Combined Ketogenic Diet and Vagus Nerve Stimulation: Rational Polytherapy?

Transcription

1 Epilepsia, 48(1):77 81, 2007 Blackwell Publishing, Inc. C 2007 International League Against Epilepsy Combined Ketogenic Diet and Vagus Nerve Stimulation: Rational Polytherapy? Eric H. Kossoff, Paula L. Pyzik, James E. Rubenstein, A. G. Christina Bergqvist, Jeffrey R. Buchhalter, Elizabeth J. Donner, Douglas R. Nordli Jr., and James W. Wheless The Johns Hopkins Hospital, Baltimore, Maryland, Children s Hospital of Philadelphia, Philadelphia, Pennsylvania, Mayo Clinic, Rochester, Minnesota, U.S.A., Hospital for Sick Children, Toronto, Ontario, Canada, Children s Memorial Hospital, Chicago, Illinois, and University of Tennessee Health Science Center, Memphis, Tennessee, U.S.A. Summary: Objective: The concept of rational polypharmacy has been associated with anticonvulsant management for decades, but the term has not been applied to nonpharmacologic therapies. Methods: We conducted a multicenter, retrospective study of children who received concurrent diet (ketogenic or modified Atkins) and vagus nerve stimulation (VNS) treatment for medically intractable epilepsy. Results: Thirty children in total from six epilepsy centers were treated over a 6-yr period. The median age at the initiation of combination therapy was 10 yr (range, 4 24 yr). Sixteen (53%) received dietary therapy followed by VNS; no differences were noted between centers. After 3 months, 21 (70%) had seizure reduced by >50% over the previous single nonpharmacologic treatment, of whom 13 (62%) had improvement within the first month. A 5-min VNS off-time correlated with >90% seizure reduction (p = 0.02). The median duration of nonpharmacologic polytherapy was 12 months (range, months); 17 (57%) remain on dual therapy at this time. No side effects were noted. Most patients who discontinued combination therapy did so because of a lack of efficacy rather than restrictiveness. Conclusions: In this small group, the combined use of diet and VNS appeared synergistic and yielded rapid benefits. It may be more effective with longer VNS off-times. Further prospective studies of this combination in refractory pediatric epilepsy are needed to help guide optimal use. Key Words: Ketogenic Atkins Vagus nerve stimulation Epilepsy Children. In the 2005 expert opinion consensus surveys of pediatric and adult epilepsy treatments, epileptologists in hypothetical clinical scenarios believed monotherapy was preferable for new-onset epilepsy (Karceski et al., 2005; Wheless et al., 2005). If initial monotherapy was unsuccessful, however, most also believed that only one to two additional anticonvulsants should be attempted as monotherapy before a polytherapy trial. This opinion is different from published recommendations to achieve monotherapy whenever possible (Perucca, 2002; Baulac, 2003; Pennell, 2003; Kanner and Balabanov, 2005). The benefits of monotherapy include reduction in adverse effects, teratogenicity, drug interactions, and improved compliance (Perucca, 2002; Baulac, 2003; Pennell, 2003; Kanner and Balabanov, 2005). For the subset of children with intractable epilepsy, the need for so-called rational polypharmacy with two Accepted August 10, Address correspondence and reprint requests to Eric H. Kossoff at The Johns Hopkins Hospital, Suite North Wolfe Street, Baltimore, MD ekossoff@jhmi.edu doi: /j x or more anticonvulsants is evident (Ferrendelli, 1995; Leppik, 2000; Deckers, 2002). In adults, more patients were seizure-free (26% vs. 17%) when a second anticonvulsant was added to rather than substituted for an ineffective one (Kwan and Brodie, 2000). Several medications seem to work well in combination, (e.g. valproate and lamotrigine, lamotrigine, and topiramate), and selecting anticonvulsants with differing mechanisms of action, low side effects, and limited drug interactions may be ideal (Stephen et al., 1998; Kwan and Brodie, 2000; Deckers, 2002). The ketogenic and modified Atkins diets are therapies for children with intractable epilepsy (Kossoff et al., 2003; Kossoff, 2004; Kossoff et al., 2006). Recent limited evidence and increased clinical use of the latter diet has demonstrated similar efficacy when ketosis is achieved (Kossoff et al., 2003, 2006). Although many families request discontinuing anticonvulsants immediately after starting these diets, the dose and number of drugs are often reduced rather than completely eliminated (Kossoff et al., 2004). Whereas no single anticonvulsant has been shown to be preferentially beneficial in combination 77

2 78 E. H. KOSSOFF ET AL. with these diets, the diets conversely have not shown any negative interactions with anticonvulsants (Kossoff et al., 2002; Takeoka et al., 2002; Lyczkowski et al., 2005; Dahlin et al., 2006). Vagus nerve stimulation (VNS) is another nonpharmacologic therapy used commonly in conjunction with anticonvulsants (Wheless and Maggio, 2002). Similar to the ketogenic diets, VNS has not been reported to have either positive or negative interactions with anticonvulsant drugs, other than a single report of improved alertness with VNS and topiramate (Labar, 2002; Kossoff and Pyzik, 2004). What about the combined use of dietary and VNS therapies? Both treatments have likely multiple mechanisms of action, low side effects, and no drug interactions; therefore, in theory, the combination is promising (Stafstrom and Spencer, 2000; Tecoma and Iragui, 2006). We hypothesized that combining these two nonpharmacologic therapies would be well tolerated and efficacious in treating refractory pediatric epilepsy. METHODS Pediatric epilepsy centers with known large ketogenic diet programs and that also offer VNS therapy were contacted via telephone and electronic mail; five centers participated in addition to the primary institution (Johns Hopkins Hospital). All participating physicians reported at least one patient in whom both a dietary therapy (ketogenic or modified Atkins) and VNS were used concurrently. There was no minimum time period of combination therapy required for inclusion in an effort to include all potential patients. Adults were not excluded, however, only pediatric epileptologists were contacted for this study. Only limited information was obtained about the prior single therapy use of diets, VNS, or anticonvulsants and time after concurrent therapy was discontinued. Deidentified records or completed data entry forms were sent to the primary institution and the information was then entered into an electronic database created specifically for this analysis. Information included patient demographics, as well as seizure type and frequency at a baseline period of at least 1 month prior to combination therapy. All records included both the duration of combination nonpharmacologic therapy and prior single nonpharmacologic therapy. Ketogenic diet ratio and calories, Atkins carbohydrate grams per day, VNS stimulation parameters, and concurrent anticonvulsants were described for the majority of the combination period. Records were examined for an estimate of seizure frequency reduction as reported by parents, (e.g., <50%, 50 90%, >90%, or seizure-free) as well as EEG changes when obtained. In occasional cases in which there was missing information, investigators were contacted by electronic mail for this specific data, which was then provided. This study was approved by the Johns Hopkins Committee for Clinical Investigation. Participating investigators obtained approval from their individual institutional review boards and removed identifying information from all records. No identifying information was subsequently maintained at the host institution on any patient. Categorical data were analyzed by Fisher s exact test for independence of rows and columns. Medians were compared by using the Wilcoxon two-sample test and means by using the Student s t-test. The significance level for all tests was p = RESULTS Patient demographics From the six institutions, 30 patients were identified with 1 14 patients per center (Table 1). Patients received combination therapy during the period between October 1999 and December No identified etiology for epilepsy was present in 16 (53%) patients. In the remaining patients, etiologies included encephalitis (4), cortical dysplasia (2), tuberous sclerosis complex (2), stroke (2), hypomelanosis of Ito (1), 15q deletion (1), Angelman syndrome (1), and hypoxic-ischemic encephalopathy (1). Prior single nonpharmacologic therapy Sixteen (53%) patients were started on the ketogenic diet before VNS implantation. The median age at diet implementation was 10 yr (range, 3 22 yr); median duration of the ketogenic diet was 18 months (range, 6 84 months) before VNS implantation to implement combination therapy. None of the patients received a modified Atkins diet before VNS implantation. The median age of VNS implantation was 10 yr (range 3 24 yr); duration of VNS was 24 months (range, 2 48 months) before the ketogenic TABLE 1. Patient characteristics (n = 30) Number Median Characteristic (percentile) (range) Age at seizure onset (yr) 0.9 (0 20) Age at combination therapy (yr) 9.5 (4 24) Seizure frequency before 25 (5 750) combination therapy(per week) Number of medications tried 7 (1 15) before first nonpharmacologic therapy Number of medications at the 2 (1 5) start of combination therapy Female gender 18 (60%) Seizure type Multiple 9 (30%) Complex partial 9 (30%) Myoclonic 4 (13%) Atonic 3 (10%) Tonic 1 (3%) Absence 1 (3%) Generalized tonic clonic 1 (3%) Simple partial 1 (3%) Infantile spasms 1 (3%)

3 RATIONAL POLYTHERAPY 79 or Atkins diet was started to implement combination therapy. Twenty-three (77%) children had >50% seizure reduction from baseline before beginning combination therapy with most (21/23, 91%) having a 50 90% improvement. The two patients with >90% seizure reductions had been on the ketogenic diet for 18 and 36 months, respectively, but persistent daily seizures led to a decision to add VNS therapy. After 6 months of combination therapy, neither of these patients had an additional >50% reduction, and their diets were discontinued. The individual investigators choices of either dietary or VNS therapy as the first treatment option was roughly equivalent among the institutions, with no clear preferences towards one particular therapy for these patients. The median age of seizure onset was identical between the patients who received the ketogenic diet or VNS first (10 months). Baseline median seizure frequency before combination therapy, gender, and seizure type were also not different between groups. Seizure reductions at the time of combination therapy were similar, 87% (diet) vs. 57% (VNS) with >50% improvement, p = Duration of combination treatment The median age when combination therapy was started was 10 yr (range, 4 24 yr). The median duration of combination therapy was 12 months (range, months). At the time of this report, 17 (57%) patients remain on combination therapy. The age when combination therapy was initiated also did not predict whether the duration of combination therapy would exceed 18 months; 9 yr for patients with more than 18 months of combination therapy duration vs. 8.5 year for those who received less than 18 months of combination therapy. Patients receiving more than 18 months of combination therapy had previously attempted a slightly greater mean number of medications than those with combination therapy of lesser duration, 8 vs. 5.7, respectively; p = Seizure reduction Efficacy of the combination therapy is presented in Table 2. At 3 months, 21 patients (70%) had a >50% seizure reduction over single therapy 10 of whom (33%) were >90% improved (intent-to-treat analysis). At 6 months, 19 patients (63%) were >50% improved, 7 TABLE 2. Efficacy of combination therapy at 3, 6, and 12 months in those remaining on the combination 3 months 6 months 12 months Seizure reduction (n = 29) (n = 25) (n = 15) Seizure-free 2 (7%) 2 (8%) 0 (0%) 90 99% 8 (28%) 5 (28%) 5 (33%) 50 90% 11 (38%) 12 (48%) 7 (47%) <50% 8 (28%) 6 (24%) 3 (20%) TABLE 3. Demographics and nonpharmacologic therapy parameters of children with >90% reduction at 3 months of combination therapy. Averages are expressed as means (standard deviation) >90% reduction <90% Factor (n = 10) (n = 20) p-value Number of medications 6.3 (4.3) 6.9 (3.2) 0.67 Prior to combination Age at cotherapy (yr) 8.6 (3.7) 11.8 (5.3) 0.11 Duration of prior single 27.2 (20.0) 20.1 (18.5) 0.34 nonpharmacologic therapy (months) Dietary therapy initially 5 (50%) 11 (55%) 0.80 Gender (female) 7 (70%) 11 (55%) 0.43 Multiple seizure types 3 (30%) 6 (30%) 0.99 Combination therapy 4:1 ratio ketogenic diet 6 (60%) 8 (40%) min VNS off-time 8 (80%) 7 (35%) 0.02 Output current <2.0 ma 7 (70%) 15 (75%) 0.77 Duration of combination 20.3 (16.0) 19.8 (21.7) 0.95 therapy (months) (23%) were >90% improved. The 10 children with >90% seizure reductions at 3 months were compared with the 20 patients with less efficacy (Table 3). All 10 patients achieved this level of improvement (>90% seizure reduction) within the first month of combination therapy and were able to maintain it for at least 3 months. Choice of initial therapy did not influence outcome. EEGs were obtained during combination in eight children; seven showed unchanged epileptiform activity. The onset of seizure reduction was extremely rapid for many children. Among the 21 patients with a 50% improvement or better at 3 months, 13 (62%) improved within the first month of combination therapy. Of these 13, 7 (54%) had VNS initially. Diet and VNS parameters Of the 30 patients, 10 (33%) received a modified Atkins diet (10 20 grams of carbohydrate per day) for some period during combination therapy; five of these 10 (50%) received this modified diet initially and solely, the other five were transitioned to this diet after the traditional ketogenic diet. Of the 25 children receiving the traditional ketogenic diet at the start of nonpharmacologic therapy, 14 (56%) were on a 4:1 (fat:protein and carbohydrate) ratio. Diet ratio or use of the modified Atkins diet did not influence outcome. VNS settings were changed as often as weekly throughout the period of combination therapy, although typically for those who were newly implanted and had begun dietary therapy first. Only one of 14 patients with a VNS already implanted when dietary therapy was started had their VNS parameters changed during combination therapy, compared to 14 of 16 (88%) in whom a VNS was added to a dietary therapy, p < Nearly all children received low output currents (<2.0 ma) but significantly more were receiving regular duty

4 80 E. H. KOSSOFF ET AL. cycles (30 s on and 5 min off) compared with rapid cycling, p = Many of these children improved within the first few weeks, as described previously, when the VNS was begun. In addition, the seizure frequency of four patients increased when shorter duty cycles (less than 5 min off-time) were used. The VNS devices were subsequently returned to regular settings, and all patients improved. Medications At the end of combination therapy, the median number of anticonvulsants was three (range, 1 5). Medications for two children were eliminated successfully during combination therapy; however, additional drugs were added to the regimens of two others. The most commonly used medications during combination therapy were valproate (11 patients), lamotrigine (9), levetiracetam (9), and zonisamide (9); 4 or fewer patients used all other drugs. No anticonvulsant seemed to have particular efficacy in combination with diet and VNS. Of the 10 patients with >90% improvement at 3 months, five were receiving valproate, and three were receiving lamotrigine, levetiracetam, or zonisamide. Side effects and discontinuation Dietary therapy was well tolerated by these 30 patients; no side effects were reported other than occasional food choice restrictiveness. No side effects of VNS therapy were described. Of the 13 patients who discontinued combination therapy, 12 stopped the diet, and one patient had the VNS pulse generator explanted and continued on the ketogenic diet. Insufficient seizure reduction was the primary reason for diet discontinuation (8/12 patients), with the remainder stopping dietary therapy due to perceived restrictiveness. DISCUSSION The use of ketogenic diets and neurostimulation for intractable epilepsy, commonly perceived as distinct therapies, which are difficult to use concurrently, may have synergistic benefits. In this limited case series, more than two thirds of the patients reported additional benefits when the therapies were combined. As neurologists selecting pharmacologic treatment strive for rational polypharmacy, these two therapies may represent rational polytherapy. According to an analysis of the Cyberonics epilepsy patient registry, combined use of the ketogenic diet and VNS, has occurred in only 14 of 740 (1.9%) children tracked in the registry (personal communication, Cyberonics, Inc.). Even in the six pediatric epilepsy centers included in this study, only 30 children were identified, and most had received single nonpharmacologic therapy for nearly 2 year before the combination therapy was eventually initiated. The reasons for this delay, as well as the infrequent use overall in most centers, are likely due to several factors. Both therapies can be time-intensive, require family commitment, and are not available in all medical centers. The absence of a longer duration of combination therapy for those children with resultant >90% seizure reduction may reflect the difficulty of using both therapies together, although patients who had attempted more anticonvulsants were slightly more willing to continue longer. Theoretical concerns of general anesthesia in combination with the diet, although largely unfounded, may lead to a discontinuation of the diet before VNS surgical implantation (Valencia et al., 2002). The long-term effects of the ketogenic diet are unknown, therefore it is frequently discontinued after 2 3 yr, and other therapies such as VNS are then attempted sequentially. Lastly, given the large variety of VNS stimulation parameters and the evidence suggesting that seizure reduction may improve over time, the consideration of adjunctive therapy to VNS frequently occurs after several years, often during adolescence (Uthman et al., 2004). As the ketogenic diet is not traditionally started in this age group, additional anticonvulsants, rather than dietary therapies, are often added to VNS therapy (Mady et al., 2003). Several findings in this study may have implications for future use of this polytherapy. For one, seizure reduction often occurred within weeks of beginning the second therapy, either VNS implantation or diet initiation. As the ketogenic and modified Atkins diets can be discontinued rapidly, one might consider a short trial of a dietary therapy for a patient with VNS and continued seizures. Because the modified Atkins diet can be started as an outpatient without a fast, relatively less dietitian and family time is required to attempt this particular dietary therapy (Kossoff et al., 2003, 2006). As has been demonstrated in another study, we found low VNS duty cycles were effective, and utilizing these settings may lead to longer VNS battery life (DeGiorgio et al., 2005). It is unclear why these settings were more beneficial compared to rapid cycling. However, as only half the children, mostly those with a VNS added onto an ongoing dietary therapy, had their VNS modified at all, any recommendations for VNS parameters are only speculative. Further prospective studies with combination therapy and different VNS parameters would be necessary to be more definitive. Acknowledgment: This study was supported by a grant from Cyberonics, Inc, manufacturer of the VNS Therapy System. REFERENCES Baulac M. (2003) Rational conversion from antiepileptic polytherapy to monotherapy. Epileptic Disorders 5: Dahlin MG, Beck OM, Amark PE. (2006) Plasma levels of antiepileptic drugs in children on the ketogenic diet. Pediatric Neurology 35:6 10. Deckers CL. (2002) Place of polytherapy in the early treatment of epilepsy. CNS Drugs 16: DeGiorgio C, Heck C, Bunch S, Britton J, Green P, Lancman M, Murphy J, Olejniczak P, Shih J, Arrambide S, Soss J. (2005) Vagus nerve

5 RATIONAL POLYTHERAPY 81 stimulation for epilepsy: randomized comparison of three stimulation paradigms. Neurology 65: Ferrendelli JA. (1995) Relating pharmacology to clinical practice: the pharmacologic basis of rational polypharmacy. Neurology 45(suppl 2):S12 S16. Kanner AM, Balabanov AJ. (2005) The use of monotherapy in patients with epilepsy: an appraisal of the new antiepileptic drugs. Current Neurology Neuroscience Report 5: Karceski S, Morrell MJ, Carpenter D. (2005) Treatment of epilepsy in adults: expert opinion, Epilepsy Behaviour 7(suppl 1):S1 S64. Kossoff EH, Pyzik PL, Furth SL, Hladky HD, Freeman JM, Vining EP. (2002) Kidney stones, carbonic anhydrase inhibitors, and the ketogenic diet. Epilepsia 43: Kossoff EH, Krauss GL, McGrogan JR, Freeman JM. (2003) Efficacy of the Atkins diet as therapy for intractable epilepsy. Neurology 61: Kossoff EH. (2004) More fat and fewer seizures: dietary therapies for epilepsy. Lancet Neurology 3: Kossoff EH, Pyzik PL, McGrogan JR, Rubenstein JE. (2004) The impact of early versus late anticonvulsant reduction after ketogenic diet initiation. Epilepsy Behaviour 5: Kossoff EH, Pyzik PL. (2004) Improvement in alertness and behavior in children treated with combination topiramate and vagus nerve stimulation. Epilepsy Behaviour 5: Kossoff EH, McGrogan JR, Bluml RM, Pillas DJ, Rubenstein JE, Vining EP. (2006) A modified Atkins diet is effective for the treatment of intractable pediatric epilepsy. Epilepsia 47: Kwan P, Brodie MJ. (2000) Epilepsy after the first drug fails: substitution or add-on? Seizure 9: Labar DR. (2002) Antiepileptic drug use during the first 12 months of vagus nerve stimulation therapy: a registry study. Neurology 59(suppl 4):S38 S43. Leppik IE. (2000) Monotherapy and polypharmacy. Neurology 55(suppl 3):S25 S29. Lyczkowski DA, Pfeifer HH, Ghosh S, Thiele EA. (2005) Safety and tolerability of the ketogenic diet in pediatric epilepsy: effects of valproate combination therapy. Epilepsia 46: Mady MA, Kossoff EH, McGregor AL, Wheless JW, Pyzik PL, Freeman JM. (2003) The ketogenic diet: adolescents can do it, too. Epilepsia 44: Pennell PB. (2003) The importance of monotherapy in pregnancy. Neurology 60(suppl 4):S31 S38. Perucca E. (2002) Overtreatment in epilepsy: adverse consequences and mechanisms. Epilepsy Research 52: Stafstrom CE, Spencer S. (2000) The ketogenic diet: a therapy in search of an explanation. Neurology 54: Stephen LJ, Sills GJ, Brodie MJ. (1998) Lamotrigine and topiramate may be a useful combination. Lancet 351: Takeoka M, Riviello JJ Jr, Pfeifer H, Thiele EA. (2002) Concomitant treatment with topiramate and ketogenic diet in pediatric epilepsy. Epilepsia 43: Tecoma ES, Iragui VJ. (2006) Vagus nerve stimulation use and effect in epilepsy: what have we learned? Epilepsy Behaviour 8: Uthman BM, Reichl AM, Dean JC, et al. (2004) Effectiveness of vagus nerve stimulation in epilepsy patients: a 12-year observation. Neurology 63: Valencia I, Pfeifer H, Thiele EA. (2002) General anesthesia and the ketogenic diet: clinical experience in nine patients. Epilepsia 43: Wheless JW, Maggio V. (2002) Vagus nerve stimulation therapy in patients younger than 18 years. Neurology 59(suppl 4):S21 S25. Wheless JW, Clarke DF, Carpenter D. (2005) Treatment of pediatric epilepsy: expert opinion, Journal of Child Neurology 20(suppl 1):S1 S56.