Ketogenic Diet and Epilepsy: What We Know So Far

1Epilepsy is a disabling and common neurological disease, which can be controlled successfully in most patients with one or more antiepileptic drugs. Approximately 30% of patients with epilepsy have refractory epilepsy, that is, have a failure of adequate trials of two tolerated, appropriately chosen and used antiepileptic drug schedules to achieve sustained relief of seizures. Some of these patients are not surgery candidates, so it is necessary to search for alternative treatments for epilepsy such as palliative surgery, neuromodulation, and a ketogenic diet (KD).

2The classic ketogenic diet (CKD) consists of a high-fat and low-protein and carbohydrate diet, with restricted calories and fluids. The diet mimics the fasting state, altering the metabolism to use fats as a primary fuel source; catabolism of fatty acids in the liver produces ketone bodies (KB), which induces urinary ketosis.

3Recent studies have found a significantly positive outcome with the use of the KD for treatment of refractory epilepsy in children and adults.

4Regardless of the efficacy of the KD, most patients discontinue the diet because of its unpalatable and restrictive features. In the last 20 years, new variants of the KD diet have emerged, including the Modified Atkins diet (MAD), a low-glycemic-index diet, which although it has a high fat content, allows more protein and does not restrict calories and fluids. Several studies have shown that the new variants of the KD have a similar efficacy to the CKD. As presently understood, the KD is involved in multiple mechanisms responsible for biochemical alterations, including cellular substrates and mediators responsible for neuronal hyperexcitability. However, it is not yet known with certainty whether the success of the KD is due to a single or several mechanisms.

5Because epilepsy is a metabolic disease, interest in studies of alterations of metabolism by anticonvulsants such as the KD has increased, as has their importance for the treatment of drug-resistant epilepsy. This contribution reviews the use and effects of the KD and its variants for the treatment of adults and children with intractable epilepsy.

6Dietary treatments for diseases have probably been used for over 2000 years. Fasting is the only therapeutic measure against epilepsy recorded in the Hippocratic collection. Two Parisian physicians, G Guelpa, and A Marie, recorded the first modern use of starvation as a treatment for epilepsy in 1911. The modern use of this form of therapy began in the early 1920s, when Drs. Stanley Cobb and W.G. Lennox of Harvard at Harvard Medical School observed the effects of starvation as a treatment for epilepsy, noting that seizure improvement typically occurred after 2–3 days. In the same period, Dr. Russel M. Wilder a physician at the Mayo Clinic in Minnesota, suggested that a specific diet could produce similar benefits to fasting, and proposed a diet that produced ketonemia. He studied a series of patients with epilepsy and demonstrated a result equivalent to fasting and that was maintained for a much longer period. This new concept of diet was designated the “KD.” Peterman, also at the Mayo Clinic, described a composition of the KD similar to that used today.

7In 1970, Robert C. Atkins developed a weight-loss diet that restricted the intake of carbohydrates, and this diet was later evaluated for seizure treatment. The first patient was a 7-year-old girl with intractable epilepsy due to a left parietal cortical dysplasia, who used the Atkins diet for a week in order to acclimate to the CKD. After 3 days, her seizures stopped, and she remained seizure-free for 3 years with continued dietary treatment. In 2006, this diet was first formally referred to as the “MAD” to distinguish it from the Atkins diet. The MAD has three significant differences from the first version: the induction phase of limiting carbohydrates is maintained indefinitely; high-fat foods are not only allowed, but encouraged; and the primary goal of the diet is seizure control.

8Nowadays, despite the new generation of anti-epileptic drugs, 35% of patients remain refractory. Interest in dietary therapy continues as a means of treatment for this group, even more with advances in knowledge regarding the association of gut microbiota and neurological diseases.

9The CKD is rich in lipids (90%) and low in carbohydrates and protein, in order to produce ketosis, and simulates a starvation state. It is a rigid diet, mathematically and individually calculated, and medically monitored. It must also provide adequate vitamins and minerals. The shift in the energy metabolism from glycolytic energy production to energy generation through oxidative phosphorylation (fatty acid b-oxidation and ketone-body production) is part of the anticonvulsant mechanism of the KD. This is discussed in more detail in the section on the mechanism of action.

10Traditionally, the KD has been considered the gold standard for the treatment of metabolic diseases such as Glucose Transporter Protein 1 (GLUT-1) deficiency syndrome and Pyruvate Dehydrogenase Deficiency. At present, the KD has been consistently reported as more beneficial, with more than 70% patients showing positive responses, as opposed to the average 50% response in several conditions such as infantile spasms (Table 1). The KD has also been used in other conditions with less evidence, but possible benefits (Table 2). Additionally, the KD is an important alternative treatment for patients with refractory epilepsy that are not surgery candidates.

11proposed that dietary therapy should be considered earlier as an option for treatment of intractable epilepsy, because of its proven efficacy, the poor chance of improvement with further anticonvulsant administration, and the possibility of using the MAD and low-glycemic-index treatment (LGIT), which are easier to manage in adults.

12In contrast, some pathologies are considered contra-indicated for KD. Absolute contraindications have been described and summarized by (Table 3). The surgical epilepsies, whenever the patient or caregivers are having difficulty maintaining compliance with the diet, are relative contra-indications for KD (Table 4).

13To obtain the optimum engagement of the family and the patients, providing information and training is essential because the diet is difficult to maintain. Counselors should talk with the family about their expectations and make clear the efficacy rate and adverse events (AE), to reduce the abandonment of the diet. Websites, videos and publications, especially from support groups, can be very helpful and should be encouraged. It is also important to review the medications and change from oral solutions (carbohydrate content) to tablets. The KD counseling, evaluation and follow-up should be done by a multidisciplinary team. A pediatric neurologist or neurologist and a nutritionist are the minimum team requirements.

14Before starting the diet, the patient should maintain a seizure diary to establish a frequency parameter. Also needed are a laboratory evaluation including selenium and carnitine levels (Table 5), electroencephalogram (EEG), and a magnetic resonance image (MRI) of the brain. A renal ultrasound should be done in case of kidney stones; an electrocardiogram and carotid ultrasound are considered optional. The nutritional evaluation includes a nutritional anamnesis including a 3-day food report, food habits, allergies, aversions, and intolerances. Baseline weight, height, and the ideal weight for stature and body mass index (BMI) are needed to calculate the ketogenic ratio, calories, and fluid intake. The diet formulation should be established according to the patient’s age and the administration route.

15The goal is to reach a ratio of four portion of fat to one portion of protein plus carbohydrate, described as “4:1.” To achieve this level, one of two approaches, with or without fasting, may be used. In the former approach, the patient must be hospitalized for 12–48 h, or when ketones are present in the urine, to prevent the development of hypoglycemia and dehydration. This method tends to accelerate the development of ketosis although it can generate more stress on the patient. When ketosis is reached, the meals are calculated to maintain a constant KD ratio, while calories are added until full-calorie meals are tolerated. The latter approach requires no hospitalization and the KD ratio increases weekly, from 1:1, 2:1 and 3:1 to 4:1. Most of the literature suggests that there is no significant difference between the two approaches in terms of the time needed to reach ketosis and the occurrence of hypoglycemia, so nowadays patients tend to not fast.

16Taking into account that the KD provides only small amounts of fruits, vegetables, grains, milk and cheese, supplementation is essential. Low-carbohydrate multivitamin and mineral supplements should be taken daily.

17Patients on the KD should be seen regularly every 3 months, and the family should be able to easily contact the diet team to resolve possible doubts and discuss adverse effects. In each evaluation, the seizure dairy and the child’s cognitive development and behavior should be observed. It has been noted that it is possible to improve the cognitive development and behavior even without a change in the seizure frequency. Although some authors have reported no relationship between the efficacy and the level of ketosis, it is still recommended to measure the urine ketosis several times a week.

18For efficacy, the KD requires a period of at least 3 months from the time that the patient reaches ketosis, so it is important to encourage the patient and the family to continue with the diet for this period without changing the medication.

19Because KD is not a physiological diet, it is necessary to recognize and closely manage AE. Acute AE include dehydration, hypoglycemia, lethargy, metabolic acidosis, and gastrointestinal symptoms. However, most of the side effects involve weight loss, high levels of low-density lipoprotein, and elevated total cholesterol. Other important AE are gastrointestinal symptoms, which include constipation, diarrhea, vomiting, and abdominal pain.

20The family should also be informed about how to recognize the symptoms of hypoglycemia and be advised to administer a small amount of juice or other forms of dextrose. Nephrolithiasis may also develop, and an abdominal ultrasonography should be requested.

21The understanding of the mechanisms of action of KD is incomplete; however, some theories have been advanced about how it modifies the neuronal metabolism and excitability in order to reduce the seizure frequency. Possibly, the real mechanism of reduction of cortical hyperexcitability involves multiple factors. Some of the systems involved in seizure reduction are related to metabolic changes in the blood and cerebrospinal fluid (CSF), including a decrease in glucose levels and an increase in KB. The mitochondria function and energy reserve may also play a role in the KD mechanisms, resulting in synapse stabilization and excitatory decrease.

22Ketone bodies, acetoacetate, and β-hydroxybutyrate (βOHB), are byproducts of fatty acid oxidation in the mitochondrial matrix of the hepatocytes. There are many theories about the role of KB, but the existence of an anticonvulsant effect is controversial. Some authors have found no relationship between KB and synaptic transmission and seizure control.

23Experimental studies in an animal model showed that in rats exposed to KD there was no change in synaptic plasticity, using paired-pulse modulation and long-term potentiation. Similarly, did not detect any anticonvulsant effects in either ketone body. In spontaneously epileptic Kcna1-null mice, KB supplementation resulted in attenuation of electrographic seizure-like events. These authors also observed an inhibitory effect of KB on mitochondrial permeability transition related to apoptotic and necrotic death. Moreover, in experimental models, acetoacetate exerted a broad-spectrum anticonvulsant effect. In another study, described a relationship among KB, neurotransmitter release and ATP-sensitive potassium channels. Similarly, to these studies, injection of KB led to the reduction of seizure susceptibility. found a decrease of the spontaneous firing rate in sections of mouse tissue, which was eliminated in the absence of ATP-sensitive potassium channels (KATP). In addition, KB can exert a direct inhibitory effect on the vesicular glutamate transport. It is possible that these divergent results are related to the different concentrations of KB used in these studies and the diverse seizure thresholds of the animal models. These conflicting results can be also explained by differences in diet composition.

24Another hypothesis regarding the function of the KD is related to changes in neuronal metabolism, mitochondrial function and energy reserve, and the environment. In normal conditions, the usual substrate for the neurons is glucose. To facilitate its diffusion through the brain-blood barrier, glucose transports are present in the brain capillary endothelial layer. The glucose metabolism produces the rapidly available energy that is necessary for seizure activity. Therefore, in patients on the KD, the blood glucose energy levels are low, and the brain begins to use KB for energy. This anaerobic metabolism slows the energy availability, which reduces seizures. The anticonvulsant propriety of a decrease in glucose metabolism has been shown in experimental models in which the administration of 2-Deoxy-D-glucose elevates the seizure threshold. The anticonvulsant effect of the KD can be quickly reversed after glucose infusion. Based on these data, we can postulate the influences not only of the KB, as discussed above, but also the reduction in glucose levels as a mechanism of action of the KD.

25Chronic ketosis may play a role in the KD anticonvulsant properties, since it has been shown that chronic ketosis elevates the brain energy reserve via stabilization and reduction of excitability of synapses. The energy reserve is directly associated with mitochondria, which is an important element to consider in the antiepileptic effect of KD. demonstrated an increase in mitochondria biogenesis in an experimental model of rats fed with KD, indicating an increase in the energy stores. The increase in mitochondrial metabolism leads to an increase in ATP production, which activates KATP, in turn attenuating neuronal excitability. This activation may be associated with adenosine A1 receptors and GABAB receptors.

26In this process, we can postulate that modifications of the metabolism are associated with an increase of ATP, and improve mitochondrial capacity and cell energy, with an increase in metabolic resilience.

27The KD-induced synaptic stabilization is additionally related to changes in critical amino acids as a result of ketone metabolism. It has been proposed that KD interferes with the concentration of gamma-aminobutyric acid (GABA), the major inhibitory neurotransmitter. There is evidence in clinical practice of increased GABA levels in the CSF of patients on the KD diet. The decrease in aspartate levels promoted by KB lead to the synthesis of GABA. This occurs because of the inhibitory effect of aspartate on glutamate decarboxylase and the facilitation of the conversation of glutamate to glutamine in the astrocytes. Not only can GABA be increased, but also other neurotransmitters such as adenosine A1 can be implicated in the anti-seizure effect of the KD. However, more evidence is needed.

28The role of gut microbiota has recently been studied for its effect on several diseases, especially those with some inflammatory involvement. Several metabolic pathways are known to be modulated by the gut microbiota. demonstrated the impact of gut microbiota on the anti-seizure effect of KD. She found that KD modifies the gut microbiota, with a decrease in alpha-diversity and increases in the putatively beneficial bacteria Akkermansia muciniphila and Parabacteroides spp. This microbiota transformation leads to changes in the colonic luminal metabolome, with a decrease in gamma-glutamyl amino acids. This increases the GABA/glutamate content in the brain by decreasing gamma-glutamyl amino acids in the blood. In an acute electroshock model, it is reported that KD confers protection against seizures. Moreover, KD decreases the frequency of spontaneous seizures in Kcna1 knockout mice. In summary, changes in the gut microbiota seem to be important for the KD-mediated seizure protection.

29The role of inflammatory cytokines in epilepsy is well known, and there is evidence that KD also interferes with pro-inflammatory cytokines. showed a peripheral and brain reduction of interleukin 1β and other pro-inflammatory cytokines in rats treated with KD in the LPS model.

30Notably, there is a relationship between metabolic and epigenetic modifications. observed that βOHB inhibits class I histone deacetylases. During the KD, the elevation of βOHB causes changes in large-scale gene transcription but particularly those linked to oxidative-stress resistance factors. This result emphasizes that the KD has a potential role as a disease-modifying treatment in epilepsy.

31In conclusion, all the mechanisms described above lead to systemic modifications and a dynamic metabolic homeostasis, in which the interplay among KB, glucose levels, mitochondrial function, synaptic neurotransmitters, and channel modifications can lead to changes in the seizure threshold and hyperexcitability. These changes contribute to the final antiseizure mechanism of KD.

32Multiple mechanisms of action may explain why the modification of the KD can be effective even without ketosis. Importantly, the KD systemic action can have a broad spectrum of effects that may be beneficial in the treatment of different types of epilepsy and associated comorbidities such as cognition impairment, psychiatric disturbance, and sudden unexplained death.

33The MAD aims to provide increased flexibility and palatability, with a 1:1 ratio of fat to carbohydrates and protein, and contains around 65% fat, 25% protein, and 10% carbohydrate. Fat is encouraged and the carbohydrate intake is limited to 10–20 g/day in children and 15–20 g/day in adults. Because of carbohydrate restriction, the MAD can also produce urinary ketones. The MAD does not require weighing food on a gram scale, or restriction of calories, protein or liquids, and may be a good option for patients who are unable to tolerate a more restrictive diet such as the classical ketogenic diet (KD). Low-carbohydrate multivitamin and calcium carbonate supplementation is recommended in the MAD.

34Several studies have shown that the MAD, besides being more palatable, is as effective as the KD in the treatment of drug-resistant epilepsy in children. A study performed using 20 children receiving 10 g of carbohydrates daily showed that 65% of the children had a >50% seizure reduction, 35% of the children had >90% improvement, and four children were seizure-free at 6 months. In a study in South Korea, 36% of 14 children treated with the MAD showed improvement of >50% in seizures and 12% were seizure-free. A recent meta-analysis performed using 70 studies concluded that the MAD and classical KD do not differ in reduction of seizure frequency at month 3 and month 6, with ≥50% and ≥90% reductions, respectively. A retrospective study showed >50% of seizure reduction in 65% of the 10 children who remained on the diet for up to 6 months, and 20% of them were seizure-free.

35Treatment with MAD was shown to be more effective in seizure control when the MAD was started with lower carbohydrate limits. In a randomized study with 20 children with drug-resistant epilepsy, 60% of them showed fewer seizures in the first 3 months on the MAD, with 10 g/day of carbohydrate intake against 10% of reduction with 20 g/day (p = 0.03). In the same study, after 3 months, an increase in carbohydrate intake to 20 g/day, maintained seizure control and improved tolerability, suggesting that a lower carbohydrate limit is important only in the first 3 months.

36The efficacy of the MAD is also proven for the treatment of drug-resistant epilepsy in adults and adolescents. In this patient group, carbohydrate intake is generally around 15–20 g/day and the rates of seizure reduction and adherence are lower compared to those of the child population.

37In a recent meta-analysis, eight studies were identified that used the MAD in adult patients with refractory epilepsy, aged between 15 and 86 years, with treatment times ranging from 3 to 36 months. In these studies, the proportion of patients who showed >50% seizure reduction ranged from 20 to 70% and the rate of seizure freedom ranged from 7 to 30%. The rate of abandonment of the diet varied between 12.5 and 82% of the patients.

38The MAD has been shown to be better tolerated than the classical KD, but some typical side effects such as gastrointestinal complaints, dyslipidemia and weight loss are reported. Beneficial effects have also been reported, such as mood improvement.

39The use of the LGIT in the treatment of drug-resistant epilepsy was initially reported in 2005 by. This alternative diet treatment is based on a ratio 0.6:1 of fat to carbohydrates and protein, containing 60% fats, 30% protein, and 10% carbohydrates with a low glycemic index (GI) (GI<50). The GI measures the tendency of a food to raise the blood glucose levels, compared to an equivalent amount of the reference carbohydrate, usually glucose. Compared to classic the KD, the LGIT produces a smaller increase in ketone body levels, but has comparable efficacy, better tolerability and easier implementation.

40The LGIT has proven to be effective in the treatment of focal and generalized epilepsies, with a reduction in seizure frequency occurring at 3–14 months and seizure control continuing for at least 1 year after the end of treatment. reported the use of LGIT in 20 drug-resistant epilepsy patients aged 5 to 34 years. After an average of 20 weeks of treatment, 50% of the patients had a >90% reduction in seizures. studied 15 children, adolescents and young adults with refractory epileptic encephalopathies treated with LGIT. After 12 months they found 75–90% seizure reduction in 6 patients (40%), 50% reduction in 2 patients (13.3%), and the seizure frequency unchanged in 7 (46.7%). In a retrospective review of LGIT in 76 children, found an overall >50% reduction in seizure frequency in 50% of the patients at 3 months, which reached 66% at 12 months.

41However, according to a recent systematic review, the positive results for LGIT efficacy in epileptic patients are doubtful because of the low number of high-quality studies. In this review, which included all electronic literature databases until July 2017, the authors found only eight studies with good or fair quality (69).

42Constipation and vomiting are the most common adverse effects reported in patients on the LGIT.

43The CKD and its variants should be considered as an alternative for non-surgical pharmacoresistant patients with epilepsy, of any age. Each patient must have an individually designed diet; however, adult patients have more difficulty in maintaining the CKD. It is essential to inform the patient and the family about the efficacy and AE related to the KD, and the use of websites and videos may help in this education. Although several theories exist regarding the mechanisms of action, further study is needed nevertheless the positive results are probably due to several mechanisms.

44All the authors contributed substantially to the writing and revising of the manuscript. ID’A, HP, TR, MP, PC, and LK participated in the conception and design of the study, collected the literature, prepared the tables, and wrote the manuscript. ID’A, TR, and HP reviewed and edited the manuscript and approved the final version.

45The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.