Population-based reference equations and Z-scores for blood biomarkers of neurodegenerative diseases

Abstract

Background Blood biomarkers are increasingly used to support the diagnosis and monitoring of neurodegenerative diseases. However, their interpretation is complicated by physiological determinants, including age, sex, body-mass index, and renal function, and by differences in absolute concentrations between analytical methods. We aimed to develop population-based reference equations allowing individualized interpretation of the main blood biomarkers used in neurology. Methods In this cross-sectional study, we analysed plasma samples from cognitively unimpaired participants selected from the French CONSTANCES and Three-City population-based cohorts. Generalized additive models for location, scale, and shape were used to model neurofilament light chain (NfL), glial fibrillary acidic protein (GFAP), phosphorylated tau 181 (p-tau181), amyloid-{beta}40, amyloid-{beta}42, and their ratios according to age, sex, body-mass index, and renal function. The resulting equations provided individualized expected concentrations, percentiles, and Z-scores. Previously established disease-specific concentrations were converted into Z-score. Cross-calibration equations were developed for NfL measurements across analytical methods and sample matrices. Findings The final reference populations comprised 5123 participants for amyloid biomarkers and p-tau181 and 5122 for NfL and GFAP; median age was 52.3 years and half were women. Between ages 40 and 80 years, expected NfL and GFAP concentrations increased by an average of 2.6% and 2.2% per year, respectively. Renal function, body-mass index, and sex had additional biomarker-specific effects. Application of the equations to clinical cohorts preserved distinct disease-associated profiles: NfL Z-scores were increased across disorders characterised by neuroaxonal injury, whereas p-tau181 and GFAP showed its greatest increase in Alzheimer disease. NfL cross-calibration equations showed excellent agreement between methods and matrices, with intraclass correlation coefficients greater than 0.90. Interpretation This population-based multibiomarker framework enables blood biomarker concentrations to be interpreted relative to individuals with similar physiological characteristics. Publicly available equations, reference curves, and standardized Z-scores could improve individualized interpretation and comparability across biomarkers, laboratories, and clinical populations.