Safe and Effective Use of the Ketogenic Diet in Children with Epilepsy and Mitochondrial Respiratory Chain Complex Defects

Authors

  • Hoon-Chul Kang,

    1. Department of Pediatrics and Epilepsy Center, Sanggye Paik Hospital, Inje University College of Medicine, Seoul, Korea
    Search for more papers by this author
  • Young-Mock Lee,

    1. Department of Pediatrics, Pediatric Epilepsy Clinics, Severance Child's Hospital, Brain Research Institute, Yonsei University College of Medicine, Seoul, Korea
    Search for more papers by this author
  • Heung Dong Kim,

    1. Department of Pediatrics, Pediatric Epilepsy Clinics, Severance Child's Hospital, Brain Research Institute, Yonsei University College of Medicine, Seoul, Korea
    Search for more papers by this author
  • Joon Soo Lee,

    1. Department of Pediatrics, Pediatric Epilepsy Clinics, Severance Child's Hospital, Brain Research Institute, Yonsei University College of Medicine, Seoul, Korea
    Search for more papers by this author
  • Abdelhamid Slama

    1. Laboratoire de Biochimie 1, APHP, Hopital de Bicetre, Le Kremlin-Bicetre, France
    Search for more papers by this author

Address correspondence and reprint requests to Dr. H.D. Kim at the Department of Pediatrics, Yonsei University of Medicine, Severance Hospital, 134, Shinchon Dong, Seodaemun Gu, Seoul, 120-749, Korea. E-mail: hdkimmd@yumc.yonsei.ac.kr

Abstract

Summary: Purpose: To evaluate the clinical efficacy and safety of the ketogenic diet (KD) for patients with intractable childhood epilepsy and mitochondrial respiratory chain (RC) complex defects.

Methods: A retrospective analysis evaluated outcomes in 14 children with intractable epilepsy and RC complex defects who were treated with the classic KD involving a 4:1 lipid to nonlipid ratio (% by weight), but without an initial fast and fluid restriction. Outcome measures included seizure frequency, electroencephalography (EEG) findings, the number of antiepileptic drugs, and adverse reactions.

Results: Of the 14 patients, 9 had Complex I defects, 1 had a Complex II defect, 3 had Complex IV defects, and 1 had combined Complex I and IV defects. Two patients with Complex IV defects showed clinical progress compatible with the Leigh disease. The epileptic diagnoses were as follows: 5 patients were diagnosed with infantile spasms, 4 with the Lennox–Gastaut syndrome, 1 with the Landau–Kleffner syndrome, 1 with nonspecific generalized seizure disorder, and 3 with partial seizure disorder. The study found that 7 patients became seizure-free after commencing the KD, three of whom successfully completed the diet without relapse. One patient with a greater than 90% seizure reduction, and 2 patients with seizure reductions between 50% and 90%, remained on the diet. Four patients, including two diagnosed with the Leigh disease, did not show any favorable responses to the diet or ceased the diet due to complications.

Conclusions: The KD was a safe and effective therapy for seizures in children with intractable epilepsy and RC complex defects.

Epileptic seizures can frequently occur as a presenting sign of mitochondrial encephalopathy (Kunz, 2002). Such seizures can be caused by defects in adenosine triphosphate synthesis and mitochondrial substrate oxidation, and also by direct partial inhibition of enzymes in the mitochondrial respiratory chain (RC) complex (Yamamoto and Tang, 1996; Urbanska et al., 1998; Kunz, 2002).

In patients with impaired aerobic glucose oxidation, ketones may provide alternative fuels for the central nervous system and other tissues (DeVivo et al., 1991). For example, the ketogenic diet (KD) can be used to treat epilepsy associated with pyruvate dehydrogenase (E1) deficiency (Wexler et al., 1997). However, conditions involving fatty acid oxidation defects, pyruvate carboxylase deficiency, and organic acidurias (e.g., 3-hydroxy-methylgluraryl Co-A lyase deficiency) would be adversely affected by the KD (DeVivo et al., 1977, Nordli and DeVivo, 2001). Although use of KD therapy has been avoided in patients with RC complex defects (Nordli and DeVivo, 2001), reports have suggested its potential for use in the treatment of patients with heteroplasmic mitochondrial DNA disorders (Santra et al., 2004). In addition, we have found that the KD was associated with favorable outcomes in several epilepsy patients with RC complex defects (Kang et al., 2005).

The present report investigated whether the KD was clinically effective and safe in patients with intractable childhood epilepsy associated with RC complex defects. We also discuss rationale for use of this diet in patients with RC complex defects.

METHODS

The study retrospectively reviewed the outcomes in 14 childhood epilepsy patients diagnosed with RC complex defects. The earliest case was diagnosed in July 1995, and all patients were monitored for at least 6 months from the commencement of the KD. RC complex defects were confirmed using mitochondrial RC complex enzyme analysis of biopsied muscle in 12 patients, skin fibroblasts in one patient (patient 1) and lymphocytes in one patient (patient 12). Prior to enzyme activity analysis, baseline studies were indicated in patients who were presented with repeated seizures of unknown cause and suspected clinical features of mitochondrial disease. Stepwise investigations included assays of serum and cerebrospinal fluid lactate/pyruvate, urine organic acids, and plasma amino acids; muscle histology and ultrastructure; genetic analysis; magnetic resonance imaging (MRI); and MR spectroscopy. Serum lactate and pyruvate levels were measured in arterial blood samples, with a lactate to pyruvate ratio >20 and a 3-hydroxybutyrate to acetoacetate ratio >2 being indicative of RC complex defects. Urine organic acid assays were performed using the first urine sample after non per os for at least 6 h. Screening for urinary lactate, citric acid cycle intermediates, and/or 3-methylglutaconic aciduria were used to diagnose RC complex defects. Ketolytic defects, which show similar metabolic patterns of mitochondrial disease, were screened using repeated analysis after protein restriction (<2 g/kg per day). Blood samples obtained after breakfast were used for plasma amino acid assays, and secondary hyperalaninemia derived from lactic acidemia was considered a marker of RC complex defects. Plasma acylcarnitine profiles were used to screen for mitochondrial fatty acid disorders. Biopsies were taken from the vastus lateralis muscle in all patients except two (patients 1 and 12); tissue samples were obtained in the operating room and were snap-frozen on dry ice. Routine morphological and histochemical staining included periodic acid-Schiff, modified Gomori trichrome, ATPase 9.4, nicotinamide adenine dinucleotide tetrazolium reductase, and succinate dehydrogenase stains. All samples underwent electron microscopy examination. MR spectroscopy was performed using a 3.0 Tesla Vision/Siemens system. The voxel size was 8 cm3 with the location determined from T1 images. Voxels were positioned adjacent to the lesion and in the contralateral hemisphere for focal lesions and the bilateral hemisphere for diffuse lesions. Mitochondrial impairment was diagnosed based on a definite lactate peak and a reduction in the N-acetyl-aspartate to creatinine ratio compared with the matched controls. Mitochondrial RC enzyme complex activities were evaluated as described previously (Rustin et al., 1994). Standard spectrophotometric assays were used to assess the activities of NADH–coenzyme Q (CoQ) reductase (complex I), succinate-CoQ reductase (complex II), succinate–cytochrome c reductase (complex II–III), cytochrome c reductase (complex III), cytochrome c oxidase (complex IV), and citrate synthase enzymes in isolated mitochondria from freshly prepared muscle tissue. Pyruvate dehydrogenase and pyruvate carboxylase activities were measured in cultured skin fibroblasts. In two patients who refused a muscle biopsy, RC enzyme complex activities were assessed in either skin fibroblasts (patient 1) or lymphocytes (patient 12). For patients other than patient 1, the diagnosis of an RC complex defect was made based on a reduction in RC complex enzyme activity of 80% or more compared to control tissue samples. The diagnosis of an RC complex defect in patient 1 was made based on a reduction in cultured skin fibroblast enzyme activity greater than two standard deviations. In addition, confirmation of a commonly distinguished mitochondrial disease was determined by means of identifying mtDNA deletions/duplications and common point mutations (A3243G and T3271C for mitochondrial encephalomyopathies, lactic acidosis, and stroke like episodes; A8344G and T8356C for myoclonus epilepsy with ragged-red fibers).

Prior to KD therapy, all 14 patients had experienced more than four seizures per month as reported by parents/caregivers, and in 12 patients the reported seizure types were confirmed using continuous electroencephalography (EEG) monitoring. The seizures were not controllable by an initial combination of three or more antiepileptic drugs. For at least 3 months prior to KD treatment, patients received a high dose of multivitamins and supplementary mitochondrial cocktail therapy including L-carnitine (100 mg/kg per day), coenzyme Q10 (5 mg/kg per day) and riboflavin (5 mg/kg per day). All patients were treated with the classic KD, which had a 4:1 lipid to nonlipid ratio (% by weight), with the only modification being the initial 36–48 h fast and fluid restriction was replaced by a gradual increase in calories. All 14 patients were hospitalized for about 7 days for diet adaptation and to monitor any acute complications. All patients were followed for at least 6 months after commencing the diet. Regular outpatient visits at monthly intervals were recommended to evaluate efficacy and adverse events. The primary efficacy outcome measures included seizure frequency, number of antiepileptic drugs, and EEG findings. Enzyme activity levels and specific types of complex defects were also monitored to determine any relationships with KD efficacy or adverse events. EEG findings included background or, where available, generalized or focal epileptiform discharges. Two authors independently assessed EEG findings and judged them to indicate either an improvement, a worsening, or no change in condition. Cognitive assessments were scheduled prior to the commencement of diet therapy and at 12 month intervals thereafter. Additional outcome measures included cognitive and behavioral changes based on global assessments by physicians and caregivers. All 14 patients used seizure diaries, and any omitted data were retrospectively collected using telephone interviews.

RESULTS

Patient characteristics

Of the 14 patients, 5 were diagnosed with infantile spasms, 4 with the Lennox–Gastaut syndrome, 1 with the Landau–Kleffner syndrome, 1 with nonspecific generalized seizure disorder, and 3 with a partial seizure disorder. Patients experienced their first seizure at a mean (± SD) age of 22 ± 21.6 months. Nonepileptic-associated neurological disabilities included mental retardation in all patients, hypotonia in five patients, cerebral palsy in seven patients, behavioral problems in four patients, and peripheral neuropathy in one patient. Nine patients had Complex I defects, one had a Complex II defect, three had Complex IV defects, and one had combined Complex I and IV defects. Of the three patients with Complex IV defects, two (patients 13 and 14) showed clinical progression compatible with the Leigh disease. The detailed clinical profiles and the results of stepwise investigations for diagnosing RC complex defects are summarized in Tables 1 and 2.

Table 1. Clinical profiles of 14 patients with intractable childhood epilepsy and mitochondrial respiratory chain defects
Sex (male:female)5:9
Age of seizure onset1 year 10 months
 (±SD, ±21.6 months)
Type of respiratory chain defects 
  I 9
  II 1
  III 0
  IV 3
  I and IV 1
Epileptic diagnoses 
  Infantile spasms 5
  Lennox–Gastaut syndrome 4
  Landau–Kleffner syndrome 1
  Partial seizure disorder 3
  Generalized seizure disorder 1
Associated neurologic symptoms 
  Mental retardation14
  Behavioral problems 4
  Cerebral palsy 7
  Hypotonia 5
  Peripheral neuropathy 1
Table 2. The results of diagnostic workups for assessing the mitochondrial respiratory chain complexes enzyme defects in 14 patients
No. of ptSex/ageSorts of RC complex defects (enzyme activity,%)aSerum lactate (ratio of lactate/pyruvate)CSF lactateUrine organic acidPlasma amino acidMRSMuscle histology
  1. Pt, patient; F, female; M, male; RC, respiratory chain; CSF, cerebrospinal fluid; MRS, magnetic resonance spectroscopy; ND, not done;a, percentages of control means of ratio of residual complex activity to citrate synthase activity;b, 19.8 mg/dl<;c, 20<;d, 22.0 mg/dl<;e, lactic aciduria accompanied with an elevation of Kreb cycle metabolites;f, definite lactate peak and reduction in N-Acetyl-Aspartate/ Creatinine ratio compared to matched controls,g, Secondary hyperalaninemia;h, >2% ragged-red fibers and/or subsarcolemmal accumulation of mitochondria, widespread electron microscopic abnormalities of mitochondria.

Pt 1F/5.6I (30.3)Increaseb (increasec)Increased+eNDND
Pt 2M/11.8I (9.7)Normal (normal)ND+f
Pt 3M/1.0I (11.7)Normal (normal)Normal+
Pt 4F/8.7I (12.9)Increase (increase)ND++g
Pt 5F/8.5I (11.1)Increase (increase)Normal+
Pt 6F/2.1I (9.3)Increase (increase)ND+++h
Pt 7M/6.3I (12.2)Increase (increase)ND++
Pt 8F/2.5I (10.8)Increase (increase)NDNDND+
Pt 9F/0.9I (10.7)Normal (normal)ND++
Pt 10M/8.8I (5.9), IV (2.9)Increase (increase)ND+ND+
Pt 11F/4.0II (10.8)Increase (increase)Normal+NDND
Pt 12F/11.8IV (8.5)Increase (increase)NDNDNDNDND
Pt 13M/1.7IV (8.6)Increase (increase)Increase+++
Pt 14F/5.1IV (10.1)Increase (increase)Normal+

KD efficacy and safety

The mean (± SD) age of patients at the commencement of the KD was 45 (± 36) months, and the mean (± SD) duration of the KD was 18 (± 15) months. Seven patients became seizure-free after commencing the KD. Three of those patients (patients 1, 7, and 12) successfully completed the diet regime without a relapse, whereas three maintained the diet for 13 (patient 6), 15 (patient 8), and 36 (patient 11) months. One patient (patient 10), who had combined Complex I and IV defects and who was seizure-free for over 4 years while on the KD, experienced an acute episode of status epilepticus combined with pneumonia. The status epilepticus was controlled, but the pneumonia progressed to acute respiratory distress syndrome, and the patient later died due to respiratory and multiorgan failure. Three patients who experienced 50–90% seizure reductions while on the diet maintained the diet for 6 (patient 3), 19 (patient 4), and 8 (patient 14) months. One patient (patient 9), who did not experience seizure reduction while on the KD, died due to aspiration pneumonia and respiratory failure after 3 months on the diet.

Symptomatic and recurrent hypoglycemia (below 40 mg% blood sugar) occurred in two patients (patients 2 and 3). One of those patients (patient 2) was taken off the diet after 3 months, while the persistent hypoglycemia in the other patient (patient 3) was successfully controlled with a low dose of prednisolone.

Of the two patients diagnosed with the Leigh disease caused by Complex IV defects, one (patient 13) was advised to cease the diet after 3 months due to unfavorable outcomes and persistent metabolic acidosis, while the second (patient 14) maintained the diet for 8 months despite intermittent, recurrent seizures.

All patients showed high blood ketosis (>3.0 mmol/L) within 2–3 days of diet commencement, and these levels were maintained throughout the KD treatment.

Most patients had enzyme activities ranging between 8% and 13% of those of matched controls. Patient 10 had 5.9% of Complex I activity and 2.9% of Complex IV activity compared to controls. Of the eight patients who showed a >90% reduction in seizure frequency while on the KD, five (patients 1, 5, 7, 8, and 11) had >10% enzyme activity, and three (patients 6, 10, and 12) had <10% activity compared to controls. The six patients who did not have favorable outcomes while on the KD had enzyme activities of approximately 10% that of controls.

Of the eight patients (patients 1–8) with Complex I defects, five (patients 1, 5, 6, 7, and 8) experienced a >90% reduction in seizure frequency while on the KD, and three (patients 2, 3, and 4) experienced a >50% reduction. Of the three patients who had Complex IV defects, two diagnosed with the Leigh disease (patients 13 and 14) did not experience favorable seizure outcomes, whereas the third (patient 12) showed a dramatic improvement and completed the KD. One patient with combined Complex I/IV defects (patient 9) and one with a Complex II defect (patient 10) remained seizure-free during KD treatment.

Routine scalp EEG was performed at least once in all but two patients (patients 9 and 13) 3 to 6 months after commencement of the diet. One patient (patient 1), who had an electrographic status epilepticus of slow-wave sleep, and two (patients 8 and 11) who had hypsarrhythmias, showed normal EEG features after becoming seizure-free. Of the four patients who showed EEG features characteristic of the Lennox–Gastaut syndrome, two (patients 5 and 12) who were seizure-free while on the KD showed normal background activities in an awake state, and their pre-KD generalized epileptiform discharges were no longer detectable while on the KD.

Four patients (patients 1, 7, 11, and 12) who responded favorably to the KD in terms of seizures were subsequently able to discontinue antiepileptic drug treatment, and another four patients (patients 5, 6, 8, and 10) were able to decrease the frequency and/or dose of drug administration.

Eight patients showed improvements in cognitive and behavioral functions following KD commencement. This was particularly obvious in patient 1, who initially could not be assessed using the Korean Version of the Wechsler Intelligence Scale for Children (K-WISC) prior to commencing the KD due to complete unresponsiveness, but showed an above average score in a K-WISC assessment 15 months after KD commencement. As expected, EEG findings, the number of medications, and neurobehavioral functions correlated with the seizure frequency.

Of the 14 patients, patient 10 with combined Complex I and IV defects was one who clearly showed a dramatic benefit from the KD but died during the diet. The patient history showed that other than slightly delayed speech development, this patient developed well until reaching 4 years and 4 months of age, at which time frequent partial seizures with secondary generalization began to occur, which were accompanied by an abrupt onset of progressive cognitive deterioration and an episodic ataxia. A brain stem auditory-evoked potential study, stimulating both ears individually with 70 dB and recording from both ear lobes, showed delayed latency of waves I, III, and V. A visual-evoked potential study, stimulating both eyes individually using a goggle stimulator and recording from the scalp, showed a delayed latency on both sides. Brain MRI showed diffuse brain atrophy that included the cerebellum. The patient had an increased level of serum lactate, an increased ratio of lactate/pyruvate, and lactic aciduria accompanied by an elevation in Kreb's cycle metabolites. Muscle histology suggested a mitochondrial myopathy, although there were no ragged-red fibers. KD was commenced 6 months after the first seizure and was maintained for 4 years, during which time the patient responded remarkably and remained seizure-free. Tapering of the KD was unsuccessful. After 4 years, the patient experienced an acute episode of status epilepticus and pneumonia. The pneumonia progressed into acute respiratory distress syndrome from which the patient did not recover.

Of the 14 patients, patient 13 with a Complex IV defect and the Leigh disease was one who clearly showed no benefit from the KD. This patient's levels of pyruvate dehydrogenase and other RC enzyme complexes were within normal ranges. The patient was born at gestational age 37 weeks to a healthy mother without any fetal or perinatal problems. However, a neurological examination showed decreased spontaneous movement from the newborn period and frequent focal seizures. Postural tremors began at 4 months. Repeated brain MRI showed a progressive signal change in the deep gray nuclei and brain stem. MR spectroscopy showed a definite lactate peak and a significant decrease in the N-acetyl-aspartate to creatinine ratio. The serum lactate and lactate/pyruvate ratio, cerebrospinal fluid lactate, and urine organic acid results were compatible with an RC complex defect. Although the seizures were not frequent, the KD was cautiously introduced to improve the clinical progress of the Leigh disease. However, seizure outcomes and hypotonia did not significantly improve. Rather, a persistent metabolic acidosis below 12 mmol/l bicarbonate persisted during the 3 months KD duration. The KD was ceased, and the patient was then treated with supplementary mitochondrial cocktail therapy only.

The detailed experiences of all 14 patients are summarized in Table 3.

Table 3. The experiences of the ketogenic diet in 14 patients with mitochondrial respiratory chain complexes enzyme defects
No. of ptEpilepsy classification/seizure type*Age/duration(year/month)A reduction of seizure frequency after the KD (%)State of the KDEEG features (result after the KD)No. of AEDs (result after the KD)Adverse events
3 mo6 mo12 mo24 mo<
  1. Pt, patient; LKS, Landau–Kleffner syndrome; LGS, Lennox–Gastaut syndrome; IS, infantile spasms; TC, tonic–clonic; C, clonic; PS, partial seizure; GS, generalized seizure; KD, ketogenic diet; EEG, electroencephalography; ESES, electrical status epilepticus in a slow wave sleep; I, improved; sl & dis, slow & disorganized background; GPFA, generalized paroxysmal fast activities; GSSW, generalized slow sharp and waves; NI, not improved; Hyp, hypsarrythmia; MI, moderately improved; SW, sharp waves; AEDs, antiepileptic drugs; Disc, discontinuation of drugs; Dec, decrease of the number and/or the dosage of drugs; MA, metabolic acidosis; Asp., aspiration; * main seizure type; at the age of beginning the ketogenic diet; duration of the ketogenic diet.

Pt 1LKS/(−)7.5/2.6100100100100CompletionESES (I)4 (Disc) 
Pt 2LGS/Atonic9.0/0.3 50 Stop d/t complicationsl && dis, GPFA, GSSW (NI)5Hypoglycemia
Pt 3IS/Spasms0.5/0.6 50 50 MaintainHyp (MI)4Hypoglycemia
Pt 4LGS/Atonic7.0/1.7 75 50 50 Maintainsl && dis, GPFA, GSSW (NI)4 
Pt 5LGS/TC8.0/0.6100100 Maintainsl && dis, GPFA, GSSW (I)3 (Dec) 
Pt 6IS/Spasms1.0/1.1100 90 90 MaintainHyp (I)3 (Dec) 
Pt 7PS/TC5.0/2.0 90100100100Completionsl && dis, focal SW (MI)3 (Disc) 
Pt 8IS/Spasms1.3/1.3100100100 MaintainHyp (I)3 (Dec) 
Pt 9IS/Spasms0.7/0.3  0 ExpireHyp (−)3Asp. pneumonia
Pt 10PS/C5.0/4.8 90100100100Expiresl && dis, focal SW (I)3 (Dec)Pneumonia
Pt 11IS/Spasms1.0/3.0100100100100MaintainHyp (I)5 (Disc) 
Pt 12LGS/Atonic3.5/2.5100 90100100Completionsl && dis, GPFA, GSSW (I)3 (Disc) 
Pt 13PS/C0.8/0.2 50 25 25 Stop d/t complicationsl && dis, focal SW (NI)3Persistent MA
Pt 14GS/TC4.8/0.8 50 25 MaintainGPSW (NI)3Asp. pneumonia

DISCUSSION

The present study detailed outcomes following KD therapy in 14 patients with intractable childhood epilepsy associated with RC complex defects. The study found that following the commencement of the KD, 10 patients responded favorably in terms of seizure frequency, and 8 patients were able to either cease or lower their use of antiepileptic medication. Furthermore, eight patients showed improvements in cognitive and behavioral functions. The effect of the KD on seizures was consistent with its reported overall efficacy (Freeman et al., 2006). The present study also found that the KD was not directly associated with any specific complications. In addition, the study found that the effect of the KD was not linked to enzyme activity levels or specific types of complex defects.

RC complex defects manifest considerable phenotypic diversity, and should be considered when unexplained progressive encephalopathies with seizures are observed (Schmiedel et al., 2003). Various biochemical assays, muscle histology results, and neuroradiological findings are useful in screening for RC complex defects. The diagnosis of RC complex defects can be confirmed by analysis of enzyme activity in isolated mitochondria from muscle tissue using standard spectrophotometric assays (Rustin et al., 1994). Measurements of individual enzyme activities in muscle are sensitive and useful (Fadic and Johns, 1996), especially in children suspected of having mitochondrial disease (McFarland et al., 2002).

The KD has been used worldwide since its resurgence in the mid 1990s (Kossoff and McGrogan, 2005), and is a valuable adjunct therapy for intractable childhood epilepsy. In particular, ketones may be an alternative energy source for the central nervous system and other tissues in patients with impaired aerobic glucose oxidation (McFarland et al., 2002). In theory, the KD can help to sustain the activities of the tricyclic acid cycle and RC, as well as minimize the dietary exacerbation of lactic acidosis (Munnich et al., 2001). The KD has been recognized as first-line therapy in patients with seizures associated with pyruvate dehydrogenase (E1) deficiency and glucose transporter 1 deficiency syndrome (DeVivo et al., 1991, Wexler et al., 1997).

There is evidence supporting the use of the KD for RC complex defects. Treatment with ketone bodies has been shown to cause “heteroplasmic shifting” not only between cells (i.e., intercellular selection) but also within cells (i.e., intracellular selection) (Santra et al., 2004). The observation that ketone bodies can distinguish between normal and respiration-compromised cells suggests that the KD may be useful in treating patients with heteroplasmic mtDNA disorders (Santra et al., 2004). Clinically, initial fasting and prolonged calorie restrictions can cause acute metabolic decompensation in patients with metabolic disorders related to energy production (Kang et al., 2004). Thus, it is important to monitor patients with RC complex defects when commencing the KD, especially during the initial fasting period. To reduce the adverse effects of fasting, some studies have omitted the initial fasting period and substituted it with a gradual increase in calories (Kim et al., 2004; Gergqvist et al., 2005). This revised protocol can provide better tolerability while maintaining efficacy, and reduce the potential risk of an acute metabolic breakdown.

Despite evidence indicating the benefits of the KD, this diet can metabolically stress patients with RC complex defects (Nordli and DeVivo, 2001; Freeman et al., 2006). For example, seizures can be precipitated by oxidative stress (Liang and Patel, 2004), and the KD has been shown to reduce the formation of reactive oxygen species through oxidation of the mitochondrial coenzyme Q couple and induction of several brain-specific mitochondrial uncoupling proteins (Sullivan et al., 2004). Activation of the latter can reduce the proton gradient across the inner mitochondrial membrane, which reduces ATP synthesis (Sullivan et al., 2004; Freeman et al., 2006). However, similar to the effects observed in cardiovascular disease models, the KD reduces the mitochondrial NAD while oxidizing the mitochondrial Q couple. Reduced NAD coupling increases the redox span between sites I and III, increasing the energy available for ATP synthesis and leading to a 28% increase in the efficiency of ATP hydrolysis (Veech, 2004).

In the present study, several patients experienced adverse events while on the KD, including symptomatic persistent hypoglycemia in two patients, persistent metabolic acidosis in one patient, aspiration pneumonia in one patient, and pneumonia followed by respiratory failure in two patients (patients 9 and 10), both of whom had pneumonia complicated with respiratory failure and died during KD treatment. However, the pneumonia in patient 9 was derived from an aspiration without any other abnormal laboratory findings, and patient 10 was able to maintain the diet for over 4 years despite episodes of pneumonia. Therefore, it appears these complications were not directly related to the RC complex defects. In the present study, most episodes of these and other KD complications were well controlled.

The Leigh disease is often derived from a partial generalized defect in Complex IV (Lombes et al., 1991). Patients with the cytochrome c oxidase-associated Leigh disease have a slower clinical course, and seizures are uncommon (DeVivo and DiMauro, 1999). In the present study, the seizures were not severe in patients diagnosed with the Leigh syndrome and Complex IV defects. The KD has been shown to induce dramatic improvements in children with the Leigh disease caused by pyruvate dehydrogenase deficiency (Wexler et al., 1997). However, in the current study, the KD was not associated with favorable outcomes in the two patients with the cytochrome c oxidase-associated Leigh syndrome.

In conclusion, the present findings indicate that the KD can control seizures in children with intractable epilepsy associated with RC complex defects. However, further studies are required to more precisely determine the efficacy and safety of the KD and the long-term prognosis of patients treated with this diet.

Ancillary