 0.51 to 0.56. Using the Hosmer-Lemeshow test sta s c we found no signs of poor calibra on. Using the results of two analyses in combina on with the Hunault model did not signifi cantly increase goodness-of-fi t compared to using a single semen analysis. The Hunault model with the addi on of the semen parameters fi t the data signifi cantly be er than the Hunault model itself (diff erence in -2 Log likelihood: 13; 3 df; p=0.002). Using the Hosmer-Lemeshow test sta s c we found no signs of poor calibra on. IMPACT OF THE MALE FACTOR ON THE PREDICTION OF NATURAL CONCEPTION