The Application Of AMH In Ovarian Response And The Individualization Of FSH Starting Dose In High Ovarian Response

Objective:To calculate an AMH predictive value for low and high ovarian response,explore the characteristics of high ovarian response.To elaborate a multiple linear regression formula based on ovarian response markers for the calculation of the appropriate starting dose of Gn(Gonadotrophins)especially FSH(follicle-stimulation hormone)in in vitro fertilization cycles/intracytoplasmic sperm injection(IVF/ICSI)cycles.Methods:We retrospectively analyzed cycles in the database of the 2nd affiliated hospital of Kunming Medical University between January of 2015 to December of 2016 which contain clinical and laboratory information on IVF/ICSI treatment cycles.Cycles were selected if all the following inclusion criteria were satisfied:(ⅰ)fresh IVF/ICSI cycle attempt,(ⅱ treatment with a long gonadotrophin-releasing hormone agonist(GnRH-a)protocol or antagonist(GnRH-A)protocol,(ⅲ)complete records of patients on IVF cycle characteristics and clinical/laboratory data collected.Cycles were excluded if one of the following criteria met:(ⅰ)previous ovarian surgery,(ⅱ)presence of ovarian cysts,(ⅲ)use of hormonal medicine in the previous 3 months,(ⅳ)hyperprolactinemia and thyroid function abnormality.Based on the above inclusion and exclusion criteria,a total of 1667 cases were selected for the study.All patients included were divided into three groups as low ovarian response(Bologna diagnosis criteria),normal ovarian response and high ovarian response based on oocytes retrieved.The levels of AMH,FSH,E2 and T were measured 1-3 days of menstruation before controlled ovulation stimulating.The age,weight,infertility was recorded.Then we find out the AMH predictive value of high and low ovarian response through ROC curve.Next,we regrouping the cases into low(Bologna diagnosis criteria),high and normal ovarian response group according to the AMH predictive value above,analyzing the differences between the different groups.Finally,multiple liner regression were applied for a formula in predicting FSH starting dose of high ovarian response people,and the differences between the formula result and the actual clinical dosage was examined in 707 high ovarian response cases.Results:An AMH of 3.86ng/ml and 0.99ng/ml was the cut-off value by ROC curve,which indicated that the prediction value of AMH in high and low ovarian response may be 3.86ng/ml and 0.99ng/ml respectively.Differences were observed among three groups in age,basal FSH,AMH,Gonadotrophin days,number of oocytes retrieved,pregnancy outcome and the cancellation of transfer(P<0.05).Moreover,OHSS was the top cause of cancellation in embryo transfer of high ovarian response people.No clear differences were found in BMI,the thickness of endometrium in HCG day,the maturation of oocytes and the rate of high quality embryo(P>0.05).707 cycles of high response group were analyzed for multiple linear regression.The results turned out that age,FSH were positively correlated with FSH dose while AMH were negatively correlated.T showed a predictive value for FSH starting dose in high ovarian response population.However,there were no correlation between BMI and FSH starting dose as the dependent variable.On the basis of analysis result above,we defined age,FSH,T and AMH as independent variable and FSH starting dose as dependent variable to conduct the multiple linear regression.We finally elaborated a predictive formula as the FSH starting dose=4.034*age+1.237*FSH-0.124*T-2.442*AMH+52.755,comparing the actual dose and predictive dose.Conclusion:1.Serum AMH can be an isolated predictor of ovarian response and the result of this research can be applied on evaluating ovarian response of patients in this center.2.Patients with high ovarian response have an increasing cycle cancellation rate due to OHSS.3.Multiple liner regression formula can be used for predicting and calculating the FSH starting dose of potential high ovarian response population,which may help the individualization and standardization of COH doses.4.This study method can be extended for the research of low and normal ovarian response.