[Establishment of age- and sex-specific reference intervals for cerebrospinal fluid total protein based on dry...

Objective: To establish age- and sex-specific reference intervals (RIs) for cerebrospinal fluid total protein (CSF-TP) in a single-center Han Chinese population, providing a reference basis for the diagnosis of neurological diseases. Methods: A retrospective cross-sectional study was conducted on CSF routine and biochemical data from Huashan Hospital, Fudan University, between January 1, 2018, and July 30, 2024. A total of 40 198 CSF test records were retrieved from the Laboratory Information System. After excluding cases with conditions known to elevate CSF-TP levels, 6 173 "apparently healthy" samples were ultimately included. Spearman's rank correlation coefficient (r) was used to assess the correlation between CSF-TP and age. Continuous variables are presented as M(Q₁, Q₃). The Mann-Whitney U test was employed to compare CSF-TP levels between sexes. Quantile regression was performed to establish age-and sex-specific RIs for CSF-TP (P_2.5-P_97.5). Subgroup analyses were conducted to compare CSF-TP levels among five categories of neurological diseases: brain tumors, demyelinating diseases, hemorrhagic diseases, infectious diseases, and rare diseases. Results: CSF-TP concentration increased with age (r=0.369, P<0.001). The overall median CSF-TP level was 462 (363, 547) mg/L, with males 481 (387, 574) mg/L exhibiting significantly higher levels than females 424 (338, 513) mg/L (U=3 664 330, P<0.001). The established RIs rose progressively with age, ranging from 170-580 mg/L in young adults (18-25 years) to 220-810 mg/L in the 61-65 years group, and further to 250-900 mg/L in the elderly (>76 years). Within the same age strata, the upper limit of the RI was consistently higher in males than in females, whereas it was slightly higher in females than in males in the >76 years group. Subgroup analysis revealed distinct distribution patterns. In the tumor group, craniopharyngioma patients had significantly higher CSF-TP levels 2 249 (1 350, 3 643) mg/L than those with meningioma 1 453 (1 056, 2 254) mg/L (U=6 151), lung cancer brain metastasis 1 085(1 005, 1 853) mg/L (U=1 177), pituitary adenoma 1 689 (1 079, 2 935) mg/L (U=26 518), and primary central nervous system lymphoma 1 079 (902, 1 710) mg/L (U=2 242) (all P<0.001). In the demyelinating group, multiple sclerosis patients exhibited significantly lower levels 693.5 (612.0, 726.5) mg/L compared to chronic inflammatory demyelinating polyneuropathy 1 080 (939, 1 494) mg/L (U=23), myelitis 1 064 (842, 1 580) mg/L (U=155), and Guillain-Barré syndrome 1 355 (975, 2 210) mg/L (U=58) (all P<0.001). In the infectious disease subgroup, the median CSF TP levels were 1 032 (863, 1 670) mg/L for cryptococcal infection, 1 235 (904, 1 724) mg/L for neurosyphilis, and 1 334 (1 019, 2 614) mg/L for tuberculous meningitis. In the hemorrhagic disease subgroup, the corresponding values were 2 803 (1 286, 5 733) mg/L for intracerebral hemorrhage, 2 118 (1 149, 3 284) mg/L for traumatic intracranial hemorrhage, and 2 984 (1 721, 5 308) mg/L for subarachnoid hemorrhage. There were no significant differences in CSF TP levels among subtypes within either the infectious or hemorrhagic disease subgroups (all P>0.05). Conclusions: This study established large-sample, age- and sex-specific reference intervals for CSF-TP. Given the absence of significant differences among subtypes of intracranial infections and hemorrhagic diseases, these RIs can effectively adjust for confounding effects from physiological factors (age, sex) and certain pathological conditions (infection, hemorrhage), thereby providing a more accurate laboratory basis for the early identification of space-occupying lesions such as craniopharyngioma and demyelinating diseases such as multiple sclerosis.