Loading ...

Estimating the risk for unprovoked seizures in the year following an ischemic stroke with the SeLECT 2.0 prognostic model may be helpful in guiding decisions about driving, new research suggested.

Data from a multicenter international cohort showed that the personalized baseline SeLECT 2.0 score strongly predicted the chance of seizure in the next year (COSY), whereas poststroke seizure-free intervals (SFIs) of less than 12 months had a minor impact on COSY.

"We were surprised by the large variability of seizure risks in stroke survivors with acute symptomatic seizures — range 2%-92% following a 3-month seizure-free interval," study investigator Marian Galovic, MD, PhD, University Hospital Zurich, Zurich, Switzerland, told Medscape Medical News. "This shows that a one-size-fits-all approach is not appropriate in this regard, and a more personalized approach is needed."

However, the investigators noted that other stroke-related deficits are not considered within the model and need to be considered on an individual basis when assessing the ability to drive.

"A key point is that neurological deficits such as hemianopia, neglect, or weakness can be assessed with a neurological examination, but the risk of seizures is much more abstract and may be more difficult to assess. Therefore, a prognostic model of poststroke seizure is helpful," said Galovic.

The findings were published online on May 15 in the Journal of Neurology, Neurosurgery & Psychiatry.

Driving Guidelines Vary Widely

Recommendations on driving for people with an acute symptomatic seizure following stroke primarily rely on SFIs, which can range from 3 to 12 months and typically do not consider other individual characteristics, the authors noted.

As shown previously, the investigators developed the SeLECT score to predict seizures after ischemic stroke using clinical variables obtainable within the first 7 days after stroke. The SeLECT 2.0 version stratifies acute symptomatic seizures into brief seizures and status epilepticus, which carries a much higher risk of developing epilepsy and mortality after stroke than brief acute symptomatic seizures, Galovic explained.

The present study analyzed data from a multicenter registry of poststroke seizures created as part of the SeLECT study and included 4552 individuals from nine centers with acute ischemic stroke and no history of seizures or epilepsy.

Investigators calculated the SeLECT 2.0 score for each participant and then modeled the risk for unprovoked seizures for the cohort using Cox proportional hazards regression. Acute symptomatic seizures during the first 7 days of poststroke were not counted as unprovoked seizures.

COSY was estimated according to each SeLECT 2.0 score value (range, 0-13 points) and a range of poststroke SFIs between 0 and 24 months. A COSY below 20% was considered safe for private driving and below 2% for professional driving.

Easy-to-Use Bedside Model

COSY ranged from 0.6% (SeLECT 2.0 score, 0 points) to 94% (SeLECT 2.0 score, 13 points) after an SFI of 1 month and from 0.2% to 53%, respectively, after an SFI of 24 months, the authors, led by Kai Michael Schubert, MD, PhD, University Hospital Zurich, reported.

A COSY below 20% — viewed as an acceptable risk for private driving by many European regulatory agencies — was already achieved at baseline for those with SeLECT 2.0 values of 0-7 points, with the upper border of the 95% CI exceeding 20% for a SeLECT 2.0 value of 7.

Stroke survivors with SeLECT 2.0 values of 8-10 points had a COSY below 20% after an SFI of 5-14 months, while COSY remained above 20% at 2 years after stroke with higher values of 11-13 points.

Those with acute status epilepticus (SeLECT 2.0 score, 7-13 points) had higher COSYs, ranging from 14% to 92% after an SFI of 3 months.

Stroke survivors without acute symptomatic seizures (SeLECT 2.0 score, 0-6 points) had the lowest COSY at baseline, ranging from 0.7%-11%.

"The model is based on data from nine international cohorts and was robust in cross-validation, showing the generalizability of the findings," Galovic said. "However, it may take some time until these findings are considered within local guidelines that regulate the legal aspects of driving ability following stroke."

For neurologists and other clinicians, she noted that a free smartphone app and web calculator are already available to facilitate bedside calculations. "The SeLECT model is very easy to use and can be integrated seamlessly into clinical practice."

More Validation Needed

Reached for comment, Nishant K. Mishra, MD, PhD, assistant professor of neurology, Yale University School of Medicine, and stroke director, US Department of Veterans Affairs, New Haven, Connecticut, said the data are useful but need to be validated in another cohort.

"If someone has never had a symptomatic seizure or status epilepticus, then based on this model, if the SeLECT 2.0 score is less than 7, it's really safe for the doctor to suggest that at least the risk of late seizure is minimal and potentially the patient can drive poststroke," he said. "If there are any other elements like vision and apraxia, more investigation needs to be done, and that's a totally different question."

Mishra pointed out that there is a lot of heterogeneity among stroke survivors who had early symptomatic seizures, and "that's the subgroup which really needs more analysis."

In an accompanying editorial, Mishra called for more research to determine the sources of heterogeneity and identify patient-specific features that reliably identify those with higher COSY. "One source of heterogeneity is hemorrhagic transformation risk, and an increasing greater risk of COSY is expected in patients with hemorrhagic infection and parenchymal hematoma (PH) 1 and PH2."

Mishra also suggested interictal discharges need focused attention in the stroke population and that bioinformatics offers opportunities to assess the biofluid biomarkers of epileptogenicity.

Galovic told Medscape Medical News that stroke is an excellent model of epileptogenesis in humans and that there are exciting new developments in the field of antiepileptogenesis. "The SeLECT model is relevant for antiepileptogenic treatment trials because it allows the recruitment of an enriched population and the selection of stroke survivors that could benefit from antiepileptogenic compounds when they become available."

Galovic reported receiving fees and travel support from Arvelle, Advisis, Bial, and Nestlé Health Science outside the submitted work. Mishra reported no competing interests.