Modified Atkins diet is an effective treatment for children with Doose syndrome

Transcription

1 FULL-LENGTH ORIGINAL RESEARCH Modified Atkins diet is an effective treatment for children with Doose syndrome *Adelheid Wiemer-Kruel, Edda Haberlandt, Hans Hartmann, Gabriele Wohlrab, and *Thomas Bast SUMMARY Adelheid Wiemer- Kruel is a consultant at the Epilepsy Centre Kork, clinic for children and adolescents. Objective: Children with myoclonic astatic epilepsy (MAE; Doose syndrome) whose seizures do not respond immediately to standard antiepileptic drugs (AEDs) are at high risk of developing an epileptic encephalopathy with cognitive decline. A classic ketogenic diet (KD) is a highly effective alternative to AEDs. To date, there are only limited data on the effectiveness of the modified Atkins diet (MAD), which is less restrictive and more compatible with daily life. We report findings from a retrospective study on 30 MAE patients treated with MAD. Methods: Four participating centers retrospectively identified all patients with MAE in whom a MAD had been started before June Seven children were recruited from a cohort included in an open prospective controlled trial. A retrospective review of all available charts was performed in the other patients. Results: Thirty patients (24 boys) were included. Mean age at epilepsy onset was 3.1 years (range ). MAD was started at a mean age of 4.5 years (range ) after the children had received an average of six different AEDs (range 2 15). Mean MAD observation time was 18.7 months (range ). Twenty of 30 patients were still on MAD at the end of study (duration range , mean 18.5 months). MAD was stopped without relapse in three patients after sustained seizure freedom for >2 years. For the other seven cases, ineffectiveness (three patients), loss of efficacy (two), or noncompliance (two) led to termination. No severe adverse effects were noted. By the end of the observation period, 25 (83%) of 30 patients experienced a seizure reduction by 50% and 14 (47%) of 30 were seizure-free. None of the evaluated factors differed significantly between the groups of seizure-free and non seizure-free children. Significance: MAD is an effective treatment for MAE. It should be considered as an alternative to AEDs or the more restrictive classic ketogenic diet. KEY WORDS: Pharmacoresistance, Myoclonic astatic epilepsy, Ketogenic diet. Doose syndrome or myoclonic astatic epilepsy (MAE) is a rare electroclinical syndrome accounting for 1 2% of childhood epilepsies. 1 Onset is between 7 months and Accepted January 17, 2017; Early View publication February 23, *Epilepsy Centre Kork, Clinic for Children and Adolescents, Kehl- Kork, Germany; Department for Neuropediatrics, Children s University Hospital, Innsbruck, Austria; Department of Pediatrics, Hannover Medical School, Hannover, Germany; and Department for Neuropediatrics, Children`s University Hospital, Z urich, Switzerland Address correspondence to Adelheid Wiemer-Kruel, Epilepsy Centre Kork, Clinic for Children and Adolescents, D Kehl-Kork, Germany. awiemer@epilepsiezentrum.de Wiley Periodicals, Inc International League Against Epilepsy 6 years (typically 1 5 years), with a clear male preponderance. 2 4 Diagnostic criteria are normal development preceding the onset of epilepsy, absence of structural or metabolic cerebral abnormalities, and presence of 2 3/s spike- and polyspike-wave complexes on electroencephalography (EEG) without focal discharges. 2 5 Severe or benign myoclonic epilepsy in infancy and Lennox-Gastaut syndrome have to be excluded. 2 Myoclonic atonic seizures typically result in drop attacks. Alongside this mandatory seizure type, generalized tonic clonic, clonic and myoclonic seizures, and absences occur in various combinations. 5 The children are at risk of repetitive nonconvulsive status epilepticus. 3 5 Tonic seizures are rare, but are a known risk factor 657

2 658 A. Wiemer-Kruel et al. Patients and Methods Key Points Myoclonic astatic epilepsy (MAE) is difficult to treat and often leads to encephalopathy and cognitive decline Compared with antiepileptic drugs, ketogenic diets are very effective; to date, there are only limited data for the alternative modified Atkins diet (MAD) Our data for 30 children with MAE on the MAD verify seizure freedom in 47% and seizure reduction by 50% in 83% MAD is an effective treatment for MAE and could be an alternative to drug therapy or the more restrictive classic ketogenic diet for a complicated course. 3 5 The clinical course may, however, be diverse. Early seizure control and complete remission can be expected in about two thirds of the children. 2,3,6 Other patients have severe pharmacoresistant epilepsy and epileptic encephalopathy with cognitive decline. Seizure control, presence and/or duration of an epileptic encephalopathy, and drug effects may be relevant factors regarding cognitive outcome, 3 but this correlation is debatable. 6,7 Combinations of valproate (VPA), ethosuximide (ESM), lamotrigine (LTG), and clobazam (CLB) are the most effective drug treatments. 3 6,8 Classic ketogenic diet (KD) has been shown to be a very effective therapy in MAE. 3,8,9 Although there had not previously been any studies exclusively analyzing the efficacy of KD in MAE, based on results from mixed epilepsy populations an international expert group assessed MAE as one of the most promising syndromes for treatment with this diet. 10 This appraisal was supported by a recent retrospective study of nine patients with MAE where treatment with high-fat diets resulted in seizure freedom in seven and seizure reduction by 90% in the remaining two children. 11 Establishing a KD in preschool children is challenging because the diet is strict and requires careful calculation of daily food intake. According to our own experiences, compliance and long-term retention are limited. The modified Atkins diet (MAD) is a less restrictive alternative. 12 In general, it is better tolerated and accepted by patients and their families. 12,13 With MAD, only the total amount of daily carbohydrate intake is limited, with the diet typically initiated at 10 g carbohydrate per day for children. Calories, fluids, and proteins are not restricted, and the intake of fats is encouraged. The efficacies of MAD and KD were found to be almost equal. 14,15 In this current retrospective study, we report the experience of four centers with MAD in 30 children with MAE. To the best of our knowledge, this is the largest studied MAE population treated with MAD. Patients were identified and included in the study retrospectively at the four participating centers in June Inclusion criteria were a diagnosis of MAE based on electroclinical data and any attempt to establish a MAD. Criteria for the diagnosis of MAE included uneventful history until epilepsy onset, normal magnetic resonance imaging (MRI), generalized spike- and poly-spike-wave discharges on EEG, and the presence of typical clinical seizures. 2 5 In the case of more than one MAD trial in the same patient, only the first attempt was included in the study and analyzed. At Epilepsy Centre Kork, a standardized in-house protocol has been used since 2007 for the introduction and follow-up documentation of MAD. In 2013, the ethics committee of the Landes arztekammer Stuttgart approved the protocol, and since May 2013 all patients on MAD were studied in the context of an open prospective controlled trial. The patients with MAE were included in this current retrospective study because the number of prospectively evaluated children with this rare syndrome was too small to allow for separate analysis. All relevant clinical and paraclinical information was extracted from the patients charts. Children were examined by laboratory tests, kidney ultrasound, electrocardiography (ECG), and echocardiography before and regularly during the MAD course (kidney ultrasound at least twice per year and ECG/echocardiogram at least once per year). The following data were documented: age at epilepsy onset, age at start of MAD, seizure types, number and types of administered AEDs before and concomitant to MAD, AED changes during MAD, daily carbohydrate intake at start and during MAD, potential side effects, MAD duration, and when applicable reasons for its termination. Seizure frequency was based on documentation by parents and caregivers. Follow-up examinations, including regular EEG recordings and documentation of seizure frequency and severity, were performed every 3 months. The effects of MAD on epilepsy and potential side effects are reported descriptively. Statistical analysis comprised a comparison of all documented parameters between the groups of seizure-free and non seizure-free patients by twotailed t-test. The correlation between seizure-free outcome and gender was examined with chi-square test in combination with Cramer s V. Results We retrospectively identified 30 patients (24 male) with MAE treated with MAD between February 2007 and June 2015 (Epilepsy Centre Kork, 22; University Children s Hospital Zurich, 4; Innsbruck, 3; and Hannover, 1). Seven of the MAE patients from Kork included in this study had participated in an open-label prospective study.

3 659 Modified Atkins Diet and Doose Syndrome Epilepsy started at a mean age of 3.1 years (range ). Patients had a mean age of 4.5 years (range ) at MAD introduction. The mean epilepsy duration prior to starting the diet was 1.4 years (range ). As glucose transporter 1 (GLUT1) deficiency may manifest itself as MAE, it was ruled out by lumbar puncture in 23 patients. Analysis of the SLC2A1 gene was negative in another patient. Six patients were not assessed for GLUT1 deficiency. Further genetic tests showed negative results for SCN1A analysis in two patients. Except for one girl, all children had daily seizures when MAD was started. Most children had from various seizure types. Individually disabling seizure types included myoclonic atonic seizures in 17, only myoclonic seizures without falling in 8, generalized tonic clonic seizures in 14, atypical absences in 5, and tonic seizures in 2. Two children had a history of nonconvulsive status epilepticus within the 6 months prior to MAD introduction (Table 1). Before starting MAD, patients had received a mean of 5.9 different antiepileptic drugs (AEDs; range 2 15) (Table S1). Seizure-free patients had received an average of five AEDs before beginning the diet, compared to 6.5 for children who continued having seizures. The difference was not statistically significant. Four patients were on monotherapy at MAD introduction (VPA, 2; LTG, 1; ESM, 1). All other patients received combination therapies with two or more AEDs. VPA, LTG, and ESM were the most frequently used drugs administered in 21, 18, and 18 children, respectively. Ten children were treated with CLB, and other AED were used in single cases (Table S1). MAD was initiated with a mean amount of 13.5 g carbohydrate per day (range g). Depending on individual factors, carbohydrate amount was modified during diet course within a range of g/day. Twenty of 30 children were still on MAD at the end of the observation period (Fig. 1, Tables 2 and S2). MAD had been stopped in 10 patients. Three of these patients had become continuously seizure-free for at least 2 years. Lack of efficacy (3), loss of efficacy (2), and noncompliance (2) were further reasons for stopping the diet. At the end of the observation period, mean MAD duration was 18.5 months (range ) in patients who either continued (range months) or stopped the diet (range ). In total, 581 months of MAD were analyzed. The mean follow-up time was 37.8 months (range ) in the three children in whom MAD had been stopped after achieving seizure freedom. There were no relapses. All therapies, including drug treatment, had been stopped successfully in two children. Sustained seizure freedom was achieved in 14 (47%) of 30 patients 2 days to 11.4 months (mean 3.7 months, median 2.3 months) after starting MAD (Tables 2 and S2). EEG background activity normalized in all seizure-free children and generalized spike-wave discharges disappeared. In addition to seizure-free patients, five children experienced a seizure reduction of 75%, and six a reduction of 50% to <75%. EEG findings were variable in patients with continuing seizures. Total responder rate ( 50% seizure reduction) was 83% (25/30) at the end of the observation period. Table 2 displays the response to MAD and overall responder rate for defined time points (Table S2). Loss of efficacy occurred in five children, including one child who experienced a relapse after 14 weeks of seizure freedom. An initial seizure reduction of 75% had been observed in three children before seizure frequency increased again. For patient 10, the father decided to stop MAD after 13 months, although a 50% seizure reduction had been achieved. After a further 7 months, MAD was reintroduced and seizures were initially reduced by 75%. However, a loss of efficacy occurred over time. Finally, the diet was stopped again after 26 months of treatment. Although patient 17 initially experienced significant seizure reduction on MAD, the family became noncompliant over time, resulting in an increase in seizure frequency. The parents were not willing to optimize the diet, which was finally stopped due to noncompliance. No positive effects on seizure frequency were noted in four children, including one child with only 1.5 months of observation time. MAD was stopped in two patients within Table 1. Predominant seizure types before starting the MAD Predominant seizure type Seizure-free (N = 14) Patients Not seizurefree (N = 16) Generalized tonic clonic 7 7 Myoclonic atonic 6 10 Myoclonic 4 6 Absences 3 2 Nonconvulsive status epilepticus 1 1 Tonic 1 1 Figure 1. Effects of the MAD on seizure frequency at the end of the observation period. Epilepsia ILAE

4 660 A. Wiemer-Kruel et al. Table 2. Response to the modified Atkins diet (MAD), and total responder rate Seizure reduction (%) by months on modified Atkins diet Patients on MAD Seizure-free (N) Responder (N) Responder rate (%) MAD stopped All patients with known outcome Responder (N) Responder rate (%) Responder = patients with 50% seizure reduction. Patients on MAD, outcome in all documented patients on MAD; MAD stopped, number of patients in whom MAD had been stopped; All patients with known outcome, outcome in all documented patients on MAD plus known outcome in patients in whom the diet had been stopped. the first 3 months, and the remaining child was switched to a classic KD without any benefit. AED doses were reduced in 12 children during MAD (Table S1). Ten AEDs were discontinued in eight seizurefree patients. Three AED were halted in two responders with ongoing seizures. Only one child received a new AED (ESM) while on MAD, with subsequent seizure freedom after 10 days. Significantly, a first trial with ESM had failed previously in this patient. Girls were younger at epilepsy onset (2.4 years compared to 3.2 years in boys, p = 0.03 in univariate analysis). There was no difference between the seizure-free and non seizure-free groups in all analyzed epilepsy and MAD-related parameters. Reduction of seizure frequency was not correlated with any specific subtype of seizures. In general, MAD was well tolerated, with only mild adverse effects. These included unintended weight gain in two patients. Cholesterol levels between 200 and 300 mg/dl were transiently measured during the first 12 months of MAD without clinical consequences in 16 children. Kidney ultrasound, ECG, and echocardiography remained normal in all patients. Only one child had been tested for IQ before epilepsy onset, with a normal result, and none had been tested immediately before starting the diet. Most of the patients underwent cognitive testing during follow-up, but time points and types of tests differed markedly. Two of the seizure-free patients had an IQ < 60 at the end of the observation period, whereas development was normal in the other 12 seizurefree children. Discussion In this retrospective study on the efficacy and tolerability of MAD in 30 children with MAE, almost half of patients became seizure-free (14/30, 47%) and, in addition, 11 (36%) of 30 experienced a seizure reduction of 50%. The total responder rate was 83%. These findings are in line with those of previous reports of a high efficacy of KD in children with MAE. 3,8 11,16 18 To the best of our knowledge, this report comprises the largest population of MAE patients treated with a high-fat diet and is the only study exclusively focussing on MAD. To date, there is no evidence from prospective studies on the efficacy of KD/MAD in MAE, most probably because of the low prevalence of this rare syndrome. At Epilepsy Centre Kork, only seven children with MAE had been included in an open controlled trial, and because the authors did not feel able to draw any conclusions from this small group, the results were included in the current study. In fact, all 22 patients from this center had been treated and followed by similar in-house protocols, resulting in a comparably high quality of retrospectively analyzed data. Although there is no randomized controlled trial comparing diets with other treatment strategies in MAE, the results from the current and previous studies on diets are more than promising. As for other epilepsy syndromes, 15 MAD appears to be an effective treatment option for seizures in MAE. This is an important finding because MAD is comparably less restrictive and more compatible with the daily lives of children and their families. 12,13 The proportion of seizure-free patients and responders in our study was lower than in a recent report by Simard-Tremblay et al., 11 but the study populations differed regarding age of children (at epilepsy onset and introduction of the diet), epilepsy-related findings, and applied diet types. It is likely that the children in our study had a more severe subtype of MAE. For example, the mean number of AEDs administered before MAD was 5.9 (2 15) in our cohort, compared to 3.8 (1 7) in the previous report. SLC2A1 mutations may present with MAE phenotype, 19 as was the case in one of nine children in the abovementioned study by Simard-Tremblay et al. 11 GLUT1 deficiency had been ruled out by lumbar puncture and/or molecular genetic testing in 24 of 30 patients of our cohort. It cannot be excluded in the six remaining patients, four of whom became seizure-free. The latter subgroup did not differ significantly in epilepsy- or diet-related parameters, and further genetic tests for SCN1A mutations (n = 2) revealed normal results. A strong genetic correlation in MAE has not

5 661 Modified Atkins Diet and Doose Syndrome been demonstrated, 20 with evidence suggesting only genetic influences and heterogeneity. A crucial shortcoming of retrospective studies is that progression after onset is unpredictable and simple seizure control and complete remission would be expected in about two thirds of children with MAE. 2,3,6 These children seem to represent a subgroup with good prognosis regardless of treatment selection. Early introduction of a KD/MAD may therefore be more likely to be effective because patients with a more benign subtype of MAE will be included. Only a prospective randomized controlled trial would be adequate to rule out this biologic effect. Therefore, the study by Simard-Tremblay et al. 11 may have included more patients with an a priori better prognosis. The patients in our study presented with highly pharmacoresistant MAE of longer duration. The value of MAD was increasingly recognized in our daily practice over the study period. Although a mean number of 9.6 (range 5 15) AEDs had been tried before diet introduction in seven patients treated before 2012, the average dropped to 4.7 (range 2 10) in the 23 children treated more recently, reflecting our experience of high diet efficacy. Seizure frequency improved within a few weeks after MAD introduction in responders, suggesting therapeutic effect. Although time in reaching seizure freedom differed widely in our study population (2 days up to 11.4 months, median 2.3 months), positive effects were obvious within the first 2 6 weeks. Seizures were reduced by 75% in all whom later became seizure-free, with only one exception. The overall responder rate ( 50% seizure reduction) was 83% after 3 months (Table 2). After 12 months, 80% of children on MAD were responders. This responder rate remained high (70%), even taking into consideration those patients with known outcomes after stopping the MAD (Tables 2 and S2). Overall, AED doses were reduced in 12 children following introduction of MAD (Table S1). One or two AEDs were stopped in eight seizure-free patients (total AEDs: 10) and three AEDs were discontinued in two responders. Fewer AEDs were discontinued in the group with continuing seizures. A new AED was added for only one child (Table S1), when ESM was reintroduced after a previous failure and added to MAD after 3 months. The patient then became seizure-free after 1.5 weeks. This retrospective study cannot rule out a role for drug treatment changes as a relevant factor in the outcome. Seizure freedom was achieved by increasing LTG doses in three early responders even after 9.7, 10.4, and 11.7 months, respectively. In a recent retrospective study on 71 children treated with classic KD for different epilepsy types, van der Louw et al. 21 found a negative correlation of LTG use with response. We did not find any negative impact of concomitant LTG therapy in our cohort. In contrast, we favor the combination of MAD with the first-line AEDs for MAE, which are VPA, LTG, and ESM. At study conclusion, only 7 of 30 families had stopped the MAD because of factors other than sustaining seizure freedom. A lack of efficacy was noted as the reason for cessation in three children, and loss of efficacy after initial improvement occurred in two patients. Noncompliance was noted in only two cases, stressing the practicability of MAD in pre-school and school-aged children. We observed no severe therapy-related adverse effects, and no child discontinued the MAD due to side effects. Conclusion KD and MAD are important and efficient therapeutic options for children with MAE. This applies even to previously pharmacoresistant and encephalopathic cases. MAD should be considered early in the course as an alternative to classic AEDs because of the generally good tolerability and the high rate of seizure-free patients and responders. Acknowledgments The authors are grateful for the support from Dr. Rolf Kruel regarding data documentation and analysis and Jim Livingston and Liz Stevens for providing editorial assistance. Conflict of Interest Dr. Adelheid Wiemer-Kruel received financial support from Nutricia and Novartis for performing investigational clinical trials, and speaker honoraria from Nutrica, Novartis, Desitin, and UCB. Dr. Edda Haberlandt received speaker honoraria from Desitin, Eisai, Novartis, Shire, and UCB. Dr. Gabriele Wohlrab received speaker honoraria from Nutricia and Novartis. Dr. Hans Hartmann received speaker honoraria and travel grants from Novartis, Nutricia, and Shire. PD Dr. Thomas Bast received speaker and/or consultant honoraria from Bial, Biocodex, Desitin, Eisai, and UCB. 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. References 1. McTague A, Cross JH. Treatment of epileptic encephalopathies. CNS Drugs 2013;27: Oguni H, Fukuyama Y, Tanaka T, et al. Myoclonic-astatic epilepsy of early childhood clinical and EEG analysis of myoclonic-astatic seizures, and discussions on the nosology of the syndrome. Brain Dev 2001;23: Oguni H, Tanaka T, Hayashi K, et al. Treatment and long-term prognosis of myoclonic-astatic epilepsy of early childhood. Neuropediatrics 2002;33: Stephani U. Myoclonic astatic epilepsy. In Nikanorova M, Genton P, Sabers A (Eds) Long-term evolution of epileptic encephalopathies. Esher (Surrey, UK): John Libbey Eurotext, 2009: Stephani U. The natural history of myoclonic astatic epilepsy (Doose syndrome) and Lennox-Gastaut syndrome. Epilepsia 2006;47(Suppl. 2): Kilaru S, Bergqvist C. Current treatment of myoclonic astatic epilepsy: clinical experience at the Children s Hospital of Philadelphia. Epilepsia 2007;48: Von St ulpnagel C, Coppola G, Striano P, et al. First long-term experience with the orphan drug rufinamide in children with myoclonic-astatic epilepsy (Doose syndrome). Eur J Paediatr Neurol 2012;16:

6 662 A. Wiemer-Kruel et al. 8. Caraballo RH, Cersosimo RO, Sakr D, et al. Ketogenic diet in patients with myoclonic-astatic epilepsy. Epileptic Disord 2006;8: Bergqvist AGC. Myoclonic astatic epilepsy and the use of the ketogenic diet. Epilepsy Res 2012;100: Nangia S, Caraballo RH, Kang HC, et al. Is the ketogenic diet effective in specific epilepsy syndromes? Epilepsy Res 2012;100: Simard-Tremblay E, Berry P, Owens A, et al. High-fat diets and seizure control in myoclonic-astatic epilepsy: a single center s experience. Seizure 2014;25: Kossoff EH, Cervenka MC, Henry BJ, et al. A decade of the modified Atkins diet ( ): results, insights, and future directions. Epilepsy Behav 2013;29: Kang HC, Lee HS, You SJ, et al. Use of a modified Atkins diet in intractable childhood epilepsy. Epilepsia 2007;48: Sharma S, Sankhyan N, Gulati S, et al. Use of the modified Atkins diet for treatment of refractory childhood epilepsy: a randomized controlled trial. Epilepsia 2013;54: Kim JA, Yoon JR, Lee EJ, et al. Efficacy of the classic ketogenic and the modified Atkins diets in refractory childhood epilepsy. Epilepsia 2016;57: Weber S, Mølgaard C, Taudorf K, et al. Modified Atkins diet to children and adolescents with medical intractable epilepsy. Seizure 2009;18: Kossoff EH, Bosarge JL, Miranda MJ, et al. Will seizure control improve by switching from the modified Atkins diet to the traditional ketogenic diet? Epilepsia 2010;51: Kumada T, Miyajima T, Oda N, et al. Efficacy and tolerability of modified Atkins diet in Japanese children with medication-resistant epilepsy. Brain Dev 2012;34: Mullen SA, Marini C, Suls A, et al. Glucose transporter 1 deficiency as a treatable cause of myoclonic astatic epilepsy. Arch Neurol 2011;68: Tang S, Pal DK. Dissecting the genetic basis of myoclonic-astatic epilepsy. Epilepsia 2012;53: van der Louw EJ, Desadien R, Vehmeijer FO, et al. Concomitant lamotrigine use is associated with decreased efficacy of the ketogenic diet in childhood refractory epilepsy. Seizure 2015;32: Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Antiepileptic drug treatment before and concomitant with the modified Atkins diet in seizure-free and not seizure-free patients. Table S2. Individual response to the modified Atkins diet (MAD), reasons for its discontinuation, and total responder rate.