- Email: [email protected]
Early biochemical and EEG correlates of the ketogenic diet in children with atypical absence epilepsy
Early Biochemical and EEG correlates of the Ketogenic Diet in Children with Atypical Absence Ep epsy Diana L. Ross, MD*, Kenneth F. Swaiman, MD*, Fernando Torres, MD~, and Jessie Hansen, MS
Early changes in blood chemistry and the electroencephalogram were monitored during the first three hours after initiating the medium chain triglyceride (blCT) diet in nine children with intractable atypical absence seizures. Serum glucose, insulin, triglycerides, cholesterol, free fatty acids, ketone bodies concentrations, and venous pH were assayed before and at timed intervals after MCT oil was administered orally. The concentration of serum ketones rose progressively over three hours, 3hydroxybutyrate proportionately higher than acetoacetate. A statistically significant decrease in the group mean number of epileptiform discharges occurred following blCT therapy. Seizure frequency decreased by more than 50 percent in two-thirds of the children during the 10 week treatment period.
fective antiepileptic drugs, has relegated the use of the ketogenic diet to the rol~ of a final attempt at seizure management after failure of appropriate antiepileptic drugs. The development of the medium chain triglyceride (MCT) modification of the diet [2] greatly simplified diet administration. The mechanism of action of the ketogenic diet remains controversial. We exami~aed the mechanism by studying the time course of changes in blood chemistry and cerebral electrical activity in the first three hours after initiating the ketogenic diet and correlating these changes with clinical response over the subsequent 10 weeks of treatment. Methods Nine children were selected who had a history, often long, of uncontrolled absence seizures associated with synchronous slow spike
Ross DL, Swaiman KF, Torres F, Hansen J. Ketogenic diet in children with atypical absence epilepsy. Pediat Neurol 1985; 1:104-108.
and wave abnormalities on their electroencephalogram (EEG). Most of the children were developmentally delayed (Table 1). They were receiving a variety of antiepileptic drugs, which were maintained at the same dosage throughout the study period; plasma levels were monitored. Prior to initiation of the diet, for three consecutive days EEGs of
Introduction Atypical absence seizures with slow spike and wave complexes are often difficult to control. The ketogenic diet is one therapy of documented efficacy in some patients. When Wilder introduced the ketogenic diet in 1921 [1], the only available antiepileptic medications were bromides and phenobarbital. The diet was employed for a variety of seizure types during the 1920's and demonstrated some effectiveness against all seizure types. Particularly good results were obtained in preschool children with myoclonic, atonic, or tonicclonic seizures. The introduction of a number of el-
From the *Division of Pediatric Neurology~ the t Department of Neurology; and the ~ Department of Clinical Chemistry; University of Minnesota Medical School; Minneapolis, MN. 104
PEDIATRIC NEUROLOGY
Vol. 1 No. 2
about 10 minutes duration were obtained at the same hour. A single differential montage covering parasagittal and temporal areas, using the 10 - 20 system for electrode placement was utilized. The EEGs were scored by counting the number of bursts of generalized synchronous spike or spike-and-slow-wave activity per 20 second epoch. The frequency of bursts on the three baseline recordings were compared to the frequency of bursts on the day of initiating the ketogenic diet as described below. The MCT portion of the study was begun after an overnight fast. The third baseline EEG recording was obtained on the day of the MCT study. Blood was then obtained for baseline values of glucose, insulin, triglycerides, cholesterol, free fatty acids, venous pH, lactate, pyruvate, osmolality, acetoacetate, and 3-hydroxybutyrate~ The concentrations of ketones were assayed enzymaticalty [3]. These tests were repeated 30, 60, 90, 120, and t80 minutes after administration
Communications should be addressed to: Dr. Ross; 111 Wellington Place; Cincinnati, OH 45219.
Table 1. Patient description
Patient
Age Yrs
Sex
P.W.
9
M
M.B.
10
M
S.N.
4
J..B.
11
D.T.
J.L.
13
Duration of Seizures Yrs
Functional Level
Anticonvulsants Maintained During Study Period
Baseline Seizure Frequency By Type
Seizure Type
EEG
Learning disability
Trimethadione Carbamazepine Mepbobarbital
tonic
Controlled 3/minute
Synchronous 2-3 Hz spike waves
6
EMH
Phenytoin Etbosuximide Acetazolamide
partial motor
3-4/day
Bursts of synchronous polyspike and slow
M
3
Loss of milestones at seizure onset, nonverbal
Primidone Pbenytoin Diazepam
myoclonic spasms provoked by hyperventilation
2-4/day
Continuous, synchronous, spike wave 1.5-2.5 Hz
M
10
EMH
Clonazepam Ethosuximide
tonic-clonic (night); absence
2-3/week
Frequent bursts of high voltage rhythmic delta and polyspikes
M
l0
11
3-4/day
Nonverbal, nonambulatory
Ethosuximide Carbamazepine Primidone
clonic (day) (night); atonic; head drops
2/week 1-2 / night 6/day numerous
Frequent bursts of polyspike and slow
Learning disability
Phenytoin Phenobarbital
partial motor; absence
2-3/day 3-4/day
Sychronous bursts of 4-5 Hz activity and left central spikes
11
M
P.H.
7
F
61/2
TMH, nonverbal
Ethosuximide Carbamazepine
atonic; absence
12/day 3-4/day
Synchronous bursts of polyspike and slow
M.C.
4-6/12
M
2 1/2
Loss of milestones at seizure onset
Phenytoin CIonazepam
tonic-clonic; myoclonic
controlled 4- 5 / day
Synchronous bursts of polyspike and slow
S.S.
3-8/12
F
3
Nonverbal, ataxic
Valproate Carbamazepine Primidone
myoclonic
15-20 / day
Synchronous bursts of polyspike and slow
of MCT oil. Each patient drank a bolus of 30 grams/m 2 body surface of MCT oil (Mead Johnson Co.) mixed in diet soda. The children tolerated this mixture well, although one child had a single bout of diarrhea soon after MCT administration. Ten minute EEG recordings were made at 15, 30, 45, 60, 90, 120, 150, and 180 minute intervals following the administration of MCT oil. Epochs of 20 seconds were inspected for bisynchronous bursts and compared to baseline. All but one patient had epileptiform abnormalities in more than 50 percent of the 20 second epochs in the 10 minute recording periods at baseline; therefore, the results were quantified by counting epochs which were devoid of epileptiform activity. At the conclusion of the three-hour study period, the children received a ketogenic diet which provided the recommended total calories for age. The daily protein intake was normally 1.0-1.5 grams/kg/day). The remaining calories were distributed in a 3:1 ratio of fat:carbohydrate. Sixty percent of the total calories were provided as MCT oil, which allowed sufficient additional lipid allotment to provide the necessary essential free fatty acids. This diet was supplemented with sugar-free multivitamins and calcium. The children remained in the hospital until they tolerated the diet and the family was confident that the diet could be continued at home. The children were seen every two weeks during the 10 week monitoring period. Daily seizure records were kept by the family on standardized forms, recording the frequency of each seizure type.
Urine ketones were assessed and recorded at least twice a day. Diet management and seizure recording were discussed at each visit. Following the 10 week study period, repeat antiepileptic drug blood levels were obtained and an EEG was performed to document longterm electrical changes.
Results
Following the
initial
MCT
ingestion,
each child
m a n i f e s t e d a progressive rise in s e r u m k e t o n e bodies o v e r t h e t h r e e h o u r study p e r i o d , w i t h a p r o p o r t i o n a t e l y greater
increase
in
3-hydroxybutyrate
than
in
acetoacetate (Table 2). In s o m e o f the subjects baseline fasting s e r u m k e t o n e c o n c e n t r a t i o n s were h i g h e r t h a n o u r l a b o r a t o r y ' s a d u l t n o r m a l values (0.036
-- 0.020
m M for 3 - h y d r o x y b u t y r a t e a n d 0.026 +- 0.011 m M for acetoacetate),
probably
ketogenesis in c h i l d r e n . changes
in
blood
reflecting
greater
fasting
N o consistent or significant
glucose,
insulin,
triglycerides,
cholesterol, free fatty acids, lactate, p y r u v a t e , venous p H , or o s m o l a l i t y occurred. C o m p a r i s o n o f baseline E E G paroxysmal spikes ot s p i k e - a n d - w a v e c o m p l e x e s to those p r e s e n t d u r i n g M C ' f Ross et al: Laboratory Studies of Ketogenic Diet Therapy
105
Table 2. Hyperketonuria following MCT diet Intervals (min) Following MCT Ingestion
Patients 0 P.W. M.B. S.N. J.B. D.T. J.L. P.H. M.C. S.S.
30
60
90 1.33
120
180
3-OH*
0.32
0.45
AA~
0.22
0.30
1.96
3-OH AA
0.05 0.04
3-OH
0.78
2.01
2.23
2.55
AA
0.28
0.44
0.49
0.57
3-OH AA
0.03 0.04
0.24 0.08
0.08
0. t0
3-OH
0.01
0.02
0.08
0.18
AA
0.08
0.04
0.06
0.09
0.49 0.75
0.57 0.14
I). 69 0.78
0.08
0.10
0.20
0.53
3-OH
0.02
0.33
0.46
0.55
0.17
AA
0.02
0.08
0.10
0.19
0.06
3-OH
0.05
0.57
0.75
AA
0.04
0.14
0.08
3-OH
0.63
1.20
1.46
1.83
2.14
AA
0.10
0.16
0.16
0.38
0.34
3-OH
0.02
0.48
0.66
1.02
1.17
1.28
AA
0.03
0.12
0.18
0.20
0.28
0.32
078 0. ~0
*3-OHbutyrate (mmol/L) tAcetoacetate (mmol/L)
instillation is depicted in Table 3. There was only one baseline EEG for patient D.T. because the patient could not be kept awake during the other two recording sessions; therefore, the tracings could not be compared with the waking records. Patient M.B. had several clinical seizures during one of the recordings; therefore, the record was eliminated. In only two patients during administration of MCT oil, were all eight of the EEG samples included. The other patients had some EEGs which were either sleep or ictal records. Table 3 reflects the number of epochs without bursts. For records that were either shorter or longer than the intended 10 minutes, the figures were adjusted accordingly. The mean for the entire group of patients was found to be significantly different between treatment and baseline by Student's t test analysis. However, EEGs obtained at the end of the 10-week study period were unchanged from baseline despite significant improvement in seizure control in two-thirds of the subjects. All of the families understood and managed the diet well. All of the children accepted the diet initially and all became strongly ketonuric. The major difficulty was with gastrointestinal symptoms (abdominal pain, diarrhea, and infrequent vomiting) which usually responded favorably to avoidance of large boluses of
106
PEDIATRIC NEUROLOGY
Vol. 1 No. 2
MCT oil during a meal; the problem became less prominent over time. Two of the patients developed transient lethargy Blood gases and glucose were monitored during the lethargic period and remained within the normal range. The clinical results of the ketogenic diet treatment over the lO-week observation period are summarized in Table 4. Two-thirds of the patients had a favorable response, showing greater than 50 percent decrease in seizure frequency, improved alertness and behavior. and satisfactory tolerance of the diet. Discussion The mechanism of action of the ketogenic diet is not understood despite considerable research. Wilder's original postulate was that ketosis was the most important factor [1]. This theory received support from the work of Keith [4], which demonstrated that mfusion of acetoacetate decreased seizures in a rabbit model. It is known that ketone bodies can be extracted from blood by the brain and used as an energy source as well as a source of carbon residues for biosynthesis of phospholipids, cholesterol and amino acids including y-aminobutyric acid (GABA) [5]. Huttenlocher's study of metabolic effects of the MCT and standard ketogenic
Table 3. Epochs w i t h o u t spikes or spike waves
Treatment: Mean2 ± SE
Baseline: Mean, _+ SE
Difference
Mean No. Epochs w i t h o u t Spikes
Mean No. Epochs w i t h o u t Spikes
Patient
N
(10 m i n recordings)
N
tO rain recordings)
Mean~-Mean,
P.W.
3
7.9 m 1.9
6
15.9 + 1.4
8.0
M.B
2
23.5 ± 1.5
5
26.0 -+ 0.5
2.5
SN.
3
o
5
0
1)
J.B.
3
0
5
1.2 _+ 0.8
1.2
D.T.
1
0
3
0
0
J.L.
3
15.0 ± 3.5
7
23.3 _+ 1.2
8.3
P,H.
3
0
7
o
0
M.C.
3
14.8 ± 6.0
8
20.1 _+ 2.0
5. ~,
S.S.
3
0
8
0
0 Mean Differem e
2.8*
*p<0.05 N = n u m b e r of EEG recordings utilized
diet demonstrated a gradual rise in serum ketones over one month which correlated closely with antiepileptic effect. The degree of ketonemia on the MCT diet was comparable to a 3:1 high fat diet. The ketonemia could be eliminated by glucose infusion within one hour, with subsequent precipitation of a clinical seizure in a patient well controlled on the diet [6]. Guisard and Debry studied the acute metabolic response to an intravenous infusion of MCT in adults, which included ketosis, with greater rise in acetoacetate than /3hydroxybutyrate, a rise in lactate/pyruvate ratio, and bimodal peaks of insulin in response to medium chain fatty acid administration and later to ketone bodies with accompanying decrements in blood glucose concentrations [7]. Our results, in children receiving an oral bolus of MCT in a higher dosage than Guisard and Debry's subjects, differed considerably. Our patients demonstrated only progressive rise of ketones, primarily the oxidized form, and no change in insulin, glucose, lactate, or pyruvate. This difference may relate to the route of administration or to differences in oxidative metabolism between adults and children. In our study, all children manifested a prompt and significant rise in blood /3-hydroxybutyrate and acetoacetate concentrations regardless of their antiepileptic response to the diet. We found no meaningful correlation between the magnitude of rise in concentration of ketone bodies and the effectiveness of the diet. In contrast, another theory which emphasizes acidosis
was put forth by Lennox, based initially on circumstantial evidence, and later supported by experimental evidence [8]. Intracellular pH in the brains of mice receiving medium chain triglycerides was studied [9]. Although the intracellular pH was decreased by acute loading with MCT, chronic administration did not sustain the intracellular acidosis. The authors suggested that the antiepileptic effect of MCT was due to increased activity of the membrane proton pump mechanism necessary to maintain a stable intracellular pH. Our results corroborate Huttenlocher's finding that children on the MCT diet do not demonstrate acidosis [2,6}. Attention has also been directed to changes in lipid profiles. Dekaban studied plasma-partitioned lipids in children beginning on the standard ketogenic diet [10]. He found elevated ketones by the third day, but delay in the rise of lipid content until 10 - 20 days. The degree of rise appeared to correlate with effectiveness of the diet and paralleled the time course of seizure control. Huttenlocher studied lipid profiles in children on the MCT diet [6]. He found that there was no rise in cholesterol and only slight elevation of total fatty acids in contrast to the marked elevations of these lipid constituents engendered by the standard ketogenic diet. This finding does not support the theory of a major contribution of lipid profile to the therapeutic mechanism of the MCT diet. We found no abnormalities of the lipid profile in the acute loading
Ross et al: Laboratory Studies of K e t o g e n h Diet Therapy
107
Table 4. Results of ketogenic diet treatment
Patient
Seizure frequency during the study period
Complications
P.W.
Absence decreased by 50% by 1 week
Became drowsy, resisted diet which was discontinued
Offdiet befbre end of study period
M.B.
Seizure frequency decreased to less than 50%
Diet discontinued due to intolerance of sugar-free ethosuximide
Offdiet before end of study period
S.N.
Seizure frequency and duration decreased 50% by 1 month
None
Maintained on diet with sustained improvement m seizures
J .U.
Calmer behavior by 2 weeks; seizure-free by 10 weeks
Transient lethargy. Glucose, pH and ketones were normal
Maintained on diet. Clonazepam disccmtinued
D.T.
More alert and nocturnal seizures 2/week by 2 weeks. Seizure frequency decreased by 50% by 6 weeks. Clonic seizures fully controlled by 10 weeks
None
Maintained on diet. Valproate added with no further benefit
J.L.
50% decrease in partial seizure frequency. Absence fully controlled
Loose stools
Maintained on diet. Vatproate added with further decrease in seizures to 1-2/month
P.H.
Seizure frequency decreased by 50 % ; improved alertness and hyperactivity
Persistent vomiting
Discontinued diet at end of study. Valproate added with further decrease in seizure frequency
M.C.
Seizure-free on diet within 2 weeks
None
Maintained on diet
S.S.
No change in seizure frequency
Single episode of pallor, diaphoresis and loss of consciousness on day 1 of diet. Glucose and electrolytes were normal
Offdiet at 7 weeks. Valproate increased
experiments. We did not repeat lipid analyses on the two children who became transiently lethargic. In retrospect, this study may have been of interest, because some short chain fatty acids are known to have central nervous system depressant activity [11]. At present, the strongest case can be made for ketosis as the dominant therapeutic factor in the MCT diet. The rise in ketones was the only prominent early biochemical change. The significant change in group mean frequency of epileptiform bursts following MCT administration suggests that this method of quantifying acute EEG effects may be of value for predicting the efficacy of the ketogenic diet. The favorable response to MCT therapy in two-thirds of our patients should encourage more extensive use of MCT treatment. This research was supported in part by National Institutes of Health contract No. 1-NS-5-2327.
References
[1] Wilder RM. The effect of ketonuria on the course of epilepsy. Mayo Clin Bull 1921 ;2:307.
108
Course following the study period
PEDIATRIC NEUROLOGY
Vol. 1 No. 2
[2] Huttenlocher, PR, Wilbourn, AJ, Signore,JM Medium chain triglycerides as a therapy for intractable childhood epilepsy. Neurology 1971; 21 : 1097-1103. [3] Hanson JL, Freier EF. Direct assays of lactate, pyruvate, 3hydroxybutyrate and acetoacetate with a centrifugal analyzer. Clin Chem 1978;24:475-9. [4] Keith HM. Convulsive disorders in children: with reference to treatment with ketogenic diet. Boston: Little, Brown and Co., 1963. [5] Sokolnff L. Metabolism of ketone bodies by the brain. Ann Rev Med 1973;24:271-80. [6] Huttenlocher PR. Ketonemia and seizures: Metabolic and anticonvulsant effects of two ketogenic diets in childhood epilepsy, Pediatr Res 1976; 10: 536-40. [7] Guisard D, Debry G. Metabolic effects of a medium chain triglyceride emulsion injected intravenously in man. Horm Metab Res 1972;4:509. [8] Lennox WB. Ketogenic diet in the treatment of epilepsy. N EnglJ Med 1928;199:74. [9] Davidian NM, Butler Poole DT. The effect of ketosis induced by medium chain triglycerides on intracellular pH of mouse brain. Epilepsia 1978; 19: 369-78. [10] Dekaban AS. Plasma lipids in epileptic children treated with the high fat diet. Arch Neurol 1966;15:177-84. [11] Trauner D, Sweetman L, Holm J, Kulovich S, Nyhan WL. Biochemical correlates of illness and recovery in Reye's syndrome. Ann Neurol 1977; 2:238-41.
Early changes in blood chemistry and the electroencephalogram were monitored during the first three hours after initiating the medium chain triglyceride (blCT) diet in nine children with intractable atypical absence seizures. Serum glucose, insulin, triglycerides, cholesterol, free fatty acids, ketone bodies concentrations, and venous pH were assayed before and at timed intervals after MCT oil was administered orally. The concentration of serum ketones rose progressively over three hours, 3hydroxybutyrate proportionately higher than acetoacetate. A statistically significant decrease in the group mean number of epileptiform discharges occurred following blCT therapy. Seizure frequency decreased by more than 50 percent in two-thirds of the children during the 10 week treatment period.
fective antiepileptic drugs, has relegated the use of the ketogenic diet to the rol~ of a final attempt at seizure management after failure of appropriate antiepileptic drugs. The development of the medium chain triglyceride (MCT) modification of the diet [2] greatly simplified diet administration. The mechanism of action of the ketogenic diet remains controversial. We exami~aed the mechanism by studying the time course of changes in blood chemistry and cerebral electrical activity in the first three hours after initiating the ketogenic diet and correlating these changes with clinical response over the subsequent 10 weeks of treatment. Methods Nine children were selected who had a history, often long, of uncontrolled absence seizures associated with synchronous slow spike
Ross DL, Swaiman KF, Torres F, Hansen J. Ketogenic diet in children with atypical absence epilepsy. Pediat Neurol 1985; 1:104-108.
and wave abnormalities on their electroencephalogram (EEG). Most of the children were developmentally delayed (Table 1). They were receiving a variety of antiepileptic drugs, which were maintained at the same dosage throughout the study period; plasma levels were monitored. Prior to initiation of the diet, for three consecutive days EEGs of
Introduction Atypical absence seizures with slow spike and wave complexes are often difficult to control. The ketogenic diet is one therapy of documented efficacy in some patients. When Wilder introduced the ketogenic diet in 1921 [1], the only available antiepileptic medications were bromides and phenobarbital. The diet was employed for a variety of seizure types during the 1920's and demonstrated some effectiveness against all seizure types. Particularly good results were obtained in preschool children with myoclonic, atonic, or tonicclonic seizures. The introduction of a number of el-
From the *Division of Pediatric Neurology~ the t Department of Neurology; and the ~ Department of Clinical Chemistry; University of Minnesota Medical School; Minneapolis, MN. 104
PEDIATRIC NEUROLOGY
Vol. 1 No. 2
about 10 minutes duration were obtained at the same hour. A single differential montage covering parasagittal and temporal areas, using the 10 - 20 system for electrode placement was utilized. The EEGs were scored by counting the number of bursts of generalized synchronous spike or spike-and-slow-wave activity per 20 second epoch. The frequency of bursts on the three baseline recordings were compared to the frequency of bursts on the day of initiating the ketogenic diet as described below. The MCT portion of the study was begun after an overnight fast. The third baseline EEG recording was obtained on the day of the MCT study. Blood was then obtained for baseline values of glucose, insulin, triglycerides, cholesterol, free fatty acids, venous pH, lactate, pyruvate, osmolality, acetoacetate, and 3-hydroxybutyrate~ The concentrations of ketones were assayed enzymaticalty [3]. These tests were repeated 30, 60, 90, 120, and t80 minutes after administration
Communications should be addressed to: Dr. Ross; 111 Wellington Place; Cincinnati, OH 45219.
Table 1. Patient description
Patient
Age Yrs
Sex
P.W.
9
M
M.B.
10
M
S.N.
4
J..B.
11
D.T.
J.L.
13
Duration of Seizures Yrs
Functional Level
Anticonvulsants Maintained During Study Period
Baseline Seizure Frequency By Type
Seizure Type
EEG
Learning disability
Trimethadione Carbamazepine Mepbobarbital
tonic
Controlled 3/minute
Synchronous 2-3 Hz spike waves
6
EMH
Phenytoin Etbosuximide Acetazolamide
partial motor
3-4/day
Bursts of synchronous polyspike and slow
M
3
Loss of milestones at seizure onset, nonverbal
Primidone Pbenytoin Diazepam
myoclonic spasms provoked by hyperventilation
2-4/day
Continuous, synchronous, spike wave 1.5-2.5 Hz
M
10
EMH
Clonazepam Ethosuximide
tonic-clonic (night); absence
2-3/week
Frequent bursts of high voltage rhythmic delta and polyspikes
M
l0
11
3-4/day
Nonverbal, nonambulatory
Ethosuximide Carbamazepine Primidone
clonic (day) (night); atonic; head drops
2/week 1-2 / night 6/day numerous
Frequent bursts of polyspike and slow
Learning disability
Phenytoin Phenobarbital
partial motor; absence
2-3/day 3-4/day
Sychronous bursts of 4-5 Hz activity and left central spikes
11
M
P.H.
7
F
61/2
TMH, nonverbal
Ethosuximide Carbamazepine
atonic; absence
12/day 3-4/day
Synchronous bursts of polyspike and slow
M.C.
4-6/12
M
2 1/2
Loss of milestones at seizure onset
Phenytoin CIonazepam
tonic-clonic; myoclonic
controlled 4- 5 / day
Synchronous bursts of polyspike and slow
S.S.
3-8/12
F
3
Nonverbal, ataxic
Valproate Carbamazepine Primidone
myoclonic
15-20 / day
Synchronous bursts of polyspike and slow
of MCT oil. Each patient drank a bolus of 30 grams/m 2 body surface of MCT oil (Mead Johnson Co.) mixed in diet soda. The children tolerated this mixture well, although one child had a single bout of diarrhea soon after MCT administration. Ten minute EEG recordings were made at 15, 30, 45, 60, 90, 120, 150, and 180 minute intervals following the administration of MCT oil. Epochs of 20 seconds were inspected for bisynchronous bursts and compared to baseline. All but one patient had epileptiform abnormalities in more than 50 percent of the 20 second epochs in the 10 minute recording periods at baseline; therefore, the results were quantified by counting epochs which were devoid of epileptiform activity. At the conclusion of the three-hour study period, the children received a ketogenic diet which provided the recommended total calories for age. The daily protein intake was normally 1.0-1.5 grams/kg/day). The remaining calories were distributed in a 3:1 ratio of fat:carbohydrate. Sixty percent of the total calories were provided as MCT oil, which allowed sufficient additional lipid allotment to provide the necessary essential free fatty acids. This diet was supplemented with sugar-free multivitamins and calcium. The children remained in the hospital until they tolerated the diet and the family was confident that the diet could be continued at home. The children were seen every two weeks during the 10 week monitoring period. Daily seizure records were kept by the family on standardized forms, recording the frequency of each seizure type.
Urine ketones were assessed and recorded at least twice a day. Diet management and seizure recording were discussed at each visit. Following the 10 week study period, repeat antiepileptic drug blood levels were obtained and an EEG was performed to document longterm electrical changes.
Results
Following the
initial
MCT
ingestion,
each child
m a n i f e s t e d a progressive rise in s e r u m k e t o n e bodies o v e r t h e t h r e e h o u r study p e r i o d , w i t h a p r o p o r t i o n a t e l y greater
increase
in
3-hydroxybutyrate
than
in
acetoacetate (Table 2). In s o m e o f the subjects baseline fasting s e r u m k e t o n e c o n c e n t r a t i o n s were h i g h e r t h a n o u r l a b o r a t o r y ' s a d u l t n o r m a l values (0.036
-- 0.020
m M for 3 - h y d r o x y b u t y r a t e a n d 0.026 +- 0.011 m M for acetoacetate),
probably
ketogenesis in c h i l d r e n . changes
in
blood
reflecting
greater
fasting
N o consistent or significant
glucose,
insulin,
triglycerides,
cholesterol, free fatty acids, lactate, p y r u v a t e , venous p H , or o s m o l a l i t y occurred. C o m p a r i s o n o f baseline E E G paroxysmal spikes ot s p i k e - a n d - w a v e c o m p l e x e s to those p r e s e n t d u r i n g M C ' f Ross et al: Laboratory Studies of Ketogenic Diet Therapy
105
Table 2. Hyperketonuria following MCT diet Intervals (min) Following MCT Ingestion
Patients 0 P.W. M.B. S.N. J.B. D.T. J.L. P.H. M.C. S.S.
30
60
90 1.33
120
180
3-OH*
0.32
0.45
AA~
0.22
0.30
1.96
3-OH AA
0.05 0.04
3-OH
0.78
2.01
2.23
2.55
AA
0.28
0.44
0.49
0.57
3-OH AA
0.03 0.04
0.24 0.08
0.08
0. t0
3-OH
0.01
0.02
0.08
0.18
AA
0.08
0.04
0.06
0.09
0.49 0.75
0.57 0.14
I). 69 0.78
0.08
0.10
0.20
0.53
3-OH
0.02
0.33
0.46
0.55
0.17
AA
0.02
0.08
0.10
0.19
0.06
3-OH
0.05
0.57
0.75
AA
0.04
0.14
0.08
3-OH
0.63
1.20
1.46
1.83
2.14
AA
0.10
0.16
0.16
0.38
0.34
3-OH
0.02
0.48
0.66
1.02
1.17
1.28
AA
0.03
0.12
0.18
0.20
0.28
0.32
078 0. ~0
*3-OHbutyrate (mmol/L) tAcetoacetate (mmol/L)
instillation is depicted in Table 3. There was only one baseline EEG for patient D.T. because the patient could not be kept awake during the other two recording sessions; therefore, the tracings could not be compared with the waking records. Patient M.B. had several clinical seizures during one of the recordings; therefore, the record was eliminated. In only two patients during administration of MCT oil, were all eight of the EEG samples included. The other patients had some EEGs which were either sleep or ictal records. Table 3 reflects the number of epochs without bursts. For records that were either shorter or longer than the intended 10 minutes, the figures were adjusted accordingly. The mean for the entire group of patients was found to be significantly different between treatment and baseline by Student's t test analysis. However, EEGs obtained at the end of the 10-week study period were unchanged from baseline despite significant improvement in seizure control in two-thirds of the subjects. All of the families understood and managed the diet well. All of the children accepted the diet initially and all became strongly ketonuric. The major difficulty was with gastrointestinal symptoms (abdominal pain, diarrhea, and infrequent vomiting) which usually responded favorably to avoidance of large boluses of
106
PEDIATRIC NEUROLOGY
Vol. 1 No. 2
MCT oil during a meal; the problem became less prominent over time. Two of the patients developed transient lethargy Blood gases and glucose were monitored during the lethargic period and remained within the normal range. The clinical results of the ketogenic diet treatment over the lO-week observation period are summarized in Table 4. Two-thirds of the patients had a favorable response, showing greater than 50 percent decrease in seizure frequency, improved alertness and behavior. and satisfactory tolerance of the diet. Discussion The mechanism of action of the ketogenic diet is not understood despite considerable research. Wilder's original postulate was that ketosis was the most important factor [1]. This theory received support from the work of Keith [4], which demonstrated that mfusion of acetoacetate decreased seizures in a rabbit model. It is known that ketone bodies can be extracted from blood by the brain and used as an energy source as well as a source of carbon residues for biosynthesis of phospholipids, cholesterol and amino acids including y-aminobutyric acid (GABA) [5]. Huttenlocher's study of metabolic effects of the MCT and standard ketogenic
Table 3. Epochs w i t h o u t spikes or spike waves
Treatment: Mean2 ± SE
Baseline: Mean, _+ SE
Difference
Mean No. Epochs w i t h o u t Spikes
Mean No. Epochs w i t h o u t Spikes
Patient
N
(10 m i n recordings)
N
tO rain recordings)
Mean~-Mean,
P.W.
3
7.9 m 1.9
6
15.9 + 1.4
8.0
M.B
2
23.5 ± 1.5
5
26.0 -+ 0.5
2.5
SN.
3
o
5
0
1)
J.B.
3
0
5
1.2 _+ 0.8
1.2
D.T.
1
0
3
0
0
J.L.
3
15.0 ± 3.5
7
23.3 _+ 1.2
8.3
P,H.
3
0
7
o
0
M.C.
3
14.8 ± 6.0
8
20.1 _+ 2.0
5. ~,
S.S.
3
0
8
0
0 Mean Differem e
2.8*
*p<0.05 N = n u m b e r of EEG recordings utilized
diet demonstrated a gradual rise in serum ketones over one month which correlated closely with antiepileptic effect. The degree of ketonemia on the MCT diet was comparable to a 3:1 high fat diet. The ketonemia could be eliminated by glucose infusion within one hour, with subsequent precipitation of a clinical seizure in a patient well controlled on the diet [6]. Guisard and Debry studied the acute metabolic response to an intravenous infusion of MCT in adults, which included ketosis, with greater rise in acetoacetate than /3hydroxybutyrate, a rise in lactate/pyruvate ratio, and bimodal peaks of insulin in response to medium chain fatty acid administration and later to ketone bodies with accompanying decrements in blood glucose concentrations [7]. Our results, in children receiving an oral bolus of MCT in a higher dosage than Guisard and Debry's subjects, differed considerably. Our patients demonstrated only progressive rise of ketones, primarily the oxidized form, and no change in insulin, glucose, lactate, or pyruvate. This difference may relate to the route of administration or to differences in oxidative metabolism between adults and children. In our study, all children manifested a prompt and significant rise in blood /3-hydroxybutyrate and acetoacetate concentrations regardless of their antiepileptic response to the diet. We found no meaningful correlation between the magnitude of rise in concentration of ketone bodies and the effectiveness of the diet. In contrast, another theory which emphasizes acidosis
was put forth by Lennox, based initially on circumstantial evidence, and later supported by experimental evidence [8]. Intracellular pH in the brains of mice receiving medium chain triglycerides was studied [9]. Although the intracellular pH was decreased by acute loading with MCT, chronic administration did not sustain the intracellular acidosis. The authors suggested that the antiepileptic effect of MCT was due to increased activity of the membrane proton pump mechanism necessary to maintain a stable intracellular pH. Our results corroborate Huttenlocher's finding that children on the MCT diet do not demonstrate acidosis [2,6}. Attention has also been directed to changes in lipid profiles. Dekaban studied plasma-partitioned lipids in children beginning on the standard ketogenic diet [10]. He found elevated ketones by the third day, but delay in the rise of lipid content until 10 - 20 days. The degree of rise appeared to correlate with effectiveness of the diet and paralleled the time course of seizure control. Huttenlocher studied lipid profiles in children on the MCT diet [6]. He found that there was no rise in cholesterol and only slight elevation of total fatty acids in contrast to the marked elevations of these lipid constituents engendered by the standard ketogenic diet. This finding does not support the theory of a major contribution of lipid profile to the therapeutic mechanism of the MCT diet. We found no abnormalities of the lipid profile in the acute loading
Ross et al: Laboratory Studies of K e t o g e n h Diet Therapy
107
Table 4. Results of ketogenic diet treatment
Patient
Seizure frequency during the study period
Complications
P.W.
Absence decreased by 50% by 1 week
Became drowsy, resisted diet which was discontinued
Offdiet befbre end of study period
M.B.
Seizure frequency decreased to less than 50%
Diet discontinued due to intolerance of sugar-free ethosuximide
Offdiet before end of study period
S.N.
Seizure frequency and duration decreased 50% by 1 month
None
Maintained on diet with sustained improvement m seizures
J .U.
Calmer behavior by 2 weeks; seizure-free by 10 weeks
Transient lethargy. Glucose, pH and ketones were normal
Maintained on diet. Clonazepam disccmtinued
D.T.
More alert and nocturnal seizures 2/week by 2 weeks. Seizure frequency decreased by 50% by 6 weeks. Clonic seizures fully controlled by 10 weeks
None
Maintained on diet. Valproate added with no further benefit
J.L.
50% decrease in partial seizure frequency. Absence fully controlled
Loose stools
Maintained on diet. Vatproate added with further decrease in seizures to 1-2/month
P.H.
Seizure frequency decreased by 50 % ; improved alertness and hyperactivity
Persistent vomiting
Discontinued diet at end of study. Valproate added with further decrease in seizure frequency
M.C.
Seizure-free on diet within 2 weeks
None
Maintained on diet
S.S.
No change in seizure frequency
Single episode of pallor, diaphoresis and loss of consciousness on day 1 of diet. Glucose and electrolytes were normal
Offdiet at 7 weeks. Valproate increased
experiments. We did not repeat lipid analyses on the two children who became transiently lethargic. In retrospect, this study may have been of interest, because some short chain fatty acids are known to have central nervous system depressant activity [11]. At present, the strongest case can be made for ketosis as the dominant therapeutic factor in the MCT diet. The rise in ketones was the only prominent early biochemical change. The significant change in group mean frequency of epileptiform bursts following MCT administration suggests that this method of quantifying acute EEG effects may be of value for predicting the efficacy of the ketogenic diet. The favorable response to MCT therapy in two-thirds of our patients should encourage more extensive use of MCT treatment. This research was supported in part by National Institutes of Health contract No. 1-NS-5-2327.
References
[1] Wilder RM. The effect of ketonuria on the course of epilepsy. Mayo Clin Bull 1921 ;2:307.
108
Course following the study period
PEDIATRIC NEUROLOGY
Vol. 1 No. 2
[2] Huttenlocher, PR, Wilbourn, AJ, Signore,JM Medium chain triglycerides as a therapy for intractable childhood epilepsy. Neurology 1971; 21 : 1097-1103. [3] Hanson JL, Freier EF. Direct assays of lactate, pyruvate, 3hydroxybutyrate and acetoacetate with a centrifugal analyzer. Clin Chem 1978;24:475-9. [4] Keith HM. Convulsive disorders in children: with reference to treatment with ketogenic diet. Boston: Little, Brown and Co., 1963. [5] Sokolnff L. Metabolism of ketone bodies by the brain. Ann Rev Med 1973;24:271-80. [6] Huttenlocher PR. Ketonemia and seizures: Metabolic and anticonvulsant effects of two ketogenic diets in childhood epilepsy, Pediatr Res 1976; 10: 536-40. [7] Guisard D, Debry G. Metabolic effects of a medium chain triglyceride emulsion injected intravenously in man. Horm Metab Res 1972;4:509. [8] Lennox WB. Ketogenic diet in the treatment of epilepsy. N EnglJ Med 1928;199:74. [9] Davidian NM, Butler Poole DT. The effect of ketosis induced by medium chain triglycerides on intracellular pH of mouse brain. Epilepsia 1978; 19: 369-78. [10] Dekaban AS. Plasma lipids in epileptic children treated with the high fat diet. Arch Neurol 1966;15:177-84. [11] Trauner D, Sweetman L, Holm J, Kulovich S, Nyhan WL. Biochemical correlates of illness and recovery in Reye's syndrome. Ann Neurol 1977; 2:238-41.