Prediction Models for IUI Success Incorporating Male Age are More Accurate among Older Women

Introduction
Intrauterine insemination (IUI) is commonly offered as a first step in fertility treatment. However many couples will not benefit from IUI and should be treated with IVF. Tools to predict IUI success rate were used in the past but they do not include many new tests.

Aim
To study which factors are important to predict who will not benefit from IUI and who should be recommended for IVF directly.

Methods
We evaluated the predictive capacity of different parameters in predicting IUI success, and constructed a generalized linear regression model (LRM). A total of 467 patients (1342 IUI cycles) were studied retrospectively. The following parameters were used in the prediction model: male age, duration of infertility, AMH, antral follicle count (AFC), DNA fragmentation index (DFI), number of follicles and E2 levels at HCG trigger, and total motile sperm count (TMC).

Results
Our prediction model was able to identify couples with low probability for pregnancy (20.5% of couples), upfront, (had less than 5% chance for pregnancy) that should be referred directly to IVF without any IUI cycle. Later after the first IUI cycle a second predictive score was calculated and identified couples who should be referred to IVF after one IUI cycle (26% of patients). Our predictive model was most accurate for females older than 40 years old (AUC=0.775).

Conclusions
Based on our results 20% of couples can be referred directly to IVF and 26% can be referred after the first cycle to save time, as pregnancy rate is low.