epilepsy diet

Epilepsy Research 68 (2006) 145180

Conference review

The ketogenic diet: From molecular mechanismsto clinical effects

John Freemana, Pierangelo Veggiottib, Giovanni Lanzib,Anna Tagliabuec, Emilio Peruccab,d,

a The John M. Freeman Pediatric Epilepsy Center, Johns Hopkins Medical Institutions, Baltimore, MD, USAb Institute of Neurology, IRCCS C. Mondino Foundation, Pavia, Italy

c Department of Health Sciences, Human Nutrition and Dietetics, University of Pavia, Pavia, Italyd Clinical Pharmacology Unit, Department of Internal Medicine and Therapeutics,

University of Pavia, Piazza Botta 10, 27100 Pavia, Italy

Received 18 October 2005Available online 5 December 2005

Abstract

nd generalen andternational

nt effects,ies on thecacy

related tonsiderableential roleludingdomizedh neededRecent years have witnessed an increased interest from pediatric neurologists, neuropediatricians, epileptologists aneurologists in the use of the ketogenic diet (KD) for the management of refractory epilepsies, particularly in childradolescents. This article summarizes current knowledge on various issues related to its use, as discussed at a recent inworkshop. Aspects discussed in some detail include (i) the putative mechanisms responsible for the diets anticonvulsabased on results of biochemical and neurophysiological studies in experimental models; (ii) consensus and controversmodalities of initiation of the diet, and different protocols of implementation; (iii) indications and contraindications; (iv) effidata, also in relation to seizure type, syndromic form and patients age; (v) adverse effects; (vi) methodological aspectsassessment of the diets clinical effects, and perspectives for future research. Overall, the data reviewed indicate that coadvances have been made in understanding the modes of action of the diet, its efficacy and tolerability profiles and its potin different types of epilepsy. Although clinical studies performed to date have important methodological limitations, incsuboptimally characterized patients populations and an uncontrolled design, a number of innovative, prospective ranstudy protocols have been recently proposed and are being implemented. The results of these will hopefully provide muchigh-quality information to better define the role of the diet in the treatment algorithms in different epilepsy syndromes.

Keywords: Ketogenic diet; Epilepsy; Mechanisms of action; Efficacy; Safety; Review

Summary of a Workshop held at the Institute of Neurology IRCCS C. Mondino Foundation, Pavia, Italy, 24 September 2005. Corresponding author. Tel.: +39 0382 986360; fax: +39 0382 22741.E-mail address: perucca@unipv.it (E. Perucca).

0920-1211/$ see front matterdoi:10.1016/j.eplepsyres.2005.10.003

mailto:perucca@unipv.it (E. Perucca)

146 J. Freeman et al. / Epilepsy Research 68 (2006) 145180

1. Introduction

Recent years have witnessed an increased interestfrom pediatric neurologists, neuropediatricians, epilep-tologists and general neurologists in the use of theketogenic diet (KD) for the management of refractoryepilepsies. In parallel with this interest, a number ofstudies have been completed addressing a wide rangeof issues related to this diet, from its mechanisms ofaction in animal models to its efficacy and tolerabilityprofiles in different populations of patients. In recog-nition of this, the Institute of Neurology IRCCS C.Mondino Foundation in Pavia, Italy, took the initiativeof organizing, on 24 September 2005, an internationalworkshop whose purposes were to (i) discuss the ratio-nale for the use of the KD, and bridge basic sciencedata with evidence from clinical experience; (ii) reviewcritically and update current acquisitions on the use ofthe diet in the epilepsies; (iii) highlight controversiesconcerning indications and modes of application; (iv)identify areas and methodology for future research.

The workshop, which was attended by many lead-ing experts from both sides of the Atlantic, provided aforum for constructive interaction between many disci-plines which contribute to our knowledge in this area.The purpose of this manuscript is to summarize infor-m wella iatedd

2p

J

TH

uresa versa ed.T enm asea tom1 es,s alss esa nd

consider the affected individual to have divine, shamin-istic abilities (Fadiman, 1997). This can result in animportant example of cross-cultural misunderstand-ing. Seizures have also been ascribed to: humors,toxins, and infections (Lennox, 1960). Treatmentshave been empirical and have ranged from magic,to the ingestion of, or abstinence from certain foodsand even of certain animal and human parts. Irritationand counter-irritation, trephination, prayer, and fastinghave been prescribed. The removal of various organsand even death because of demonic possession has beenused as treatment.

Bromide, the first useful antiepileptic drug (AED),was used in the 1850s for hystero-epilepsy and laterused successfully in other forms of seizures. Phenobar-bital introduced in 1917 and phenytoin in 1937 werebut the first of what has become a plethora of newAEDs. Lack of understanding of the cause of seizuresand epilepsy has, over time, seemed to be the motherof therapeutic intervention and invention. While wehave learned much about the physiology of neuronalexcitation and inhibition, even today we understandlittle about what causes most seizures or what triggerstheir occurrence. Our treatments, therefore, continueto remain largely empirical. The KD is only one of themore recent, empirically derived, effective therapies.

medH edf m-i inga 20d .,2 ivet d tob ntf dieth o-h thiss s ontt fectso t the1 toina Med-i KD.

arly1 cen-t ve oration that was presented at this workshop, ass the main points that emerged from the associscussions.

. Dietary treatments for epilepsy: a historicalerspective

.M. Freeman

he JohnM. Freeman Pediatric Epilepsy Center, Johnsopkins Medical Institutions, Baltimore, MD, USAIndividual seizures, as well as the repeated seiz

ssociated with epilepsy are frightening to obsernd may affect the quality of life of those affecthroughout the millennia their etiology has beysterious. Epilepsy, termed The Sacred Disend The Falling Sickness, has been ascribedany causes, including demonic possession (Tempkin,976). Even among the modern day Hmong tribeizures are ascribed to problems with the individuoul (Fadiman, 1997). The Hmong consider seizurs indication of possession of a healing spirit aIn the early 1920s, an osteopathic physician naugh Conklin discovered that epilepsy, result

rom intoxication of the brain from substances cong from the intestine, and devised a cure by fastnd having the patient imbibe only water for up toays thus putting the intestine at rest (Freeman et al002). At a time when there were few other effect

reatments for seizures, the diet was documentee effective. This was termed The Water Treatme

or epilepsy. It was found that a calorie-restrictedigh in fats, with sufficient protein and limited carbydrates could mimic the biochemical changes oftarvation and could preserve its beneficial effecthe seizures (Freeman et al., 2002). In modified formhis has been became todays KD. The beneficial eff this dietary treatment were confirmed throughou930s, and but then largely forgotten as first phenynd then other anticonvulsants became available.

cation was far easier to prescribe and take than theInterest in the KD was re-awakened during the e

990s and since then multiple studies from manyres have concluded that the diet can be as effecti

J. Freeman et al. / Epilepsy Research 68 (2006) 145180 147

even more effective in the management of children withdifficult-to-control seizures than other currently avail-able anticonvulsant medications (Freeman et al., 1998;Freeman and Vining, 1998). At Johns Hopkins, 150consecutive children with multiple seizure types andhaving at least two seizures per day were started on theKD and followed up (Hemingway et al., 2001a; Madyet al., 2003). Slightly more than half of the children whohad averaged >400 seizures per month at diet onset con-tinued the diet for 1 year, and 27% of those children,had had a >90% decrease in their seizures. Three to6 years later a similar percentage of this same grouphad continued, now off the diet, with few seizures, andmany had decreased or eliminated their medications.There are no studies of AEDs documenting similarlygood outcomes over comparable time periods.

The KD has been reported to be effective across awide variety of ages, seizure types and severities (Madyet al., 2003; Kossoff et al., 2002a) as well as differentetiologies (Kossoff et al., 2002a, 2005). It can be usedin liquid form (Kossoff et al., 2004a,b), for bottle ortube feeding, and perhaps even in a modified form asthe Atkins diet (Kossoff et al., 2003). The diet has fewsignificant side effects and when properly superviseddoes not cause weight gain or substantial growth abnor-malities (Vining et al., 2002) and has minimal effectso

isu rumlI sid-e dR s ofa at-m

3p

J

DU

or-m ast-i ents

are met by lipolysis and-oxidation of fatty acids ratherthan by the breakdown of carbohydrates. Thus, the KDmaintains an anabolic nutritional state in a metabolicsituation of fasting. Any diet providing nutritional fatfor the generation of ketones that serve as an alternativefuel to body tissues may be called ketogenic. Ketonesmay produce an anticonvulsant effect, presumably dueto changes in (a) cerebral energy metabolism; (b) cellproperties decreasing excitability; (c) neurotransmitterfunction; (d) circulating factors acting as neuromodula-tors, and (e) brain extracellular milieu (Schwartzkroin,1999).

For the treatment of intractable childhood epilepsytwo types of KDs have thus far been developed. Thediets differ in their composition of free fatty acids.The traditional long-chain triglyceride (LCT) diet ismost commonly used (Swink et al., 1997; Lefevre andAronson, 2000). A second type, the medium-chaintriglyceride (MCT) diet is more ketogenic but lessused as it often causes gastrointestinal side effects. Anongoing clinical trial at Great Ormond Street Hospi-tal (London, UK) is currently comparing both waysof giving the diet. Finally, several infant formulas arenow available in the USA and in Europe (Nordli etal., 2001; Klepper et al., 2002). Despite the world-wide use of the diet (Kossoff et al., 2005) (Fig. 1) anda cha-n vel-o ons,t l asa tingp ctantt nes.A tresf ared na-t os-p y/P -s ducea op-k tas orldw ndw ardC n-t

n serum lipid profiles (Kwiterovich et al., 2003).How the diet exerts its anticonvulsant effects

nknown, but is thought possibly to relate to seevels of beta-hydroxy butyrate (Gilbert et al., 2000).ts mechanisms of action are the subject of conrable research, and of a recent book (Stafstrom anho, 2004). Perhaps understanding the mechanismction of the KD will lead to new approaches to treent of the epilepsies.

. The ketogenic diet: definitions and standardrotocols

. Klepper

epartment of Pediatrics and Pediatric Neurology,niversity of Essen, Essen, Germany

The KD is a high-fat, low carbohydrate, nocaloric diet that mimicks the metabolic state of f

ng. During a prolonged fast, body energy requiremn increasing understanding of the underlying meisms, no international protocols have yet been deped. Age, groups, indications and contraindicati

he initiation, ratio, and duration of the diet, as welnticonvulsant co-medications are left to the treahysicians. Medical associations have been relu

o promote the diet and the development of guidelis a result, both in the USA and in Europe cen

ollow individual protocols and the data obtainedifficult to evaluate. A substantial number of inter

ional clinical studies follow the Johns Hopkins Hital protocol (http://www.neuro.jhmi.edu/Epilepseds/ketodiet.htm) (Swink et al., 1997). In Germanpeaking countries, a consensus on how to intrond maintain the diet that closely follows the Hins protocol was introduced in 2003 (Klepper el., 2004) (http://www.neuropaediatrie.com). Free KDoftware and a list of pediatric centres on the wide web providing a KD program in the USA aorldwide is available on the website of the Packhildrens Hospital Stanford University Medical Ce

er (http://www.stanford.edu/group/ketodiet/).

http://www.neuro.jhmi.edu/epilepsy/peds/keto_diet.htmhttp://www.neuro.jhmi.edu/epilepsy/peds/keto_diet.htmhttp://www.neuropaediatrie.com/http://www.stanford.edu/group/ketodiet/


Fig. 1. Countries providing the ketogenic diet (highlighted in grey) as identified in a recent survey. Updated fromKossoff et al. (2005). Reproducedwith permission from the author and from the publisher.

Indications for the KD are intractable child-hood epilepsy, glucose transporter type 1 (GLUT1)deficiency syndrome, and pyruvate dehydrogenasecomplex deficiency. Before initiating a KD contraindi-cations such as-oxidation defects, liver or metabolicdisease interfering with glucose or ketone homeostasisneed to be excluded. A full history as well as a clinicaland nutritional assessment is necessary. Finally, thecompliance of patient and caretakers is essential. In theJohns Hopkins Hospital protocol and in the Germanconsensus protocol patients are admitted to the hospitaland ketosis is initiated by fasting. Reasons for hospital-isation are the difficulties of fasting in children and thepotential hazards of prolonged fasting as an outpatient.Others have described benefits of the non-fasting KDcompared with the initial fasting KD but prospective,randomized trials are not yet available to settle thiscontroversial issue (Kim et al., 2004a; Wirrell et al.,2002). During fasting, basic laboratory parameters, inparticular blood glucose and ketones in blood and urineneed to be monitored closely, preferably at the bedside.Once ketosis is established the diet will be introducedstarting with one-third of calories per meal. Calories areincreased to two thirds of calories per meal after every24 h and patients are discharged when the total amount

of calories per meal is well tolerated. During admissionpatients, parents, and caretakers are educated aboutsuitable food items, the calculation, and the preparationof ketogenic meals by the dietician. Multivitamins,minerals, and trace elements are supplemented.Follow-up demands regular outpatient admissions andeasy access to the treating pediatrician and dietician.Side effects of the diet such as kidney stones and bloodlipids need to be monitored (Kwiterovich et al., 2003;Ballaban-Gil et al., 1998). The time course of theintroduction of a KD and essential investigations asoutlined in the German consensus protocol is shown inFig. 2.

Despite the resurgence of the diet several controver-sial issues about this treatment remain. KD protocolsdiffer in the necessity of the initial fast as outlinedabove, the duration of the diet, fluid restriction, theuse of MCT versus LCT fatty acids, and the needfor carnitine supplements. Other issues such as thepotential long-term adverse effects or the potentialinteractions of the diet with AEDs remain largely unan-swered. Future progress in KD protocols will requirean international expert panel to clarify definitions anddevelop standard protocols. Recommendations shouldbe authorized by an international medical society to


Fig. 2. Algorithm of the introduction of the ketogenic diet as agreed by the consensus in German speaking countries (Klepper et al., 2004).

ensure quality control and to provide an interactive plat-form for physicians using the KD.

4. Implementation and monitoring: nutritionalaspects

A. Tagliabue

HumanNutrition andDietetics, Department of AppliedHealth Sciences, University of Pavia, Pavia, Italy

4.1. Introduction

To obtain a level of ketosis adequate for seizurecontrol, low amounts of carbohydrates and highamounts of fat are administered in a rigidly con-trolled ratio. The use of such an unbalanced diet inchildren requires particular attention in implementa-tion and monitoring. The main nutritional problemsthat require careful monitoring arise from marginalor overt nutritional deficits (energy, proteins, mineralsand vitamins) or from excess nutrients (lipids, satu-rated fat and cholesterol) which are peculiar to thisdiet.

4.2. Fasting versus non-fasting at initiation

Ketosis is usually induced with a short period offasting (Freeman et al., 1996), followed by a gradualintroduction of the high-fat diet. From a biochemicaland physiological point of view, however, a period offast is notnecessary to induce and maintain a con-dition of metabolic ketosis. The key issue is whetherfasting can beusefulwhen initiating the diet in childrenwith epilepsy. Advantages of fasting include a shortertime to onset of ketosis, improved screening condi-tions for identifying underlying metabolic disorders,and the need for a hospital admission period whichallows to teach the family about the preparation ofmeals and the monitoring of ketosis (Katyal et al., 2000;Kossoff et al., 2005). On the other hand, disadvantagesof fasting include psychological stress, the risk of hypo-glycemia or dehydration, and the cost or inconveniencyof hospitalisation and repeated blood tests. In a world-wide survey, fasting was found to be preferred in mostcountries, though it was shortened or eliminated bymost physicians in Canada, UK, India, Sweden, TheNetherlands and Korea (Kossoff and McGrogan, 2005).Unfortunately, only few groups that use the non-fasting


protocol have reported on their experience (Wirrell etal., 2002; Vaisleib et al., 2004; Kim et al., 2004a) andonly one centre has compared the two protocols intheir patients (Kim et al., 2004a). In the latter study,no significant difference was found in time to strongurinary ketosis between the two groups. As for earlyside-effects, moderate dehydration was less frequent innon-fasted subjects, but the incidence of hypoglycemiawas not different between groups.Wirrell et al. (2002)reported that time to ketosis and diet efficacy in eightinpatients managed without fasting were similar tothose described with fasting protocols. Groups usingthe non-fasting protocol have also compared inpatientversus outpatient initiation of the diet, and found no sig-nificant differences in seizure control between the twoapproaches (Vaisleib et al., 2004). More informationfrom well-designed prospective studies, however, areneeded to assess optimal non-fasting protocols and out-patient initiation of the diet. In any case, when the dietis initiated without hospitalisation, adequate facilitiesmust be available to instruct families on meal prepara-tion and monitoring techniques.

4.3. Providing adequate energy supply

It is extremely important to provide adequate energyi asf thec r 7-d , orb ces( ge(e uals asalm ingt eds.G foodr atek ightl satef

4a

ro-t ins

which can be included in the diet, but it is generallypossible to give 1 g protein per kg of body weight. Thisis slightly less than the RDA for infants (1.5 g/kg/day)and for children aged 13 years (1.1 g/kg/day), but it isadequate for older children. Proteins must be of highquality (i.e. rich in essential amino acids), which is usu-ally the case because foods which are rich in animalproteins usually are also rich in fat.

A marginal intake of proteins, calories or other nutri-ents can affect growth. Two prospective (Vining et al.,2002; Liu et al., 2003) and several retrospective studies(Couch et al., 1999; Williams et al., 2002) have shown adecrease in growth percentiles andz scores in childrenon the KD. These children were subjected to caloricrestriction at initiation of the diet with adjustments dur-ing the course of the treatment to match individualneeds, and therefore no weight or height impairmentwas expected. In the latest study, a retrospective chartreview byPeterson et al. (2005), children were giventheir usual caloric intake since the beginning of treat-ment. A decrease in height-for-agez scores was foundat 6 and 12 months, which was significant only in sub-jects with high urinary ketosis (80160 mg/dl). As apoint of interest, subjects who stayed on the diet for12 months tended to consume fewer kilocalories. At 6months the weight change was predicted by kilocalo-r h isn wthi ac-t thm teinl thata ea-s sessc ucedb y ber ncea icalt lin-i

4

ain-t itor-i henp cor-r uredntake to avoid protein (muscle mass) utilizationuel. To this purpose, energy requirements ofhild need to be determined by means of 3- oay food diaries to estimate usual caloric intakey referral to the recommended dietary allowanRDA) for healthy children of the same age ranFood and Nutrition Board, 2002). However, groupstimates can give inaccurate results in individubjects. In our protocol, we measure individual betabolic rate by indirect calorimetry before start

he diet and every 3 months, to adjust caloric neastrointestinal disturbances, poor appetite,

efusal, or intermittent fasting to restore adequetosis may determine a low energy intake and weoss. We use a liquid formulated snack to compenor poor caloric intake when needed.

.4. Ensuring appropriate intake of proteins andmino acids

The maintenance of a constant ratio of fat to peins + carbohydrates limits the amount of proteies intake, but at 12 months it was not. Researceeded to identify factors that permit normal gro

n some children but not in others. In clinical price, all children on the KD should have their growonitored by growth charts, and energy and pro

evels should be adjusted to assist growth at a ratellows seizure control to be maintained. Because murement of body weight alone does not permit to ashanges in body protein content (muscle mass) indy the diet, other estimates of body composition maequired (arm anthropometry, bioelectrical impedanalysis or dual X-ray absorptiometry). Biochem

ests (plasma proteins) may aid in identifying subccal protein deficiency.

.5. Monitoring intake of carbohydrates

A low intake of carbohydrates is essential to main ketosis. Measuring urine ketones allows monng of adherence to the low carbohydrate intake. Wossible, blood ketones, which have been shown toelate better with seizure control, should be meas


Table 1Nutritional monitoring protocol

Nutritional risk Clinical effect Monitoring method

Energy deficits Weight loss; growth impairment Anthropometry (weight, height, skinfold thickness);resting metabolic rate measurements

Protein deficits Muscle mass loss, bone metabolismimpairment

Arm muscle circumference; plasma proteins (albumin,pre-albumin, transferrin)

Lipid excess (saturated) Atherosclerosis Plasma lipids profileCalcium deficits Osteopenia and osteoporosis Dual X-ray energy absorptiometry, bone metabolism

markersSelenium and potassium deficits Cardiac complications Whole blood or serum levelsWater-soluble vitamins deficits Dermatitis, anemia, neurological symptoms,

etc.Plasma and urine levels, blood count

(Gilbert et al., 2000). The low content of starchy foodand vegetables makes the KD extremely poor in fiber,which can cause constipation.

4.6. Optimizing intake of lipids

In a recent study of 141 children, the classic dietwas found to cause a significant increase in athero-genic apoB-containing lipoproteins and a decrease inthe anti-atherogenic HDL cholesterol after 6 months;significant but less marked changes persisted in thosechildren who were followed up for 12 and 24 months(Kwiterovich et al., 2003). To reduce cardiovascularrisk, the ratio of poly-unsaturated + mono-unsaturatedfatty acids to saturated fatty acids needs to be increasedin ketogenic meals, and blood lipid profile should bechecked regularly.

4.7. Minerals and vitamins

Due to the limited food choice, the classic KD is alsodeficient in minerals and vitamins, and needs to be sup-plemented with sugar-free products. Inadequate cal-cium intake and limited sun exposure can impair bonemineralization in children already at risk of osteopeniaand osteoporosis due to long-term AED therapy (Shet,2 in-e

4

il-d itionas re-s ded.

Careful monitoring of nutritional status during treat-ment is necessary to prevent malnutrition. The risk ofmalnutrition is particularly high in children with poornutrition at baseline (Bertoli et al., 2002). Main nutri-tional risks and monitoring methods are summarizedin Table 1.

5. Maximizing acceptability and compliance

S. Bertoli

International Center for the Assessment of NutritionalStatus (ICANS), University of Milan, Milan, Italy

The composition of the KD is expressed by a ratioof fat to protein and carbohydrates weights equal to 4(fat):1 (protein + carbohydrates); proteins are providedin order to meet but not to exceed the recommendedamount of 1 g/kg of body weight, whereas total caloricintake is based on age-adjusted standards. This impliesthat 90% of dietary calories are derived from fat andno more than 23% are derived from carbohydrates.To maintain a stable ketosis, strict adherence to thedietary plan is required, and even small deviations infood preparation or the intake of small amounts of foodoutside the dietary plan may reduce blood ketones,causing a substantial reduction in the diets efficacy( 98

ctso andu tinga effi-c tings ck-i oneh n 1004). Therefore, routine examinations of bone mral density and/or bone enzymes are required.

.8. Conclusions

Implementation and monitoring of the KD in chren needs a dedicated team to assess basal nutrtatus in order to improve caloric and nutrient pcription. Adequate supplementation is always neel

Casey et al., 1999; Carroll and Koenigsberger, 19).The potential short- and long-term side effe

f the diet, together with the diets harshnessnpalatability are the most important factors affecdversely compliance and, consequently, clinicalacy. Although large multicentre studies investigapecifically long-term compliance on the diet are lang, a review of the literature indicates that less thanalf of the patients remain on the diet for more tha


year (Hemingway et al., 2001a; Freeman et al., 1998;Lightstone et al., 2001). Discontinuations are due tomedical and non-medical reasons.

Our experience in a total of 32 children placed onthe KD between September 1994 and August 2005showed that 24 discontinued the diet after 133 months(18 within the first 12 months). Of these, 18 (75%)discontinued for medical reasons, specifically due tolack or loss of efficacy (16 patients) or adverse compli-cations (3 patients, including 2 with severe metabolicacidosis and one with repeated probably diet-unrelatedchest infections). Six children discontinued the diet fornon-medical reasons; these included caregivers issuesand patients unwillingness to continue on the diet. Infact, the diets acceptance and compliance are stronglydependent on the collaboration of both the child and theparents. The preparation of meals can be burdensomefor parents, who must pay great attention in choosing,weighing and cooking every dietary component. Onthe other hand, the direct involvement of parents in the

management of their childs epilepsy often makes thediet more acceptable than expected. The poor palatabil-ity of some ketogenic meals is not a common cause ofinadequate compliance, and it is almost never a causefor discontinuation. By contrast, ingestion of additionalforbidden foods is one of the most common causes ofan insufficient level of ketosis, and may be due to inter-ference from a sympathetic relative or friend or theresult of the childs own industry. The latter possibilityis more frequent in older children and in children withmild cognitive impairment. To maximize acceptability,at our institution the nutritional and neurological teamshave established a collaborative, highly interactive pro-tocol (Table 2).

The recent introduction of nutritionally completepowdered formulas has provided an additional tool toimprove compliance when suitable food is unavailableor cannot be conveniently administered, for exampleduring illness. Regular monitoring of ketone bodieswith urinary sticks or with capillary measurements

Table 2Protocol for maximizing compliance with the ketogenic diet

Operator

Before starting the diet

the diensider

tients

impai

of a se

I, cooke of thesistanc

At 2 wee

Nm rela ietician

ements

M

D ationEnrolment motivated patients/familyDetailed description to family and patients of both the role of

epilepsy (history, efficacy, side effects, etc.), palatability coadherence to meal preparation instructions

Dietary history. Seven-day diary with careful evaluation of paorder personalize at best the dietary plan

Clinical evaluation to identify possible chewing or swallowingdietary strategies

Compilation of a questionnaire concerning the acceptabilityfoods (lecithin, butter, high fat cream, olives, etc.)

nduction period (hospital)Training of parents to instruct them on how to choose, weighContinuous dieticians presence and assistance during intakContinuous neurologists and nutritionists monitoring and as

hospital admissionfter discharge from the hospitalDaily phone contacts with dietician and nutritionist in the firs

eurological and nutritional evaluation every 3 months24 h/day availability of nutritionist and dietician for any probleDietary adjustments according to changing nutritional requir

appetite, and to meet any other potential requirementonitoring of ketones every day with urinary sticks

elivery of a short manual containing the most important inform

of the diet. The manual is directed to all people involved in the m(including school teachers, educators, physiatrists, grandparentsNeurologistt in the treatment ofations, and need for strict

Neurologist, nutritionist

food habits and tastes in Dietician

rment requiring special Nutritionist

lected list of ketogenic Parents

and serve ketogenic meals Dieticianfirst ketogenic meals Dietician

e during the period of Neurologist, nutritionist

ks Nutritionist

Nutritionistted to diet management Neurologist, nutritionist, d, to changing taste and Nutritionist, dietician

Parents

required for maintenance Nutritionist, dietician

anagement of the patient, etc.)


with user-friendly instruments that determine simul-taneously blood ketones and blood glucose, and thecompilation of daily diary of seizures, are essential tothe physician not only to assess compliance but also toevaluate the relationship between ketosis and seizurecontrol.

A synergic collaboration between the patients fam-ily and the epilepsy team is essential for a successfuluse of the diet. In contrast with using AEDs, managingepilepsy with the diet requires very active caregiversinvolvement, and families are prone to balance theirefforts to the extent of perceived results. If there is goodefficacy, acceptance and compliance are greatly facil-itated. In the long-term management of responders,we observed a reduced efficacy in 25% of patients andthis was almost invariably associated with a reducedacceptance of the diet and with reduced plasmaketones. A reasonable hypothesis is that the reductionof compliance was the cause, and the decreasedplasma ketones and loss of the diets efficacy werethe consequence. However, some kind of metabolicadaptation leading to reduced ketosis despite sub-stantial compliance with the diet cannot be excluded.Reduced compliance is frequently a late consequenceof the familys frustration when it becomes difficult orimpossible to attain stable ketosis or effective seizuresc

6i

Ja

a

Do

6

ofe un-d thed il toa ;F int bout

its effects on brain function, but few studies addressedmechanisms of the diets action. Indeed, it was not untilthe mid-1990s that a resurgence in interest led to sus-tained research efforts into this non-medicinal therapy.

That dietary factors can influence brain functionshould not be surprising, and the concept of treatinga chronic disorder such as epilepsy with nutritionalapproaches holds an intrinsic appeal. During the historyof the diets use, clinicians have attempted to modify itsclassic formulation to improve tolerability, and in thehopes of enhancing efficacy. Recent variations on thistheme include the Atkins diet (Kossoff et al., 2003), thelow glycemic-index diet (Pfeifer and Thiele, 2004), anddiets supplemented or enriched in polysaturated fattyacids (PUFAs) (Schlanger et al., 2002; Fuehrlein et al.,2004). This interest has brought together investigatorsfrom diverse fields to focus on how systemic metabolicchanges induced by a high-fat, low-carbohydrate diet which results in prominent ketosis can dampen neu-ronal excitability and/or hypersynchrony, the presumedend product(s) of the diets action.

There have been numerous mechanistic hypothe-ses to explain the anticonvulsant activity of the KD.These include (1) changes in brain pH (e.g., acido-sis which would favour neuronal inhibition throughproton-sensitive ion channels) (Al-Mudallal et al.,1 nce(a .,2 ta( rt bykD -i withw ucho e ofk evi-d moree

6a

thep oneb ace-t dietontrol.

. How does the ketogenic diet work? A looknto the mechanisms of action

.M. Rhoa, P.G. Sullivanb

Barrow Neurological Institute, St. Josephs Hospitalnd Medical Center, Phoenix, AZ, USA;

bSpinal Cord and Brain Injury Research Center,epartment of Anatomy and Neurobiology, Universityf Kentucky, Lexington, KY, USA

.1. Introduction

One of the perplexing mysteries in the fieldpilepsy is the question of how the KD works. This famental question is all the more compelling sinceiet may yield considerable success when AEDs fadequately control seizure activity (Vining et al., 1998reeman et al., 1998). Ever since its introduction

he early 1920s, investigators have been curious a996); (2) changes in electrolyte and water balaMillichap and Jones, 1964); (3) direct inhibitoryctions of fatty acids (i.e., PUFAs) (Cunnane et al002); (4) neurotransmitter alterations (Erecinska el., 1996; Szot et al., 2001; Yudkoff et al., 2001a,b); and5) changes in energy metabolism, reflected in paetone body production and metabolism (Appleton andeVivo, 1974; Pan et al., 1999). Because of the strik

ng ketosis associated with the diet, and the easehich beta-hydroxybutyrate can be measured, mf the attention in this field has focused on the roletone bodies as mediators. However, little directence in this regard has emerged to date (for axtensive review, seeStafstrom and Rho, 2004).

.2. Anticonvulsant effects of ketone bodies andctions on inhibitory neurotransmission

When fats are metabolized preferentially asrimary energy source, the liver produces ketodies (beta-hydroxybutyrate, acetoacetate and

one). The prominent ketosis accompanying the


led Keith (1933) to consider whether ketone bodieswere in fact anticonvulsant compounds. Seizuresacutely induced by thujone (the active constituent ofwormwood oil) in rabbits were blocked by pretreat-ment with acetoacetate (Keith, 1933), an observationthat was recently confirmed in an audiogenic seizuresusceptible mouse model (Rho et al., 2002). Therehave also been several historical reports of acetonepossessing anticonvulsant effects (Helmholz andKeith, 1930; Keith, 1931; Jenney and Pfeiffer, 1958;Yamashita, 1976; Vodickova et al., 1995). The mostextensive characterization of acetones anticonvulsantproperties was conducted byLikhodii et al. (2003),who demonstrated dose-dependent suppression ofseizures by acetone administered intraperitoneally infour separate animal models of seizures and epilepsy:(1) the maximal electroshock (MES) test, which mod-els human tonicclonic seizures; (2) the subcutaneouspentylenetetrazole (PTZ) test, which models humantypical absence seizures; (3) the amygdala kindlingtest, which models human complex partial seizureswith secondary generalization; and (4) the AY-9944test, which models chronic atypical absence seizures.Despite what appears to be broad-spectrum activityexhibited by acetone, the precise anticonvulsant mech-anism(s), however, remain(s) unclear. Interestingly,e dietw vels( M)a scopy( noc ct oft ithe arei eart( calt onso nota Ps)a onsa ouse alc ntsea

ionsa ised

at the centre of molecular bioenergeticsresponsiblefor cellular homeostasis but also the critical determi-nant of cell survival or death. As fatty acids comprisethe bulk of the KD composition, and undergo beta-oxidation in mitochondria, it is presumed that thisimportant metabolic shift alters the balance of impor-tant substrates and neurotransmitters, and at the sametime, affects respiratory chain function with concomi-tant effects on mitochondrial respiration and controlof oxidative stress, among other actions. Additionally,ketone bodies may be juxtaposed not simply as passivebystanders. Ketone bodies are generated by liver mito-chondria as alternative substrates for fuel consumption(when the availability of glucose is low during fast-ing states), especially by the brain. However, recentevidence suggests that ketone bodies may also servea teleological function in preserving brain function asneuroprotective agents (Kashiwaya et al., 2000; Tieu etal., 2003; Reger et al., 2004; Veech et al., 2001; Veech,2004).

Yudkoff and colleagues have proposed that in theketotic state, there are major shifts in brain aminoacid handling, the most significant of which is thereduction of aspartate relative to glutamate, reflectinga shift in the equilibrium of the aspartate aminotrans-ferase reaction (Yudkoff et al., 2001a, 2004a,b). Thisa uro-t rateo bep oxy-l it-t thes ta 5A bee -i tentp hens ck-e one( dyd ildk iso-f ndG uc-t ta ionm rela-pileptic patients who responded favourably to theere found to have elevations in brain acetone le

with estimated concentrations upwards of 1 ms measured by magnetic resonance spectroSeymour et al., 1999). Curiously, there has beenlear demonstration of a direct anticonvulsant effehe principal ketone body, beta-hydroxybutyrate (wither stereoisomer). While these in vivo studies

ntriguing, at a cellular level, ketones do not appo acutely influence synaptic transmission.Thio et al.2000)utilized standard cellular electrophysiologiechniques and found that millimolar concentratif beta-hydroxybutyrate and acetoacetate didffect: (1) excitatory post-synaptic potentials (EPSnd population spikes in CA1 pyramidal neurfter Schaffer collateral stimulation; (2) spontanepileptiform activity in the hippocampalentorhinortex slice seizure model; and (3) whole-cell currevoked by glutamate, kainate, and-aminobutyriccid (GABA) in hippocampal neurons.

Further mechanistic considerations focus on actt the level of the mitochondrion, the organelle podaptation in the metabolism of the excitatory neransmitter glutamic acid (i.e., a decrease in thef glutamate transamination to aspartate) wouldredicted to increase the rate of glutamate decarb

ation to GABA, the major inhibitory neurotransmer, since more glutamate would be available forynthesis of both GABA and glutamine (Erecinska el., 1996; Yudkoff et al., 1996, 1997, 2001a,b, 200).n increase in brain GABA levels would thenxpected to dampen seizure activity (Fig. 3). Interest

ngly, in a variety of seizure models, the most consisrotection afforded by the KD was observed weizures were evoked by the GABA receptor blors picrotoxin, bicuculline and gamma-butyrolactBough et al., 2003). Further, a recent animal stuemonstrated that caloric restriction, resulting in metosis, enhanced expression in the brain of bothorms of glutamic acid decarboxylase (GAD65 aAD67, the biosynthetic enzymes for GABA prod

ion), suggesting an increase in GABA levels (Cheng el., 2004). However, while increased GAD expressight suggest enhanced GABA synthesis, the


Fig. 3. Diagram illustrating the metabolic inter-relationships between brain metabolism of glutamate, ketone bodies and glucose. In ketosis,3-OH-butyrate (beta-hydroxybutyrate) and acetoacetate contribute heavily to brain energy needs. A variable fraction of pyruvate (1) is ordinarilyconverted to acetyl-CoA via pyruvate dehydrogenase. In contrast, all ketone bodies generate acetyl-CoA which enters the tricarboxylic acid (TCA)cycle via the citrate synthetase pathway (2). This involves the consumption of oxaloacetate, which is necessary for the transamination of glutamateto aspartate. Oxaloacetate is then less available as a reactant of the aspartate aminotransferase pathway which couples the glutamateaspartateinterchange via transamination to the metabolism of glucose through the TCA cycle. Less glutamate is converted to aspartate and thus, moreglutamate is available for synthesis of GABA (3) through glutamic acid decarboxylase (GAD). Adapted fromYudkoff et al. (2004c)withpermission.

tionship between GAD expression and GABA levelsis not entirely clear, and depends on numerous vari-ables (Rimvall and Martin, 1992, 1994; Rimvall etal., 1993); increased GAD expression might actuallyreflect decreased GABA production. GAD expressionis more accurately interpreted in the context of mRNAexpression, GAD protein levels, GABA levels andturnover.

Clinically, few measures of excitatory or inhibitoryneurotransmission have been reported in patientstreated with the KD, or who became ketotic throughother means.Dahlin et al. (2005)recently reportedchanges in several amino acids in the cerebrospinalfluid of patients on a KD, the most notable observa-tion being that GABA levels were higher in respondersthan in nonresponders. Moreover, in very good respon-ders (i.e., >90% seizure reduction), GABA levels weresignificantly higher at baseline as well as during thediet. On the other hand, CSF aspartate levels remainedunchanged. The collective evidence from laboratoryand clinical observations strongly suggests that theKD may enhance brain GABAergic inhibition, but not

via the more established mechanism of post-synapticGABAA receptor modulation shared by several anti-convulsant medications.

6.3. The role of PUFAs and modulation ofmitochondrial uncoupling proteins (UCPs)

Ketone bodies are not the only substrates thatincrease during treatment with the KD. The diet ismostly comprised of fats, constituting90% of thetotal calories by weight, and poly-unsaturated fattyacids (PUFAs) may represent a key fat constituent.Certain PUFAs are known to inhibit fast, voltage-dependent sodium channels (Xiao et al., 1995, 1998)and L-type calcium channels (Xiao et al., 1997).Such actions would be expected to dampen membraneexcitability. More relevant, however, are reports thatPUFAs may help reduce seizure activity (Yehuda etal., 1994; Voskuyl et al., 1998; Schlanger et al., 2002;Fuehrlein et al., 2004) and are elevated in patientstreated with the diet (Fraser et al., 2003). Similar tothe effects seen in cardiovascular disease models, it is


Fig. 4. Uncoupling action of an uncoupling protein (UCP). The electron transport chain builds up a proton gradient across the inner mitochondrialmembrane (IMM) which provides the proton-motive force for ATP production (conversion of ADP to ATP through ATP synthase [ComplexV]). The uncoupling proteins provide an alternative pathway for proton re-entry into the mitochondrial matrix, thereby generating heat insteadof ATP. Two major hypotheses for how this occurs have been proposed: (1) the UCP transports protons directly; or (2) electroneutral, protonatedfatty acids diffuse across the inner mitochondrial membrane and the proton is released in the matrix. The fatty acid anion is transported back tothe intermembrane space by the UCP, and the cycle repeats. Adapted fromMattiasson and Sullivan (in press), with permission.

hypothesized that PUFAs possess anticonvulsant activ-ity through modulation of voltage-gated ion channels(Vreugdenhil et al., 1996; Xiao and Li, 1999).

Fatty acids are also known to induce the activity ofUCPs (seeFig. 4), several of which have been identifiedin the mammalian brain (Horvath et al., 2003; Mattsonand Liu, 2003; Sullivan et al., 2003, 2004a; Krauss etal., 2005; Mattiasson and Sullivan, in press). Activationof mitochondrial UCPs results in a reduction of theproton gradient across the inner mitochondrial mem-brane, which dissipates heat and reduces ATP synthe-sis, calcium influx into the mitochondrial matrix, andreactive oxygen species (ROS) production (Horvath etal., 2003). Mitochondrial ROS production is intimatelylinked to the mitochondrial membrane potential ( ),and a high promotes ROS production throughincreased electron shunting. UCPs dissipate byincreasing proton conductance, resulting in decreasedROS formation. Even a modest reduction of cansignificantly reduce ROS production (Votyakova andReynolds, 2001). The relevance of these observationsto the KDs actions was recently illustrated bySullivan

et al. (2004b), who demonstrated that the diet increasedmitochondrial UCP activity and decreased ROS forma-tion in mouse hippocampus.

Reduction of calcium influx and ROS productionare actions that are highly relevant to neuroprotec-tion. However, how do such changes relate to ananticonvulsant effect? It is now well known that thedramatic metabolic and bioenergetic changes in thebrain that accompany seizure activity can contributeto neuronal injury, in part through oxidative mecha-nisms (Liang et al., 2000; Frantseva et al., 2000; Uedaet al., 2002; Cock, 2002; Cock et al., 2002). Lessappreciated is the notion that oxidative stress itselfcan provoke seizure activity. Increased mitochondrialfree radical load, occurring through a variety of mech-anisms, has been shown to produce seizure activity(Willmore et al., 1978; Patel, 2004) or enhance sus-ceptibility to seizure provocation (Liang and Patel,2004; Patel, 2004). Electrophysiological studies havealso confirmed the seizure-provoking effects of oxida-tive stress. Exogenous H2O2 administration can mimicacute oxidative stress in vitro, and several studies have


shown that H2O2 exposure can induce neuronal hyper-excitability. In acute thalamo-cortical and hippocampalslices from rodents, H2O2 caused epileptiform activityboth during infusion as well as after washout (Frantsevaet al., 1998; Avshalumov and Rice, 2002).

Thus, there are emerging data that seizures may beprecipitated by oxidative stress, and that a reductionin free radical formation may prevent seizure activity.While none of the clinically available anticonvulsantmedications are believed to act principally throughantioxidant mechanisms, the KD has been shownto reduce ROS formation (Sullivan et al., 2004b),through induction of several brain-specific mitochon-drial UCPs. Additionally, in a preliminary study,Kimet al. (2004b)demonstrated that the ketone bodies beta-hydroxybutyrate and acetoacetate (in millimolar con-centrations) prevented H2O2-induced hyperexcitabil-ity in acute neocortical brain slices and dissociatedneurons, and that ketones were capable of signifi-cantly reducing ROS formation in isolated mitochon-dria. These observations suggest that the diet mayprotect against seizure activity (and may also be neuro-protective) through antioxidant mechanisms activatedby fatty acids and ketones.

6.4. Actions mediated by the ATP-sensitive (K )a

on-d es ine neu-r P-s lyc aticb them d byh er-p .

s asw ip-p orsalvi ich,i edt tion(e sn est

that ketone bodies block repetitive firing of SNPR neu-rons through activation of KATP channels (G. Yellen,personal communication). It remains unclear, however,whether the biochemical changes induced by the dietactually favour or negate activation of KATP channels(Vamecq et al., 2005). Certainly, prior observationshave suggested that the diet increases the level of bioen-ergetic substrates in the brain (De Vivo et al., 1978;Pan et al., 1999). Additionally, it is known that acuteseizure activity is accompanied by increases in glucosemetabolism (Chugani et al., 1994). These findings maysuggest an inhibition, not activation, of KATP channels.

In addition to the plasmalemmal KATP channel,there exists a mitochondrial form, localized to theinner mitochondrial membrane, that is likely to be dis-tinct in terms of molecular composition, pharmacologicregulation, and biological function(s) (Busija et al.,2004). While the specific role(s) of mitochondrial KATP(mitoKATP) channels has (have) yet to be elucidated,published data suggest that activation of these channelsmay reduce ROS formation (Ferranti et al., 2003), andby extension, potentially reduce neuronal excitability.While there remains some controversy regarding theeffects of mitochondrial KATP channels on ROS pro-duction (Krenz et al., 2002; Xu et al., 2004), publishedreports may reflect tissue-specific and/or methodolog-i rtedta tmP

etst ta theK s-s per-f ranee lsm eenm ilitys smo-l ensi-t orsh theo Kc thatt (V

ATP

nd the two pore domain (K2P) potassium channels

When one considers the pivotal role that mitochria play in energy production, and asks how changnergy metabolism might translate to alterations inonal membrane excitability, the involvement of ATensitive potassium (KATP) channels must be strongonsidered. KATP channels, first described in pancreeta cells, regulate insulin release by modulatingembrane potential. These channels are inhibiteigh ATP/ADP ratios, and activated (resulting in hypolarization) when intracellular ATP levels are low

KATP channels have been identified in neuronell as glia in different brain areas, including hocampus, hypothalamus, substantia nigra and dagus nerve. The highest expression of KATP channelss in the substantia nigra pars reticulata (SNPR) whn addition to other well-known functions, is believo be a subcortical modulator of seizure propagaIadarola and Gale, 1982; Depaulis et al., 1994). Whilexperimental data linking the KD to KATP channels iot yet forthcoming, preliminary observations suggcal differences. Interestingly, ROS has been repoo activate mitochondrial KATP channels (Avshalumovnd Rice, 2003; Lebuffe et al., 2003), an action thaay contribute to ischemic preconditioning (Gross andeart, 2003).In considering the multitude of molecular targ

hat are likely to be affected by the KD,Vamecq el. (2005)have hypothesized that an alternative toATP channel is the K2P (i.e., two pore domain) potaium channel. K2P channels represent a diverse suamily of channels that can regulate baseline membxcitability (Franks and Honore, 2004). These channeay represent a more likely physiological link betwetabolic activity and plasma membrane excitab

ince they can be modulated by changes in pH, oality, temperature, mechanical pressure, and are sively activated by certain PUFAsall of these factave been explored to one extent or another. Onther hand, while the link between PUFAs and2Phannels is tantalizing, there is no evidence to datehe KD induces stable changes in pH or osmolalityDeivo et al., 1978; Al-Mudallal et al., 1996).


6.5. Conclusions

Within the last decade, several putative moleculartargets and novel insights have emerged. Not surpris-ingly, given all of the metabolic, physiological andhormonal changes induced by the KD, there is a mul-tiplicity of relevant mediators involved and it is likelythat no single mechanism, however well substantiated,will explain all of the diets clinical effects. The com-plex mechanisms that are involved in seizure genesisand epileptogenicity, as well as the diverse treatmentapproaches taken clinically, strongly suggest that suc-cessful approaches, including the KD are likely basedon fundamental mechanisms that are multiple, paralleland possibly synergistic.

7. Clinical evidence of efficacy: an overviewfrom major trials

E. Kossoff

The JohnM. Freeman Pediatric Epilepsy Center, JohnsHopkins Medical Institutions, Baltimore, MD, USA

7.1. Introduction

Despite 20 AEDs on the market, many of which haveb e ofp dlyr nglyuf ska gerdW ane ticep ardr ized,c

lueS d toaa rei nsH1 e( d,t ateda hat

it is unlikely that this degree of benefit can result froma placebo response and/or spontaneous remission.In all studies combined, 16% were seizure-free, 32%greater than 90% improved, and 56% with a 50% orhigher response.

Three years later, with more data available, theCochrane Library performed their own review of the lit-erature from 1966 to 2003 (Levy and Cooper, 2003). Inthis review, utilizing more stringent criteria, they wereless convinced about the diets efficacy. Their statedobjective was to find randomized, controlled studies toprove the diets efficacy. None were found. They con-cluded there is no reliable evidence from randomizedcontrolled trials to support the use of KDs for peoplewith epilepsy. However, the reviewers still classifiedthe diet as a possible option because a small num-ber of observational studies lend some support for abeneficial effect.

With this confusing data, how does the diet fit into aphysicians treatment algorithm? In 2001,Epilepsy andBehavior published The Expert Consensus Guide-lines, in which 100 adult and pediatric epileptologistsreviewed clinical case scenarios and provided theirtreatment options in order of preference (Karceski,2001). Among this group, the diet did not fare well.For idiopathic generalized epilepsy, the diet ranked in1 oft , buto er-a ), thed

7

re-s ti-p noxG syn-d neo eter-m in1 ryt gly,1 0%i dietb toinat nlyeen introduced in the past decade, the incidencharmacological intractability has not been markeeduced. Partly as a result of this, the KD is increasised in every continent (Kossoff et al., 2005). Many

amilies of children with intractable epilepsy now about the KD at clinic visits, and the diet is no loniscouraged by physician (Hemingway et al., 2001b).hat is, however, the evidence for its efficacy? In

ra in medicine of increasing guidelines and pracarameters, many neurologists want straightforwecommendations based on evidence from randomontrolled trials.

In 2000, an article written for the Blue Cross Bhield insurance company in the U.S. attemptenalyze all studies available at the time (Lefevrend Aronson, 2000). In this article, 11 studies we

dentified; 9 of which were retrospective from Johopkins. Only two were prospective (Vining et al.,998; Freeman et al., 1998), one was multicentrVining et al., 1998). Although none were controllehe authors of this review felt that the diet demonstrsignificant reduction in seizure frequency and t5th position in a list of possible therapies; 74%he physicians considered it an appropriate choicenly as third-line therapy. Even in symptomatic genlized epilepsies (e.g., Lennox Gastaut syndromeiet ranked no higher than 9th to 11th place.

.2. Evidence from retrospective studies

Information on studies with over 20 patients is pented inTable 1. All studies generally describe mulle and varied seizure types, most commonly Lenastaut syndrome and other generalized epilepsyromes typically with difficult-to-control seizures. Of the earliest retrospective studies was by Dr. Pan from the Mayo Clinic in Rochester, Minnesota925 (Peterman, 1925). He reported on a new dieta

reatment in 17 children aged 115 years. Amazin0 (60%) were seizure-free, and 4 (23%) were >9

mproved. Relatively few papers followed and theecame less popular with the introduction of phenynd other AEDs.Hopkins and Lynch (1970)reported

he treatment of 34 patients with a 3:1 ratio KD. O


three (8.8%) were seizure-free, but 71% had eithermuch or moderate improvement.

In 1986, the modern era of KD studies began withSills report (Sills et al., 1986) of 50 patients treatedwith an MCT diet. Of these, 44% had a >50% improve-ment in seizures, and 8 (16%) were seizure-free. Inter-estingly, 86% of the children with absence epilepsyand 50% of those with myoclonicastatic epilepsy hada >50% improvement.

Kinsman et al. (1992)reviewed 58 children fromJohns Hopkins Hospital; 67% had a >50% improve-ment, with 29% having a near complete response.Almost two thirds were able to reduce medications andmore than one-third were reportedly more alert andinteractive. Due to the interest elicited by this publi-cation, many other studies at Johns Hopkins Hospital,including four important prospective studies in the pastdecade, were subsequently completed (see below).

Two studies in infants have been recently published.One reported that 54% of 32 infants had a >50%response (19% seizure-free); in this population, thediet was remarkably well tolerated and ketosis wasattained easily (Nordli et al., 2001). A retrospectivereview of 23 children specifically with intractableinfantile spasms found that 6 (46%) of the 13 remain-ing on the diet at 12 months had a >90% spasmr ara erl

pi-t rtedo facedmt r ana oler-a cals

thel uingt .,1 01;C .,2

7

ta of

three diets including the classic, MCT, and a modifi-cation, 81% showed a >50% response (Table 1). Theclassic diet had the highest efficacy and was the besttolerated. In 1998, two prospective studies from JohnsHopkins were published. The first (Vining et al., 1998)was a multicentre (seven hospitals) study of 51 chil-dren aged 18 years with intractable epilepsy, who hada mean of 230 seizures per month and had failed anaverage of seven AEDs. At 6 months, 53% had a >50%response, 6 (12%) were seizure-free and 73% remainedon the diet. The second prospective study from JohnsHopkins was a single-centre study of 150 consecutivechildren aged 116 years with more than two seizuresper week that had failed at least two AEDs and werefollowed-up for 1 year (Freeman et al., 1998). At base-line, children had a mean of 410 seizures per month anda history of exposure to 6.2 AEDs. In this group, 71%were able to remain on the diet, and 51% had a >50%improvement, 32% with a >90% reduction at 6 months.In general, those who discontinued the diet before 1year did so due to inefficacy rather than restrictiveness.No specific seizure type preferentially improved, butage older than 8 years at diet initiation was slightlyless likely to show an improvement. In comparing out-come on the diet to many new add-on medications,the authors considered the diet as potentially moree

ate( edi 0%ra 20c ni-ta ada tra-d ithn nes.I osisa ortedf nda der-w

tud-iO un-t ofp

eduction (Kossoff et al., 2002a). Age less than 1 yend trials of two or fewer AEDs correlated with high

ikelihood of improvement.A retrospective study from Johns Hopkins Hos

al and the University of Texas at Houston repon 45 adolescents aged 1219 years who hadany years of epilepsy (Mady et al., 2003). Despite

he concerns that the diet might be too restrictive fodolescent population, seizure improvement and tbility were remarkably similar to those in more typichool-age children.

The KDs use worldwide has blossomed sinceate 1990s, with many retrospective studies contino report high response rates (Table 1) (Hassan et al999; Vaisleib et al., 2004; Kankirawatana et al., 20oppola et al., 2002; Francois et al., 2003; Kim et al004a; Klepper et al., 2004).

.3. Prospective uncontrolled studies

In the first modern prospective study,Schwartz el. (1989)reported that, of 63 patients given oneffective.A small study on a modified low-carbohydr

Atkins) diet in six children and adults publishn 2003 reported that four patients had a >5esponse, correlating with level of ketosis (Kossoff etl., 2003). A follow-up prospective study assessedhildren with intractable epilepsy for 6 months, iially with 10 g of carbohydrate per day (Kossoff etl., in press). Nearly two thirds of the patients hsignificant seizure reduction, a similar to the

itional KD. The diet was very well tolerated wo significant weight loss nor cases of kidney sto

nterestingly, the presence of large urinary ketppeared to be less important than commonly rep

or the KD. Further studies in both children adults with this novel ketogenic therapy are unay.A summary of outcome data in prospective s

es including at least 20 patients is given inTable 3.verall, results show some disparity between co

ries, with 6-month efficacy rates ranging from 27%atients with a 50% seizure reduction in Italy (Coppola


Table 3A summary of retrospective and prospective efficacy studies of the ketogenic diet including over 20 patients and published since 1970 (notincluding abstract presentations)

Reference No. of patients Age range (years) >90% improvement rateat 6 months

>50% improvement rateat 6 months

Hopkins and Lynch (1970) 34 113 71Sills et al. (1986) 50 215 24 44Schwartz et al. (1989)a 59 90% reductionin Thailand (Kankirawatana et al., 2001).

7.4. Prospective randomized studies

Four randomized studies are being completed in theUSA and UK. In a study from Johns Hopkins, 20 chil-dren with intractable Lennox Gastaut syndrome wereadmitted, fasted, slowly advanced onto the KD over3 days and randomized to either receive an additionaldaily solution containing saccharin or glucose. After 1week, they were crossed to the opposite arm before dis-charged to home. The study, currently being analyzed,was the first to prove that the KD could be studied in adouble-blinded, randomized, placebo-controlled man-ner. Dr. Helen Cross at Great Ormond Street Hospital inLondon is completing a prospective randomized studycomparing the classical KD with an MCT oil diet, witha control period of 416 weeks incorporated for eachindividual patient. Dr. Christina Bergqvist at the Chil-drens Hospital of Philadelphia is currently analyzinga third randomized study of the value of a fasting com-pared to a gradual initiation of the diet. A fourth studyby our group is examining the Atkins diet in children,randomized to either 10 or 20 g of carbohydrate perday, with a crossover period after 3 months for eachp

7.5. Conclusions

The efficacy of the KD was initially assessed mostlyin small, single-centre, retrospective studies. In recentyears, reports of the diets efficacy worldwide have con-tinued to show that approximately half of patients willhave half their seizures, and about one-third in totalwill have one-tenth their baseline number of seizures.Prospective studies have confirmed these data, and fourrandomized studies are in the process of completion.Although clearly a need exists for further prospectivestudies, the quality of evidence for the diets efficacy isimproving over time.

8. Seizures type and epilepsy syndrome aspredictors of response

P. Veggiotti, R. Epifanio, F. Longaretti, G. Lanzi

Department of Child Neurology and Psychiatry, Insti-tute of Neurology, IRCCS C. Mondino Foundation,University of Pavia, Pavia, Italy

8.1. Introduction

Few studies have investigated clinical outcome afters es.atient. tarting the KD in well-defined epileptic syndrom


Moreover, there are no randomized trials comparing thediet with placebo or with AEDs, and most of the largescale studies conducted to date only reported resultsafter short term treatment. For these reasons, definingthe comparative efficacy of the KD in relation to spe-cific syndromes or seizure types is no easy task. Inthis report, this aspect will be addressed by reviewingbriefly data from the literature and by presenting ourpersonal experience.

8.2. Literature data

With respect to response to the diet in relation toseizures type, several studies have reported particu-larly favourable outcomes in children with myoclonicseizures (Wheless and Sankar, 2003); these data, how-ever, refer to observations in large series of patientswith drug resistant epilepsies, and there are no pub-lished studies that document the comparative efficacyof the diet in different types of myoclonic epilepsies.A study byMaydell et al. (2001)did not find any sig-nificant difference in the diets efficacy when patientswith partial seizures were compared with those withgeneralized seizures.

Studies relating response to specific epilepsy syn-dromes are even fewer. Efficacy of the KD in LennoxG ;M i-t eses as fter5 he1 nics

.( insti dif-f Ds.H lo-ge ntsw ibil-i ralb e ofc tor ofa

t vere

myoclonic epilepsy of infancy (Dravet syndrome). Inthis report, 20 patients meeting the diagnostic criteriafor the syndrome were enrolled; 10 of the 13 childrenwho remained on the diet for a duration of follow-upof 14 years had a significant reduction of seizures.These results are of particular interest because of thehomogeneity of the cohort and the long-term follow-up.

8.3. Our experience

We took part in an Italian multicentre studythat assessed 56 children, adolescents and youngadults with refractory epilepsy; and we also reportedseparately our personal experience in a total of 36patients (including 16 who had taken part in themulticentre study) (Coppola et al., 2002; Bertoli etal., 2005). In these studies, which included patientswith partial and generalized symptomatic epilep-sies, myoclonicastatic epilepsy, Lennox Gastautsyndrome and others epileptic encephalopaties, wedid not identify a statistically significant relationshipbetween outcomes on the diet and either seizure typeor epilepsy syndrome. Subsequent evaluation of ourexperience, however, have led us to hypothesize agood response of the diet can be especially observed inp they tients( ffusem at 9m asei ta mayr sisf verem tiont somer re ise turet diesiC et nsc eirc thats ainst psyc entsastaut syndrome has been reported (Wheless, 2004ckay et al., 2005), but diagnostic electro-clinical cr

eria do not appear to have been always fulfilled in thtudies.Freeman and Vining (1999)have describedignificant reduction in seizure frequency (>50%) adays of the introduction of the diet in all of t

7 Lennox Gastaut patients with atonic or myocloeizures included in their study.

Recently,Nordli et al. (2001)and Kossoff et al2002a)reported a beneficial effect of the diet aganfantile spasms in patients with West syndrome oferent etiologies unresponsive to conventional AEowever, efficacy rates in relation to specific etioies, or specific EEG patterns, were not specified.Thant al. (2005)also reported favourable results in patieith infantile spasms, and did not exclude the poss

ty of a good outcome even in patients with structurain abnormalities. In the latter report, the presencomplex partial seizures was regarded as a predicn unfavourable response to the KD.

A recent study byCaraballo et al. (2005)reported onhe effectiveness of the KD in the treatment of seatients with diffuse migrational disorders. Duringears between 1994 and 2004 we treated seven pamean age 8.2 years, range 2.816.1) showing diigrational disorders; four of them experienced,onths since the introduction of the diet, a decre

n seizure frequency between 50% and 90%.Than el. (2005)have also suggested that these patientsespond particularly well to the diet. A biological baor putatively better responses in patients with seigrational disorders may be found in the observa

hat the cerebral structure of these patients bearsesemblance to the foetal brains structure. Thevidence that, as in suckling rats, the more immahe cerebral cortex is, the more it uses ketone bonstead of glucose as an energetic substrate (Crino andhou, 2000; Morris, 2005); accordingly, it is possibl

hat patients with diffuse cortical malformatioould utilize more efficiently ketone bodies in therebral metabolism. While it is acknowledgedurgical intervention, whenever possible, remhe treatment of choice in drug refractory epileaused by focal cortical dysplasia, some pati


may benefit from the diet, at least before surgery isapplied.

It has been suggested that patients with Laforabody disease may also respond favourably to the diet(Minassian, 2001). Given the pathophysiology of thisdisorder, it is reasonable to hypothesize that the diet,by promoting the brain utilization of ketones insteadof glucose, might reduce glycogen synthesis and pos-sibly decrease polyglucosan accumulation and slowdown the progression of the disease. In a pilot study(Cardinali et al, submitted for publication) we assessedthe effects of the diet in five patients with Laforabody disease (average age at our first observation:13.23.2 years), in order to determine the feasibilityof its long-term term use in adolescents with this condi-tion, and its potential impact on clinical outcome. Thediet was well tolerated in all five patients for a periodof 28.48.8 months, but the illness showed a progres-sive deterioration, with worsening of the myoclonus,appearance of new seizure types, and increased fre-quency of generalized tonicclonic seizures. Moreover,four of the five patients showed a significant declinein cognitive functions, with special reference to mem-ory, long lasting attention and performance abilities.These findings seem to argue against a protective roleof the diet on the progression of the disorder, but fur-t sionsc rtedo dis-e ronald bser-v ases ouldb develo sible( ofe llyd asep

8

andl oft ec-t s ofp aren racti-

cal implications for the selection of the best candidatesfor treatment.

9. Is there a role for the ketogenic diet beyondchildhood?

E. Kossoff

John M. Freeman Pediatric Epilepsy Center, JohnsHopkins Medical Institutions, Baltimore, MD, USA

In one of the earliest publications (Peterman, 1924)reporting significant benefits from ketosis, patientstreated with the diet ranged from 3.5 to 35 years,although no information was provided as to the influ-ence of age on outcome. In subsequent years, the dietwas increasingly used for adults; the largest seriesreported to date, published in 1930 (Barborka, 1930)included 100 patients aged 1651 years. Only 27 wereaged 30 or above, and 38 were under age 20 andcould be more appropriately classified as adolescents.Ketosis typically occurred, but interestingly severalpatients had improved seizure control without persis-tent ketosis. Overall, 56% had a >50% response and12% were seizure-free, outcomes remarkably similarto those reported for children in the next 80 years. Inter-estingly, the authors commented that there seems tob rdedt unga yt -t eralp e 56w ecentsy allyf ringt l.( e oft

D in1 izuref Jef-fw ted4 , 6( 7%)h s ofc eveher studies are necessary before definite concluan be drawn. Moreover, our patients were stan the diet years after the clinical onset of thease, when Lafora bodies accumulation and neuegeneration had already become established. Oations in a knockout mouse model of the diseuggest that whatever treatment is applied, it she started as early as possible, because alreadyped pathological changes do not appear to reverGanesh et al., 2002). This raises the possibilityxploring the potential value of the KD in geneticaiagnosed, pre-symptomatic siblings of Lafora diseatients.

.4. Conclusions

There is inadequate information on how short-ong-term outcome on the KD relates to the typehe epileptic disorder and/or ictal semiology. Prospive studies with large and homogeneous cohortatients with well-defined electro-clinical featureseeded to address this issue, which has obvious p-

e no question but that the patient who can be affohe best opportunity for treatment is the child or yodult. . .whereas older patients. . .are the least likel

o be benefited (Barborka, 1930). Although not sysematically monitored, cholesterol increased in sevatients; amenorrhea occurred in 12 (21%) of thomen treated, an occurrence also seen in a rtudy of adolescents (Mady et al., 2003). For nearly 70ears, thereafter, the KD was used only sporadicor adults in some centres. In a 1989 study compahe MCT oil diet to the classical KD,Schwartz et a1989)described 4 adults out of 63 patients. Nonhe 4 adults had any significant improvement.

In 1999, a retrospective assessment of the K1 adults (9 women, age 1945 years, baseline se

requency 230 per month) was reported from theerson centre in Philadelphia (Sirven et al., 1999). Allere treated with a typical calorie and fluid restric:1 ratio diet after a 2472 h fast. At 8 months55%) had a >50% seizure reduction and 3 (2ad a >90% improvement, similar to most studiehildren. Only two patients were unable to achi


and maintain ketosis; all patients reported constipa-tion and all women described menstrual irregularities.Cholesterol increased over 6 months from a baseline of208 to 291 mg/dl and triglycerides were only slightlyincreased from 190 to 203 mg/dl. A follow-up studyfrom the same group with 15 additional patients (total26) was presented at the 2003 meeting of the Ameri-can Epilepsy Society in poster form (Nei et al., 2003).Similarly to the previous study, 54% had a >50%response, and the authors were able to differentiateoutcome based on seizure type. They indicated that the12 patients with symptomatic generalized epilepsy hadthe best outcome, with 73% having a >50% response,compared with only 27% of the 11 adults with partialepilepsy. Improvement, when it occurred, tended to bein the initial 2 weeks of the diet. Diet duration wasrelatively short, only 7 months on average. Only 11%continued the KD for 2 years, and all of these patientshad symptomatic generalized epilepsy. Weight loss wasalso described, with a mean decrease of 6.7 kg. Furtherinformation about these two studies are detailed in atextbook chapter on adults and the KD (Sperling andNei, 2004).

In 2001, a retrospective study of 45 highly moti-vated adolescents with intractable epilepsy (age 1219years) was reported (Mady et al., 2003). Six monthsa iet,w and8 ilart1 es ilart2 tob thed ndm ostc

fiedA orc reeo age( da tely,w thed t, a4 cess.T or 3

months but had no measurable seizure reduction. Theolder patient was unable to create ketones because hewas given carbohydrates in his chronic care facility, anddiscontinued the diet after 8 months. Encouraged by theabove results, we have started a study of adults aged 18and over using a modified Atkins diet; carbohydratesare limited to initially 15 g per day and medicationsare left unchanged for the first month. Daily multivi-tamin and calcium supplementation is provided, andfrequent serum and urine laboratory monitoring is per-formed. In approximately 10 patients to date, resultsare mixed and continue to be correlated with level ofketosis.

In conclusion, studies of the use of the KD for ado-lescents and adults are still in their infancy despiteover 80 years of clinical use for children. Efficacyappears to be similar to published studies of children.Further prospective studies, including perhaps a lessrestrictive modified Atkins diet, are warranted. Sideeffects such as weight loss, dysmenorrhea, constipa-tion, and dyslipidemia will need monitoring in anyfuture studies.

10. Discussion summaryimplementationprocedures and efficacy data

E

IaP

Q cine,t vesm olledt didn izuret ernsb

A ctives KDm ma-r m-i vend osev on-i bep im-fter diet initiation, 28/45 (62%) remained on the dith 6/28 (21%) having 5090% seizure reduction/28 (29%) having >90%. These results were sim

o those observed in younger children (Vining et al.,998; Freeman et al., 1998). Patients with multipleizure types (generalized epilepsy) did best, simo the adult study from the Jefferson centre (Nei et al.,003). The restrictions of the diet did not appeare a major problem, with only 22% discontinuingiet primarily for this reason. Weight loss (60%) aenstrual dysfunction (45% of females) were the m

ommon side effects.A small case series of the benefits of a modi

tkins diet (providing more protein and no fluidalorie restriction) was reported in 2003, with thut of six patients being older than 18 years ofKossoff et al., 2003). An 18-year-old female ha

90% improvement in seizures almost immediaith large urinary ketosis, and was maintained oniet for 20 months without side effects. In contras2-year-old man and 52-year-old man had less suche younger man maintained moderate ketosis f. Perucca

nstitute of Neurology, IRCCSC.Mondino Foundation,nd Clinical Pharmacology Unit, University of Pavia,avia, Italy

. By current standards of evidence-based medihe quality of the studies conducted with the KD leauch to be desired. We have no adequately contr

rials that demonstrate efficacy, and most reportsot characterize adequately patients in terms of se

ype and epilepsy syndrome. How can these conce addressed?

. Everyone acknowledged that we need prospetudies of better quality. While the nature of theakes blinding difficult, recent developments sum

ized in Dr. Kossoffs presentation show that randozed studies are now being initiated, and that eouble-blind protocols (e.g., by administering glucersus saccharin in addition to the diet, with the mtoring of ketosis by an external investigator) canractically and ethically feasible, at least over l


ited assessment periods. While limitations of currentevidence are acknowledged, Dr. Freeman and otherpresenters remarked that the time course and mag-nitude of improvement is hardly compatible with aplacebo response or with the natural history of theseizure disorder. It is hard to question the diets effi-cacy when children with a long standing history offrequent daily seizures improve dramatically duringthe fasting period and remain persistently improvedfor months and years thereafter. Outcome data in termsof sustained seizure freedom observed with the KDappear to be greater than those reported for histor-ical controls included in AED trials. Nevertheless,it is a fact that overall the quality of the efficacydata is not satisfactory, and that it is important infuture studies to use prospective controlled designs,and to include patients with better defined epilepsysyndromes.

Q. There seems to be controversy about what imple-mentation procedure is best. We have been presentedwith different initiation protocols, controversial data onthe need for hospital admission, even different types ofdiets.

A. We do not have adequate data to answer these issues.Again, randomized comparative studies are needed tod ndo uchk r thel pe-c pitala nta-g g off bet-t itiona , fore edm thep tkinsd dieti TheA ept-a d itm icald uret d asa atp

Q. Dr. Tagliabue described a rather comprehensive setof investigations to monitor the effects of the diet.Should these be all part of our routine monitoring pro-cedures?

A. Dr. Tagliabue clarified that some of the investiga-tions, for example some anthropometric and metabolicmeasurements, are conducted in her centre as part of aresearch protocol, and should not be regarded as nec-essary in a routine setting. Some tests, such as bonemass determinations and abdominal ultrasound, mayonly be required on a case-by-case basis, for examplewhen nutritional status and liver enzymes justify con-cerns with bone mineralization loss, or when the findingof blood in urine raises suspicions about possible renalstones. It appeared from the discussion that laboratorymonitoring at some European centres tends to be moreintensive than in North America. In any case, there wasagreement that the need for testing for nutritional defi-ciencies can be minimized by providing patients withappropriate supplementation of vitamins and minerals.

Q. How can we best monitor ketosis, by determinationsin blood or by simply checking urine ketones?

A. Dr. Kossoff commented that one study did suggestan improved outcome with monitoring blood ketones( att oodt withs hatd opti-m

Q wsa esa

A arelya loodl nts,e ther esti-g ftent mil-i

Q thed oesn forew

etermine the value of fasting, fluid restriction, ather initiation procedures, to determine how metosis is needed for seizure control, and whetheevel of ketosis should to be adjusted in relation to sific patients characteristics. On the issue of hosdmission, most participants considered it advaeous to ensure adequate involvement and trainin

amily members. Hospital admission also ensureser adjustment of dietary needs, and prompt recognnd treatment of potential early adverse effectsxample in children with a previously unrecognizetabolic defect. There was some discussion onros and cons of alternative diets, such as the Aiet. The prevailing view was that the classical

s the best documented in terms of efficacy data.tkins diet has some attractiveness in terms of accbility, particularly in adolescents and adults, anight be offered when compliance with the classiet is anticipated to be problematic. However, fail

o respond to the Atkins diet should not be regarden indication that the classical diet will not work in thatient.Gilbert et al., 2000). Overall, however, he felt thhere is still insufficient evidence to conclude that blesting ensures an improved outcome comparedimply monitoring urinary ketones. He also felt tevices for monitoring blood ketones are less thanal at the current state of technology.

. What do we do when blood biochemistry shobnormalities, particularly high levels of triglyceridnd cholesterol?

. There was consensus that dyslipidemias are rcause for discontinuing the diet. Changes in b

ipids can usually be managed by dietary adjustme.g. by altering the caloric intake and/or modifyingatio of unsaturated to saturated fat. On further invation, some cases of problematic dyslipidemias o

urn out to reflect co-morbid conditions, such as faal hypertriglyceridemia.

. If a child does well on the KD, how long shouldiet be continued? On the other hand, if the diet dot appear to work, how long should we go on bee give it up?


A. There was agreement that duration of treatmentshould be determined on a case-by-case basis. InGLUT1 deficiency syndrome and in pyruvate dehy-drogenase (E1) deficiency syndrome, the diet usuallyneeds to be continued for several years. In epilep-sies that responded well to the KD, discontinuationof the diet may be considered after about 2 years. Innon-responders, a trial of about 3 months will usu-ally be sufficient to conclude that diet is not going tohelp.

11. Adverse effects and contraindications

J. Campistol

Department of Neuropediatrics, Hospital Sant Joan deDeu, Barcelona University, Barcelona, Spain

11.1. Introduction

Before starting the KD, clinicians need to be awareof the type of epilepsy, the characteristics of thepatients, the type of potential adverse effects, and thediets medical contraindications (Kossoff, 2004; Kanget al., 2005).

1

nots enti a;C ri ntalr

TE

G

DSCHPB ,

lic

Table 5Long-term or late-onset adverse effects of the ketogenic diet

Growth retardationHepatic failureConstipationExacerbation of gastro-esophageal refluxMineral deficienciesVitamin deficienciesUrolithiasisOsteopeniaCardiomyopathyProlonged QT intervalOptic neuropathyBasal ganglia injuryAnemiaBiochemical disturbances (hypocarnitinemia, dyslipidemias,

elevated very long chain fatty acids)

Most of the early onset adverse effects of the diet(Table 4) are transient and can be carefully managedwith conservative strategies. These include gastroin-testinal disturbances such as nausea, loss of appetite,refusal to drink, and vomiting, usually associated withgastritis and fat intolerance. Another disturbance isconstipation that can be caused by a decrease in inges-tion of food and fibers; laxatives or enemas can suc-cessfully control it.

Dehydration is one of the most common com-plications that occur early after initiation of thediet (Wheless, 2001; Kang et al., 2004). It can beprevented by avoiding a fasting period and/or byencouraging oral or intravenous intake of fluidswithout dextrose (Wheless, 2001). Potentially seriouscomplications include sepsis, cardiomyopathy, lipoidaspiration pneumonia, hepatitis and acute pancreatitis.Sometimes these complications may be related toconcomitant use of some AEDs, especially valproicacid (VPA). Fatal fulminating sepsis in sporadic caseshas been attributed to pre-existing metabolic disease(Kang et al., 2004). Overall, serious complications maylead to discontinuation of the diet in 612% of patients.

The list of early onset biochemical disturbancescomprises hypoglycemia, hypertriglyceridemia (whichmay lead to pancreatitis), hypercholesterolemia, tran-s ases( ne-m bolica g eta ane

1.2. Adverse effects

Most side effects of the diet are mild and doignificantly affect the acceptability of the treatmn the majority of patients (Hemingway et al., 2001oppola et al., 2002). In general, tolerability is poore

n young patients with refractory epilepsy and meetardation than in other populations.

able 4arly onset adverse effects of the ketogenic diet

astrointestinal effects (nausea, vomiting, loss of appetite,constipation)

ehydrationepsisardiomyopathyepatitisancreatitisiochemical disturbances (hypoglycemia, hypertriglyceridemiahyperuricemia, hypertransaminemia, hypercholesterolemia,hypoproteinemia, hypomagnesemia, hyponatremia, metaboacidosis)ient hyperuricemia, increased liver transaminespecially in patients receiving VPA), hypoproteiia, hypomagnesemia, hyponatremia and metacidosis (Galvan et al., 2001; Wheless, 2001; Kanl., 2004; Coppola et al., 2002; Vining, 1999; Freemt al., 2000).


Long-term or late-onset adverse effects (Table 5)include growth retardation (mainly related to proteindeficiency), hepatic failure, vitamin and mineral defi-ciencies, immune disturbances (impaired neutrophilefunction, related to imbalanced diet and protein restric-tion) and cardiomiopathy, which may also be seen inthe early phases of treatment (Vining, 1999; Coppolaet al., 2002; Wheless, 2001; Galvan et al., 2001; Furthet al., 2000; Vining et al., 2002). Another late-onsetcomplication is nephrolithiasis, which may result fromchronic acidosis, dehydration, and fat malabsorption,and may be more common in patients receiving car-bonic anhydrase inhibitors such as topiramate (TPM)and/or zonisamide (ZNS). Osteopenia may be causedby nutritional deficiencies, reduced sunlight exposure,immobilisation and underlying AEDs.

Late-onset biochemical disturbances include sec-ondary hypocarnitinemia, iron deficiency anemia, vita-min and mineral deficiencies, increased uric acidproduction, low serum bicarbonate levels, persistenthyponatremia, sustained metabolic acidosis, and dys-lipidemias (Vining, 1999; Kossoff, 2004; Furth et al.,2000).

Physicians should be aware of the above compli-cations, so that they can be carefully monitored andsuccessfully prevented and treated as appropriate. Ther ullyw hichm ilityo en-t za-t asedh ge

1

rali andk driald ne

ofm dur-i ne mw itined ns-

ferase I and II and carnitine translocase) or secondary.Fatty acid defects are absolute contraindications tothe diet, and this is especially true for medium chainacyl-CoA dehydrogenase deficiency (MCAD). Otherdefects in fat metabolism that can exacerbate ketotichypoglycemia are long-chain acyl-CoA dehydrogenasedeficiency (LCHAD), short chain acyl-CoA dehydro-genase deficiency (SCHAD) and mitochondrial andketone metabolism defects (Wheless, 2001; Kang et al.,2005). The KD should never be used in children withpyruvate carboxylase deficiency, in whom its adminis-tration may have fatal consequences.

The diagnosis of inborn errors of metabolism isoften difficult. Patients can present with refractoryepilepsy and non-specific symptoms that can beattributed to seizures or drugs. In some of these cases,the epilepsy may be secondary to inborn errors ofmetabolism, a possibility which should be consideredalways and particularly in children with associateddevelopmental delay o

epilepsy diet

epilepsy research

university of pavia

use of theketogenic

different epilepsy syndromes

diets clinical effects

clinical studies

adverse effects

mondino foundation

Documents

epilepsy epilepsy overview

neurobiochemical mechanisms of a ketogenic diet in...

canine epilepsy: current management and nutritional...

epilepsy o epilepsy overview

epilepsy epilepsy news - epilepsy nl | epilepsy...

natural approaches to epilepsy - chiro · natural...

ketogenic diet at children’s mercy ketogenic diet is an...

the ketogenic diet for epilepsy learn about the diet, the...

international epilepsy news · since the mid-90s, the...

crosstalk between the ketogenic diet and epilepsy:...

ketogenic diet guidelines for infants with refractory...

epilepsy and the ketogenic diet - home -...

epilepsy and the ketogenic diet: clinical - researchgate

ketogenic diet for epilepsy · in fact, the ketogenic diet,...

the ketogenic diet for epilepsy learn about the diet, the...

diet - adventhealth neuroscience institute · adult...

ketogenic diet guidelines for infants with refractory · pdf...

ketogenic diet management - rdn.pnds.orgrdn.pnds.org › ......

tracy keto diet 11.17.12 - epilepsy foundation€¦ · day...

ketogenic diet for epilepsy€¦ · a child on the diet...