ART 주기에서 AMH가 난소 반응 및 임신 결과에 미치는 이중적 영향 : 후향적 코호트 연구 = Dual Effects of AMH on Ovarian Response and Reproductive Outcomes in ART Cycles: a Retrospective Cohort Study

국문 초록 (Abstract)

가임기 여성의 항뮬러관호르몬（anti-Müllerian hormone, AMH）은 보조생식술（assisted reproductive technology, ART）에서 난소 예비력과 난소 반응을 예측하는 대표적인 지표로 활용되고 있다. 그러나 AMH가 난자 회수량 뿐만 아니라 이후 초기 배아 발달 과정에서의 체외수정 효율 및 임신낭（gestational sac, GS） 확인을 기준으로 판단되는 임상 임신율（clinical pregnancy rate, CPR） 결과까지 예측할 수 있는지는 아직 명확하지 않다. 본 연구에서는 ART 주기에서 AMH, 연령 및 체질량지수（body mass index, BMI）가 난소 반응, 정상 수정 배아 수 및 GS 기준 임상 임신 결과에 미치는 영향을 분석하고, AMH의 양적 효과와 생식 결과 간의 불일치 가능성을 평가하고자 하였다.
본 연구는 2018년부터 2023년까지 부산마리아 난임병원에서 시행된 난임 환자 대상 보조생식술 중 IVF/ICSI 주기를 대상으로 한 후향적 코호트 연구이다. 총 6,312건의 주기 중 AMH 또는 BMI 결측, 비ART 주기, BMI 〈18 또는 ≥45 kg/m²인 주기 등을 제외한 최종 4,399건을 본 연구에 분석하고자 활용하였다. 난소 반응과 정상 수정 배아 수의 관련 요인은 다변량 선형 회귀분석으로, 임상 임신의 예측 요인은 다변량 로지스틱 회귀분석으로 평가하였다. 분석 대상자의 평균 연령은 37.91 ± 5.24세, 평균 AMH는 2.09 ± 2.47 ng/mL, 평균 BMI는 22.70 ± 3.55 kg/m²였다. 평균 채취 난자 수는 10.67 ± 10.30개, 평균 정상 수정 배아 수는 6.80 ± 6.43개였으며, 전체 임상 임신율은 20.8%였다. 다변량 분석에서 AMH는 채취 난자 수와 강한 양의 연관성을 보였으나（β = 2.348, P 〈 0.001）, 정상 수정 배아 수와는 작지만 유의한 음의 연관성을 보였다（β = −0.044, P = 0.027）. 임상 임신에 대한 로지스틱 회귀분석에서는 연령 증가（OR = 0.920, P 〈 0.001）와 AMH 증가（OR = 0.955, P = 0.010）가 임상 임신 가능성 감소와 유의하게 관련되었으며, BMI는 유의한 관련성을 보이지 않았다.
특히, AMH 10 ng/mL은 매우 높은 난소 예비력 또는 과도한 난소 반응을 시사할 수 있는 기준으로, 본 연구에서는 AMH의 양적 효과와 생식 결과 간의 차이를 보다 명확히 평가하기 위해 AMH 〈10 ng/mL군과 AMH ≥10 ng/mL군으로 구분하여 하위 그룹 분석을 시행하였다. AMH 수치에 따른 하위 그룹 분석에서 AMH ≥10 ng/mL군은 AMH 〈10 ng/mL군보다 채취 난자 수가 많았으나（34.41 ± 19.57 vs. 11.85 ± 9.50, P 〈 0.001）, 수정률은 낮았고（0.561 ± 0.214 vs. 0.694 ± 0.531, P 〈 0.001）, 임상 임신율도 낮았다（CPR: 5.7% vs. 23.1%, P 〈 0.001）. 연령에 따른 분석에서는 35세 미만군이 35세 이상군보다 채취 난자 수, 정상 수정 배아 수 및 임상 임신율이 유의하게 높았다.
결론적으로, AMH는 ART 주기에서 난소 반응과 난자 회수량을 예측하는 유용한 양적 지표이나, 수정 효율과 임상 임신 결과를 단독으로 설명하기에는 한계가 있음을 확인하였다. 본 연구 결과는 AMH가 난자의 양적 특성을 반영하는 지표로는 유용하지만 생식 능력 또는 임신 가능성을 직접 반영하는 지표로 해석되어서는 안 되며, ART 결과 예측에는 연령, 수정 결과 및 배아 관련 요인을 함께 고려해야 함을 시사한다.

다국어 초록 (Multilingual Abstract)

The relationship between anti-Müllerian hormone （AMH）, body mass index （BMI）, and age in infertility treatment is known to be closely interrelated. In particular, AMH is used as a marker of ovarian reserve and is recognized as an important predictor of ovarian response during controlled ovarian stimulation. The independent and combined effects of AMH, age, and BMI on ovarian response and fertilization outcomes have not been fully elucidated, and this study was conducted based on the hypothesis that AMH may have a dual role-affecting both quantitative aspects （oocyte yield） and qualitative aspects （post-fertilization outcomes）. The study was conducted based on IVF outcomes performed at Busan Maria Fertility Hospital from 2018 to 2023. A total of 4,399 cycles were analyzed. The mean maternal age was 37.91 ± 5.24 years, and the mean AMH level was 2.09 ± 2.47 ng/mL. The mean BMI was 22.70 ± 3.55 kg/m². The mean number of oocytes retrieved per cycle was 10.67 ± 10.30, the mean number of normally fertilized embryos （2PN） was 6.80 ± 6.43, and the overall clinical pregnancy rate was 20.8%. In the multivariable linear regression analysis for oocytes retrieved, AMH showed a strong and independent association with ovarian response （β = 2.348, SE = 0.051, P 〈 0.001）. Maternal age showed a significant inverse association with the number of oocytes retrieved （β = −0.285, P 〈 0.001）, whereas BMI showed no significant association （P = 0.477）. In the multivariable model for normally fertilized embryo count （2PN）, AMH showed a small but statistically significant negative independent association （β = −0.044, SE = 0.020, P = 0.027）. Maternal age also showed a negative correlation （β = −0.019, P = 0.019）, while BMI showed no significant correlation. In contrast, the number of oocytes retrieved （β = 0.552, P 〈 0.001） and the day-1 embryo transfer success rate （β = 0.076, P 〈 0.001） were the strongest predictors of 2PN embryo count. Multivariable logistic regression analysis revealed that increasing maternal age was significantly associated with a reduced probability of clinical pregnancy （OR 0.920, 95% CI 0.906–0.934, P 〈 0.001）, and higher AMH levels were independently associated with a decreased probability of clinical pregnancy （OR 0.955, 95% CI 0.923–0.989, P = 0.010）. In contrast, BMI （OR 1.009, 95% CI 0.989–1.030, P = 0.366） and 2PN embryo count （OR 1.010, 95% CI 0.997–1.023, P = 0.143） were not significantly associated with clinical pregnancy. The predictive performance of the logistic regression model was moderate （AUC = 0.722）, indicating acceptable discriminatory ability but limited explanatory power （McFadden's R² = 0.119）. AMH demonstrated a dual effect in ART cycles: a strong positive association with ovarian response, but a negative association with fertilization efficiency and clinical pregnancy rate. In contrast, as maternal age increased, ovarian response, fertilization outcomes, and clinical pregnancy rate consistently declined. BMI showed no significant association with any of the evaluated outcomes. These findings demonstrate a quantitative-qualitative dissociation of AMH, meaning that an increase in oocyte number does not necessarily lead to improved reproductive potential.
In the subgroup analysis stratified by AMH level （〈 10 ng/mL vs. ≥ 10 ng/mL）, women with AMH ≥ 10 ng/mL had significantly higher oocyte yields than those with AMH 〈 10 ng/mL （34.41 ± 19.57 vs. 11.85 ± 9.50, P 〈 0.001）. However, the fertilization rate was significantly lower in the high-AMH group （0.561 ± 0.214 vs. 0.694 ± 0.531, P 〈 0.001）. The clinical pregnancy rate was also markedly lower in women with AMH ≥ 10 ng/mL compared to those with AMH 〈 10 ng/mL （5.7% vs. 23.1%, P 〈 0.001）. Women with AMH ≥ 10 ng/mL were significantly younger than those with AMH 〈 10 ng/mL. When patients were divided into two groups based on the threshold of 35 years, AMH levels were significantly lower in the older group. However, when maternal age was analyzed as a continuous variable, no clear linear relationship was observed. These findings suggest that AMH does not decline strictly linearly with age, but may vary according to individual ovarian reserve characteristics. When classified by maternal age （〈 35 vs. ≥ 35 years）, younger women had significantly higher oocyte yields （16.02 ± 11.83 vs. 10.49 ± 8.81, P 〈 0.001） and normally fertilized embryo counts （9.31 ± 7.29 vs. 6.23 ± 5.65, P 〈 0.001）. The clinical pregnancy rate was also significantly higher in women under 35 years of age （29.1% vs. 19.9%, P 〈 0.001）. In contrast, the fertilization rate was higher in women aged ≥ 35 years （0.709 ± 0.568 vs. 0.643 ± 0.419, P 〈 0.001）. The clinical pregnancy rate was significantly higher in women under 35 years compared to those aged ≥ 35 years （29.1% vs. 19.9%, P 〈 0.001）, representing an approximately 31% relative reduction in pregnancy rate in the older age group. Therefore, this study suggests that although AMH is a strong predictor of ovarian response, elevated AMH levels, particularly ≥10 ng/mL, are not associated with improved fertilization efficiency or increased clinical pregnancy rates in ART cycles. These findings support the concept of a quantitative-qualitative discordance, in that an increase in oocyte number does not necessarily indicate enhanced reproductive potential. Prospective studies with larger cohorts are needed to clarify the clinical implications of this discordance. AMH showed a strong positive association with ovarian response （β = 2.348, P 〈 0.001）, while age showed a significant negative association. In contrast, after adjusting for age, BMI, number of oocytes retrieved, and day 1 embryo transfer success rate, AMH showed a small but significant negative independent association with 2PN embryo count （β = −0.044, P = 0.027）. In the subgroup analysis, cycles with AMH ≥ 10 ng/mL showed markedly increased oocyte yields compared to cycles with AMH 〈 10 ng/mL, but significantly lower fertilization rates （0.561 vs. 0.694, P 〈 0.001） and clinical pregnancy rates （5.7% vs. 23.1%, P 〈 0.001）.
In the logistic regression analysis, increasing age and AMH levels were independently associated with a lower probability of clinical pregnancy, whereas BMI and the number of 2PN embryos were not significantly associated with clinical pregnancy. Although AMH is a strong predictor of ovarian response, its quantitative advantage did not translate into improved fertilization efficiency or higher clinical pregnancy rates. These findings suggest that, while AMH is a meaningful marker of ovarian reserve, it should not be interpreted as a direct indicator of reproductive potential or the likelihood of pregnancy. The prediction of ART outcomes should consider maternal age, fertilization outcomes, and embryo-related factors in addition to AMH levels.

목차 (Table of Contents)

• I. 서 론 1
• II. 재료 및 방법 4
• 1. 재료 4
• 1.1. 연구 설계 4
• 1.2. 연구 대상 4
• 2. 연구 방법 6
• 2.1. 난소 자극법 및 난자 회수 6
• 2.2. 수정 방법 및 배아의 생성 7
• 2.3. 배아의 판단과 이식 8
• 2.4. 임신의 판정 8
• 2.5. 통계처리 9
• Ⅲ. 결 과 10
• 1. 연구 대상 및 기본 특성 10
• 2. AMH의 양적 효과: 난소 반응 및 정상 수정 배아 수와의 연관성 11
• 2.1. AMH와 난소 반응 및 수정률의 연관성 11
• 2.2. AMH, 연령 및 BMI가 ART 결과에 미치는 단계별 영향 14
• 3. AMH의 생식 결과에 대한 상반된 효과 17
• 3.1. 임상 임신（GS） 예측인자 17
• 3.2. AMH 수치에 따른 ART 결과 비교 20
• 3.3. 연령에 따른 ART 결과 비교 23
• 3.4. AMH의 양적-질적 불일치 요약 26
• Ⅳ. 고 찰 29
• 참고문헌 32
• 영문초록 38