Abstract
Elevated blood neurofilament light chain (NfL), which indicates the loss of neuronal integrity, is increasingly implicated as a diagnostic and outcome-predicting biomarker for neurological diseases. However, its diagnostic implication for Parkinson's disease (PD) remains unclear, with conflicting data reported by several studies. This may result from the demographic heterogeneity of the studied cohorts. The present study investigated the comparability of blood NfL between a domestic, single-centered PD cohort from Shuang Ho Hospital (SHH) in Taiwan, with the large international, multi-center cohort, Parkinson's Progression Markers Initiative (PPMI). In the SHH PD cohort, with 61 people with PD (PwP) and 25 healthy non-PD controls, plasma NfL unexpectedly was significantly higher in the control group than PwP (14.42 ± 13.84 vs. 9.39 ± 6.91 pg/mL, p = 0.05). Interestingly, subgroup analysis revealed a non-significant difference of plasma NfL levels in male PwP compared with controls (8.58 ± 6.21 vs. 7.25 ± 4.43 pg/mL, p =0.575), whereas NfL levels were significantly lower in the female PwP group than in their healthy control peers (10.29 ± 7.62 vs. 17.79 ± 15.52 pg/mL, p = 0.033). Comparative analysis of the SHH and PPMI cohorts revealed a comparable gender-stratified distribution of blood NfL based on approximate theoretical quantiles. After adjusting for age and gender, no apparent difference in NfL value distribution was observed between the SHH and PPMI cohorts' control or PD groups. Significant downregulation of blood NfL levels were positively correlated with a reduced probability of having a PD diagnosis in both cohorts. These results demonstrated that the adjustment for demographic background enhances comparability between cohorts, and may be required to eliminate covariate/confounder-associated conflict in blood NfL results between different PD studies. This experience may be beneficial to other researchers around the world who are saddled with limited study participants, especially as data from small cohort sizes are often at greater risk of being skewed by specific variables.