This review aimed to outline the evidence for the use of traditional and novel treatments for seizures in individuals with ASD. Using a standardized scale, Table presents the evidence for using these treatments to improve seizures in general, to improve the behavioral and cognitive symptoms associated with ASD, and to improve seizures specifically in individuals with ASD. One of the most obvious conclusions from this review is that few treatments for seizures have been specifically evaluated in individuals with ASD. Specific genetic (93
), metabolic (276
), and immunologic (276
) abnormalities as well as cortical hyperexcitability (277
) are believed to be associated with ASD. All of these abnormalities could also drive the development of seizures and epilepsy in ASD individuals. Thus, it is important to determine whether specific treatments that address these pathophysiological mechanisms are effective treatments, as such treatments may improve seizures while also improving, or at least not worsen, core symptoms of ASD.
Anti-epileptic drug treatment for seizures in autism spectrum disorder
There are no well-controlled clinical trials which examine the effectiveness or efficacy of AEDs for the treatment of seizures in individuals with ASD, despite the fact that this issue has been pointed out almost a decade ago (278
). There is good reason to believe that certain AEDs might be more appropriate for certain individuals with ASD. AEDs can cause neurological adverse effects (e.g., ataxia, tremor, and nystagmus), gastrointestinal adverse effects (e.g., abdominal pain and nausea), and allergic reactions. Since adverse effects of AEDs tend to be more frequent in children with developmental disabilities, AEDs with greater risk of adverse effects might limit their usefulness in treating children with ASD (279
). In addition, prolonged treatment with AEDs, especially older AEDs, can cause memory and/or attention deficits as well as somnolence, psychomotor abnormalities, and dizziness (280
). Alternatively, ASD may be associated with cortical hyperexcitability, potentially due to deficits in cortical inhibitory circuits or glutamate receptor abnormalities (277
). Thus, AEDs that enhance γ-aminobutyric acid (GABA) signaling, such as valproate, gabapentin, clobazam, clonazepam, phenobarbital, primidone, tiagabine, or vigabatrin, might be relatively better treatments for individuals with ASD. Lastly, since individuals with ASD appear to have a wide range of seizure types, broad-spectrum AEDs, such as valproate, lamotrigine, or levetiracetam might be optimal treatments.
From the above literature review, it appears that there is only limited evidence to guide the selection of specific AEDs for treating seizures in ASD, although the limited evidence does appear to be consistent across studies. The majority of studies on AEDs in ASD have reported data primarily on three AEDs, valproate, lamotrigine, and levetiracetam. Some studies have also reported the use of phenytoin, carbamazepine, ethosuximide, topiramate, oxcarbazepine, gabapentin, and phenobarbital in individuals with ASD. Overall, studies suggest that broad-spectrum AEDs, specifically valproate, lamotrigine, and levetiracetam, appear to be the most effective and have a low rate of cognitive and neurological adverse effects. Therefore, these may be the most appropriate primary AEDs to start in individuals with ASD and clinical seizures. We will briefly discuss these AEDs below.
Valproate is a broad-spectrum AED which has been documented to have positive behavioral effects in several DBPC studies, has been associated with improvement in core ASD symptoms in case reports and in an open-label trial and, at therapeutic doses, has little adverse effects on cognition (283
). Yet it requires multiple blood tests, can deplete carnitine and can result in serious adverse effects, including hepatotoxicity, hyperammonemia, and pancreatitis. l
-Carnitine has been shown to decrease the hyperammonemia in patients with valproate-induced encephalopathy and is recommended in severe valproate poisoning. Co-treatment with l
-carnitine is recommended for high-risk pediatric patients receiving valproate (284
). Given that children with ASD are at high risk for mitochondrial dysfunction and, as a group, generally have low carnitine levels (107
), and that carnitine deficiency may be a risk factor for developing valproate hepatotoxicity (287
), it is probably wise to start l
-carnitine supplementation when prescribing valproate to individuals with ASD especially if they consume little beef or pork, the primary dietary sources of carnitine. The recommended oral dose varies from 30 to 100
mg/kg/day in two to three daily divided doses for high-risk individuals such as young children or those who have carnitine deficiency (284
). Thus, it is probably wise to use care when prescribing valproate in children with ASD and monitor carnitine levels during treatment. However, the fact that valproate is reported to be well tolerated and considerably beneficial in many studies examining the ASD population specifically is reassuring.
Lamotrigine is another broad-spectrum AED that is believed to have minimally adverse effects on cognition in individuals with epilepsy (36
) and may have beneficial effects on core ASD symptoms (17
), or at least not worsen behavior or ASD symptoms (35
). The use of lamotrigine as a first-line AED needs to be balanced with the prolonged time required for titration to a therapeutic dose in order to minimize the unlikely but serious adverse effect of a Stevens–Johnson reaction.
Levetiracetam is a relatively broad-spectrum AED that has minimal liver metabolism with a low incidence of serious adverse effects. The most common adverse effect of levetiracetam in the general population is behavioral, including agitation, aggressive behavior, and mood instability. However, it is reassuring that a DBPC trial suggests that there is no change in aberrant or repetitive behaviors or impulsivity or hyperactivity in children with ASD treated with levetiracetam (34
). If behavioral adverse effects arise, they may be mitigated with pyridoxine. Major et al. (289
) treated patients on an average dose of 63
mg/kg/day of levetiracetam with an average of 6
mg/kg/day of pyridoxine if behavioral problems developed after starting levetiracetam. Of the 22 patients treated, 41% improved, 36% demonstrated no change, and 18% became worse. Improvement in levetiracetam associated behavior abnormalities with pyridoxine treatment was also reported in one case study (290
). Since pyridoxine clearly does not work in all cases and since levetiracetam does not appear to cause significant behavioral problems in ASD in controlled studies (34
), it is probably best to use pyridoxine on an as needed basis. It is also important to be aware that many children with ASD may already be on a form of pyridoxine as a novel therapy (10
). In addition, this review suggests that the combination of pyridoxine with Mg may be more effective for behavior in children with ASD than either one alone, so it may be wise to add Mg to pyridoxine treatment. Levetiracetam has an intravenous and liquid formulation, so it can be titrated quickly, provided to individuals that cannot take medication orally, and given to children who cannot swallow pills.
Mitochondrial disease and dysfunction are prevalent in individuals with ASD (106
), and certain AEDs may be more appropriate for individuals with mitochondrial abnormalities. Some worry about the use of valproate in ASD given its potentially devastating effect on children with specific MDs. It is important to understand that this detrimental effect is isolated to individuals with POLG1 mutations and myoclonic epilepsy with ragged red fibers syndrome (291
). POLG1 mutations have only been reported in two children with ASD, or 2% of ASD children reported to have MD, and myoclonic epilepsy with ragged red fibers has not been reported to date in individuals with ASD (107
). Given that approximately 5% of children with ASD have classically defined MD, the estimated prevalence of the POLG1 mutation is therefore approximately 0.1% of the ASD population. One study has examined the effect of common AEDs, including phenobarbital, carbamazepine, and lamotrigine, on mitochondrial function. While carbamazepine showed a detrimental effect on mitochondrial function with chronic use, lamotrigine was found to enhance mitochondrial function (292
). Other in vitro
studies have also demonstrated that lamotrigine is mitochondrial protective (293
). Thus, there is little evidence to guide the understanding of optimal AED treatments for children with mitochondrial abnormalities with and without ASD. However, given that lamotrigine appears to be well tolerated in individuals with ASD and may have positive effects on mitochondrial function, at least in preliminary studies, lamotrigine may be the optimal AED for children with ASD and MD. This is indeed a ripe area for clinical research.
Overall, the data reviewed above supports the use of valproate, lamotrigine, and levetiracetam as the first-line treatments in children with ASD who have seizures or epilepsy. Table outlines some guidelines that might be helpful for choosing first-line AEDs. Clearly more research is needed to document the efficacy of AEDs in the ASD population.
Guidelines for selecting a first-line antiepileptic drug for children with autism spectrum disorder.
Traditional non-anti-epileptic drug treatment for seizures in autism spectrum disorder
Several traditional non-AED treatments for seizures were reviewed, specifically the KD and MAD, the VNS, standard epilepsy surgery and MST, immunomodulatory therapy, and neurofeedback. Overall, many of these therapies, except for standard epilepsy therapy (i.e., cortical resection) and the VNS, appear to have promising applications for the treatment of seizures in children with ASD.
The KD and MAD may be useful for treating several aspects of ASD, especially in individuals with ASD with seizures and/or MD. Indeed, recently there has been increased interest in using these diets in ASD (294
). The KD may be an effective treatment for MD (298
) and has been recommended for individuals with co-occurring MD and epilepsy (301
). Given that individuals with ASD and co-occurring MD have high rates of seizures (107
), the KD and MAD should be strongly considered in the subgroup of individuals with ASD and co-occurring MD. In addition, given the excellent safety profile of the KD and MAD as well as studies which suggest their effectiveness in drug-resistant epilepsy and their tolerability in ASD, the KD and MAD should be considered in children with ASD who have epilepsy that is refractory to standard treatments. Of course, children should be carefully monitored when the KD is started as the diet can worsen the metabolic acidosis associated with mitochondrial or other metabolic disorders and should be managed by a practitioner experienced with these diets.
Although standard epilepsy surgery may be helpful for controlling seizures, there does not appear to be good evidence supporting the notion that standard epilepsy surgery improves cognition or symptoms associated with ASD, and there are many cases in which standard epilepsy surgery has worsened these factors. Several case series have suggested that MST may improve both seizures and ASD related symptoms. Impressive outcomes have been demonstrated in one study in which children underwent extensive electrophysiological study and had careful surgically treated foci with MST. However, MST requires more extensive study before it can be routinely used as a treatment option for children with ASD who have refractory epilepsy.
Given the growing literature on immune dysregulation in ASD (276
), it would not be surprising if immunomodulatory treatments would be useful in children with ASD and seizures. There are promising studies for the use of both IVIG and steroids for the treatment of epilepsy, ASD related symptoms, and seizures in individuals with ASD, although none of these studies are high quality. Given that these treatments target specific pathophysiological mechanisms, the development of clear guidelines for identifying children with ASD with or without epilepsy who might benefit from such treatment would be helpful, especially given the potential adverse effect of long-term use of immunomodulatory therapies such as steroids. Clearly this is an area that is ripe for further clinical study.
Neurofeedback is an interesting emerging treatment for both seizures and ASD symptoms that has an excellent safety profile and has growing evidence for its effectiveness. However, studies have been limited to particular individuals who could cooperate with the treatment protocol and have underrepresented very young children and adults as well as lower functioning individuals and those with more severe ASD symptoms. Because of the safety of neurofeedback, this may be a promising treatment for children with ASD but further blinded studies would strengthen the evidence of effectiveness and efficacy for this therapy.
Thus, several non-AED traditional therapies demonstrate considerable promise in the treatment of seizures in individuals with ASD, including low-carbohydrate diets such as the KD and MAD, MST, immunomodulatory therapy, and neurofeedback, although more research is needed in all of these areas to gain better evidence for the efficacy and effectiveness of these therapies as well as which specific ASD subgroups might best respond to these therapies. Standard epilepsy surgery therapy raises the risk of making ASD symptoms worse and the VNS has not been shown to improve ASD symptoms in several studies. Thus, these latter two therapies should probably be reserved for epilepsy that is refractory to other epilepsy treatments.
Treatments for seizures in genetic and metabolic syndromes associated with autism spectrum disorder
This review examined treatments for specific syndromes associated with ASD and seizures, including genetic syndromes such as TSC and Fragile X and metabolic disorders such as mitochondrial disease and dysfunction, urea cycle disorder, succinic semialdehyde dehydrogenase, branched-chain ketoacid dehydrogenase kinase, creatine and biotinidase deficiency, Smith–Lemli–Opitz syndrome, pyridoxine-dependent and responsive seizures, organic acidemias, and abnormalities of folate and cobalamin metabolism.
Mitochondrial disease and dysfunction and abnormalities in cerebral folate metabolism are two metabolic abnormalities that are associated with seizures and appear to affect a substantial portion of individuals with ASD. Several therapies for these two disorders have demonstrated effectiveness in the treatment of symptoms associated with ASD in controlled studies, particular l-carnitine, multivitamins with antioxidants, N-acetyl-l-cysteine, and folinic acid. Folinic acid has also been shown to improve seizures in individuals with seizures and cerebral folate abnormalities in uncontrolled studies. Thus, these treatments may be very useful for some individuals with ASD and seizures and should be strongly considered in selected cases.
Evidence also exists for a defect is cobalamin associated metabolism in ASD although not for a cobalamin deficiency itself. Still there is limited evidence that methylcobalamin can improve ASD symptoms and glutathione metabolism in children with ASD and other evidence that cobalamin supplementation can be useful in seizures due to cobalamin deficiency. Because methylcobalamin appears to be potentially useful for individuals with ASD, this treatment could be useful for those with ASD and seizures, but further research is needed before such recommendations can be made.
Several case studies have suggested that ASD can occur in specific metabolic syndromes that have specific treatment. In limited studies, within the context of the specific syndrome, treatments for creatine deficiency, d-glyceric aciduria and pyridoxine-dependent and responsive seizures appear to be helpful for treating seizures, while treatments for creatine deficiency, d-glyceric aciduria, and urea cycle disorders appear to be helpful for ASD associated symptoms. Treatments for biotinidase deficiency, branched-chain ketoacid dehydrogenase kinase deficiency, and semialdehyde dehydrogenase deficiency do not appear to be helpful for ASD symptoms or seizures in individuals with ASD, but this is based on limited reports.
For the two genetic syndromes reviewed, there are no clinically tested treatments that address the underlying pathophysiological mechanisms believed to be involved, although bench research is actively ongoing. For TSC, vigabatrin is the AED of choice for infantile spasms and seizures starting in infancy and appears to be most effective when started early in life. For Fragile X, standard AED therapy appears to be the mainstay at this time. Hopefully ongoing clinical trials will provide evidence for novel efficacious therapies that address the underlying pathophysiology of these syndromes.
This review provides guidance for treatment of seizures in ASD for specific syndromes and suggests that certain novel treatments, including l-carnitine, multivitamins with antioxidants, N-acetyl-l-cysteine, and folinic acid, may be helpful in a wider number of individuals with ASD and seizures.
Potential novel treatments for seizures in autism spectrum disorder
This review examined novel treatments that may have potential use for individuals with ASD and seizures. Such novel treatments included Mg, pyridoxine and Mg combined, Zn, DMG, taurine, l-carnosine, omega-3 fatty acids, homeopathy, the GFCF and Feingold/elimination diets, and low-frequency repetitive transcranial magnetic stimulation.
Mg has good evidence as an adjunctive therapy for seizures but does not have evidence to support its use for ASD associated symptoms. However, there is evidence that Mg combined with pyridoxine can improve ASD associated symptoms. As neither pyridoxine nor Mg alone appear to be detrimental to ASD associated symptoms and pyridoxine itself can be a treatment for underlying metabolic deficiencies, the combination of Mg and pyridoxine for treatment of individuals with ASD and seizures could be a good treatment to investigate. For an adjunctive treatment for individuals with ASD and seizures, Mg is a good candidate, especially since it can also treat constipation which is a symptom commonly seen in individuals with ASD, particularly those with mitochondrial abnormalities.
There has been significant interest in the use of omega-3 fatty acids for epilepsy and for ASD. Preliminary reports are encouraging but it is clear that the effects are subtle and require larger clinical samples to obtain statistically significance in clinical studies. One interesting aspect of this supplement is its potential cardiovascular benefits in epilepsy patients, specifically improvement in the index of sudden unexplained death in epilepsy. This is particularly important as individuals with ASD and epilepsy appear to have higher rates of unexplained mortality than individuals with ASD without epilepsy (302
The GFCF diet appears to be useful in some children with ASD and has been rated as improving seizures in a controlled survey study of ASD patients with seizures. Interestingly, a recent case report has suggested that combining the GFCF diet with the KD may have some utility in refractory epilepsy in ASD. The GFCF diet is also a milk-free diet which can decrease the folate-receptor alpha autoantibody. Thus, the GFCF diet could have some utility for seizure control in families that are able to implement it, although controlled studies are clearly needed to determine this possibility.
l-Carnosine has evidence of usefulness in the treatment of ASD associated symptoms and bench research suggests that carnosine and its active metabolite may be therapeutic in animal models of epilepsy. Thus, l-carnosine might be a novel treatment that has usefulness in the treatment of seizures in individuals with ASD. In addition, Zn has been shown in bench research to modulate neuroexcitability and in clinical studies has been found to be abnormal in individuals with seizures. As Zn abnormalities have been described in individuals with ASD, this may be a novel treatment for individuals with ASD and seizures.
Transcranial magnetic stimulation is an emerging treatment that has shown efficacy in both seizures and ASD symptoms in limited studies and appears to be safe in children. Although this treatment requires a relatively cooperative patient, thus limiting the specific ASD patients in which it is applicable, it might be a beneficial therapy for selected patients with refractory epilepsy. Clearly more research and high-quality clinical trials would be useful to further investigate this therapy.
Certain treatments, such as taurine and homeopathy, have the potential to be detrimental, at least based on limited evidence, thus suggesting that they should be generally avoided in individuals with ASD and seizures. In limited studies, DMG does not appear to be useful for treatment of seizures or ASD symptoms, so it will probably not be generally useful for the treatment of seizures in ASD.
Thus, this review suggests that certain novel treatments, such as Mg with pyridoxine, omega-3 fatty acids, the GFCF diet, and low-frequency repetitive transcranial magnetic simulation could be useful for the treatment of seizures and ASD symptoms while other treatments such as Zn and l-carnosine hold some promise. Other treatments such as taurine, DMG, and homeopathy are probably unlikely to be of use.
Limitations of previous studies and guidelines for future studies
Many of the studies reviewed have substantial limitations, particularly in their approach to documenting the effectiveness of treatments, defining ASD and following improvement in seizures and ASD symptoms. Few studies are high-quality, controlled, and/or blinded, and even the high-quality studies have relatively small populations thus limiting the generalizability of the findings. Many studies do not have well-defined populations and do not use standardized tools such as the Autism Diagnostic Observation Schedule or the Autism Diagnostic Interview – Revised to document the diagnosis of ASD, and other important factors such as language and intellectual development are not clearly defined in most studies. Furthermore, many studies do not quantitatively measure changes in seizures and many studies use various criteria to define seizure improvement. There is also a great need for long-term studies, since many of the treatments are used for years or even decades, whereas most clinical studies are usually months in duration, and some of the medications may have long-term adverse effects. Thus, in the future, promising treatments for seizures in individuals with ASD must be evaluated in high-quality DBPC studies in order to establish effectiveness, with long-term follow-up.