| Objective:Anti-mullerian hormone(AMH)is one of the best assessment methods for evaluating ovarian function,predicting fertility,ovarian reactivity and guiding assisted reproductive drugs.In the clinical practice of assisted reproduction,for infertility patients with different pathological conditions,such as polycystic ovary syndrome(PCOS),decreased ovarian reserve function(DOR)or poor response to ovarian stimulation(POR),drug pretreatment is often used to improve the success rate of assisted reproduction and reduce the risk of relative complications.At present,more and more clinical studies have reported that the pre-treatments of assisted reproduction have effects on the serum AMH level in women,which leads to the risk of clinical misinterpretation.However,the relevant research reports are scattered,the sample size is small,the research results are biased,and there is a lack of thematic evidence-based analysis.In this paper by the method of systematic review and meta-analysis,evidence-based investigation was done on three kinds of drugs(COCs:combined oral contraceptives;MET:metformin;and DHEA:dehydroepiandrosterone)as a pretreatment before assisted reproduction,and the corresponding changes of serum AMH level to help clinically assess the AMH information obtained correctly and effectivelyMethods:Systematic search of various internet sources,Web of Science,Cochrane and Medline databases that published about the three types of assisted reproductive pretreatment drugs(COCs,MET and DHEA)and the corresponding changes in blood AMH levels clinical research articles until to Jan 20,2020.According to the internationally-required systematic review and meta-analysis(PRISMA)statement requirements to include the literature,the quality of the literature was evaluated according to the New-Ottawa Scale(NOS),the study characteristics and relevant data were extracted,and Revman5.3 and Stata 12.0 were used for analysis.Results:1.Thirty-six studies were eventually included in the systematic review and meta-analysis of this study.There were 34 self-controlled studies and 2 cohort studies.2.As to changes in serum AMH level post COCs pre-treatment in women with normal ovarian function(before assisted reproduction):Meta-analysis of all 8 groups of data(n=275)in random effect model showed that the pre-treatment(7 days to 6 months)resulted in a significant trend of serum AMH level decline(SMD:-0.38,95%CI[-0.78,0.02];P=0.06).The decline median:23.15%;95%CI[-45.4%,+4.1%].The heterogeneity among the sample data was large(I2=74%,P=0.0003).The relative subgroup meta analysis of 6 groups of self-control study data(n=91)showed that the heterogeneity of literature data(I2=65%)was significantly improved,but the serum AMH level showed no statistically significant decline(SMD:-0.37,95%CI[-0.89,0.15];Z=1.38;P=0.17).Further subgroup meta analysis of three groups of COCs data pretreated continuously(7 days to 9 weeks)showed that there was little inter-group heterogeneity(I2=0%,P=0.62)and still significant decline in serum AMH level(SMD:-1.42,95%CI[-2.57,-0.27]);Z=2.42,P=0.02).3.As to changes in serum AMH level after COCs pretreatment in PCOS patients(pre-assisted reproductive):Meta-analysis of all 9 groups of pretreatment data(n=200)in random effect model showed that serum AMH level declined after pretreatment(SMD:-1.34,95%CI[-2.04,-0.65]),which was statistically significant(Z=3.78;P=0.0002);The decline median:-26.1%;95%CI[-40.4%,-13.2%].The heterogeneity among sample data was large(I2=89%,P=0.00001).The relative subgroup meta analysis of COCs with antiandrogenic components(CPA:cyproterone acetate)in 3 groups(n=99)showed that there was no statistically significant decrease in serum AMH level(SMD:-1.64,95%CI[-4.12,0.83].Z=1.30;P=0.19).The relative subgroup analysis of 6 groups(n=101)of COCs without CPA(2-12 cycles)showed that the serum AMH level decreased significantly(SMD:-1.83,95%CI[-2.84,-0.81]Z = 3.54;P=0.0004).Furthermore,meta analysis of the subgroup with COCs(4 groups of data,n=89)in the shorter term(2-4 cycles)showed that the serum AMH level had a statistically significant decrease(SMD:-1.44,95%CI[-2.73,-0.15].I2=90%,P=0.00001;Z=2.19;P=0.03).The decline median:-25.2%;95%CI[-42.0%,-6.2%].And further meta analysis of the subgroup with COCs(5 groups of data,n=108)for a longer term(6～12 cycles)showed that the serum AMH level also decreased statistically significantly(SMD:-1.32,95%CI[-2.32,-0.33].I2=90%,P=0.00001;Z=2.60;P=0.009).The decline median:-41.8%;95%CI[-57.2%,-0.6%].Finally,based on the systematic analysis of the corresponding two groups of continuous dynamic observation data provided by Fabregues F et al.(2011),it was shown that the reduction of serum AMH level after 3 artificial cycles of COCs pretreatment(-24.2%)was significantly lower than that of 6 artificial cycles of COCs pretreatment(-52.0%).The decrease of serum AMH level in the two groups was statistically significant(P<0.05).4.As to changes in serum AMH level after MET pretreatment before assisted reproductive therapy in PCOS patients:the meta-analysis of the random effects model of all 14 groups of data(n=269)showed that MET pretreatment(2-12 months)could induce a significant decrease in serum AMH level WMD:-1.17,95%CI[-1.59,-0.76];I2=76%,P=0.00001),which was statistically significant.The decline median:-16.9%;95%CI[-25.0%,-10.8%].The subgroup meta analysis of patients who were overweight(body mass index≥25kg/m2)after corresponding pretreatment(10 groups of data,n=201)showed a significant decrease in serum AMH level(WMD:-0.96,95%CI[-1.28,-0.65];I2=57%,P=0.01),which was statistically significant(Z=5.98;P=0.00001).The decline median:-16.9%;95%CI[-23.0%,-11.8%].The meta analysis of the subgroup of the corresponding short-term(2-4 months)MET pretreatment(7 groups of data,n=131)showed that the induced decline of serum AMH level(WMD:-1.56,95%CI[-2.17,-0.94].I2=83%,P=0.00001)was statistically significant(Z=4.97;P=0.00001).The decline median:-22.0%;95%CI[-32.9%,-14.4%].The relative meta analysis in the subgroup of short-term MET pretreatment in patients with overweight PCOS(5 groups,n=88)showed that the serum AMH level induced significantly decreased(WMD:-1.10,95%CI[-1.50,-0.69].I2=59%,P=0.05;Z=5.31;P=0.00001),which was statistically significant.The decline median:-16.9%;95%CI[-29.9%,-9.2%].The heterogeneity of sample data(I2=59%)was significantly improved compared with that of all 15 groups(I2=89%)mentioned above.Finally,the meta analysis of the subgroup of PCOS patients with MET over a longer period(6-12 months;7 groups of data,n=138)showed that there was no statistically significant decrease in serum AMH level(WMD:-0.75,95%CI[-1.15,-0.35];I2=28%,P=0.22;Z=3.71;P=0.0002).Accordingly,long-term MET pretreatment was not statistically significant in overweight patients(6 groups of data,n=103),and the decrease of serum AMH level was not statistically significant(WMD:-0.79,95%CI[-1.25,-0.32];I2=44%,P=0.13;Z=3.31,P=0.0009).5.As to changes of serum AMH level after DHEA pretreatment in DOR/POR patients:Random effect model meta-analysis of all 14 groups of data(n=536)showed that serum AMH level significantly increased after DHEA pretreatment(6 weeks to 6 months)(SMD:+0.84,95%CI[+0.48,+1.21].I2=87%,P=0.00001),which was statistically significant(Z=4.58;P=0.00001).The rising median:46.1%;95CI[35.1%,94.5%].The meta analysis of the subgroup of the corresponding pretreatment(7 groups of data,n=262)of the elderly patients showed that the heterogeneity of the sample data(I2=67%)was significantly improved compared with that of all the 14 groups of DHEA pretreatment data(I2=87%)above.The serum AMH level was significantly increased(SMD:+0.54,95%CI[+0.21,+0.86].I2=67%,P=0.006),which was also statistically significant(Z=3.25;P=0.001).The rising median:26.5%;95%CI[11.3%,80.9%].Meta analysis of the subgroup of the short-term(6 weeks to 3 months)DHEA pretreatment(11 groups of data,n=431)showed that the heterogeneity of sample data(I2=53%)was significantly improved compared with that of all 14 groups(I2=87%)mentioned above.Serum AMH levels were significantly increased(SMD:+0.51,95%CI[+0.30,+0.72].I2=53%,P=0.002),which was statistically significant(Z=4.82;P=0.00001).The rising median:33.0%;95%CI[24.5%,83.8%].The meta analysis of the subgroup pretreated with DHEA over a longer period(4-6 months;3 groups of data,n=105)showed that the induced serum AMH level was significantly increased(SMD:+1.94,95%CI[+0.77,+3.11].I2=91%,P=0.0001),which was also statistically significant(Z=3.26;P=0.001).6.Meta-sensitivity analysis of relevant literature sample data:The corresponding meta-sensitivity analysis of the four sets of data of the above three types of pretreatment drugs(COCs,MET and DHEA)all showed that the point estimates of the combined effect amount after any study were removed Within the 95%confidence interval of the total combined effect;after excluding any study,there is no significant difference between the combined effect and the total combined effect7.Bias analysis of relevant literature sample data report:Bias analysis of the corresponding literature sample data report of the above three types of pretreatment medications(COCs,MET,and DHEA)shows that:except for normal ovarian function,the pretreatment serum AMH level of COCs in women The funnel chart of the change data is more symmetrical,and the corresponding Egger test bias P>0.05,the funnel charts of the other three groups of data are all asymmetrical,and the corresponding Egger test bias P<0.05Conclusion:The three types of pretreatments(COCs,MET and DHEA)commonly used before assisted reproduction had effects on serum AMH levels in different aspects.Among them● There is no certain effect of COCs pretreatment on serum AMH level in women with normal ovarian function● The effect of COCs pretreatment on serum AMH level in PCOS is the descending effect,and there is a progressive downward trend.The trend was independent of COCs components(such as CPA with anti-androgen effect)● The effect of MET pretreatment on serum AMH level in PCOS patients is also a descending effect.The lowering effect was obvious in the short term(2～4 months)of MET pretreatment.In the corresponding overweight patients,also appear.The long-term(≥6 months)pretreatment of MET didn’t show the corresponding descending effect advantage● The effect of DHEA pretreatment on serum AMH level of DOR/POR patients is the enhancement effect,and short-term(6 weeks to 3 months)conventional medication pretreatment has show a significant increase in serum AMH levels,also a good improvement effect in the use of older women(≥35 years old)womenIn the clinical practice of assisted reproduction,it is necessary to consider the obvious influence of relevant pretreatment on serum AMH level,to correctly interpret and make good use of the serum AMH detection report,and to give play to its true clinical effect. |
