Non-equivalence of anti-Müllerian hormone automated assays—clinical implications for use as a companion diagnostic for individualised gonadotrophin dosing

STUDY QUESTION: Can anti-Müllerian hormone (AMH) automated immunoassays (Elecsys(®) and Access) be used interchangeably as a companion diagnostic for individualisation of follitropin delta dosing? SUMMARY ANSWER: The Access assay gives systematically higher AMH values than the Elecsys(®) assay which results in over 29% of women being misclassified to a different follitropin delta dose. WHAT IS KNOWN ALREADY: Follitropin delta is the first gonadotrophin to be licenced with a companion diagnostic, the Roche Elecsys(®) AMH Plus assay. Alternative automated AMH assays including the Beckman Coulter Access immunoassay are considered to provide similar results, but clarification of their suitability as an off-licence companion diagnostic for follitropin delta is required. STUDY DESIGN, SIZE, DURATION: We systematically searched the existing literature for studies that had measured AMH using both automated assays in the same cohort of women. Individual paired patient data were acquired from each author and combined with unpublished data. PARTICIPANTS/MATERIALS, SETTING, METHODS: We identified five eligible prospective published studies and one additional unpublished study. A 100% response from the authors was achieved. We collected paired AMH data on samples from 848 women. Passing–Bablok regression and Bland–Altman plots were used to compare the analytical performance of the two assays. The degree of misclassification to different treatment categories was estimated should the Access AMH be used as a companion diagnostic instead of the Elecsys AMH in determining the dosing of follitropin delta. MAIN RESULTS AND THE ROLE OF CHANCE: The Passing–Bablok regression shows a linear relationship (Access = −0.05 + 1.10 × Elecsys). The Access assay systematically gave higher values by an average of 10% compared with the Elecsys assay (slope = 1.10, 95% CI: 1.09 to 1.12). The average of the difference between the two assays was 2.7 pmol/l. The 95% limits of agreement were −11.7 to 6.3. Overall 253 (29.3%) women would have received an inappropriate follitropin delta dose if the Beckman Coulter Access assay was used. Specifically, a substantial proportion of women (ranging from 49% to 90% depending on the AMH category) would receive a lower dose of follitropin delta based on the Access AMH assay. Up to 10% (ranging from 2.5% to 10%) of women with high ovarian reserve would have been misclassified to a greater dose of follitropin delta based on the Access AMH assay. LIMITATIONS REASONS FOR CAUTION: We compared the values of the two principal automated assays, extrapolation of our findings to other automated AMH assays would require similar comprehensive examination. WIDER IMPLICATIONS OF THE FINDINGS: An international standard for the calibration of the automated AMH assays is warranted to facilitate efficient use of AMH as a companion diagnostic. The variable calibration of alternative automated AMH assays may adversely impact on the performance of the follitropin delta dosing algorithm. STUDY FUNDING/COMPETING INTEREST(S): No formal funding has been received for this study. SI is funded by a UK Medical Research Council skills development fellowship (MR/N015177/1). SMN has received speakers fees, travel to meetings and participated in advisory Boards for Beckman Coulter, IBSA, Ferring Pharmaecuticals, Finox, Merck Serono, Merck and Roche Diagnostics. SMN has received research support from Ansh laboratories, Beckman Coulter, Ferring Pharmaceuticals and Roche Diagnostics. TRIAL REGISTRATION NUMBER: N/A.