The aim of this study was to evaluate the efficacy of the modified Atkins diet in a randomized controlled trial in children with refractory epilepsy.
The aim of this study was to evaluate the efficacy of the modified Atkins diet in a randomized controlled trial in children with refractory epilepsy.
Children aged 2–14 years who had daily seizures despite the appropriate use of at least three anticonvulsant drugs were enrolled. Children were randomized to receive either the modified Atkins diet or no dietary intervention for a period of 3 months. The ongoing anticonvulsant medications were continued unchanged in both the groups. Seizure control at 3 months was the primary end point. Analysis was intention to treat. Adverse effects of the diet were assessed by parental reports (ClinicalTrials.gov Identifier: NCT00836836).
Among a total of 102 children, 50 were in the diet group and 52 in the control group. Four children discontinued the diet before the study end point, and three children in the control group were lost to follow-up. The mean seizure frequency at 3 months, expressed as a percentage of the baseline, was significantly less in the diet group: 59 ± 54 (95% confidence interval [CI] 44–74.5) versus 95.5 ± 48 (95% CI 82–109), p = 0.003. The proportion of children with >90% seizure reduction (30% vs. 7.7%, p = 0.005) and >50% seizure reduction was significantly higher in the diet group (52% vs. 11.5%, p < 0.001). Constipation was the most common adverse effect among children on the diet (23, 46%).
The modified Atkins diet was found to be effective and well tolerated in children with drug-refractory epilepsy.
One third of children with epilepsy have seizures that are refractory to anticonvulsant medications. Several catastrophic epilepsies present in childhood, including West syndrome, Dravet syndrome, and Lennox-Gastaut syndrome. Seizures in these disorders are difficult to control, and sometimes only at the expense of multiple anticonvulsant medications (Patsalos & Duncan, 1993). In addition, many of these children are not good surgical candidates.
The ketogenic diet is a medically supervised high fat, low carbohydrate, and restricted protein diet that maintains a chronic state of ketosis while providing proteins and calories for adequate growth. The modified Atkins diet is a less restrictive alternative to the traditional ketogenic diet. This diet is started on an outpatient basis without a fast, allows unlimited protein and fat, and does not restrict calories or fluids (Kossoff & Dorward, 2008). Many recent studies have demonstrated efficacy and safety in refractory epilepsy (Kossoff et al., 2006; Kang et al., 2007; Weber et al., 2009; Miranda et al., 2011). However, there have been no randomized controlled trials. Hence this randomized controlled trial was conducted to investigate whether modified Atkins diet improves seizure control in children with refractory epilepsy.
This open-label, parallel-group, randomized-controlled trial was conducted in the pediatric department of a tertiary care hospital between May 2009 and March 2011. Ethical clearance was obtained from the institutional ethical committee. Written informed consent was obtained from the parents. The study was registered with the clinicaltrials.gov. (ClinicalTrials.gov Identifier: NCT00836836). The study design was similar to the randomized controlled trial of ketogenic diet in refractory epilepsy by Neal et al. (2008).
Children aged 2–14 years who had daily seizures (or more than seven seizures per week) despite the appropriate use of at least three anticonvulsant drugs were enrolled. Children with known or suspected inborn errors of metabolism, systemic illness, or motivational issues in the family that would preclude compliance were excluded.
Each child underwent detailed history and examination. The seizure type, frequency, age at onset, perinatal details, family history, developmental status, and treatment history were noted. Results of investigations such as neuroimaging and EEG were also noted. Eligible children were randomized using computer-generated random number tables in two groups: the intervention and the control arm. The allocation concealment was done using opaque sealed envelopes. Neither the treating physicians nor the parents of the children were blinded to the group allocation.
Both groups underwent a baseline 4-week observation period, during which parents were asked to maintain a daily seizure log; recording seizure type, duration, and frequency. Seizure frequency was assessed by the parent-maintained seizure records on the basis of a chart with various categories of seizure (infantile spasms, absence, myoclonic, atonic, tonic, tonic–clonic, and focal). The definitions and clinical manifestations of the different seizure types were clarified with parents.
In the intervention arm, the children started the modified Atkins diet after this 4-week baseline period. The control group received their normal diet with no additional dietary input, and remained on the same ongoing anticonvulsant medication for the 3 months. Anticonvulsant medication remained unchanged during the 3 month trial period in both groups, unless the change in the anticonvulsant regimen was medically indicated; for example drug side effects or status epilepticus; in which case, standard therapy was provided.
Children were reviewed as outpatients at 1, 2, and 3 months. A 3-day dietary intake chart was reviewed at each visit in the diet group to compute calorie and carbohydrate intake, and reinforce compliance.
The modified Atkins diet was started on an outpatient basis. Carbohydrate intake was restricted to 10 g/day. The carbohydrate content of various food items was explained to the parents, and exchange lists were provided. The intake of fat (e.g., cream, butter, oils, ghee) was actively encouraged. Protein intake was not restricted. There was no calorie restriction. A list of recipes based on common locally available foods, with prespecified carbohydrate contents (e.g., 2.5 g, 5 g, and so on) was provided. These recipes required weighing of some of the components that contained carbohydrates; the rest of the components could be added ad lib. The recipes were also planned considering the families and the child's preferences and cultural taboos such as vegetarianism and avoidance of eggs, onions, and garlic in some families. For vegetarians, soya products and milk-based products such as cottage cheese were used. Parents were also given a list of carbohydrate-free foods that they could be offered unrestricted to the children.
The ongoing anticonvulsant medications were continued unchanged. Each child also received a sugar-free, fat-soluble vitamin supplement and calcium supplementation. Parents were asked to check urine ketones by reagent sticks daily for the first week on the diet and twice weekly thereafter.
Seizure frequencies were recorded daily for the 4-week baseline period and the 3-month study period. At the end of the 3-month study period, the average seizures per week in the preceding 28 days was compared with the average weekly seizure frequency in the baseline period. This was expressed as a percentage change in the seizure frequency as compared to the baseline.
The proportion of children who had seizure freedom, >90% reduction in seizures, and >50% reduction in seizures as compared to baseline was also compared in both the groups. Tolerability of the diet and its side effects was evaluated by means of parental interview at each visit: vomiting, lethargy, poor appetite, refusal to feed and constipation. Any other parental concerns regarding side effects were also noted.
This trial was planned to test the hypothesis that the administration of the modified Atkins diet might be more efficacious than the continuation of antiepileptic medication with no other changes. A 25% difference in the primary outcome measure was thought to be of clinical importance. The sample size was based on an expected mean percentage of baseline seizures in the diet group of 75% (standard deviation [SD] 37.5%). The sample size of 48 children per group (1:1 allocation ratio) was estimated to enable the detection of a difference that was significant at 5% with a power of 90%.
Univariate analysis was done to assess the distribution of data in groups and to choose the appropriate statistical test. Mean percentage of seizures at baseline and 3 months was compared between children on the Atkins diet and controls using unpaired t-test/Mann-Whitney U test. Responder rates of seizure freedom, greater than either 50% or 90% seizure reduction were compared between groups using the Fisher's exact test. All analysis was intention-to-treat. For children whose outcome data was missing, a worst case scenario was assumed for the diet group (i.e., 0% seizure reduction), and a best case scenario was assumed for the control group, that is 100% seizure control.
A total of 102 children (78 boys) were enrolled in the study period. Fifty children were randomly assigned to the diet group, and 52 were assigned to the control group. Four children in the diet group discontinued the diet before the study end point, and three children in the control group were lost to follow-up (Fig. 1).
The baseline demographic and clinical characteristics of the two groups were comparable (Table 1). Comorbid developmental delay and cerebral palsy were commonly associated. In both the groups, tonic and myoclonic seizures were the most common seizure types. Lennox-Gastaut syndrome was the most common epilepsy syndrome (Table 1). Most children had mixed seizure types. A median of five anticonvulsant drugs (range three to nine) had been tried in the diet group, whereas four anticonvulsant drugs (range three to nine) had been tried in the control group. In both groups, the children were taking a median of three ongoing anticonvulsant medications (range two to four) during the study period. The mean daily seizure frequency was 30 (SD 33) in the diet group, and 27 (SD 29) in the control group.
|Diet group (n = 50)||Control group (n = 52)||p- Value|
|Age (years, mean ± SD)||4.7 ± 2.8||5.2 ± 3.3||0.4|
|Boys||41 (82%)||37 (71%)||1|
|Age at onset of epilepsy (years, mean ± SD)||1.4 ± 1.6||2 ± 2||0.1|
|Duration of epilepsy (years, mean ± SD)||3.4 ± 2.1||3.3 ± 1.9||0.8|
|Developmental delay||45 (90%)||45 (86.5%)||0.8|
|Cerebral palsy||32 (64%)||30 (57.7%)||0.5|
|Tonic||25 (50%)||23 (44.2%)||0.7|
|Myoclonic||23 (46%)||24 (51%)||1|
|Atonic||14 (28%)||13 (25%)||0.8|
|Absence||13 (26%)||7 (13.5%)||0.1|
|Tonic–clonic||7 (14%)||9 (17.3%)||0.8|
|Partial||10 (20%)||9 (17.3%)||0.8|
|Epileptic Spasms||9 (18%)||10 (19.2%)||1|
|Lennox-Gastaut syndrome||25 (50%)||22 (52.3%)||0.5|
|West syndrome (with persisting epileptic spasms)||9 (18%)||10 (19.2%)||1|
|Myoclonic astatic epilepsy||2 (4%)||3 (5.8%)||1|
|Partial epilepsy secondary to structural lesions||2 (4%)||3 (5.8%)||1|
|Othersb||3 (6%)||1 (2%)||–|
|Unclassified||10 (20%)||12 (23.1%)|
The most common etiology for epilepsy was structural brain injury secondary to perinatal asphyxia seen in 30 children (14 in the diet group, 16 in the control group). Other perinatal problems such as neonatal hypoglycemia, sepsis, and meningitis accounted for another 16 cases. Five children had neurocutaneous syndromes. Eighteen children had acquired brain insults after the neonatal period such as meningoencephalitis, postcardiac injury hypoxic sequelae, traumatic brain injury, and stroke. Malformations of cortical development were seen in three children. One child each had Rett syndrome and hemiplegia hemiconvulsion epilepsy syndrome. In 29 children (28%), the etiology of epilepsy was unknown.
The mean seizure frequency at 3 months, expressed as a percentage of the baseline, was 59% (95% confidence interval [CI] 44–74.5) in the diet group, as compared to 95.5% (95% CI 82–109) in the control group. The median seizure frequency at 3 months, expressed as a percentage of the baseline, was significantly less in the diet group as compared to the controls (37.3% vs. 100%, p = 0.003). The proportion of children with >90% seizure reduction (30% vs. 7.7%, p = 0.005) and >50% seizure reduction was significantly higher in the diet group (52% vs. 11.5%, p < 0.001) (Table 2). Five children in the diet group were seizure free at 3 months compared with none in the control group. No significant differences were found in the response rates among different seizure types or epilepsy syndromes.
|Diet group (n = 50)||Control group (n = 52)||p- Value|
|Mean percentage of seizures as compared to baseline||59 ± 54 (95% CI, 44–74.5)||95.5 ± 48 (95% CI, 82–109)||0.003|
|Median percentage of seizures as compared to baseline (IQR)||37 (7–107)||100 (75–123)||0.003|
|Proportion of children with >90% seizure control||30%||7.7%||0.003|
|Proportion of children with >50% seizure control||52%||11.5%||0.001|
The modified Atkins diet was generally well tolerated. Twenty-two of the children on the diet (44%) were vegetarian; 12 of them did not eat eggs either. The most common adverse effect was constipation, which was seen in 46% of the children (Table 3). Other side effects included anorexia (18%), lethargy (6%), and vomiting (10%). Two children developed frequent chest infections, and one child developed hyperammonemic encephalopathy 1 week after starting the diet. This child had refractory seizures of unknown etiology. The metabolic screen (blood ammonia, arterial blood gas, blood lactates, blood tandem mass spectrometry, and urine gas chromatography mass spectrometry) prior to starting the diet was normal. In these three children, the diet was discontinued. The patient with hyperammonemic encephalopathy improved after discontinuation of the diet. Another child discontinued the diet 1 month after starting because the child and the family found the diet too restrictive. Older children who were cognitively better preserved and able to express their feelings also complained that they found the diet too restrictive and difficult, as it did not contain the tradition staple components of Indian meals; that is, rice and chapattis (Indian wheat breads). Parents of 23 children on the diet reported improved alertness and interaction.
|Adverse effect||Number (%)|
|Lower respiratory tract infections||2 (4)|
|Hyperammonemic encephalopathy||1 (2)|
The ketogenic diet is one of the established nonpharmalogic options for control of intractable epilepsy. In this respect the ketogenic diet compares favorably with the newer anticonvulsant drugs developed for the treatment epilepsy in children (Freeman et al., 1998). The ketogenic diet has also been shown to be effective in a recent randomized controlled trial (Neal et al., 2008). However, the traditional ketogenic diet, with 4:1 ratio of fat to carbohydrate + protein has its drawbacks. It restricts calories and fluids, and requires weighing of foods. Protein is generally restricted to 1 g/kg/day, with the majority of the remaining calories coming from fats. This may lead to hypoproteinemia (Ballaban-Gil et al., 1998; Sharma et al., 2009) and growth problems (Peterson et al., 2005). Hospitalization is generally advocated for diet initiation. Long-term side effects of the diet include kidney stones, constipation, acidosis, diminished growth, weight loss, and hyperlipidemia (Kang et al., 2004). The modified Atkins diet is a less-restrictive alternative to the traditional ketogenic diet. This diet is started on an outpatient basis without a fast, allows unlimited protein and fat, and does not restrict calories or fluids (Kossoff & Dorward, 2008).
In this randomized-controlled trial, the modified Atkins diet was found to be significantly more effective in controlling seizures in children with refractory epilepsy, as compared to the continuation of anticonvulsant drugs alone. The median seizure frequency at 3 months, expressed as a percentage of the baseline, was significantly less in the diet group as compared to the controls (37.3% vs. 100%, p = 0.003). The results were comparable to a similar randomized controlled trial of the ketogenic diet in children with refractory epilepsy (Neal et al., 2008). Neal et al., found that the median seizure frequency at 3 months, expressed as a percentage of the baseline, was 47.7% in the ketogenic diet and 106% in the control group. In the present study, half (52%) of the children on the diet had >50% seizure reduction at 3 months. This proportion is comparable to the efficacy seen in previous uncontrolled studies on the modified Atkins diet (Kossoff et al., 2006; Kang et al., 2007). This proportion is also similar to our center's past experience with the ketogenic diet (Sharma et al., 2009). The preponderance of male participants in our study possibly reflects the gender bias for referral in our society. This preponderance has been noted in earlier studies on dietary treatments of epilepsy from our center as well (Sharma et al., 2009; Raju et al., 2011; Sharma et al., 2012). The majority of the children in our study had catastrophic epilepsies such as Lennox-Gastaut syndrome and West syndrome, with mixed seizure types. We had few children with focal epilepsy. Although we did not specifically exclude surgical candidates, it is possible that surgical candidates with focal seizures were referred for surgery early. This factor could account for a favorable outcome in our study, as children with surgically amenable focal epilepsy do not respond as well to the ketogenic diet as children with generalized epilepsy (Stainman et al., 2007).
We found several advantages with the use of the modified Atkins diet. Apart from the outpatient initiation and ease of administration, the counseling time was reduced to 30–45 min. This is an ideal treatment option for resource-constraint settings with a paucity of trained dieticians (Kossoff et al., 2008a). Tedious individualized calculations are not needed with the modified Atkins diet. Our way of administration of the diet differs from the studies in the West, where parents and/or patients are given broad guidelines, and given the freedom to choose from the available labeled foods. This diet is therefore considered better for older children, adolescents, and adults (Kossoff et al., 2008b). In our scenario, this approach is not possible because of the lack of availability of labeled foods, and limited understanding of parents of poor socioeconomic strata. Hence, we found that providing simple written instructions, with choice of recipes with precalculated carbohydrate contents was feasible for the parents, although it did require some weighing of foods.
The modified Atkins diet was well tolerated. In most children, the side effects of the diet were manageable with medical measures and reassurance, and did not require discontinuation of the diet. The occurrence of lower respiratory tract infections in two children could be because of exacerbation of preexisting gastroesophageal reflux. The development of hyperammonemic encephalopathy in one child reinforces the need for continued careful medical supervision, as underlying inborn errors of metabolism may be missed on metabolic screens and worsen after initiation of the diet.
The diet does become very restrictive, all the more so for vegetarians. This was not a problem in young children or cognitively impaired children. Older children who were able to express their preferences did complain about the diet's restrictiveness and absence of cereal staples of Indian food. Perhaps the carbohydrate content can later be increased in older children, to allow approximation to the usual Indian diets. Kossoff et al. (2007) performed a crossover study to compare the carbohydrate limits of 10 and 20 g in the modified Atkins diet. A significantly higher likelihood of >50% seizure reduction was noted for children started on 10 g of carbohydrate per day at 3 months. However, most parents reported no change in seizure frequency or ketosis between groups, but improved tolerability with 20 g/day. The authors concluded that a starting carbohydrate limit of 10 g/day for children starting the modified Atkins diet may be ideal, with a planned increase to a more tolerable 20 g/day after 3 months.
The limitations of this study include the unblinded design, and absence of masked outcome assessment. In addition, the use of parental seizure records runs the risk of subjective errors, especially in children who have myoclonic and absence seizures, which are likely to be missed. Nocturnal seizures also may have been missed. In children with numerous daily seizures of multiple types, it was difficult for the parents to count all the seizures, and many times, approximate estimates were given. A pre- and postobjective long-term video-EEG assessment would have been the best way to assess the outcome; however, this was not feasible and beyond the scope of the present study.
In conclusion, the modified Atkins diet was found to be effective and well tolerated in children with refractory epilepsy. It is a feasible option even in resource-constrained settings with a paucity of trained dieticians. However, the diet does have side effects, and careful medical supervision is warranted.
SS was financially supported as a Senior Research Associate in the “Scientists pool scheme” of the Council for Scientific and Industrial Research (CSIR), Govt. of India, for this study. There was no other funding source for the study.
None of the authors has any other conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.