Certain cerebrospinal fluid biomarkers may help differentiate Parkinson’s disease from other causes of parkinsonism and also predict which patients will later develop Parkinson’s-related dementia, according to a report published online Aug. 10 in JAMA Neurology.

In a prospective, population-based, longitudinal study, Dr. David C. Bäckström of Umeå (Sweden) University and his coinvestigators examined several biomarkers in cerebrospinal fluid (CSF) samples from 128 patients with new-onset, treatment-naive parkinsonism who resided in a defined geographic area of Sweden. These patients agreed to CSF collection by lumbar puncture at baseline, and some agreed to serial sampling. All were followed yearly for 5-9 years, undergoing periodic comprehensive neuropsychological testing. At their most recent follow-up visit, 104 of these participants had been diagnosed as having Parkinson’s disease, 11 as having multiple system atrophy, and 13 as having progressive supranuclear palsy. For reference, CSF samples were also obtained from 30 demographically similar but neurologically healthy control subjects.

High levels of neurofilament light chain protein (NFL) together with low levels of amyloid-beta1-42 distinguished progressive supranuclear palsy from PD at baseline as well as 1 year later, after patients had been receiving dopaminergic therapy for months. Even elevated NFL alone was useful for differentiating progressive supranuclear palsy from PD: A baseline NFL exceeding 2,020 ng/L had a sensitivity of 75% and a specificity of 83% in distinguishing the two disorders. If measured over time, the tendency of NFL to gradually increase in progressive supranuclear palsy was even more indicative: At 1 year, an NFL exceeding 2,916 ng/L had a sensitivity of 89% and a specificity of 93% for identifying progressive supranuclear palsy. However, CSF samples at baseline or could not distinguish multiple system atrophy from Parkinson’s disease.

In addition, an early pattern of high NFL, low amyloid-beta1-42, and high heart fatty-acid-binding protein (HFABP) predicted the later onset of PD-related dementia with 90% sensitivity and 71% specificity. Patients whose baseline NFL exceeded 1,100 ng/L, whose amyloid-beta1-42 was less than 626 ng/L, and whose HFABP exceeded 500 ng/L were, respectively, at 2.6, 2.8, and 2.8 times higher risk of future PD-related dementia than were other patients with parkinsonism. And low amyloid-beta1-42 alone predicted PD-related dementia for up to 9 years after initial presentation of parkinsonism, Dr. Bäckström and his associates said (JAMA Neurol. 2015 Aug 10. doi:10.1001/jamaneurol.2015.1449).

Taken together, these findings suggest that CSF testing can be clinically useful to differentiate PD from other disorders and as a presymptomatic screening tool for PD-related dementia, they added.

In contrast, other CSF biomarkers examined in this study, including total tau, phosphorylated tau, and alpha-synuclein, were not helpful in identifying PD or predicting PD-related dementia.

The Swedish Medical Research Council, Erling Persson Foundation, Parkinson Foundation in Sweden, Umeå University, Västerbotten County Council, King Gustaf V and Queen Victoria Freemason Foundation, Swedish Parkinson Foundation, Kempe Foundation, and Swedish Parkinson’s Disease Association funded the study. Dr. Bäckström reported having no conflicts of interest. One of his associates reported ties to numerous pharmaceutical companies.