THU645 Comparison Of LC-MS/MS Testosterone Assays Towards Standardization Of Lower Testosterone Ranges

Abstract

Abstract Disclosure: M.G. Cree: Consulting Fee; Self; Pollie, Inc. Research Investigator; Self; Eli-Lilly, Amino Co. S. Ottey: None. K.D. Sherif: None. O. Sugahara: None. F. Pokuah: None. A.N. Lyle: None. V. Hubert: None. Objective: Polycystic ovary syndrome (PCOS) is a common disorder in women. The diagnosis of PCOS often includes elevated concentrations of serum testosterone. Previous work has harmonized the higher range of the testosterone assay and normative cut-offs for low and high testosterone concentrations for the male gender. However, similar work is needed to better align existing Liquid Chromatography Tandem Mass Spectroscopy (LC-MS/MS) methods in the lower range of testosterone and determine upper limits of normal for females. Methods: This is a secondary analysis of a pooled cohort of 60 obese adolescents with and without PCOS, enrolled in 1 of three studies (NCT02157974; NCT03041129; NCT03717935) at the University of Colorado, Anschutz. None of the participants were taking hormonal medication to treat PCOS. Frozen plasma samples collected between 6-7 AM were analyzed by Esoterix hormone branch of LabCorp (Agoura Hills, CA) and by the Centers for Disease Control (CDC) Clinical Standardization Programs with LC-MS/MS. Samples were selected from the available participants to encompass the full range of testosterone concentrations available within the cohort. Results: Within the entire cohort (N=60, Age = 15.8, SD: ± 1.7 years, BMI 35.4, SD: ± 5.5 kg/m2), the means between the assays were statistically, but not clinically different (CDC 48.4 SD: ±24.4 ng/dL vs OL 45.4 SD: ±22.1, p=0.013).Deming regression indicated that the assays were not different with an intercept of 2.05 (-3.69 to 7.79, 95% confidence interval) and slope of 0.895 (0.753 to 1.04, 95% confidence interval). The two assays were highly correlated via linear regression (R2=0.859, p<0.001) and with a Bland-Altman comparison, the CDC assay only had a slight positive bias of 3.04, SD: ± 9.18 ng/dL relative to Esoterix. When divided by PCOS status (PCOS N=52, Hirsutism score 6.5±5.9, Control N=8, Hirsutism score 2.1±4.5), the CDC assay ROC demonstrated discrimination of PCOS status with an AUC of 0.903 (0.800-1.00), p<0.001 and the Youden index was <27.1 ng/dL with a sensitivity of 87.5% and a specificity of 86.5%. However, when the sensitivity was at 100%, the specificity was 57.7% and the cut-off <43.0 ng/dL. Conclusions: The Esoterix, a major commercial lab, assay is very similar to that of CDC LC-MS/MS, and able to discriminate PCOS status at 43.0 ng/dL. Future work is needed to 1) align more clinical assays to the CDC standard and 2) define a common cut-off to describe hyperandrogenism in females across the weight and age spectrum. Presentation: Thursday, June 15, 2023