A patient-specific model combining antimüllerian hormone and body mass index as a predictor of polycystic ovary syndrome and other oligo-anovulation disorders.

Objective: To determine whether a patient-specific predictive model combining antimüllerian hormone (AMH) levels and body mass index (BMI) can aid in the diagnosis of polycystic ovary syndrome (PCOS) and other ovulatory dysfunction disorders (OVDYS) among infertile women.

Design: Retrospective cohort study.

Setting: Academic fertility center.

Patient(s): One thousand and ten infertile women undergoing 3,160 intrauterine insemination (IUI) cycles, stratified by diagnosis in three groups: PCOS, OVDYS, and other etiologies.

Intervention(s): Ovulation induction followed by IUI or ultrasound-monitored natural cycles.

Main Outcome Measure(s): The probability of either PCOS or OVDYS diagnosis based on AMH levels alone and a patient-specific predictive model that combines serum AMH and patient's BMI.

Result(s): Median and interquartile range (IQR) for the serum AMH levels (ng/mL) were the highest in women with PCOS, and lowest in those with other infertility causes. Overall, for every 1 ng/mL increase in AMH, the odds of PCOS and OVDYS versus other causes increased by 55% and 24%, respectively. Postestimation from multivariate logistic regression models showed that PCOS diagnosis can be predicted with lower AMH values in women with a higher BMI compared with the AMH values predicting PCOS in normal-weight or underweight patients. The receiver operating characteristic curves reinforced these findings, and the best cutoffs for PCOS diagnosis were 7.5, 4.4, and 4.1 ng/mL for women belonging to the BMI groups 18.5-24.9, 25.0-29.9, and ≥30.0 kg/m, respectively.

Conclusion(s): Taking into account AMH and BMI, we developed a model that predicts the probability of an oligo-anovulation diagnosis, thus facilitating patient-specific counseling in the infertility setting.