Late-onset seizures: a subclinical cerebrovascular disorder?

KEYWORDS:

• Antiepileptic drugs
• cerebrovascular disease
• epilepsy
• epileptic seizures
• risk factors
• stroke

Stroke and epileptic seizures have a bidirectional relation [1]. A clinically overt stroke can be associated with epileptic seizures, which may occur at the time of or in close temporal association with stroke (acute symptomatic seizures), or after a variable interval (days/years) (remote symptomatic seizures) [2]. The distinction between acute and remote symptomatic poststroke seizures is relevant from an epidemiological and clinical perspective. Acute symptomatic seizures are not suggestive of an enduring predisposition of the brain to generate epileptic seizures, and therefore do not entail a diagnosis of epilepsy. Conversely, even a single remote symptomatic poststroke seizure is associated with high risk of seizure recurrence (at least 60% over the next 10 years [3]) and should be considered synonymous with poststroke epilepsy [4].

In recent years, there is increasing evidence that otherwise unexplained late-onset seizures and epilepsy might represent the first clinical manifestation of an underlying occult cerebrovascular disease. In other terms, epileptic seizures can be regarded as a marker of subclinical cerebrovascular disease that has not yet resulted in transient ischemic attack (TIA) or stroke [1].

The observation that seizures can precede stroke has been reported in several clinical trials. A case–control study conducted by Shinton and coworkers showed a much higher prevalence of epilepsy in patients younger than 70 years with first-ever acute stroke compared with controls (eight patients vs. one patient; 4.5% vs. 0.6%; p < 0.05) [5]. The reliability of this finding is however limited by the retrospective design of the study, the small sample size, and the hospital study population. A possible association between vascular risk factors and late-onset epilepsy was further confirmed by a cross-sectional study showing a significant relationship between total cholesterol levels (odds ratio [OR]: 1.3, 95% confidence interval [CI]: 1.0–1.6) or left ventricular hypertrophy (OR: 2.9, 95% CI: 1.0–8.6) and late-onset epilepsy in patients without a previous stroke [6].

The strongest evidence for otherwise unexplained late-onset seizures as predictor of subsequent stroke comes from three large cohort studies. In the first study, over 4700 patients with late-onset (>60 years) epilepsy and matched controls were prospectively assessed; late-onset seizures were associated with a striking increase in the risk of stroke (hazard ratio for stroke: 2.89, 95% CI: 2.45–3.91) [7]. This association was further confirmed by two subsequent cohort studies, which showed an increased risk of ischemic stroke (hazard ratio: 2.85, 95% CI: 2.49–3.26) [8], even in patients with otherwise unexplained seizures occurring after the age of 35 (hazard ratio: 1.6, 95% CI: 1.42–1.80) [9].

Neuroradiological studies have shown that, compared to controls, patients with late-onset epilepsy and no previous history of stroke have a higher prevalence of occult cerebrovascular lesions (focal infarctions, leukoaraiosis, large or small vessel disease) [10–12]. Hemodynamic changes and brain–blood barrier dysfunction due to occult cerebrovascular disease in patients with late-onset epilepsy and no clinically overt stroke has also been demonstrated in multimodal magnetic resonance imaging [13].

All these data suggest that late-onset seizures could reflect a subclinical cerebrovascular disease. To further emphasize the clinical relevance of this association, we proposed to consider seizures that subsequently turn out to have preceded a stroke as a distinct clinical entity using the name ‘heraldic seizures’ [1].

The pathophysiologic mechanisms for late-onset seizures as an early symptom of an underlying cerebrovascular disease are far from being fully elucidated. A disruption of cortico-subcortical circuits due to subcortical small vessel disease with subsequent imbalance between excitability and inhibitory pathways eventually leading to epileptogenicity has been proposed [5]. Furthermore, in patients with subcortical small vessel disease, the neurovascular unit dysfunction due to altered integrity of blood–brain barrier may lead to disruption of cerebral metabolism and/or perfusion with increasing risk of seizures [14]. Ultimately, subcortical small vessel disease should not be considered as a homogeneous entity, but rather as a constellation of different endophenotypes with variable risk for epileptogenesis that might be explained by genetic factors, location and type of pathological changes, and exposure to nongenetic factors (the so-called ‘exposome’) [15]. However, further research is required to understand whether and to what extent vascular comorbidity and late-onset seizures share a common basis, and to elucidate the genetic interactions between vascular risk factors and epilepsy.

Another aspect to consider is the possible association between antiepileptic drugs and the risk of subsequent stroke in epilepsy patients without clinically overt cerebrovascular disease. In patients with late-onset seizures treated with antiepileptic drugs, treatment itself might indirectly increase the risk of stroke, possibly playing a synergistic effect with the occult cerebrovascular disease. Enzyme-inducing antiepileptic drugs may actually increase total cholesterol, lipoprotein(a), C-reactive protein, and homocysteine levels [16]. Whether this pharmacological effect on surrogate markers of vascular disease is clinically relevant or not remains to be defined. A large population-based cohort study has shown that epilepsy patients without a previous stroke treated with antiepileptic drugs had a statistically significant increased risk of stroke (adjusted hazard ratio: 2.22, 95% CI: 2.09–2.36) [17]. When using carbamazepine as reference, oxcarbazepine was found to carry a higher risk of stroke (adjusted hazard ratio: 1.21, 95% CI: 1.10–1.34). Conversely, valproate (an enzyme-inhibiting drug) was associated with reduced risk of stroke (adjusted hazard ratio: 0.86, 95% CI: 0.76–0.96), whereas phenobarbital, lamotrigine, and phenytoin were not found to significantly increase the cerebrovascular risk [17]. A subsequent cohort study failed to find a significant association between use of inducing (adjusted risk ratio: 1.05, 95% CI: 0.78–1.41) or inhibiting (adjusted risk ratio: 1.12, 95% CI: 0.83–1.52) antiepileptic drugs and subsequent risk of stroke [18]. Overall, there is therefore insufficient and conflicting clinical evidence for a strong association between antiepileptic drugs (even enzyme-inducing drugs) and risk of stroke in epilepsy patients without previous clinically overt cerebrovascular disease. However, it is important to consider that epilepsy patients require medication for years, and often for life. Due to the wide availability of antiepileptic drugs devoid of potentially detrimental metabolic effects on surrogate markers of vascular disease, it seems therefore reasonable to prefer the use of non-inducing antiepileptic drugs when treating these patients.

Another question which still remains unanswered is whether late-onset seizures could serve as TIA-equivalent and may therefore help to identify patients at increased risk for stroke and who could benefit from treatment. Considering the evidence available in the literature, it seems reasonable to screen patients with otherwise unexplained late-onset seizure for the presence of modifiable vascular risk factors [1]. Further studies are nonetheless required to assess the effectiveness of this vascular work-up in reducing the subsequent risk of stroke, and the appropriateness of starting an antiplatelet treatment in these subjects. To justify a primary prevention for vascular disease with aspirin in patients with a first-ever seizure, the risk of stroke must be sufficiently high to warrant the risk associated with treatment. The American Heart Association has recently recommended aspirin as primary prophylaxis for cardiovascular disease only when the risk of a vascular event exceeds 6–10% in a 10-year period [19]. However, whether late-onset seizures infer such a risk of stroke is still unknown. To further complicate matters, one has to consider the risks associated with aspirin used as primary prevention of cerebrovascular disease, e.g. the increased risk of cerebral bleeding which may outweigh benefits of reduction in ischemic stroke [20].

In conclusion, data from the scientific literature suggest that cerebrovascular disease often underlies otherwise unexplained late-onset seizures. This is a rather recent view and a fascinating field of research which in our opinion should deserve more attention and interest in a clinical and epidemiological perspective. Patients with a first-ever late-onset seizure and no previous clinically overt cerebrovascular disease should be considered at increased risk of stroke, and may benefit from a diagnostic work-up to identify modifiable vascular risk factors. Further studies are definitely needed to elucidate the risk–benefit ratio of vascular primary prevention with antiplatelet agents in these subjects. In particular, randomized controlled trials are probably required to inform clinical practice, especially regarding the use of aspirin as primary prevention of cerebrovascular accidents in these patients. However, more robust epidemiological evidence should be obtained before undertaking adequately powered clinical trials. To this purpose, prospective population-based cohort studies and studies conducted on a national scale using stroke registers may prove extremely useful. Population-based cohort studies could also investigate a possible association between otherwise unexplained late-onset seizures and the subsequent risk of vascular disease in general, including coronary artery and peripheral vascular disease.

Declaration of interest

F. Brigo has received speakers’ honoraria from Eisai and PeerVoice, payment for consultancy from Eisai, and travel support from Eisai, ITALFARMACO, and UCB Pharma. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Acknowledgments

The authors are grateful to the anonymous reviewers for Expert Review of Neurotherapeutics whose thoughtful and critical engagement with this article greatly assisted in its revision.

Additional information

Funding

This paper was not funded.