Select plasma proteins confirm recent seizure activity, differentiate between seizure type

An analysis of 51 plasma proteins associated with immune response identified four proteins capable of confirming a recent epileptic seizure or a psychogenic non-epileptic seizure, according to findings published in Neurology.

The study results also demonstrated that the four plasma proteins differentiated between a psychogenic non-epileptic seizure (PNES) and an epileptic seizure (ES) with sensitivity and specificity.

“Accurate diagnosis of epilepsy is challenging because clinicians rarely observe the actual clinical seizure outside of the hospital,” John Gledhill, of Cognizance Biomarkers, and colleagues wrote. “Furthermore, PNES can mimic ES leading to erroneous diagnosis and inappropriate treatments. Thus, a critical gap in the diagnostic assessment of seizures is a blood test that can distinguish ES from PNES.”

Multiple biomarkers have been suggested to distinguish an ES from a PNES, according to Gledhill and colleagues, but the lack of diagnostic sensitivity, specificity and accuracy “precluded use in clinical practice.” In contrast, a growing body of evidence supports a link between neuroinflammation and ES. The researchers hypothesized that modified levels of certain plasma markers of neuroinflammation would discern an ES from a PNES.

The study included 137 patients aged 12 years and older who were assessed in an epilepsy monitoring unit for clinically indicated differential diagnosis or presurgical examination as well as 29 healthy volunteers between 2014 and 2015; the researchers enrolled patients from the epilepsy monitoring unit because of the EEG results available through the unit, the “gold standard” for diagnosing clinical events. They also recruited healthy controls aged 18 years and older among the family and friends of patients admitted to the epilepsy monitoring unit.

The researchers obtained blood samples from patients within 24 hours of an EEG-confirmed ES or PNES and isolated plasma. They collected levels of 51 candidate plasma proteins related to inflammatory response through a literature review. These proteins had been linked to seizures, proinflammatory pathways, cell adhesion and the systemic immune response and included C-reactive protein, eotaxin, nectin-4 and tumor necrosis factor-alpha, among others.

Gledhill and colleagues used a 51-protein multiplexed enzyme-linked immunosorbent assay (ELISA) to calculate plasma levels in patients with ES, patients with PNES and healthy controls. They examined 85 samples, including 31 from patients with ES, 25 from patients with PNES and 29 from healthy controls. The researchers observed no significant differences in age, sex or race between the groups.

Classifying ES, PNES

The combination of protein concentrations TRAIL, ICAM-1, MCP-2 and TNF-R1 “provided a probability that a patient recently experienced a seizure,” Gledhill and colleagues wrote. They found that TRAIL and ICAM-1 were higher in PNES than in ES; MCP-2 and TNF-R1 were higher in ES than in PNES.

The diagnostic algorithm developed by the researchers resulted in an area under the curve of 0.94 (±0.07), a sensitivity of 82.6% (95% CI, 62.9-93) and a specificity of 91.6% (95% CI, 74.2-97.7). When they extended the algorithm to include previously recognized PNES risk factors – including major depressive disorder and migraine among others – the results showed an enhanced diagnostic performance, with an AUC of 0.97 (±0.05), a sensitivity of 91.3% (95% CI, 73.2-97.6) and a specificity of 95.8% (95% CI, 79.8-99.3).

“While a risk factor only algorithm provides some clinical value, when combined with protein information, the algorithm had superior power to correctly classify patients suffering simultaneously from ES and PNES,” Gledhill and colleagues wrote.

The researchers addressed two false-negatives generated with the diagnostic algorithm, which they noted could be the result of confounding variables that were not included in the study or not sufficiently powered. They also stated that the study design may have resulted in observations of “a chronic immune response” among patients with baseline seizure frequency, such as epilepsy, and that future studies should evaluate baseline protein levels. False-negative seizures may have also been identified by the algorithm “but not the gold standard ictal EEG,” according to Gledhill and colleagues.

“To fully understand the results presented here, a sufficiently powered study with a reserved validation data set will be required to understand all aspects, including the longitudinal, demographic and pharmacological aspect of epilepsy patients,” the researchers wrote. “The results will also need to be tested against other paroxysmal events such as syncope, migraine and hypoglycemia to further understand the clinical utility of these results.”

‘Promising, but very preliminary’ results

In a related editorial, G. Bryan Young, MD, FAAN, FRCPC, of the University of Western Ontario, Canada, and Liu Lin Lin Thio, MD, PhD, of Washington University in St. Louis, wrote that, while biomarkers have so far “been largely unreliable,” they have promise as an additional tool, particularly when used in addition to clinical features. They also noted that Gledhill and colleagues “wisely evaluated the changes following individual seizures,” as some patients with ES also have PNES and vice versa.

Young and Thio said the study “should be viewed as a promising, but very preliminary, study in exploring biomarkers to differentiate ES from PNES.” They pointed to the need for larger cohorts that provide a better evaluation of how clinical factors such as age, seizure type and antiseizure medications affect blood levels of pro-inflammatory or anti-inflammatory markers.

“The specific proteins found to differentiate ES and PNES are somewhat unexpected and many good candidates were not found to be significant. Also, the origin of the plasma proteins that help to distinguish ES from PNES is unclear and could have been muscle or another tissue rather than brain,” Young and Thio wrote. “Perhaps more work needs to be done on suitable animal models with motor and non-motor seizures. Multicenter studies involving large numbers of patients with various seizure types and careful tabulation of various associated clinical variables (as done in this preliminary study) may provide definitive answers.”

According to Young and Thio, if specific proteins eventually enable clinicians to differentiate between ES and PNES in combination with clinical features, “an inexpensive, practical assay should be developed, as measuring cytokines and other inflammatory markers would otherwise not be feasible for most clinical units.”