Prostaglandins, Leukotrienes and Essential FattyAcids (2002) 67(2^3),131^135

& 2002 Elsevier Science Ltd. All rights reserved.

doi:10.1054/plef.409, available online at http://www.idealibrary.com on

Potential role of polyunsaturates in

seizure protection achieved with the
ketogenic diet
S. C. Cunnane,1,4 K. Musa,1 M. A. Ryan,1 S.Whiting,1^4 D. D. Fraser1^4
1

Department of Nutritional Sciences, Faculty of Medicine, University of Toronto,Toronto, Canada, 2Department of Anatomy and Cell Biology, Queens University,
Kingston, Canada, 3Department of Pediatrics, Childrens Hospital of Eastern Ontario, Ottawa, Canada and 4Canadian Pediatric Epilepsy Network, Canada

Summary Epilepsy is a serious neurological disease that responds to two very different treatments involving lipids. Clinically, it
respondsto a state of ketosisinduced byaveryhigh-fatketogenicdiet.Experimentally, invitro andinvivo models demonstratethat
injection or infusion of free (non-esterified) polyunsaturates such as arachidonate and docosahexaenoate also reduces seizure
susceptibility.In ourexperience, rats on a very high-fat ketogenic diet not only have mild-to-moderate ketosis, but also have raised
serum free fatty acids. Some polyunsaturates, particularly linoleate and a-linolenate, are relatively easily b-oxidized and are
therefore ketogenic.We conclude that raised levels of free plasma polyunsaturates could contribute to the beneficial effect of the
ketogenic diet in refractory epilepsy not only by helping sustain ketosis, but also by their own direct (though poorly defined)
antiseizure effects. & 2002 Elsevier Science Ltd. All rights reserved.

INTRODUCTION
A dietary source of o3 polyunsaturated fatty acids (PUFA)
is essential for normal neurological and visual function
in mammals, especially humans. Interest has overwhelmingly focussed on how docosahexaenoate (DHA,
22:6o3) fulfils this role. Despite intense research over
the past 20 years, the reasons why DHA is virtually alone
amongst the o3 PUFA in neuronal and visual membranes
are only gradually being elucidated. As the evidence for a
role of DHA in early postnatal neurological development
strengthens, the emphasis in this field is shifting towards
determining how dietary DHA and eicosapentaenoate
(EPA, 20:5o3) may help sustain normal neurological

Received 10 October 2001

Accepted 3 May 2002
Correspondence to: Dr. S. C. Cunnane, Department of Nutritional Sciences,
University of Toronto,Toronto, Canada M5S 3E2.Tel.: +1-416-978-8356;
Fax: 416-978-5882;
E-mail: s.cunnane@utoronto.ca

Grant Support : Bloorview Childrens Hospital Foundation, Dairy Farmers of

Canada, Savoy Foundation, Physicians Services Incorporated Foundation,
Queens University, American Academy of Pediatrics, and NSERC.

& 2002 Elsevier Science Ltd. All rights reserved.

function and reduce the risk of psychiatric illness and

premature memory loss in maturity and old age.
Epilepsy is a serious and often lifelong neurological
disorder which may also shed some light on how PUFA,
particularly DHA, are utilized in the nervous system. In
vitro and animal models are beginning to suggest that
seizure susceptibility may benefit from increased availability of o3 PUFA and possibly o6 PUFA such as
arachidonate (AA, 20:4o6). Some of this evidence comes
from experiments using direct addition of DHA or other
PUFA to the model system under study.

EPILEPSY
A seizure is a disordered yet rhythmic firing of a
population of neurons resulting in a behavioural change.
The behavioural change associated with seizures can be
as mild as repeated distracted staring or blinking, or could
be as severe as violent muscle spasms, or falls due to a
total loss of muscle control. Epilepsy is the brain disorder
characterized by recurrent spontaneous seizures. It
affects 12% of children and 0.51% of the population
as a whole. There are many epilepsy phenotypes, some
more severe than others. Some go into spontaneous

Prostaglandins, Leukotrienes and Essential FattyAcids (2002) 67(2^3), 131^135

132

Cunnane et al.

remission before adulthood. Other forms are persistent,

socially and developmentally debilitating, and resistant
(refractory) to available antiepilepsy medication.
About 70% of epilepsy is considered to respond
adequately or well to available medication. Refractory
epilepsy is usually defined as epilepsy that does not
adequately respond to three successive treatments with
one or more different antiepileptic drugs given alone or in
combination. Refractory epilepsy of temporal lobe origin
occasionally benefits remarkably from brain surgery to
remove the seizure-generating focus in the brain. Seizures may be well controlled by a particular drug, but
serious side effects such as antisocial behaviour, drowsiness, and impaired learning can reduce the overall benefit
of these drugs. Hence, alternative treatments and new
drugs for refractory epilepsy are in constant demand.
THE KETOGENIC DIET FOR REFRACTORY
EPILEPSY
The ketogenic diet has re-emerged in the 1990s as a
legitimate treatment for refractory epilepsy. First, developed early in the 20th century to mimic the biblical
assertion that seizures could be treated by fasting, the
ketogenic diets popularity has risen as a last gasp
treatment for refractory epilepsy. Occasionally, it is used
when there are excessive side effects of an otherwise
effective antiepilepsy drug. It is a last gasp treatment for
two reasons. First, unlike with treatments for cardiovascular disease or diabetes, epilepsy is not a field accustomed to treatments involving dietary manipulation.
Second, it is a very difficult treatment with which to
comply. However, extensive and growing documentation
of its use in several centres particularly in the USA and
Canada shows that about two-thirds of cases have
significantly reduced seizure frequency, and about onethird do not benefit significantly. Patients can often be
weaned off medications that were in use prior to starting
the ketogenic diet. In this sense, fortunate cases treated
with the ketogenic diet can achieve a lifelong cure from
seizures. This is an excellent (though uncommon)
achievement for a tertiary level of epilepsy treatment.
The classic ketogenic diet is usually 80% dairy fat by
weight and generates insufficient ketosis if carbohydrate
is greater than 45% total food intake. Hence, extreme
carbohydrate restriction (including sugar free toothpaste)
is required which has to be sustained throughout the
treatment period. Sufficient protein is given to achieve
somewhat sub-normal growth in children. Apart from
mandatory vitamin and mineral supplements (also sugar
free), the rest of the diet is fat. The fat is usually butter
and cream because they offer the most flexibility and
palatability, and are the most economical. This treatment
will induce sustained, moderate and safe ketosis with a

plasma b-hydroxybutyrate of 210 mM. It appears to be

equally effective in all forms of epilepsy. Far more
children than adults with refractory epilepsy are treated
with the ketogenic diet because of difficulties with
compliance and the increased cardiovascular risk of a
diet exceptionally high in dairy (saturated) fat. While
hypercholesterolaemia does occur in some children
and adults on the ketogenic diet, not all become
hypercholesterolaemic, perhaps because slower growth
(children) or weight loss (adults) often occurs with this
treatment. As with drugs for epilepsy, there are other side
effects of the ketogenic diet, including increased risk of
kidney stones, vomiting, anorexia, coagulopathy, and
osteoporosis.
Since its first introduction and frequent reintroduction,
there has been much speculation about the mechanism
by which the ketogenic diet induces seizure control. Little
of this speculation has found sustained experimental
support. A conundrum that may eventually relate to the
mechanism is the exquisite sensitivity of seizure control
on the ketogenic diet to a rise in plasma glucose; a candy
taken or given surreptitiously will suspend seizure
control immediately, though usually temporarily, until
ketosis is achieved again. Occasionally, however, seizure
control can be lost and remains frustratingly unachievable despite the re-establishment of ketosis. Plasma
glucose levels are usually 1020% lower than normal
while on the ketogenic diet so it may not be the actual
glucose level but, rather, spikes in this level that inhibit
ketogenesis and break seizure control. Hence, brain
glucose transport and the competition between glucose
and ketones as brain energy substrates may have a role in
seizure control.
One of the difficulties in understanding this enigmatic
epilepsy treatment is the absence of a well-accepted
model in which invasive analyses are permitted and in
which the ketogenic diet is reproducibly effective. For the
most part, the classic animal seizure models on which
pharmacological treatments are based (electroshock- or
chemically-induced seizures) do not respond to the
ketogenic diet though there has been some recent
success with the pentylenetetrazol model.13 Furthermore, rats do not achieve the same degree of ketosis as
humans on the same ketogenic diet formulation.
KETONES THEMSELVES MAY HAVE ANTISEIZURE
EFFECTS
Two groups have shown that a minimum threshold
plasma b-hydroxybutyrate level of about 2.5 mM probably needs to be reached in order achieve seizure control
in humans with refractory epilepsy on the ketogenic
diet.4,5 However, there was no correlation between degree
of seizure control and ketosis above this minimum level,

Prostaglandins, Leukotrienes and Essential FattyAcids (2002) 67(2^3), 131^135

& 2002 Elsevier Science Ltd. All rights reserved.

Potential role of polyunsaturates 133

suggesting that either b-hydroxybutyrate itself does not

have anticonvulsant properties, or that ketosis is associated with another metabolic change that actually
inhibits seizures. We are unaware of any studies assessing
whether b-hydroxybutyrate itself inhibits seizures.
However, blood acetone is raised to similar levels
as b-hydroxybutyrate while on the ketogenic diet and,
at physiological concentrations while in moderate ketosis
(10 mM), inhibits seizures in the pentylenetetrazol model
in rats (Likhodii and Burnham, unpublished). Hence,
breath acetone measurements which correlate positively
with plasma ketones,6 may reveal that seizure control is
more directly correlated to plasma (and brain?) acetone
concentrations than to b-hydroxybutyrate.

PLASMA FATTY ACID PROFILES ON THE

KETOGENIC DIET

POLYUNSATURATES AND SEIZURE CONTROL

Fraser and colleagues were amongst the first to examine
the role of exogenous PUFA in neuronal electrophysiology.7 Using electrophysiological recordings from cultured
central neurons, they demonstrated a significant inhibition of voltage-gated sodium channels with application
of 130 mM AA. AA shifted the sodium channel inactivation curve to hyperpolarizing potentials, resulting in
decreased action potential firing and subsequent neurotransmitter release. Interestingly, an identical effect on
the sodium channel inactivation curve has been seen
with several anticonvulsants including phenytoin (Dilantin) and carbamazepine (Tegretol). It is therefore possible
that some PUFA like AA may be the endogenous ligand
mimicked by some anticonvulsants.
While investigating the antiarrhythmic effects of DHA
and other PUFA in the mammalian heart, Leaf and
colleagues speculated that similar beneficial effects might
be seen in seizures or brain arrhythmias. They undertook
a series of animal and in vitro studies that supported a
role of DHA in seizure control. One such study involved
permanent brain surface electrodes used to stimulate a
mild form of seizures.8 Sequential 15 s pulse trains
increased in current strength from zero to 1000 mA.
Seizures were video recorded to assess the current at
which localized and generalized seizures occurred. A
30 min infusion of 40 mmol free DHA took 6 h to have any
effect, but increased the seizure threshold by 1520%.
Despite apparently rapid uptake of non-esterifed PUFA
into the brain from the circulation,9 this delay in the
antiseizure effect of free DHA was presumably due the
time required to integrate the DHA into the relevant
neuronal cell membranes. Voskuyl and colleagues also
noted that infused DHA esterified to phospholipid was
ineffective in controlling seizures.5 Other free PUFA
including EPA or linoleate (LA, 18:2o6) were less effective
in controlling seizures in this model. AA or a-linolenate
(ALA, 18:3o3) was not tested.
& 2002 Elsevier Science Ltd. All rights reserved.

Yehuda and colleagues have reported that intraperitoneal injection of a mixture of free ALA and LA inhibits
seizures induced by pentylenetetrazol.10 Since a mixture
was used, it is unclear whether LA or ALA or the
combination was required. These in vitro and in vivo
studies conducted by different groups suggest that PUFA,
notably DHA, can have acute antiseizure effects in
diverse models. One common feature of this effect seems
to be that the PUFA in question needs to be in the nonesterified (free) form. Only Voskuyl and colleagues appear
to have directly tested this assertion,5 but this may be an
important parameter in order to obtain antiseizure effects
of PUFA.

Plasma and liver free fatty acids roughly double in rats on

a high fat ketogenic diet containing flaxseed oil (Table 1).
Hence, another possibility besides ketones directly inhibiting seizures is that ketosis achieved on the ketogenic
diet sufficiently raises free fatty acids to, themselves, have
antiseizure effects. DHA and AA in the free (non-esterified
form) both inhibit seizures in experimental models,7,8 so
raised plasma free fatty acids may account for at least part
of the antiseizure effect of this dietary treatment. Thus,
increased transfer of free DHA from plasma to the brain
could also be an effective and direct mediator of seizure
control in humans on the ketogenic diet, a concept we are
presently evaluating. Raised concentrations of plasma
free LA, AA, ALA and EPA might potentially add to the
antiseizure effect of higher plasma free DHA.
Ketosis is a state of increased fatty acid oxidation in
which the mitochondrial capacity to oxidize acetyl CoA is
exceeded by the plasma free fatty acid concentration. At

Table 1 Plasma and liver lipids in rats on a ketogenic diet (modified

from ref. 3)
Control

MCT

FSO

Butter

Plasma
Free fatty acids (mg/ml) 54+51,a 79+20a,b 92+27b 79+21a,b
Triglycerides (mg/dl)
6+2a,b
3+1a
5+3a,b
9+3b
Cholesterol (mg/dl)
96+8
108+15
96+8
116+15
Liver
Free fatty acids (mg/g)
Triglycerides (mg/g)

18+5a
7+4a

22+5b
7+6a

48+11c
8+2a

29+11b
73+17b

1
Mean+SD, n=6/group.
Different superscripts connote statistically significant differences
(Po0.05).
ControlF10% fat from soybean oil. MCT F 80% fat diet based on
medium chain triglyceride oil (87% of fat) and soyabean oil (13%
of fat). FSO F 80% fat diet based on flaxseed oil (87% of fat) and
soyabean oil (13% of fat).Butter F 80% fat diet based on butter (87%
of fat) and soyabean oil (13% of fat).

Prostaglandins, Leukotrienes and Essential FattyAcids (2002) 67(2^3), 131^135

134

Cunnane et al.

Table 2 Adipose tissue levels of PUFA in rats on a high-fat ketogenic

diet (modified from ref. 3)
Control
18:2o6
20:4o6
18:3o3
22:6o3

MCT
1,a

28.6+2.4
0.3+0.1a
2.9+0.3a
0.05+0.06a

FSO
a

25.9+2.9
1.1+0.2b
3.2+0.5a
0.2+0.05b

Butter
c

21.7+0.6
0.1+0.02c
38.9+0.5c
0.1+0.06a

10.6+0.1b
0.3+0.03a
1.7+0.1b
0.02+0.03a

1
Mean+SD, n=6/group.
Different superscripts connote statistically significant differences
(Po0.05).
Control F10% fat from soyabean oil. MCT F 80% fat diet based on
medium chain triglyceride oil (87% of fat) and soyabean oil (13%
of fat). FSO F 80% fat diet based on flaxseed oil (87% of fat) and
soyabean oil (13% of fat).Butter F 80% fat diet based on butter (87%
of fat) and soyabean oil (13% of fat).

present, we have no data to determine what proportion of

ketosis on the ketogenic diet arises from the increased
intake of fatty acids compared to an increased flux of fatty
acids primarily from adipose tissue. Despite similar
intakes of all PUFA, rats on a ketogenic diet have lower
proportions of several PUFA in their adipose tissue than
controls on a 10% fat diet (Table 2). In rats, higher plasma
b-hydroxybutyrate is achieved with a PUFA-rich compared to a saturated fat-rich ketogenic diet.2 Nevertheless,
we still do not know the extent to which addition of more
easily b-oxidized long-chain fatty acids to the diet might
increase b-oxidation and ketogenesis in humans.
The relation between mild-to-moderate ketosis and
raised serum free fatty acids will be important to model in
the future for two reasons. The first reason is that the
main dietary fatty acids are not mobilized from adipose
tissue nor are they b-oxidized at the same rate.11,12
Broadly speaking, ALA and LA are more easily mobilized
and b-oxidized than AA or saturated fatty acids. Surprisingly, EPA and DHA are also relatively easily mobilized
and b-oxidized.12,13 Oleate (18:1o9) is somewhat intermediate in both these parameters. Second, with the
exception of AA, the dietary long-chain fatty acids which
are most easily b-oxidized are the same as those that are
most effective in inhibiting seizures (DHA, EPA, ALA,
LA).7,8,10 Therefore, if adipose tissue could be pre-loaded
with these easily b-oxidized, seizure-inhibiting PUFA
from the diet, the efficacy of seizure control with the
ketogenic diet might be improved.

mechanism. Some PUFA have antiseizure effects under

controlled experimental circumstances in vitro or in
animals. Even if these seizure experiments closely
modelled seizures or epilepsy in humans (which most
do not), it is technically and ethically difficult to
administer free PUFA to humans because these fatty
acids are highly insoluble in aqueous media such as
blood.
The fatty acid profile of the ketogenic diet heavily
emphasizes saturates and is usually low in PUFA because
it is based primarily on dairy fat (butter, cream). Devising
ways to increase the PUFA content of the ketogenic diet
would seem to be worthwhile. Current investigation of
treatments for neurological disease involving PUFA, DHA
and EPA is presently being emphasized but, in agreement
with previous in vitro data,7 our present results suggest
that AA may also be important in reducing seizure
susceptibility.
Research into how these changes in blood fatty acids
influence brain fatty acids is also merited. Our data show
that despite raised plasma free fatty acids, brain DHA and
AA composition do not change in rats on ketogenic diets
differing markedly in fatty acid content and composition.3
These results suggest that the fatty acid composition of
the mature rat brain is markedly resistant to changes in
plasma and diet fatty acids. Nevertheless, fatty acid

CONCLUSIONS
The ketogenic diet is a difficult but much needed
treatment for refractory epilepsy which demonstrates
that the body is potentially capable of correcting a major
neurological disorder without exogenous agents. This is
remarkable, all the more so because this treatment is
empirical in the sense that it arose with no insight into its

Fig. 1 Proposed new scheme (dotted arrows) by which raised levels

of polyunsaturates (PUFA) in serum could contribute to the beneficial
effects of the ketogenic diet in refractory epilepsy.The ketogenic diet
is primarily a clinical approach while PUFA injection or infusion is
primarily an experimental approach.We postulate that raised free
PUFA in serum could raise blood ketones because most PUFA are
relatively easily b-oxidized.11

Prostaglandins, Leukotrienes and Essential FattyAcids (2002) 67(2^3), 131^135

& 2002 Elsevier Science Ltd. All rights reserved.

Potential role of polyunsaturates 135

binding proteins exist at the bloodbrain barrier14 and

some models suggest that long-chain fatty acids have
ready access to the brain.9 Perhaps, the neonatal or infant
brain more readily acquires fatty acids present in the
circulation.
If the brain is resistant to changing its fatty acid
composition, how then does the increase in plasma free
PUFA reduce seizure susceptibility and what are these
fatty acid binding proteins in the brain actually accomplishing? One mechanism by which raised free PUFA in
serum could have antiseizure effects without directly
entering the brain is by being b-oxidized to ketones
which cross the bloodbrain barrier and have antiseizure
effects (Fig. 1). In animal experiments, DHA takes a
relatively long time to increase seizure threshold8 and is
relatively easily b-oxidized and recycled into brain lipids
synthesized de novo.13 If brain DHA does not change in
rats consuming diets very high in ALA,3 the antiseizure
effects of DHA could potentially arise through ketogenesis either from ALA or from DHA itself. In any event, a
pharmaceutical or nutriceutical approach based on
raising PUFA in serum may have a future role in the
treatment of epilepsy.

REFERENCES
1. Bough K. J., Eagles D. A. A ketogenic diet increases the resistance
to pentylenetetrazole-induced seizures in the rat. Epilepsia 1999;
40: 128143.
2. Likhodii S. S., Musa K., Mendonca A., Dell C., Burnham W. M.,
Cunnane S. C. Dietary fat, ketosis and seizure protection in rats
on the ketogenic diet. Epilepsia 2000; 41: 806811.

& 2002 Elsevier Science Ltd. All rights reserved.

3. Dell C., Likhodii S. S., Musa K., Ryan M. A., Burnham W. M.,
Cunnane S. C. Dietary fat and lipid profiles in rats on ketogenic
diets with different fat formulations. Lipids 2001; 36: 373378.
4. Sirven J., Whedon B., Caplan D. et al. The ketogenic diet for
intractable epilepsy in adults: preliminary results. Epilepsia 1999;
40: 17211726.
5. Huttenlocher P. R. Ketonemia and seizures: metabolic and
anticonvulsant effects of two ketogenic diets in childhood
epilepsy. Pediatr Res 1976; 10: 536540.
6. Musa K., Likhodii S. S., Cunnane S. C. Breath acetone is a reliable
indicator of ketosis in adult volunteers consuming ketogenic
meals. Am J Clin Nutr 2001; 76: 6570.
7. Fraser D. D., Hoehn K., Weiss S., MacVicar B. A. Arachidonic acid
inhibits sodium currents and synaptic transmission in cultured
striatal neurons. Neuron 1993; 11: 633644.
8. Voskuyl R. A., Vreugdenhil M., Kang J. X., Leaf A. Anticonvulsant
effects of polyunsaturated fatty acids in rats using the cortical
stimulation model. Eur J Pharmacol 1998; 341: 145152.
9. Rapoport S. I. In vivo fatty acid incorporation into brain
phospholipids in relation to plasma availability, signal
transduction, and membrane remodelling. J Mol Neurosci 2001;
16: 243262.
10. Yehuda A., Carasso R. L., Mostovsky D. I. Essential fatty acid
preparation (SR-3) raises the seizure threshold in rats. Eur J
Pharmacol 1994; 254: 193198.
11. Cunnane S. C. New developments in alpha-linolenate
metabolism with emphasis on the importance of beta-oxidation
and carbon recycling. World Rev Nutr Diet 2001; 88: 178183.
12. Raclot T., Groscolas R. Selective mobilization of adipose tissue
fatty acids during energy depletion in the rat. J Lipid Res 1995; 36:
21642173.
13. Sheaff-Greiner R. C., Zhang Q., Goodman K. J., Guissini D. A.,
Nathanielsz P. W., Brefnna J. T. Linoleate, a-linolenate
and docosahexaenoate recycling into saturated and
monounsaturated fatty acids is a major pathway in pregnant or
lactating adults and fetal or infant rhesus monkeys. J Lipid Res
1996; 37: 26752686.
14. Glatz J. F. C., Storch J. Unravelling the significance of cellular fatty
acid-binding proteins. Curr Opin Lipidol 2001; 12: 267274.

Prostaglandins, Leukotrienes and Essential FattyAcids (2002) 67(2^3), 131^135