Age- and Sex-Specific Reference Intervals for Laboratory Parameters from Aging Trajectories in Healthy Beijing Adults (2010–2026)

Background Current reference intervals (RIs) applied uniformly across ages ignore physiological aging. Aims We aimed to characterize age-dependent biomarker trajectories, identify aging inflection points, and establish age- and sex-stratified reference intervals in apparently healthy Chinese adults. Methods We analyzed laboratory data from 4,392 healthy adults (1,894 males, 2,498 females; aged 18–91 years) across four survey waves (2010–2026). A three-tier non‑circular exclusion framework (infectious disease serology, cross‑system extreme‑value thresholds, and refineR statistical decomposition) was applied. Inflection points were detected via segmented regression with bootstrap validation. RIs (P2.5–P97.5) were derived nonparametrically with 90% bootstrap CIs and validated using the 2024 elderly cohort. Results Four consensus inflection points (~ 35, 50, 60, and 70 years) were identified; the 50–55 year peri‑menopausal inflection in women was the most consistent. Age‑stratified RIs deviated substantially from uniform RIs: 42 of 160 (26.3%) indicator‑sex‑age groups had out‑of‑range (OOB) rates > 20%, with CHOL in women aged 50–59 years reaching 58.8%. Cross‑year validation showed new RIs achieved closer‑to‑target OOB rates in 60.9% of groups. Significant secular trends included GGT + 35% and UA + 17% over 16 years, while LDL declined 18%; anchor indicators (MCV, MCH) remained stable (< 2% change). Discussion Our non-circular, age-differentiated exclusion framework avoids reproducing existing reference intervals and prevents physiological aging from being misclassified as pathology, supporting age-adapted interpretation. Conclusions Uniform RIs systematically misclassify age‑related physiological changes, particularly for lipids in peri‑ and post‑menopausal women. Our age‑ and sex‑specific RIs, grounded in aging trajectories, support periodic recalibration and age‑adapted interpretation to reduce misclassification and improve clinical decision‑making.