CSF heavy neurofilament may discriminate and predict motor neuron diseases with upper motor neuron involvement

Objective: To assess whether phosphorylated neurofilament heavy chain (pNfH) can discriminate different upper motor neuron (UMN) syndromes, namely, ALS, UMN-predominant ALS, primary lateral sclerosis (PLS) and hereditary spastic paraparesis (hSP) and to test the prognostic value of pNfH in UMN diseases. Methods: CSF and serum pNfH were measured in 143 patients presenting with signs of UMN and later diagnosed with classic/bulbar ALS, UMNp-ALS, hSP, and PLS. Between-group comparisons were drawn by ANOVA and receiver operating characteristic (ROC) analysis was performed. The prognostic value of pNfH was tested by the Cox regression model. Results: ALS and UMNp-ALS patients had higher CSF pNfH compared to PLS and hSP (p < 0.001). ROC analysis showed that CSF pNfH could differentiate ALS, UMNp-ALS included, from PLS and hSP (AUC = 0.75 and 0.95, respectively), while serum did not perform as well. In multivariable survival analysis among the totality of UMN patients and classic/bulbar ALS, CSF pNfH independently predicted survival. Among UMNp-ALS patients, only the progression rate (HR4.71, p = 0.01) and presence of multifocal fasciculations (HR 15.69, p = 0.02) were independent prognostic factors. Conclusions: CSF pNfH is significantly higher in classic and UMNp-ALS compared to UMN diseases with a better prognosis such as PLS and hSP. Its prognostic role is confirmed in classic and bulbar ALS, but not among UMNp, where clinical signs remained the only independent prognostic factors.