Albert P Aldenkamp, Willem Lavrijssen and Dominique Ijff
The cognitive side effects of the antiepileptic drugs (AEDs) have emerged as an important aspect in medical decision making in childhood epilepsy (Mandelbaum et al. 2009). Children with epilepsy have a higher risk of learning disabilities and academic weakness because of the epilepsy and the seizures (Helmstaedter et al. 2014). In such a situation, even a modest adverse drug effect may have consequences because it can amplify the already existing weaknesses and ultimately may cause developmental arrest (Mula and Trimble 2009; Cross 2010; IJff and Aldenkamp 2013). Children are potentially more susceptible to the adverse effects of AEDs than adults because of the potential effect of AEDs on brain maturation and hence on neurodevelopment (Cross 2010; Ijff and Aldenkamp 2013). Early identification of AED-induced cognitive impairment in children with epilepsy is, therefore, crucial because those deficits impact on future educational possibilities and eventually the occupational and social outcomes.
The first studies on cognitive side effects of AEDs were published in the early 1970s. However, recently, most studies have focused on adults, which limits the generalization of findings to children. More importantly, most studies have been of limited duration. Most of our knowledge is based on clinical trials lasting on average 12 weeks (Vermeulen and Aldenkamp 1995). This is completely different from normal clinical practice where treatment is assessed over years rather than weeks (Mula and Trimble 2009).
We, therefore, screened the literature for studies on the relationships between chronic AED treatment and cognitive function in children. “Chronic” was here defined as treatment durations of greater than 6 months, preferably in a controlled trial. Studies were evaluated using the U.S. Preventive Services Task Force (USPSTF) evidenced-based classification levels (see Table 1). The discussion of the identified studies was limited to currently used AEDs in the general childhood epilepsy population rather than effects in specific syndromes. For example, we do not discuss the use of vigabatrin in infantile spasms or the use of levetiracetam in continuous spike-waves during slow wave sleep. Also only those studies are discussed that used formal neuropsychological testing. The add-on design studies sometimes showed interpretable and interesting results. Polytherapy studies were not used.
In Box 1, we explain the evaluation strategy, based on an evidence level approach. In Table 7.2, we provide the evaluation of all articles that passed the search strategy.
In total, 18 studies were identified that met our inclusion criteria. No level 1 studies have been found. A disappointing number of controlled studies allowed us to study chronic effects: level II-1 and—if not available—level II-2 studies. In addition, almost all studies were comparative studies using an active comparator (another AED), which allowed us only to draw relative conclusions in a nonequivalent analysis (not worse than another AED) rather than absolute conclusions (no side effects).
Evaluation per AED
Four relatively recently published level II-1 studies are available (Chen et al. 1996; Donati et al. 2006, 2007; Kang et al. 2007; Eun et al. 2012b). These are all comparative studies with an active control revealing no absolute effects (effects against no drug treatment) but only relative effects. The four studies did not show inferior performance compared to lamotrigine (LTG), valproate (VPA), topiramate (TPM), phenobarbitone (PHB), and oxcarbazepine (OXC). Three studies used a 6-month follow-up, and one (Kang et al. 2007) a 12-month follow-up. In total, 156 children had a chronic exposure (28 vs OXC; 35 vs LTG; 43 vs TPM; 25 vs PHB; and 25 vs VPA).
We may conclude that CBZ does not induce more impairment after chronic treatment than other first-line AEDs: VPA, LTG, and OXC and no worse impairment than TPM.
Four level II-1 studies were available (Vining et al. 1987; Chen et al. 1996; Donati et al. 2006, 2007; Glauser et al. 2013). No differences with CBZ or VPA were found. Compared to ethosuximide (ETS) and LTG, attentional problems were found. Cognitive global level was higher than PHB in a crossover trial (Vining et al. 1987), but this was not confirmed in the comparative trial of Chen et al. (1996). Long-term data are available for a total of 201 children in the level II-1 studies. All studies used follow-up periods of 12 months except for Donati et al. (2006, 2007) with 6-month follow-up. The study by Glauser et al. (2013) is noteworthy because of the large number of children on VPA (n=147) showing attentional problems relative to those on LTD (n=149) and ETS (n=155). As this study was performed in children with typical absences, one would not expect a confounding seizure effect. We may, therefore, cautiously conclude that VPA may induce attentional problems after 12 months of treatment.
One level II-1 study is available (Donati et al. 2006, 2007) in 55 children on OXC. The study did not reveal differences with 28 children on CBZ and 29 on VPA after 6 months of treatment. No confirmation of this finding is available.
Two level II-1 studies are available, showing no impairment relative to ETS and VPA (Glauser et al. 2013) and CBZ (Eun et al. 2012b). Again the study by Glauser has a high power because of the large number of patients included. In total, 181 patients were studied over 6 months (Eun et al. 2012b) and 12 months (Glauser et al. 2013).
No interpretable long-term studies are available.
Only one study is available, a class II-1 study in which TPM is compared with CBZ over a 6-month period (Kang et al. 2007). They found impairment on the arithmetic subscale of the Wechsler Intelligence Scale for Children at higher doses of TPM.
A number of studies are available on the long-term effect of PHB. This is possibly due to the use of PHB for febrile seizures in toddlers, raising the issue of later developmental consequences. There are two level II-1 studies (Vining et al. 1987; Chen et al. 1996). In the study by Vining et al. (1987), intelligence and memory scores were lower in the PHB condition in a crossover design with 6-months of treatment in 21 children. However, in the study by Chen et al. (1996), no differences with CBZ or VPA were found after 12 months of treatment in 23 children. Both used the same intelligence test. Vining et al. (1987), however, used in addition a large test battery and used compensation or correction for multiple testing.
No interpretable long-term studies are available.
Only one level II-1 study is available; however, the well-powered study by Glauser et al. (2013) including 155 children on monotherapy ETS in a comparative design versus VPA and LTG over 12 months revealed no long-term side effects.
One class II-1 study is available comparing high dose with low dose ZSM in 70 children, revealing cognitive impairment at higher dose concerning language development, or more specifically vocabulary acquisition (Eun et al. 2011) after 6 months of treatment.
The results are summarized in Table 7.1. For none of the AEDs confirmed cognitive adverse effects have been demonstrated after long-term treatment (≥6mo). For CBZ and LTG confirmed evidence is available that they do not impair cognitive function in the long term. This statement must, however, be considered as a relative statement as all the evidence comes from comparative design, implying noninferiority versus a number of other AEDs. For OXC and ETS, evidence is found in the same positive direction, but confirmation is needed. AEDs that seem to increase the risk for cognitive adverse effects after long-term treatment are TPM, zonisamide, and VPA. In all three cases, confirmation has yet to be found. Not surprisingly, for the newer AEDs, such as lacosamide or pregabaline, no long-term data are yet available. However, for important AEDs in current use, phenytoin and LEV, data are also lacking.
The overall conclusion is as follows:
A.Long-term treatment with most of the AEDs does not seem to endanger cognitive development in children.
B.However, because of the lack of long-term data, this conclusion must be drawn with great caution. Clearly, we need more studies, such as the study by Glauser et al. (2013), with a large number of patients, newly diagnosed, in types of epilepsy where seizures are no major confounders and preferably assessed at a no-medication baseline, at a steady state and after 6 months of treatment.
BOX 1 Levels of evidence used based on the US Preventive Services Task Force classification
Level I: Evidence obtained from at least one properly designed randomized controlled trial.
Level II-1: Evidence obtained from well-designed controlled trials without randomization.
Level II-2: Evidence obtained from well-designed cohort or case-control analytic studies, preferably from more than one center or research group.
Level II-3: Evidence obtained from multiple time series designs with or without the intervention. Dramatic results in uncontrolled trials might also be regarded as this type of evidence.
Level III: Opinions of respected authorities, based on clinical experience, descriptive studies, or reports of expert committees.