Congression and regression tree analyses on the outcome of ovulation induction in women with polycystic ovary syndrome
Wang, Y.; Li, M.; Zhao, Y.
Zhonghua fu chan ke za zhi 35(10): 594-596
2000
ISSN/ISBN: 0529-567X PMID: 11372408 Document Number: 512289
To investigate the risk factors for failure of ovulation induction in polycystic ovary syndrome (PCOS) patients and establish congression and regression tree (CART) model. Sex hormones were measured in 103 PCOS patients and 31 healthy women by radioimmunoassay, oral glucose tolerance test (OGTT) and insulin release test were done as well. Three ovulation induction protocols including clomiphene (CC)/chorionic gonadotrophin (hCG), CC/menotrophins (hMG)/hCG and gonadotrophin releasing hormone agonist(GnRH-a)/hMG/hCG were used for PCOS related infertility. Transvaginal ultrasonography was used for monitoring. The correlation between hormone levels and clinical outcomes was analyzed by CART method. The ovulation rate in patients with 2-hour blood insulin level < or = 68.9 mIU/L during OGTT was 86.9%, which was significantly higher than those with > 68.9 mIU/L (61.9%, P < 0.05). In the model, the insulin and glucose level of 0 min, 60 min during OGTT acted obviously, but estradiol and basal T level did not, the protocol of ovulation induction was not able to enter the model. Two hour insulin level during OGTT is the most important factor affecting the outcome of ovulation induction, while the others parameters of glucose and insulin level have useful action too.