Ongoing pregnancy rates in women with low and extremely low AMH levels. A multivariate analysis of 769 cycles.

BACKGROUND: The ideal test for ovarian reserve should permit the identification of women who have no real chance of pregnancy with IVF treatments consequent upon an extremely reduced ovarian reserve. The aim of the current study was to evaluate pregnancy rates in patients with low AMH levels (0.2-1 ng/ml) and extremely low AMH levels (<0.2 ng/ml) and to determine the cumulative pregnancy rates following consecutive IVF treatments. METHODS: We conducted an historical cohort analysis at a tertiary medical center. Serum AMH levels were measured at initial clinic visit and prior to all following treatment cycles in 181 women (769 cycles) with an initial AMH level ≤1 ng/ml, undergoing IVF-ICSI. Main outcome measures were laboratory outcomes and pregnancy rates. RESULTS: Seventy patients undergoing 249 cycles had extremely low AMH levels (≤0.2 ng/ml), whereas 111 patients undergoing 520 cycles had low AMH levels (0.21-1.0 ng/ml). Number of oocytes retrieved per cycle, fertilized oocytes and number of transferred embryos were significantly lower in the extremely low AMH levels group compared to the low AMH levels (P<0.003). Crude ongoing pregnancy rates were 4.4% for both groups of patients. Among 48 cycles of women aged ≥42 with AMH levels of ≤0.2 ng/ml no pregnancies were observed. But, in patients with AMH levels of 0.2-1.0 ng/ml, 3 ongoing pregnancies out of 192 cycles (1.6%) were observed. However, in a multivariate regression analysis adjusted for age and cycle characteristics, no significant differences in ongoing pregnancy rates per cycle between the two groups were evident. Cumulative pregnancy rates of 20% were observed following five cycles, for both groups of patients. CONCLUSIONS: Patients with extremely low AMH measurements have reasonable and similar pregnancy rates as patients with low AMH. Therefore, AMH should not be used as the criterion to exclude couples from performing additional IVF treatments.

Abstract licence: CC BY

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.