| 【OBJECTIVE】This study is to identify the predictors of ovarian hyperstimulation syndrome through analysis of relationship between the predictors of the mild-server ovarian hyperstimulation syndrome including patient’s age, antral follicle count(AFC), serum follicle-stimulating hormone levels(FSH), basic E2 level, serum oestradiol concentration on the day of human chorionic gonadotrophin(HCG) administration, oocytes retrieved with ovarian hyperstimulation syndrome among the patients who undergone in vitro fertilization and embryo transfer/intracytoplasmic sperm injection(IVF/ICSI). By exploring the predictive ability of them by ROC curve and analyzing the predictive value of the combination of indicators, we can work out an individualized ovarian stimulation method for patient, improve the clinical pregnancy rate and reduce the occurrence of ovarian hyperstimulation syndrome.【MATERIALS AND METHODS】A total of 218 patients who conducted in vitro fertilization(IVF) or intracytoplasmic sperm injection(ICSI) were selected between 2009 January and 2013 December in Reproductive Medicine Center of Shanghai First Maternity and Infant Hospital in this retrospective analyzed study. The patients were divided into case group and control group and compared the predictors among the two groups. On the day 2~4 of menses cycle, AFC was measured using transvaginal sonography and a total of 1 tube with 3~4ml peripheral venous blood was collected in patients, then sent Shanghai First Maternity and Infant Hospital laboratory for the measurement of FSH, LH, E2 level. Record the patient’s age, body mass index(BMI), the days of Gn, the total amount of Gn, E2 levels on the day of HCG administed, the number of occytes collected and so on. All data were processed by SPSS19.0 software; compare the basic characteristics and the treatment of the patients among the different groups using Chi-Square test. Then compare the diagnostic utility of the predictors by receiver operating curve and analyze the relationship between the predictors of ovarian hyperstimulation syndrome by Spearman correlation and Logistic Regression.【RESUITS】1. Clinical characteristics and treatment results of infertility patients among different groups.We have a total of 115 cycles in case group and 103 cycles in control group. Basal E2 level in the different groups was not significant difference(P>0.05). There was significant difference in the patient’s age, body mass index(BMI), AFC, the level of serum FSH, LH/FSH levels among the two groups(P<0.001). Gn total dose, E2 on the HCG day, normal fertilized oocytes in the two groups were also significant difference(P<0.001). Age, body mass index(BMI) and basal FSH levels of case group were significantly lower than that of control group, while AFC and LH/FSH levels were significantly higher in control group. And E2 on the HCG day and the number of occytes collected of case group were significantly higher than that of control group.2. Logistic regression analysis showed that AFC, E2 on the HCG day and the number of retrieved occytes were included into model to predict ovarian hyperstimulation syndrome.3. Relationship between the predictors of ovarian reserve, E2 on the HCG day and oocytes retrieved.Spearman correlation analysis showed that the level of E2 on the HCG day, AFC and serum LH/FSH were negatively correlated with the number of retrieved occytes, patient’s age and basal FSH levels were negatively correlated with the number of retrieved occytes, the correlation coefficients(r) were 0.615, 0.351, 0.266,-0.145,-0.227, P<0.001;Basic E2 levels had no significant correlation with retrieved occytes, r=-0.129, P=0.057.4. The evaluation of predictive value of ovarian hyperstimulation syndrome using indicators of ovarian reserve and treatment results.ROC curve analysis showed that AFC was good parameter to predict the ovarian response, followed by the basal LH/FSH. For the prediction of ovarian response, the area under the ROC curve(AUC) were 0.799(0.736-0.862), 0.688(0.617-0.759), respectively(all P<0.001). The cut-off value of AFC and the basal LH/FSH for ovarian response were AFC≥13.5,the basal LH/FSH≥0.865.E2 on the HCG day and the number of retrieved occytes were good parameters to predict the ovarian hyperstimualtion syndrome, the area under the ROC curve(AUC) were 0.924(0.891-0.958), 0.847(0.797-0.897), respectively(all P<0.001). The cut-off value of E2 on the HCG day and the number of retrieved occytes for ovarian hyperstimulation syndrome were E2 on the HCG day≥2966.60pg/ml, the occytes retrieved≥11.5.5. The evaluation of predictive value of ovarian hyperstimulation syndrome using indicators of AFC, E2 on the HCG day and the number of retrieved occytesCombining the three parameters can improve the prediction efficiency, and the area under the ROC curve(AUC) were 0.961(0.939-0.982), P<0.001; Combing E2 on the HCG day and AFC were good parameters to predict the ovarian hyperstimualtion syndrome, the area under the ROC curve(AUC) were 0.951(0.926-0.975), P<0.001; Combing E2 on the HCG day and the number of retrieved occytes were also good parameters to predict ovarian hyperstimulation syndrome, the area under the ROC curve(AUC) were 0.947(0.92-0.974), P<0.001;Combing the number of retrieved occytes and AFC were good parameters to predict ovarian hyperstimulation syndrome, the area under the ROC curve(AUC) were 0.895(0.855-0.936), P<0.001.In a word, combing the three parameters can improve the prediction efficiency, and the sensitivity and specificity were respectively 86.1% and 90.3%.【CONCLUSIONS】1. The level of serum FSH, LH/FSH, AFC, patient’s age, E2 on the HCG day and the number of retrieved occytes, all of them can predict ovarian hyperstimulation syndrome.2. As a single parameter, AFC, E2 on the HCG day and the number of retrieved occytes, which is better than serum LH/FSH, FSH and patient’s age. The cut-off value for AFC, E2 on the HCG day and the number of retrieved occytes to predict ovarian hyperstimulation syndrome were 13.5, 2966.60pg/ml, and 11.5 respectively.3. Combining the three parameters(AFC, E2 on the HCG day and the number of retrieved occytes) can improve the prediction efficiency of ovarian hyperstimulation syndrome, and the sensitivity and specificity were respectively 86.1% and 90.3%. |
PDF Full Text Request