Home Publications Research Data Research Grants Inventions & Patents Awards Additional Research Activities Faculties & Institutions Research Areas

Multicenter comparison of LC-MS/MS, radioimmunoassay, and ELISA for assessment of salivary cortisol and testosterone

Dlugash G, Rauh M, Carré JM, Marcellus A, Plachecki S, Schultheiss O (2026)

---

Publication Type: Journal article

Publication year: 2026

Journal

Book Volume: 183

Article Number: 107618

DOI: 10.1016/j.psyneuen.2025.107618

Abstract

Introduction Salivary steroid assessment has become an essential part of human social neuroendocrinology, offering non-invasive, easy, and cost-effective measurements. Despite researchers’ preference for methods like enzyme-linked immunosorbent assay (ELISA) and radioimmunoassay (RIA) over the complex and costly liquid chromatography-mass spectrometry (LC-MS/MS), concerns about the validity of immunoassays (IAs) remain. The present study examines the convergence of LC-MS/MS, ELISA, and RIA in measuring salivary cortisol (C) and testosterone (T) and explores the contributions of intra-lab and inter-lab factors. Method Samples were collected from 81 men and 39 women in the morning and evening. Moreover, women provided samples for both the follicular and the luteal cycle phases. Natural hormone fluctuations (e.g., diurnal C decrease in the evening, T male-to-female ratio, influence of hormonal cycle phase, and hormonal contraceptive intake) and quality control samples were used as validity criteria for method evaluation. Over 336 samples and quality control samples were assayed by one RIA, two ELISA, and two LC-MS/MS methods across four labs. Correlational analyses were conducted to examine inter-lab x inter-method reliability, intra-lab x inter-method reliability, and inter-lab x intra-method reliability. Results For C and T, all methods demonstrated sufficient validity in detecting well-known natural fluctuations, with LC-MS/MS performing consistently best across all evaluated criteria. Nevertheless, ELISA did not achieve the expected male-to-female T ratio and tended to inflate estimated C and T levels, especially in the lower concentration range. Further, we found for all methods highly significant correlations with r ≥ .92 for C and with r ≥ .85 for T. However, when samples were divided by sex, correlations stayed comparable for C but decreased for T to r ≥ .71 in men and r ≥ .41 in women. Discussion LC-MS/MS was the best-performing method for both C and T across all criteria examined. RIA, despite showing slightly higher variance, can still be considered a reliable analytical technique as it met most of the set criteria for C and T. On the other hand, ELISA overestimated values, especially at low T levels. Therefore, caution should be exercised when selecting an appropriate method for the specific need.

Authors with CRIS profile

Involved external institutions

How to cite

APA:

Dlugash, G., Rauh, M., Carré, J.M., Marcellus, A., Plachecki, S., & Schultheiss, O. (2026). Multicenter comparison of LC-MS/MS, radioimmunoassay, and ELISA for assessment of salivary cortisol and testosterone. Psychoneuroendocrinology, 183. https://doi.org/10.1016/j.psyneuen.2025.107618

MLA:

Dlugash, Gelena, et al. "Multicenter comparison of LC-MS/MS, radioimmunoassay, and ELISA for assessment of salivary cortisol and testosterone." Psychoneuroendocrinology 183 (2026).

BibTeX: Download