Background: Clomiphene citrate (CC) is known to improve semen quality for men with infertility, but there are no published algorithms to predict the changes in an individual’s semen parameters following CC therapy. Since there are multiple options to treat men with infertility, a model to predict the outcomes of CC therapy would allow men to make a more informed decision on their treatment choices. Methods: This is a prospective study on a cohort of 121 infertile men being treated empirically with CC 25 mg every other day for a minimum of three months. Men were included if they did not have other active fertility therapies other than the use of supplements. Semen samples and a hormone profile (including total testosterone, follicle stimulating hormone (FSH) and leutinizing hormone (LH)) were provided prior to and at least three months after initiation of therapy. The patient age, pre-CC hormone values and semen parameters were used to develop a nomogram to predict the post-CC hormone values and semen parameters. The model was developed with predictors selected using backward selection methods by minimizing root mean square error evaluated on 500 bootstrap runs. A zero-inflated negative binomial modeling (ZINB) approach was used to model sperm concentration and sperm motility. A generalized linear model for the Gamma distribution was used to model testosterone. A linear model was used to model log transformed LH, FSH and semen volume. Results: Post-CC sperm parameters and hormone values were predicted by the pre-CC sperm concentration, hormone values and patient. Nomograms were developed to predict the outcomes of CC therapy based on pre-CC parameters. Conclusion: These new models will help physicians personalize care by predicting the outcomes of therapy and allow clinicians to tailor the treatment to the individual couple.
