Epilepsy Research (2011) 96, 96—100
journal homepage: www.elsevier.com/locate/epilepsyres
Efficacy of 4:1 (classic) versus 2.5:1 ketogenic ratio
diets in refractory epilepsy in young children: A
randomized open labeled study
K.N. Vykunta Rajua, Sheffali Gulatia,∗, Madhulika Kabraa, Anuja Agarwalaa,
Suvasini Sharmaa, Ravindra Mohan Pandeyb, Veena Kalraa
aDepartment of Pediatrics, All India Institute of Medical Sciences, New Delhi 110029, India
bDepartment of Biostatistics, All India Institute of Medical Sciences, New Delhi 110029, India
Received 19 November 2010; received in revised form 1 May 2011; accepted 8 May 2011
Available online 28 May 2011
Low ratio diets;
Modified Atkins diet
Purpose: The ketogenic (lipid to non-lipid) ratio may play an important role in the efficacy
and tolerability of ketogenic diets (KD). This study was planned to compare the efficacy and
tolerability of 2.5:1 versus 4:1 lipid:non-lipid ratio KD in young children with refractory epilepsy.
Methods: Children aged 6 months to 5 years with refractory epilepsy were enrolled. They
were randomized to receive either a 4:1 or 2.5:1 ketogenic ratio diet, which was introduced
using a non-fasting protocol. Seizure frequency, biochemical profile (liver and kidney function
tests, fasting lipid profile, and spot urinary calcium—creatinine ratio), and adverse effects were
recorded at three months in both groups.
Results: Thirty eight children were enrolled, 19 in each group. At three months, 11 children
(58%) in the 4:1 group and 12 (63%) in the 2.5:1 group had more than 50% reduction in seizures
(p=0.78). Five children (26%) in the 4:1 group and four (21%) in 2.5:1 group became seizure free.
There was no significant difference in the biochemical parameters between the two groups.
Conclusion: 2.5:1 ratio KD is possibly as effective as 4:1 KD in controlling seizures and has fewer
© 2011 Elsevier B.V. All rights reserved.
∗Corresponding author at: Child Neurology Division, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi 110029,
India. Tel.: +91 11 26593209; fax: +91 11 26588641.
E-mailaddresses:firstname.lastname@example.org (K.N.V. Raju), email@example.com (S. Gulati), firstname.lastname@example.org (M. Kabra),
email@example.com (A. Agarwala), firstname.lastname@example.org (S. Sharma), email@example.com (R.M. Pandey),
firstname.lastname@example.org (V. Kalra).
0920-1211/$ — see front matter © 2011 Elsevier B.V. All rights reserved.
The ketogenic diet (KD) is a medically supervised high fat,
low carbohydrate diet that has been used to treat refractory
epilepsy (Freeman et al., 1998). The reported effectiveness
of KD matches or exceeds that of anti-epileptic drugs (AED)
in many cases. The ketogenic ratio is defined as the grams
of fat to the grams of carbohydrate plus protein. Tradition-
ally, ratios of 4:1 in older children and 3:1 in infants have
been used. Higher ratios result in greater degrees of keto-
sis (Wirrell, 2008). While such high ratios may confer better
seizure control, they may cause poorer tolerability of the
diet (Nylen et al., 2005).
Higher ratios are especially difficult for Indian patients
to accept, as the traditional Indian diets are cereal based,
predominantly vegetarian, and much higher in carbohydrate
content as compared to Western diets. Reducing the pro-
portion of fat would be a helpful option to make KD more
acceptable, palatable and tolerable. Preliminary data on
modifications of KD such as the modified Atkins and the low
glycemic index diets have shown that efficacy is not compro-
mised by reducing the proportion of fat in the diet (Kossoff
et al., 2003; Pfeifer and Thiele, 2005). There have been few
studies on the use of lower ketogenic ratio diets (Seo et al.,
2007; Nathan et al., 2009). Seo compared the efficacy of
3:1 versus 4:1 lipid to non-lipid ratio ketogenic diets in chil-
dren with refractory epilepsy (Seo et al., 2007). There have
been no other comparative studies of low ratio diets with the
classic ketogenic diet. Therefore, this study was planned to
compare the efficacy and tolerability of a 2.5:1 ratio diet
with the classic 4:1 diet.
This was a randomized, open labeled trial which was conducted
in the Pediatric Department of a tertiary care hospital between
January 2008 and June 2009. Ethical clearance was obtained from
the Institute ethical committee (ref no. A-61/23.01.2008). Writ-
ten informed consent was obtained from the parents. The study
was registered with the clinical trial registry of India (CTRI), No-
Children aged six months to five years who had at least two seizures
per month despite the appropriate use of at least two antiepileptic
drugs (AED) including one newer AED were enrolled. Children with
known or suspected inborn errors of metabolism, systemic illness,
or surgically remediable causes of epilepsy were excluded.
Each child underwent detailed clinical evaluation. Seizure type,
frequency, age at onset, perinatal events, family history, develop-
mental status and treatment history were noted. Medications were
ticosteroids or ACTH were tapered off four weeks before starting
diet. A four week baseline period of seizure recording was used prior
to initiation of KD without any changes in anti-epileptic medication.
Seizure frequency was assessed by the parent-maintained seizure
records on the basis of a chart with various categories of seizure
(infantile spasms, absence, myoclonic, atonic, tonic, tonic—clonic,
and focal). The definitions and clinical manifestations of the differ-
ent seizure types were clarified with parents.
Baseline investigations included: blood urea, serum creatinine,
liver function tests (serum albumin, bilirubin, serum glutamate
oxaloacetate transaminase [SGOT], serum glutamate pyruvate
transaminase [SGPT]), fasting serum lipid profile (including low
density lipoprotein (LDL), high density lipoprotein (HDL), very low
density lipoprotein (VLDL), total cholesterol and triglycerides) and
urinary spot calcium—creatinine ratios.
Enrolled patients were randomized using a computer generated
random number table into 2.5:1 or 4.1 arm. Allocation conceal-
ment was done by opaque sealed envelopes. Neither the ketogenic
diet team nor the parents of the children were blinded as to the
The children were admitted to the hospital for KD initiation using
a non-fasting gradual initiation protocol. KD was started with a
full calorie, ketogenic ratio (ratio of fat:protein+carbohydrate) by
weight of 1:1. This was built up over a period of three days to
2.5:1 (1:1 on the 1st day, 2:1 on the 2nd day, and 2.5:1 on the
3rd day) in the 2.5:1 group and over four days to 4:1 (1:1 on the
1st day, 2:1 on the 2nd day, 3:1 on the 3rd day, and 4:1 on the
4th day) in the 4:1 group. Blood sugars, urine ketones, serum elec-
trolytes, weight, and intake—output was monitored. The recipes
were planned in-house and calculated considering the families and
the child’s preferences and cultural taboos such as vegetarianism
and avoidance of eggs, onions and garlic in some families. Diets
were based on Indian recipes and prepared with common locally
available foods. The AED were continued unchanged.
All outcomes were assessed at three months after KD initiation.
The primary outcome variable was the proportion of children with
>50% reduction in seizure frequency from baseline at 3 months in
both groups. Seizure frequency was checked according to parent
maintained seizure records. The seizure control was categorized
as: seizure free, >50 seizure reduction, or <50% seizure reduction.
Adverse effects of the diet were noted as per parental reports.
The biochemical profile including a fasting lipid profile, urinary spot
calcium—creatinine ratios and urine ketones by reagent strips were
tested at three months. Oral potassium citrate supplementation was
started in children with urinary spot calcium creatinine ratio more
Data was analyzed using SPSS software version-15. Data were
expressed as mean (SD) or median (range) for continuous variables
and as proportions for categorical variables. Due to non-parametric
distribution of data, the Mann—Whitney U-test was used for contin-
uous variables and Fischer’s exact test for categorical variables. We
evaluated our findings against a two-tailed significance level of 5%
(p<0.05). Given the exploratory nature of this trial, and the lack of
data on the use of 2.5:1 diet, no sample size estimation was done.
It was decided to enrol all the consecutive patients with refractory
epilepsy who presented during the predetermined study period.
During the study period, 60 children were screened, and 38
found eligible. Nineteen (16 boys, 3 girls) children were ran-
domly assigned to the 4:1 group and 19 children (15 boys,
4 girls) to the 2.5:1 group (Fig. 1). The median age at KD
98 K.N.V. Raju et al.
Baseline characteristics of study population.
Characteristic Ketogenic diet ratio
4:1 (n=19) 2.5:1 (n=19)
Age at initiation of KD (months) [median (range)]
Male sex (percentage)
AED tried (no.) before KD institution [median (range)]
Generalized tonic clonic
aAll the children other than those with infantile spasms had mixed seizure types.
bWest syndrome etiology: 4:1 group — symptomatic-6 (perinatal asphyxia sequelae-4, neonatal meningitis sequelae-1, tuberous
sclerosis-1), cryptogenic-1, idiopathic-2; 2.5:1 group — symptomatic-5 (perinatal asphyxia sequelae-3, neonatal hypoglycemia sequelae-
1, tuberous sclerosis-1), cryptogenic-1, idiopathic-1.
60 patients assessed for eligibility
45 met inclusion
7 excluded (3- not willing, 2-
disorders, 2- surgical
38 patients enrolled
19 patients enrolled in 4:1group
19 patients enrolled in 2.5:1 group
19 received treatment 19 received treatment
3 discontinued3 discontinued
19 included in final
19 included in final
Flow of patients in the study.
Ketogenic diet 99
Biochemical and lipid profile on ketogenic diet: base line and at 3 months.
ParameterBaselineAt 3 months
4:1 (n=19)2.5:1 (n=19)
p Value4:1 (n=16)2.5:1 (n=16)
Blood urea (mg/dl)
S. creatinine (mg/dl)
S. albumin (g/dl)
S. calcium (mg/dl)
Urine spot calcium-creatinine
Total cholesterol (mg/dl)
26.2 ± 6.7
0.50 ± 0.08
38 ± 10
27 ± 17
4.17 ± 0.4
9.3 ± 0.6
0.05 ± 0.04
32.1 ± 10.2
0.55 ± 0.08
37 ± 16
26 ± 9
4.42 ± 0.2
9.3 ± 1.1
0.045 ± 0.06
26.6 ± 6
0.57 ± 0.09
31 ± 13
30 ± 9
9.1 ± 0.5
0.14 ± 0.05
25.1 ± 5
0.48 ± 0.12
37 ± 5
28 ± 9
4.38 ± 0.3
8.9 ± 0.7
0.18 ± 0.16
134 ± 20
126 ± 74
26 ± 13
75 ± 15
36 ± 5
138 ± 17.7
138 ± 86
26 ± 11
79 ± 18
36 ± 7
132 ± 31
114 ± 41
26 ± 36
107 ± 41
36 ± 7
143 ± 17
133 ± 53
26 ± 12
117 ± 18
43 ± 10
Values are in mean±SD. VLDL, very low density lipoprotein; LDL, low density lipoprotein; HDL, high density lipoprotein.
initiation, the male—female ratio and co-morbidities were
comparable in both groups (Table 1). All the patients had
daily seizures, with a median seizure frequency of 12 per
day (range 5—220) in the 4:1 KD group, and 10 per day (range
3—96) in the 2.5:1 group. Multiple AED had been tried in both
groups. The epilepsy syndromes included infantile spasms,
Lennox Gastaut syndrome and myoclonic astatic epilepsy.
There was no significant difference in the seizure types and
epilepsy syndromes in both the groups.
Three months after initiation, 16 children (88%) contin-
ued on the diet in each arm. Eleven children (58%) in the
4:1 group and 12 (63%) in the 2.5:1 group had more than 50%
reduction in seizures at three months. Five children (26%) in
the 4:1 group and four (21%) in 2.5:1 group became seizure
free. These differences were not statistically significant.
Constipation was the most common adverse effect, noted
in five children in 4:1 group, and three children in the
2.5:1 group. Three children in the 4:1 group and one child
in the 2.5:1 group lost weight. This improved on increas-
ing the caloric intake. Two children in the 4:1 group and
one child in 2.5:1 group were hospitalized for lower inter-
current respiratory tract infections. In both children, KD
was continued both during hospital stay and after dis-
The biochemical profile including the fasting lipid profile
of the patients at baseline and three months after starting
the KD in both groups is detailed in Table 2. There was no sta-
tistically significant difference at baseline and at 3 months
on diet in both groups. At the three month spot check test,
all the 16 children in the 4:1 group had large ketosis (urine
ketones 80—160mg/dl), while 15 children in the 2.5:1 group
had large ketosis, and one child had moderate ketosis (urine
The reasons for discontinuation in the 4:1 arm included
unsatisfactory seizure control, child’s refusal to eat and
non-acceptance by family members in one child each. The
reasons in the 2.5:1 arm were unsatisfactory seizure control
in two children and refusal to eat in one child.
The association between ketogenic ratios and seizure
control is not well understood. Studies in animals have
shown that higher ratios correlate with greater efficacy.
Bough et al. (2000) conducted a study to determine the
dose—response relationship by feeding rats with ketogenic
diets with ratios varying from 1:1 to 9:1. Animals who were
fed diets exceeding a 6:1 ratio had a significantly greater
resistance to pentylenetetrazole (PTZ)-induced seizures
compared to those fed 4:1 or 5:1 ratios (p=0.009 and
p=0.02). Nylen et al. (2005) demonstrated that a 6.3:1 diet
significantly elevated seizure threshold to PTZ as compared
to a 4:1 diet in young rats. However, the 6.3:1 KD group
showed poorer weight gain than the 4:1 KD group.
To the best of our knowledge, our study is the first study
comparing a 2.5:1 ratio KD with the classic 4:1 ratio keto-
genic diet. Though the number of patients was small, we
found a comparable efficacy and tolerability of both the
diets. The biochemical profile of the two groups was also
comparable, except for a trend (p=0.06) towards higher
cholesterol and HDL in the 2.5:1 group, which was surpris-
ing. This finding needs to be clarified in a larger cohort of
Seo et al. (2007) compared the efficacy and tolerability
of 3:1 with 4:1 diet in 76 children with refractory epilepsy.
The authors found a better seizure control on the 4:1 diet
(55% in the 4:1 versus 31% in the 3:1 group were seizure
free, p<0.05), but a better tolerability of the 3:1 diet,
with fewer gastrointestinal side effects. One of the rea-
sons for a better seizure control in the 4:1 group could
possibly be the fact that there were more patients with
Lennox Gastaut syndrome and fewer patients with partial
seizures in the 4:1 as compared to the 3:1 group. There
was no significant difference in the blood beta hydroxy-
butyrate levels between the two groups, suggesting that
seizure control may not correlate well with blood ketone
100K.N.V. Raju et al. Download full-text
Nathan et al. (2009) reported on the use of ketogenic
ratios ranging from 2:1 to 4:1 in 105 children with refrac-
tory epilepsy. Though they did not compare the ratios, they
reported an overall efficacy of seizure freedom in 37% and
a better than 50% reduction in 44% of the patients. Recent
studies on the modified Atkins diet, which is a carbohydrate
restricted diet with unlimited protein and fat, with a keto-
genic ratio between 1:1 and 2:1, have shown efficacy similar
to that of the classic ketogenic diet (Kossoff et al., 2006;
Kang et al., 2007).
Asians and Indians may be particularly suited for low ratio
diets. The traditional diet of Indians is very carbohydrate
rich, with carbohydrates forming almost two thirds of the
diet intake. The present study and the study by Nathan have
shown that lower ratio ketogenic diets are effective in con-
trolling seizures in Indian children. Children in both groups
in our study had large urinary ketosis. Indians are perhaps
able to switch to ketogenic metabolism even with a mod-
est reduction in carbohydrate and increase in fat content.
Also, the use of lower ketogenic ratios permits increased use
of carbohydrates in the diet, allowing approximation to the
traditional Indian diets, leading to better acceptability.
The limitations of this study include the small sample size
and the unblinded design. Also the use of parental seizure
records runs the risk of subjective errors, especially in chil-
dren who have myoclonic and absence seizures, which are
likely to be missed. Despite these limitations, this study
demonstrates the potential efficacy of a 2.5:1KD, with pos-
sibly lesser side effects as compared to the classic 4:1KD.
Conflict of interest
Bough, K.J., Yao, S.G., Eagles, D.A., 2000. Higher ketogenic diet
ratios confer protection from seizures without neurotoxicity.
Epilepsy Res. 38, 15—25.
Freeman, J.M., Vining, E.P.G., Pillas, D.J., Pyzik, P.L., Casey, J.C.,
Kelly, L.M., 1998. The efficacy of the ketogenic diet: a prospec-
tive evaluation of intervention in 150 children. Pediatrics 102,
Kang, H.C., Lee, H.S., You, S.J., Kang du, C., Ko, T.S., Kim, H.D.,
2007. Use of a modified Atkins diet in intractable childhood
epilepsy. Epilepsia 48, 182—186.
Kossoff, E.H., Krauss, G.L., McGrogan, J.R., Freeman, J.M., 2003.
Efficacy of the Atkins diet as therapy for intractable epilepsy.
Neurology 61, 1789—1791.
Kossoff, E.H., McGrogan, J.R., Bluml, R.M., Pillas, D.J., Rubenstein,
J.E., Vining, E.P., 2006. A modified Atkins diet is effective for
the treatment of intractable pediatric epilepsy. Epilepsia 47,
Nathan, J.K., Purandare, A.S., Parekh, Z.B., Manohar, H.V., 2009.
Ketogenic diet in Indian children with uncontrolled epilepsy.
Indian Pediatr. 46, 669—673.
Nylen, K., Likhodii, S., Abdelmalik, P.A., Clarke, J., Burnham, W.M.,
2005. A comparison of the ability of a 4:1 ketogenic diet and a
6.3:1 ketogenic diet to elevate seizure thresholds in adult and
young rats. Epilepsia 46, 1198—1204.
Pfeifer, H.H., Thiele, E.A., 2005. Low-glycemic-index treatment: a
liberalized ketogenic diet for treatment of intractable epilepsy.
Neurology 65, 1810—1812.
Seo, J.H., Lee, Y.M., Lee, J.S., Kang, H.C., Kim, H.D., 2007. Efficacy
and tolerability of the ketogenic diet according to lipid: non-
lipid ratios—–comparison of 3:1 with 4:1 diet. Epilepsia 48, 801—
Wirrell, E.C., 2008. Ketogenic ratio, calories, and fluids: do they
matter? Epilepsia 49 (Suppl. 8), 17—19.