Real-world daily clinical use of liquid chromatography-tandem mass spectrometry (LC-MS/MS) for gonadal sex steroid measurement: comparison with chemoluminescent immunoassays in health and disorders of steroidogenesis

BACKGROUND: Immunoassay (IA) remains the most widely used method for serum testosterone (T) determination, requiring low sample volume and offering a simple, rapid, cost effective approach. However, liquid chromatography-tandem mass spectrometry (LC-MS/MS) is regarded as the gold standard for steroid hormone quantification. Real-world data comparing these methods for identifying low serum T in large, unselected clinical populations remain limited. AIM: To compare IA and LC-MS/MS for the diagnosis of biochemical male hypogonadism and to assess the proportion of patients correctly classified as hypogonadal by IA versus LC MS/MS. METHODS: We conducted a prospective real-world observational study including 1062 adult men (median age 53.4 [38.6–65.1] years) consecutively referred to the Endocrinology Unit of the University of Modena and Reggio Emilia (Baggiovara Hospital) between January 2022 and December 2025 for any clinical indication. Serum total testosterone (TT) was measured on the same fasting morning sample by IA (Access Total Testosterone, DxI800, Beckman Coulter) and LC-MS/MS (Chromsystems MassChrom® Steroids). SHBG was measured by IA, and calculated free testosterone (cfT) was derived by the Vermeulen formula using TT from both methods. Biochemical hypogonadism was defined as TT ≤3.5 ng/mL and cfT ≤6.5 ng/dL, per SIAMS (Società Italiana di Andrologia e Medicina della Sessualità) criteria. Data are reported as median (IQR). RESULTS: TT was significantly higher by LC-MS/MS [4.46 (2.98–6.16) ng/mL] than IA [3.60 (2.40–4.80) ng/mL] (p<0.001), with a strong correlation between methods (R²=0.842, p<0.001). Similarly, cfT was higher by LC-MS/MS [8.16 (5.52–11.68) ng/dL] than IA [6.30 (4.20–9.00) ng/dL] (p<0.001; R²=0.810, p<0.001). Bland-Altman analysis showed a proportional bias, larger for cfT (mean bias +2.17 ng/dL, LoA −6.63 to 10.97) than TT (+0.97 ng/mL, LoA −1.54 to 3.46). Applying SIAMS thresholds, IA classified 49.4% (n=524) of patients as hypogonadal by TT versus 33.5% (n=355) by LC-MS/MS (χ²=464.86, p<0.001); for cfT, IA classified 51.8% (n=405) versus 33.8% (n=264) by LC-MS/MS (χ²=324.75, p<0.001). Hypogonadal patients by IA had significantly higher odds ratio of also being classified as such by LC-MS/MS (OR 13.07, 95%CI 7.59–24.38 for TT; OR 18.63, 95%CI 9.21–43.94 for cfT), yet discordance remained substantial (McNemar's test, both p<0.001). 1 CONCLUSION: In this large real-world cohort, we demonstrate that IA systematically underestimated both TT and cfT compared with LC-MS/MS, leading to a substantially higher rate of hypogonadism diagnoses. IA overestimates male hypogonadism relative to LC-MS/MS, a discrepancy that may result in overdiagnoses and unnecessary testosterone replacement therapy in routine clinical practice, underscoring the need for method-specific diagnostic thresholds.