Kristin Miller, MD
Post-stroke seizures are a well-described potential complication of stroke. However, the long-term risk of seizures in stroke survivors in relation to the characteristics of stroke subtypes and demographics has not been fully explored. Authors Merkler et al endeavored to explore this particular long-term risk in a large, multistate retrospective review using administrative claims data and Medicare claims data.
Using ICD-9 codes from multistate claims data for ischemic stroke, intracerebral hemorrhage (ICH), and subarachnoid hemorrhage (SAH), the investigators identified 777,276 patients who were hospitalized with an index stroke. Among those patients, 45,708 (5.88%) developed a new seizure with an annual incidence of 1.68% in the all-stroke group compared to 0.15% in the general population. The annual incidence of seizures and the cumulative rate of seizures were highest in ICH subtypes. Younger patients (<65 years) and nonwhite patients were more likely to develop seizures after stroke. A similar pattern was seen when analyzing Medicare beneficiary claims data in 81,984 patients; however, there was no difference in the annual incidence of post-stroke seizure based on race and no analysis was performed on age (due to the age of Medicare patients).
This study confirmed the significant long-term risk of post-stroke seizures with data showing a 7-fold higher risk than in the general population. It also demonstrated the relative risk of post-stroke seizures to be higher in younger patients, nonwhite patients (in one of the two cohorts), and following hemorrhagic strokes.
Limitations in this study include the reliance on ICD-9 coding to identify appropriate patients; however, this use had been previously validated. Additionally, in the multistate analysis, only seizures resulting in emergency room visits and/or hospitalizations were used, which is likely greatly underestimating the number of patients with a diagnosis of post-stroke seizures that present in the outpatient setting. While the subtype of stroke was evaluated, there was no data on further characteristics of the strokes, including severity, location, and size, which likely play a significant role in the risk of developing post-stroke seizures. Similarly, there was no mention of the seizure characteristics. Lastly, while prophylactic antiepileptics are not recommended in stroke care, the possible use of antiepileptics due to different practice habits may have influenced the association between stroke and post-stroke seizures and could not be evaluated based on the ICD-9 coding. Further investigation of race and its impact on the long-term risk of post-stroke seizures may be warranted in the future as the association between nonwhite race and a higher incidence of post-stroke seizure was only seen in the multistate analysis and not in the Medicare beneficiaries analysis.
In regards to changing practice and management, this data may not have much of an impact. Certainly, understanding a patient’s risk for developing post-stroke seizures and epilepsy is important to help guide the discussion and education of patients and their caregivers regarding possible future complications of stroke; however, this data and knowledge should not influence our decision on the use of prophylactic antiepileptics.