Quantifying the intraindividual variation of antimüllerian hormone in the ovarian cycle

Abstract

To quantify intraindividual variability of antimüllerian hormone (AMH) as analytical and biological coefficients of variation and assess the effects of variation on clinical classification. Retrospective cohort study. Not applicable. Thirty-eight women referred by general practitioners. None. Total intraindividual variability (CVW), analytical (CVA) and biological variability (CVI) for each woman and for AMH ranges: low (<5 pmol/L), reduced (5-10), moderate (>10-30) and high (>30 pmol/L), with calculation of proportion of women crossing clinical cutoffs and expected variability around each cutoff. Cycling women (n = 38) contributed 238 blood samples (average 6 samples each). The average total intraindividual AMH variability was 20% (range: 2.1% to 73%). Biological variation was 19% (range: 0 to 71%) and at least twice the analytical variation of 6.9% (range: 4.5% to 16%). Reclassification rates were highest in women with low (33%) or reduced AMH (67%) levels. Expected variations around the 5, 10, and 30 pmol/L cutoffs were 3-7, 7-13, and 20-40 pmol/L, respectively. In a woman with mean AMH in the 10-30 pmol/L range, the span of results that could occur was 7-40 pmol/L. Total variation in AMH was 20%, and the majority of this was biological. Changes in AMH resulted in reclassification in 29% of women and occurred most frequently in those with low and reduced AMH. In cycling women, the variability in AMH should be considered by clinicians, especially if a result is close to a clinical cutoff.