Reference intervals for intact FGF 23 in healthy Korean adults: Lower concentrations in young adulthood require age-specific partitioning

Yonggeun Cho¹Hanmil Jang²Hyun-June Nam²Jaehyeok Jang²Hyein Kang³John Hoon Rim^2*Sang-Guk Lee²Jong-Baeck Lim^2*

• ¹Hallym University Sacred Heart Hospital, Anyang-si, Republic of Korea
• ²College of Medicine, Yonsei University, Seoul, Republic of Korea
• ³Keimyung University School of Medicine, Dalseo-gu, Republic of Korea

Background: Fibroblast growth factor 23 (FGF23) is a bone-derived phosphaturic hormone that is essential for phosphate homeostasis. Elevated FGF23 levels underlie FGF23-related hypophosphatemic rickets and tumor-induced osteomalacia. Despite its clinical importance, population-based reference intervals (RIs) for intact FGF23 using the widely deployed LIAISON XL automated chemiluminescent immunoassay platform (DiaSorin) are lacking for East Asian populations. Methods: We established method-specific RIs for intact FGF23 (iFGF23) in 386 healthy Korean adults (193 males and 193 females; age, 20–79 years) following the Clinical and Laboratory Standards Institute EP28-A3c guidelines. After the Box–Cox transformation and Horn's outlier detection, the RIs were derived using nonparametric methods (2.5th–97.5th percentiles). The necessity for partitioning was assessed using the Harris–Boyd method. Associations between iFGF23 levels and demographic, anthropometric, and biochemical parameters were examined using Pearson's correlation coefficients. Results: The overall nonparametric RI was 28.04–100.33 pg/mL (90% CI: 25.77–31.91 to 96.29–109.20). Age emerged as the primary determinant requiring partitioning, with young adults (20–29 years) exhibiting significantly lower concentrations than older adults (≥30 years): 25.73–78.76 versus 32.01–107.00 pg/mL. A sex-stratified analysis confirmed that this age effect persisted independently in both males and females. Although males had higher median iFGF23 than females (65.03 vs. 51.98 pg/mL, p < 0.001), Harris–Boyd analysis did not support sex-based partitioning (z = 5.17, z* = 5.38). Intact FGF23 was significantly correlated with age (r = 0.278), estimated glomerular filtration rate (r = –0.254), and alkaline phosphatase (r = 0.143; all p ≤ 0.005), but not with traditional mineral metabolism parameters (phosphate, calcium, parathyroid hormone, and 25-hydroxyvitamin D). Conclusions: This study provides the first population-and method-specific RIs for intact FGF23 in an East Asian population and establishes critical age-stratified benchmarks for clinical interpretation. The distinct RI in young adults underscores the necessity of age-appropriate reference standards for diagnosing and monitoring phosphate homeostasis disorders. These findings highlight the importance of population-specific reference data in the absence of assay harmonization.