A 14-year longitudinal study of neurofilament light chain dynamics in premanifest and transitional Huntington's disease

Abstract

Background: Growing evidence supports the value of neurofilament light (NfL) as a prognostic biomarker in premanifest Huntington's disease (HD). To date, however, there has been no longitudinal study exceeding 3 years examining either its serial dynamics or predictive power in HD. We aimed to conduct the first such study.

Methods: Serum NfL was sampled using ultrasensitive immunoassay at four timepoints across a 14-year period in a cohort of HD gene carriers (n = 21) and controls (n = 14). Gene carriers were premanifest at baseline. Clinical features of HD were evaluated by Unified Huntington's Disease Rating Scale (UHDRS TMS), Montreal Cognitive Assessment (MoCA), Trail A/B task, Symbol Digit Modalities Task and semantic/phonemic fluency tasks.

Results: 14/21 HD gene carriers converted to prodromal or manifest disease by the final timepoint ("converters"). At baseline and each subsequent timepoint, NfL levels were higher in converters than in non-converters and controls (p = < 0.001-0.03, η_p² = 0.25-0.66). The estimated rate of change in NfL was higher in converters than in non-converters (p = 0.03) and controls (p = 0.001). Baseline NfL was able to discriminate converters from non-converters (area under curve = 1.000, p = 0.003). A higher rate of change in NfL was predictive of more severe motor (UHDRS-TMS p = 0.007, β = 0.711, R² = 0.468) and cognitive deficits (MoCA p = 0.007, β = - 0.798, R² = 0.604; Trail B, p = 0.007, β = 0.772, R² = 0.567; phonemic fluency p = 0.035, β = - 0.632, R² = 0.345).

Conclusions: Our data suggest that (1) NfL longitudinal dynamics in premanifest/transitional HD are non-constant; rising faster in those closer to disease onset, and (2) NfL can identify individuals at risk of conversion to manifest disease and predict clinical trajectory, > 10 years from disease onset.

Keywords: Biomarker; Dementia; Huntington’s disease; Longitudinal; Neurodegeneration; Neurofilament light chain.

Scheme 1

Clinical conversion trajectory of gene carriers. X denotes timing of conversion to manifest disease, where conversion could be confirmed to within 12 months (n = 5/8)

Fig. 1

A Time 4 NfL concentrations. Group differences assessed by ANOVA adjusted for age, sex, CAG repeat length and BMI. ***p < 0.001; **p < 0.01; *p < 0.05. B Linear regression analyses of Time 4 NfL concentrations versus Time 4 clinical assessment scores among all gene carriers. Linear regressions adjusted for age, sex, CAG repeat length and BMI. p values corrected for multiple comparisons by false discovery rate (Benjamini–Hochberg correction). Shaded areas reflect 95% confidence interval. Results meeting significance threshold depicted

Fig. 2

A Baseline, Time 2, and Time 3 NfL concentrations by group. Group differences assessed by ANOVA adjusted for age, sex, CAG repeat length and BMI. ***p < 0.001; **p < 0.01; *p < 0.05. B Longitudinal dynamics of NfL concentrations by group. Top = individual participants. Left shaded markers indicate sampling in first half of a given timepoint period, right shaded markers indicate sampling in second half of a given timepoint period. Bottom = pooled data. Error bars = SEM. **p < 0.01; *p < 0.05 in group*time interaction adjusted for age, sex, CAG repeat length and BMI

Fig. 3

A ROC curve analysis of Baseline NfL concentrations (left) and annualised rate of change in NfL (right) versus discrimination of converter from non-converter gene carriers. AUC area under curve. B Linear regression analysis of annualised rate of change in NfL concentrations and Time 4 clinical assessment scores among all gene carriers. Linear regression adjusted for age, sex, CAG repeat length and BMI. p values corrected for multiple comparisons by false discovery rate (Benjamini–Hochberg correction). Shaded areas reflect 95% confidence interval. Results meeting significance threshold depicted