Serum NFL discriminates Parkinson disease from essential tremor and reflect motor and cognition severity

All participants were recruited from the Second Affiliated Hospital of Soochow University. This study included 288 participants: 146 patients with idiopatic PD, 82 patients with ET, and 60 healthy controls. All patients with PD met the following inclusion criteria: (1) age 40–80 years old, Han Chinese; (2) All patients were diagnosed independently by two neurologists, according to the United Kingdom PD Society Brain Bank Criteria [13]. patient with atypical parkinsonism such as the multiple system atrophy, progressive supranuclear palsy, and vascular or secondary parkinsonism was excluded. (3) > 2 years diagnosed with PD. (4) receiving stable doses of L-Dopa/carbidopa administered at least 3 times per day for at least 4 weeks before the initial screening visit.

Patients with ET met the diagnostic criteria by the Consensus Statement of the Movement Disorder Society on Tremor [14]. Additionally, 60 age- and sex-matched healthy controls with no history of any neurological or psychiatric diseases were recruited from their spouses and friends of patients with PD at the same period. All subjects were Chinese Han population and at least 8 years of formal education.

We received approval from the ethics committee of the Second Affiliated Hospital of Soochow University. All participants gave written informed consent before inclusion in this study.

We assessed motor symptom severity of the PD patients using the Hoehn & Yahr (H-Y) stage and the Unified Parkinson’s Disease Rating Scale part III (UPDRS III) score as previously literature described. Which was assessed only during “on” condition [15, 16].

We measured Tremor severity and functional effects of the ET patients using the Fahn-Tolosa-Marin Tremor Rating Scale (TRS) subscales A, B and C score as previously literature described [17]. Each subscale element is rated from 0 to 4 (none to severe tremor) giving a maximum score of 16, 36 and 32 for each subscale.

According to the Jankovic et al.’s criteria (1990) [18], based on the UPDRS, we subdivided PD patients into the tremor-dominant (TD) and postural instability/gait disorder (PIGD) subtypes.

PD with mild cognitive impairment (PD-MCI) and with dementia (PDD) were diagnosed with the diagnostic criteria of the Movement Disorder Society Task Force [19, 20]. An MMSE score of 26 to 28 was used as a possible PD-MCI diagnostic feature as previously literature described [9, 21]. An MMSE score of ≤25 was used as a possible PDD diagnostic feature as previously literature described [9, 21]. The diagnostic criteria of ET and control with cognitive impairment was consistent with those of PD. Structural MRI and routine laboratory tests were performed for all patients to exclude non-PDD causes of dementia.

We collected 10 ml peripheral blood from each participant between 8:00 and 9:00 AM prior to clinical assessment and following an overnight fast. Blood samples were centrifuged (2500 g for 15 min) within 1 h of collection and then kept frozen at − 80 °C until assay. Serum NFL concentrations were determined by researchers who were blinded to the clinical diagnosis, and serum samples were transferred onto the single molecule array (Simoa) NF-light Advantage Kit from Quanterix (Lexington, MA), as previously described [10].

Demographic and clinical variables between groups were compared using analysis of variance (ANOVA) for continuous variables, and chi-squared for categorical variables. Since serum NFL levels were normally distributed in in all three groups (Kolmogorove-Smirnov one-sample test, for PD group: p = 0.22; for ET group: p = 0.14 and for control group: p = 0.36), the parametric tests were used. For variables that violated the assumptions of normality, the groups were compared with nonparametric Mann-Whitney U test (for 2 groups) or Kruskal-Wallis test (for > 2 groups). We compared NFL levels among the three groups using ANOVA, and Fisher’s least significant difference (LSD) test was performed for Bonferroni post-hoc pairwise comparisons. Where there was a significance in ANOVA, the effect of sex, age, education, BMI and clinical variables was tested by adding these variables to the analysis model as covariates. To determine the diagnostic accuracy, a receiver operating characteristic (ROC) curve was constructed to determine the area under the curve (AUC) and the values of sensitivity and specificity with 95% confidence interval (CI), and Youden index was determined (sensitivity + specificity - 1.0) to find the optimal cutoff value. Relationships between variables were assessed with Pearson correlation and Spearman Rank correlation, where appropriate. Bonferroni corrections were applied to each test to adjust for multiple testing. A multivariate regression analyses was used to assess association of serum NFL with UPDRS III, H-Y and MMSE while adjusting for potentially confounding demographic and clinical variables, including sex, age, education, BMI, alcohol, age of onset and duration of disease. Effect sizes (0.2 = small effect, 0.5 = medium effect, 0.8 = large difference effect) were calculated using Cohen’s d method for the two-way comparisons and represented the mean difference, in standard deviation units, between the groups of interest. The SPSS Statistics (version 19.0, IBM Corp., Armonk, NY, USA) was utilized for all statistical analyses. All data are expressed as mean and standard deviation (mean ± SD). All p-values of < 0.05 was considered significant.

The study included 288 participants: 146 patients with PD, 82 patients with ET, and 60 were healthy controls. Demographic and clinical information and serum NFL concentrations of all participants are shown in Table 1. The three groups were matched for age, sex and education. PD patients and ET patients were not significantly different in body mass index (BMI), age of onset and disease duration (all p > 0.05). The TRS score of ET patients was 33.2 ± 16.7. The MMSE scores were significantly lower in patients with PD (28.4 ± 2.9) than in patients with ET (29.4 ± 1.4) and in healthy controls (29.8 ± 0.4) (both p < 0.01; effect sizes = 0.44 and 0.68, respectively; Kruskal-Wallis test; Table 1). Table 1 shows the mean H-Y stages (2.5 ± 1.0) and UPDRS part III scores (23.2 ± 8.9) during the “on” condition.

To determine whether older age might influence serum NFL levels, we examined the correlation between the age of the subjects and NFL levels. Consistent with a previous report, higher serum NFL levels were associated with older age in healthy controls, patients with PD and patients with ET (control: r = 0.30, 95% CI 0.05–0.52, p = 0.02; PD: r = 0.25, 95% CI 0.07–0.42, p = 0.002; ET: r = 0.30, 95% CI 0.13–0.46, p = 0.006; Pearson correlation analysis). The serum NFL concentrations were significantly higher in patients with PD (16.6 ± 3.5 pg/ml) than in patients with ET (12.2 ± 2.4 pg/ml) and in healthy controls (11.8 ± 2.4 pg/ml) (both p < 0.01; effect sizes = 1.47 and 1.60, respectively; Table 1 and Fig. 1A). As shown in Table 2, PD patients were classified as having the TD and PIGD motor subtypes. There were no significant differences in demographic and NFL level variables in PD patients (all p > 0.05). The ROC analysis showed that a serum NFL cutoff value of 13.75 pg/ml had a sensitivity of 76% and a specificity of 85% for distinguishing between PD and healthy controls, with an area under the curve of 0.869 (Fig. 1B). Because serum NFL was higher in patients with PD than in patients with ET (p < 0.01, Fig. 1A), the serum NFL cutoff value of 13.65 pg/ml had a 76.7% sensitivity and a 84.1% specificity for distinguishing between PD and ET, with an area under the curve of 0.854 (Fig. 1C).

Motor disorder is the most typical dysfunctionis of PD patients. In assessing the motor symptom severity of PD, we observed an rose in serum NFL levels with the increasing motor disorder as reflected by the H-Y stage (Fig. 2A). The Specific serum NFL concentrations were 12.2 ± 1.8, 15.1 ± 2.1, 18.0 ± 1.5, 20.6 ± 1.5, and 25.1 ± 1.5 pg/ml for H-Y stage I through V (n = 27, n = 46, n = 47, n = 23, n = 3), respectively (F = 104.1, p < 0.001). Cognitive impairment is one of the most disabling nonmotor features of PD. We then examined the relationship between serum NFL concentrations and various levels of cognitive function among PD patients and healthy controls. We observed that serum NFL significantly increased as the cognitive impairment severity increased (Fig. 2B). The specific NFL concentrations were 11.8 ± 2.4, 15.0 ± 2.6, 19.6 ± 1.1, and 22.2 ± 1.9 pg/ml for the control, PD with normal cognition, PD-MCI, and PDD groups (n = 60, n = 102, n = 31, n = 13), respectively (F = 114.8, p < 0.001).

In patients with PD, correlation analysis showed that serum NFL levels were significantly positively correlated with motor symptom severity, measured with UPDRS III score and H-Y stage (UPDRS III score: r = 0.79, 95% CI 0.70–0.86, p < 0.001; Bonferroni corrected p < 0.01; Pearson correlation analysis; H-Y stage: r = 0.86, 95% CI 0.78–0.91, p < 0.001; Bonferroni corrected p < 0.01; Spearman correlation analysis) (Fig. 3A, Fig. 2A). At the same time, serum NFL levels were significantly negatively correlated with MMSE scores in PD patients (MMSE scores: r = − 0.70, 95% CI − 0.63 to − 0.77, p < 0.001; Bonferroni corrected p < 0.01; Pearson correlation analysis) (Fig. 3B).

Further multivariate regression analysis showed that serum NFL was an independent contributor to motor symptom and cognition decline severity in PD patients (H-Y stage: t = 8.75, VIF = 2.93, p < 0.001; UPDRS III score: t = 3.17, VIF = 2.92, P = 0.002; MMSE score: t = − 4.81, VIF = 1.73, p < 0.001), while adjusting for potentially confounding demographic and clinical variables, including age, sex, education, BMI, age of onset, and disease duration. Our data has been normalized and there was no multicollinearity.