Epileptic Seizures in Alzheimer’s Disease: What Are the Implications of SANAD II?

Clinician awareness of seizures in Alzheimer’s disease (AD) has increased in recent years [1–5]. Prospective studies suggest that as many as 25%of AD patients may have epileptic seizures, that seizures predict more severe impairment in activities of daily living in these patients [6], and that seizures may be a marker of more rapid decline and worse prognosis [7]. However, how best to manage these seizures remains unclear.

A Cochrane review [8] found only one randomized controlled trial of anti-seizure drugs (ASDs) in AD, reported more than 10 years ago [9], which in-cluded only 95 patients and provided insufficient evidence to support the use of lamotrigine, phenobarbital, or levetiracetam. Expert opinion has suggested that many ASDs are best avoided in AD patients because of pharmacodynamic and tolerability issues in older persons, including topiramate, zonisamide, phenobarbitone, benzodiazepines, vigabatrin, tiagabine, carbamazepine, oxcarbazepine, eslicarbazepine, gabapentin, pregabalin, sodium valproate, and phenytoin [2]. As there is some evidence for lamotrigine and levetiracetam [9], albeit low quality [8], they have emerged as the least worst options.

In the absence of dedicated randomized controlled trials of ASDs in AD patients, it has been suggested that results from pragmatic studies of treatments for epilepsy in other clinical settings might be extrapolated to AD to assist with these clinical judgments [10]. In this context, the recent publication of the findings of the second Standard And New Anti-epileptic Drugs (SANAD II) trial may be informative [11, 12].

The general classification of seizure types includes focal seizures, which include focal aware seizures (previously known as simple partial seizures), focal seizures with impaired awareness (formerly complex partial seizures), and focal to bilateral tonic-clonic seizures (formerly secondarily generalized tonic clonic seizures), and generalized seizures, which inc-lude generalized tonic clonic seizures and myoclonic seizures. Seizure types in AD are not well-defined, but the evidence base [1–6] (and our clinical experience) suggests that in the earlier stages of disease focal seizures with impaired awareness and focal to bilateral tonic clonic seizures predominate, whereas in the later stages myoclonic and generalized tonic clonic seizures are seen (as in Down syndrome, which shares the same neuropathological basis as AD).

Hence both the SANAD II studies, of focal epilepsy [11] and of generalized and unclassifiable epilepsy [12], may be pertinent to considerations of AD. In focal epilepsy, levetiracetam and zonisamide were compared to lamotrigine [11], and in generalized and unclassifiable epilepsy, valproate was compared to levetiracetam [12]. The primary outcome measure was time to 12-month remission from seizures. Secondary outcomes were time to treatment failure overall and time to treatment failure due to either inadequate seizure control or unacceptable adverse reactions.

In the SANAD II focal epilepsy study [11], lev-etiracetam did not meet the definition of non-inferi-ority to lamotrigine for time to 12-month remission from seizures (we focus on these data since zonisamide has previously been considered as not suitable for use in AD as potentially cognitively harmful [2]). Analysis of overall time to treatment failure for any reason found lamotrigine to be significantly less likely to fail than levetiracetam. Levetiracetam was more likely to fail than lamotrigine because of adverse reactions but not inadequate seizure control. The main difference in adverse reaction profiles related to psychiatric symptoms (13%in patients who initiated with lamotrigine, 30%in those initiated with levetiracetam). In conclusion, the existing recommendation that lamotrigine should be a first line standard treatment for patients with focal epilepsy, based in part on the results of the original SANAD study which found it to be non-inferior to carbamazepine and less likely to fail because better tolerated [13], was endorsed by SANAD II.

In the SANAD II generalized and unclassifiable epilepsy study [12], levetiracetam did not meet the definition of non-inferiority to valproate for time to 12-month remission from seizures. Time to achieve 12-month remission was shorter for valproate. Overall time to treatment failure found valproate to be superior to levetiracetam, with inadequate seizure control for levetiracetam but no difference for unacceptable adverse reactions. However, more adverse psychiatric symptoms were reported in participants allocated to levetiracetam.

It was previously questioned whether the findings of the pragmatic SANAD II study could be extrapolated to the AD population with seizures [4]. There are reasons for caution when generalizing the trial results in this way. The mean age of patients in SANAD II was 39.3 years. Patients with known dementia were excluded from the study. The observed adverse reaction profile, for example psychiatric symptoms, may be more significant in older persons with AD. Whether the cost-effectiveness data, favoring lamotrigine [11] and valproate [12], would also translate to AD patients is unknown. Generally, trial participants may have fewer underlying conditions and less medication use than non-participants [14].

Nevertheless, in the current absence of any more compelling data from dedicated randomized controlled trials in patients with AD and acknowledging the limitations of extrapolating from trials in which AD patients are absent, we suggest that the following possible recommendations emerge in light of the findings of SANAD II.

In AD patients with focal epilepsy in whom the decision has been made to initiate ASDs, clinicians should unequivocally prefer lamotrigine to levetiracetam, as previously suggested [10], since the former is associated with fewer adverse reactions, particularly psychiatric symptoms.

In AD patients with myoclonic and generalized seizures requiring ASDs, consideration should be given to valproate rather than levetiracetam. We acknowledge that this is a potentially controversial recommendation since valproate has previously been considered unsuitable for use in AD [2], in part because of adverse cognitive effects. Furthermore, a Cochrane review found valproate to be ineffective for agitation in dementia, with a higher rate of adverse effects (sedation, gastrointestinal effects, urinary tract infections) than in controls [15]. Clearly any decision regarding use of valproate needs to be made on an individual patient basis, but the evidence from SANAD II of better seizure control and fewer adverse psychiatric symptoms than levetiracetam should inform both clinician and patient considerations. The original SANAD study showed valproate to be superior to lamotrigine for seizure control in generalized epilepsies [16].

Much remains to be learned about the optimal treatment of seizures in AD. Any possible role for other ASDs such as lacosamide, perampanel, and brivaracetam has yet to be defined, as is the case for focal epilepsy in general. The need for separate pragmatic trials of ASDs in AD patients, as previously advocated ten years ago [10], remains.