Comparative analysis of population-based and personalized reference intervals for biochemical markers in peri-menopausal women: population from the PALM cohort study

Abstract Objectives Significant changes in clinical biochemical markers occur during the peri-menopausal period. Traditional population-based reference intervals (popRIs) may not reflect individual physiological variability, limiting clinical interpretation. This study aimed to establish personalized reference intervals (prRIs) for menopausal women and compare them with popRIs. Methods We analyzed 899 healthy women aged 35–64 from the Peking Union Medical College Hospital Aging Longitudinal Cohort of Women in Midlife (PALM) cohort. 13 biochemical markers were evaluated across reproductive, menopausal transition, and postmenopausal stages. Six key biomarkers were selected through Kruskal–Wallis tests and ranked by their importance in menopausal status classification using a Random Forest model. Biological variation (BV) data were used to calculate total variation (TV) and index of individuality (II). The prRIs were constructed based on BV estimates, and the reference interval index (RII) was applied to compare popRIs and prRIs. Results ALT, TG, and FSH showed significant differences across menopausal stages and ranked highly in the Random Forest model. These markers also had large BV and differed across three menopausal stages. Most II values ranged from 0.6 to 1.4, and all median RII values were below 1.0, suggesting limited utility of popRIs. Crea in reproductive women had the highest proportion of RII>1.0, while FSH showed RII<0.5 in over 90 % of women in the menopausal transition. Conclusions For women in the menopausal transition with high BV estimates, combining popRIs with prRIs improves interpretation. Larger, more diverse cohorts are needed to validate and optimize prRIs for clinical application.