RESUMEN
STUDY QUESTION: Is a low (<1.0 µg/L) or moderately low (1.0-1.9 µg/L) serum anti-Müllerian hormone (AMH) level a risk factor for early pregnancy loss in IVF/ICSI with a fresh or frozen-thawed embryo transfer (ET)? SUMMARY ANSWER: A low or moderately low serum AMH level does not associate with miscarriage, non-visualized pregnancy loss or overall early pregnancy loss rate in the IVF/ICSI treatment. WHAT IS KNOWN ALREADY: Low AMH predicts poor ovarian response and small oocyte yield in IVF/ICSI treatment, but its value in the evaluation of live birth rate (LBR) is modest. Little is known about the risk of early pregnancy loss in ART among women with low AMH. STUDY DESIGN, SIZE, DURATION: A retrospective cohort study on 1383 women undergoing their first oocyte retrieval for IVF/ICSI in Helsinki University Hospital in Helsinki, Finland, between 2012 and 2016, with all associated fresh (n = 1315) and frozen-thawed (n = 1418) ET cycles finished by August 2018. AMH was measured within 12 months before the IVF/ICSI stimulation. PARTICIPANTS/MATERIALS, SETTING, METHODS: Of all the women, 235 (17.0%) had low (<1.0 µg/L), 278 (20.1%) had moderately low (1.0-1.9 µg/L) and 870 (62.9%) had normal (≥2.0 µg/L) AMH. The primary outcomes were miscarriage, non-visualized pregnancy loss and early pregnancy loss (miscarriage and non-visualized pregnancy loss combined) after fresh or frozen-thawed ET. The impact of AMH on these outcomes was calculated in three populations: among all women who became pregnant, among women with AMH ≤6.0 µg/L and in a population weighted by the inverse probability of becoming pregnant (inverse probability weighting, IPW). The impact of AMH was also assessed on the secondary outcomes, cumulative pregnancy rate (cPR) and cumulative live birth rate (cLBR) across all ET cycles in the woman's first IVF/ICSI. Potential confounders (the woman's age, overweight, smoking, history of endometriosis and underlying medical conditions) adjusted the final results. MAIN RESULTS AND THE ROLE OF CHANCE: Of 1123 pregnancies, 285 (25.4%) ended in non-visualized pregnancy loss and 143 (12.7%) in miscarriage. The LBR was 24.6% per ET (673/2733). Low or moderately low AMH, compared with normal AMH, did not associate with miscarriage or non-visualized pregnancy loss in analyses among all women who became pregnant (adjusted relative risk (RR) for miscarriage vs live birth, 0.70 and 95% CI 0.42-1.17 in low AMH and adjusted RR, 1.00 and 95% CI, 0.68-1.49 in moderately low AMH; adjusted RR for non-visualized pregnancy loss vs live birth, 0.90 and 95% CI, 0.65-1.23 in low AMH and adjusted RR, 1.09 and 95% CI 0.85-1.41 in moderately low AMH), nor did low or moderately low AMH associate with the overall early pregnancy loss rate (adjusted RR for early pregnancy loss vs live birth, 0.86 and 95% CI, 0.68-1.10 in low AMH and adjusted RR, 1.01 and 95% CI, 0.86-1.27 in moderately low AMH). Results remained similar after restricting the analysis to women with AMH ≤6.0 µg/L. Women with low or moderately low AMH had fewer pregnancies and live births than women with normal AMH in their first IVF/ICSI (cPR/cLBR in women with low AMH 50.6/34.0%, moderately low AMH 59.0/36.3% and normal AMH 68.3/49.2%). When the lower probability for pregnancy was considered by using IPW, women with low or moderately low AMH did not have a higher risk for miscarriage, non-visualized pregnancy loss or overall early pregnancy loss compared with women with normal AMH. LIMITATIONS, REASONS FOR CAUTION: The number of miscarriages in women with low AMH was moderately small, limiting the power of the study. The real-world clinical setting of the study restricted the ability to control for all factors causing selection bias. WIDER IMPLICATIONS OF THE FINDINGS: The cLBR was higher among women with normal AMH than among women with low or moderately low AMH in their first IVF/ICSI treatment because these women had more oocytes and embryos. Women with low or moderately low AMH did not have an increased risk for early pregnancy loss. This information is reassuring for couples and useful in counseling. These results are also valuable when assessing the overall effectiveness of IVF/ICSI treatment. STUDY FUNDING/COMPETING INTEREST(S): Research funds from Helsinki University Hospital (no. TYH2018232), Hyvinkää Hospital (no. M3080TUT18) and the Emil Aaltonen Foundation for P.P. Grants from the Paulo Foundation and the Finnish Medical Foundation for H.H. The authors report no conflicts of interest. TRIAL REGISTRATION NUMBER: HUS/138/2017.