Epilepsy & Behavior 29 (2013) 437–442

Contents lists available at ScienceDirect

Epilepsy & Behavior

journal homepage: www.elsevier.com/locate/yebeh

Targeted Review

A decade of the modified Atkins diet (2003–2013): Results, insights, and

future directions
Eric H. Kossoff ⁎, Mackenzie C. Cervenka, Bobbie J. Henry, Courtney A. Haney, Zahava Turner
Johns Hopkins University School of Medicine, Baltimore, Maryland USA

a r t i c l e i n f o a b s t r a c t

Article history: The modified Atkins diet has been used since 2003 for the treatment of children and adults with refractory epi-
Received 17 September 2013 lepsy. This “alternative” ketogenic diet is started in clinic, without fasting, hospitalization, and restriction of pro-
Accepted 21 September 2013 tein, calories, or fluid intake. Now after 10 years of continued use, approximately 400 patients have been reported
Available online 26 October 2013
in over 30 studies of the modified Atkins diet as treatment for intractable seizures, with results demonstrating
similar efficacy to the ketogenic diet and improved tolerability. The modified Atkins diet is being increasingly
Keywords:
Atkins
used in the adult population. Clinical trials have provided insight into the mechanisms of action of dietary ther-
Ketosis apies overall. This review will discuss the past decade of experience with the modified Atkins diet as well as pre-
Ketogenic dictions for its role in the treatment of epilepsy a decade from now.
Diet © 2013 Elsevier Inc. All rights reserved.
Epilepsy

Included in the variety of therapeutic options are dietary therapies,

Key questions
1. How is the modified Atkins diet different from the traditional
classic ketogenic diet (or low glycemic index treatment)?
2. Is the modified Atkins diet effective?
3. Which patients should receive the modified Atkins diet in-
stead of the ketogenic diet?
4. What clues does the modified Atkins diet give us into the
mechanisms of action of all dietary therapies?
5. What is the future of the modified Atkins diet 10 years from
now?
1. Introduction
Patients with refractory epilepsy today have more treatment options
than ever before. New anticonvulsant drugs with novel mechanisms of ac-
tion such as ezogabine, lacosamide, perampanel, and rufinamide have been
introduced in recent years. Epilepsy surgery is becoming more widely avail-
able and safer, with new techniques and earlier resections leading to im-
proved quality of life even for patients with the most medically resistant
seizures. Vagus nerve stimulation is beneficial for many patients, and
other forms of neurostimulation, such as deep brain stimulation (targeting
the anterior nucleus of the thalamus) and responsive neurostimulation
(targeting the cortex), are in various stages of development and licensure.
now often being referred to by basic scientists as “metabolism-based
therapies” [1]. The classic ketogenic diet (KD) is one of the oldest
therapies for epilepsy, having been introduced to the world formally
in June 1921 [2], and now popular worldwide for the treatment
of children with refractory epilepsy. The KD is an extremely high fat
(approximately 90% of calories), low carbohydrate, moderate protein
diet that is calorie- and fluid-limited and started typically in the hospital
by a dietitian familiar with its use [1]. In the 92 years since its introduc-
tion, there have been changes to the initiation period of the KD to make
it more tolerable, but the composition itself has not been significantly
altered since its original conception.
As KD usage has increased, neurologists, dietitians, and parents alike
have sought alternatives with better tolerability and fewer adverse ef-
fects and have considered utilizing these treatments for a broad range
of neurologic conditions other than epilepsy. In 2003 at Johns Hopkins
and in 2005 at Massachusetts General Hospital, two “alternative” diets
were created to make dietary therapies easier to implement, and studies
demonstrated that the restrictiveness of the KD was not necessary in all
circumstances [3,4]. These diets, the modified Atkins diet (MAD) and
the low glycemic index treatment (LGIT), have changed the mindset
of many neurologists and parents regarding the feasibility of dietary
therapies and appear to have opened the door for their wider use. The
LGIT has been shown to be effective, but for the purposes of this review,
the MAD only will be discussed in detail.
2. History

⁎ Corresponding author at: 200 North Wolfe Street, David M. Rubenstein Child Health
Building, Suite 2158, The John M. Freeman Pediatric Epilepsy Center, The Johns Hopkins
Hospital, Baltimore, MD 21287, USA. Fax: +1 410 614 2297.
E-mail address: ekossoff@jhmi.edu (E.H. Kossoff).
In 2003, the first case series of 6 patients (three of whom were
adults) who were treated with the Atkins diet for seizures was pub-
lished [3]. The first patient was a 10-year-old boy (Patient 2) whose

1525-5050/$ – see front matter © 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.yebeh.2013.09.032
438 E.H. Kossoff et al. / Epilepsy & Behavior 29 (2013) 437–442
parent started the Atkins diet unsupervised, loosely following recipes
from a KD regimen he had received 2.5 years prior, and this led to sei-
zure freedom. Seizures remained completely controlled on the Atkins
diet. The child, however, that truly raised our awareness and led to fur-
ther studies was a 7-year-old girl (Patient 1) with intractable epilepsy
due to a left parietal cortical dysplasia. In May 2003, her mother placed
her on the Atkins diet to restrict carbohydrates a week in advance of the
classic KD, primarily in order to acclimate her child prior to starting. She
achieved large urinary ketosis, and her 80 seizures per day stopped after
3 days and did not return for 3 years of continued dietary treatment.
These children and 4 other patients were presented in a poster at the
2003 American Epilepsy Society Annual Meeting in Boston and received
media attention. The article was published that same month (December
2003) in Neurology [3].
In 2006, the diet was first formally referred to as the “modified
Atkins diet” to distinguish it from the Atkins diet [5]. There are 3
primary differences between these diets: 1) the “induction phase”
of the Atkins diet (20 g of carbohydrates per day) is maintained in-
definitely on the modified Atkins diet (MAD) instead of being grad-
ually increased, 2) high-fat foods are mandatory on MAD (not just
acceptable as with the Atkins diet), and 3) weight loss is not the pri-
mary goal of the MAD as with the Atkins diet. Families that were
following the Atkins book often reported insufficient ketosis and
efficacy, thus the decision to change the name to emphasize these
differences [6].
3. Key questions
3.1. How is the modified Atkins diet different from the traditional classic
ketogenic diet (or low glycemic index treatment)?
The guiding principle of the MAD is to be easy to grasp and imple-
ment (Table 1). As such, there are no formal guidelines for the composi-
tion of the foods eaten, with the exception of carbohydrates [7].
Upon reviewing the food records of patients on the MAD, they receive
an approximate 1–2:1 ketogenic ratio (grams of fat:carbohydrate and
Table 1
Modified Atkins diet protocol (2013).
Initiation
– Baseline laboratory studies (complete blood count, liver and kidney function
tests, electrolytes, fasting lipid profile) and 2-day food record obtained before
beginning the diet
– High-fat foods added liberally
– Carbohydrates: 10 g/day for children, 15 g/day for adolescents, 20 g/day for adults
(net carbohydrates counted = total carbohydrates minus fiber; no subtraction of
sugar alcohols)
– Multivitamin supplement and calcium supplement with vitamin D daily
– If possible, KetoCal® 4:1 — 400-calorie shake daily (2/3 cup powder with 8 oz of
water or one “tetrapak” liquid container) for the first month
– Clear carbohydrate-free fluids encouraged
Monitoring and maintenance (1st month)
– Weight checked weekly
– Ketones checked twice a week for at least the first month
– Carbohydrate counting guides reviewed (from dietitians, books, Internet, apps)
– Avoid anticonvulsant changes
– Avoid low carbohydrate store-bought products
Maintenance (after 1st month)
– May increase net carbohydrates (by 5–10 g/day), change anticonvulsants, use
low-carb products (as deemed necessary)
– Follow-up laboratory studies (including lipid profile) after 3 months
– Urine ketones checked weekly
– Food record at 3 months
Discontinuation
– Increase carbohydrate limit by 10 g/day/week until 60 g/day, then substitute
regular meals, one per week. A quicker titration off may be acceptable in some
circumstances.
protein combined), but it is only approximate and may change day by
day. Therefore, calculating a ketogenic ratio to provide to patients is
not recommended [5]. Fig. 1 demonstrates the differences in composi-
tion between diets. The MAD is taught to patients and/or families in an
outpatient clinic visit lasting 1 h, during which materials are provided,
including recipes, carbohydrate lists, food label explanations, and details
regarding monitoring ketones and possible side effects (Table 2).
Net carbohydrates, calculated by subtracting fiber from total carbo-
hydrates, are limited to 10g/day from age 2 to 12years. There is no spec-
ification on which types of carbohydrates are eaten, unlike the LGIT
which tracks the glycemic indices of carbohydrates; however, carbohy-
drates with lower glycemic indices (e.g., those found in berries and
whole grains) typically allow for higher quantities to be eaten [4]. Ado-
lescents are limited to 15 g/day and adults to 20 g/day. A multivitamin
supplement and calcium supplement with vitamin D are recommended.
Based upon the results from a 2011 prospective clinical trial, we recom-
mend a 400-calorie KetoCal® 4:1 shake during the initial month for
children, but recognizing the potential out-of-pocket expense, other
high-fat supplements (e.g., whipping cream or oils) may be substituted
[8]. If the MAD has led to seizure improvement after 1 month of use, we
then suggest arranging a follow-up with a dietitian and neurologist
along with obtaining laboratory studies at 3 months after initiation.
The MAD differs from the LGIT primarily in regard to the nature
and quantity of carbohydrates provided. The LGIT specifies that only
carbohydrates (approximately 40–60 g/day) with glycemic indices
b50 can be eaten, whereas the MAD does not [4]. However, most
families do eat carbohydrates with lower glycemic indices on the
MAD in order to allow for larger quantities (e.g., several strawberries
versus a very tiny piece of chocolate). The exact composition of the
LGIT, however, is largely similar to the MAD with 60% fat, 30% protein,
and 10% carbohydrates and approximates a 1:1 ketogenic ratio [4].
One major reported difference is the lack of urinary ketosis with the
LGIT [4]. We use the LGIT at our center for patients who cannot tolerate
the MAD (or KD) because of the high fat content, or who have persistent
difficulties with hyperlipidemia or are symptomatic from overketosis
(e.g., vomiting or fatigue).
3.2. Is the modified Atkins diet effective?
The first formal, prospective, open-label study was designed for
children with refractory epilepsy and was funded graciously by the
Dr. Robert C. Atkins Foundation [5]. Twenty children were treated
from September 2003 to May 2005, with 16 (80%) completing the
6-month trial [5]. Using an intent-to-treat analysis, 13 (65%) had N 50%
seizure reduction at 6 months, including 7 (35%) with N 90% improve-
ment. These results were strikingly similar to most studies of the classic
KD, justifying its continued use and study. Interestingly, although the
presence of large urinary ketosis at 1 month correlated with success,
this did not persist at 3 and 6months. This suggests that strict adherence
to the diet during the first month may improve efficacy, but this would
need to be confirmed with subsequent studies.
A second prospective study randomized children to 10 or 20 g of net
carbohydrates per day for 3 months, followed by a crossover to the op-
posite treatment arm [9]. Similar to our 2006 study, it was found that a
stricter carbohydrate limit (10g/day) was more effective, and switching
later (in either direction) did not influence seizure control [9]. The
most recent study to test this “strict first month” hypothesis used a
400-calorie KetoCal® 4:1 shake as a supplement to a 10-g/day MAD
during the first month in an open-label manner [8]. When compared
to historical MAD outcomes to date, this approach was more successful
(80% with N50% seizure reduction including 37% with N90% after
1 month). Again, loosening the restrictiveness (in this case, by elimi-
nating the shakes at 1 month) did not lead to increased seizures. Con-
trary to our prediction, KetoCal® did not appear to boost urinary
ketosis; rather, it increased the daily fat intake and, thus, the ketogenic
ratio to 1.8:1 [8].
 E.H. Kossoff et al. / Epilepsy & Behavior 29 (2013) 437–442 439

8% 2% 6%

35% Legend
30%
50% Fat
Protein
64%
Carbohydrates

90% 15%

Ketogenic Diet Modified Atkins Diet Standard Diet

Fig. 1. Differences between the macronutrient compositions among the diets.

There have been many other studies examining the MAD, which are Lastly, as the MAD has been used for 10 years, we are starting to ob-
listed in Table 3 [10–37]. Several of these publications in particular have tain long-term outcomes from patients who have received it for longer
demonstrated some findings of particular note. Two studies in 2007 and than the 6 months typically reported in clinical trials to date. At a mean
2012 from Korea confirmed the benefit of the MAD, with the first study of 20 months, 30 of 54 (55%) children with long-term MAD duration
suggesting that stable blood ketones over time (beta-hydroxybutyrate maintained N50% seizure reduction [40]. Side effects continued to
of N 3 mmol/L) may be important [10,11]. Dr. Sharma and her team in be mild and predominantly consisted of gastrointestinal discomfort,
India have studied the MAD twice, initially for children with refractory fatigue, and constipation [40].
infantile spasms, using the MAD in children as young as 6 months of
age [12]. Most recently, in the only randomized controlled study of the 3.3. Which patients should receive the modified Atkins diet instead of the
MAD for the treatment of childhood epilepsy, published this year in ketogenic diet?
Epilepsia, it was found that the MAD arm (n = 50) was more likely to
have N 50% seizure reduction than the control arm of standard medical In our opinion, the classic KD is more appropriate than the MAD for
management (n=52) (52% vs. 11.5%, p b 0.001) [13]. Similarly, the like- infants under age 2 years and for those receiving formula-only nutrition
lihood of N 90% seizure reduction was higher (30% vs. 7.7%, p = 0.005). (e.g., gastrostomy tube-fed). However, the decision becomes difficult
This trial followed the general study design of a previous trial by Neal when determining which diet is best to begin in most children ages
et al. of the classic KD, published in 2008 [38]. 2–12 years. As several of the MAD studies suggest that a strict first
Other recent pediatric studies include a pilot series of the MAD for month improves efficacy, a reasonable strategy would be to start with
Sturge–Weber syndrome and a case series treating childhood absence the KD and then switch to the MAD later [8,9]. Some families truly re-
epilepsy [31,32]. An attempt at a direct comparison was performed in quire the supervision and detailed meal plans provided with the KD
France in 2009 and found retrospectively that the KD led to better sei- by dietitians, and the MAD is deemed too variable. In this age range,
zure reduction than the MAD but only after 3 months (not at 6 months) children whom we consider for the MAD before the KD include the
[21]. In a series comprised of patients from 5 countries, retrospective following: 1) children in busy, large families in which sharing food is
data suggested that switching from the MAD to the KD led to a 30% highly important; 2) children with limited tolerance for extremely
chance of additional improvement, in many ways similar to raising the high-fat foods; 3) children with a need for more protein than allowed
dose of an anticonvulsant [39]. Interestingly, the only children in this se- on the KD; 4) children in families in which hospital admission or a
ries who became seizure-free after changing to the KD had myoclonic- fasting period would be problematic because of cost, location, or other
astatic epilepsy (Doose syndrome), indicating that patients with that factors; 5) children with idiopathic generalized epilepsies (e.g., absence
syndrome may respond preferentially to either higher fat, greater keto- epilepsy, juvenile myoclonic epilepsy); and 6) children who must be
sis, or, possibly, calorie limitation. started on treatment urgently (e.g., same day), which can be done in
the outpatient clinic.
Table 2 Even as far back as 2003 in the original case series on the MAD, 4 of
Differences between the ketogenic and the modified Atkins diet. 6 patients were over the age of 12 years, suggesting that the “ideal”
population for this alternative diet may be adolescents and adults [3].
Ketogenic diet Modified Atkins diet
Historically, these patients were not typically offered dietary therapy,
Calories (% recommended Restricted (75%) Unrestricted
although this is changing with the MAD. The first formal prospective
daily allowance) or matched
Fluids (“) Measured to RDA Unrestricted study was published in 2008 using the same study design and criteria
Fat 85%–90% ~60% as the first pediatric study except that the patients enrolled were
Protein 15% ~30% 18 years of age or older [15]. Thirty adults, ages 18–53 years, were
Carbohydrates 5% ~10% started on a 15-g/day net carbohydrate MAD. Most who responded
Fasting period Occasionally done Not used
did so within 2 weeks, including 47% by 1 and 3 months and 33% after
Admission to hospital Typically done No
Meal plans computer-created Yes No 6 months [15]. Surprisingly, weight loss correlated with efficacy at
Foods weighed and measured Yes No (carbs monitored) 3 months, although we are not aware of any other study reporting this
Sharing of food at family meals No Yes observation. The MAD was very well-tolerated in adults, with only a
Ability to eat foods made in restaurants No Yes
slight increase in mean total cholesterol (187 to 201 mg/dL) and an
“Low carbohydrate” store-bought Not used Allowed sparingly
products associated increase in HDL cholesterol, although the latter was not sta-
Intensive education provided Yes No tistically significant.
Dietitian involvement Yes Typically yes Other studies on adults have followed, generally with equivalent
Multiple studies over many years Yes Yes, recently outcomes [18,26,33–35]. We now recommend a net carbohydrate
showing benefits
limit of 20 g/day, which is well-tolerated and leads to similar levels
440 E.H. Kossoff et al. / Epilepsy & Behavior 29 (2013) 437–442

Table 3
Modified Atkins diet studies, 2003–2013, with intent-to-treat analyses at 3–6 months (primarily studies in which the MAD was used first, not as a transition from KD).

Author [Ref] Year Number of patients Prospective or retrospective? Age range N50% seizure reduction N90% seizure reduction Seizure-free
(years)

Kossoff et al. [3] 2003 6 Retrospective 7–52 3 (50%) 3 (50%) 2 (33%)

Kossoff et al. [5] 2006 20 Prospective 3–16 13 (65%) 7 (35%) 3 (15%)
Kossoff et al. [9] 2007 20 Prospective 3–16 10 (50%) 7 (35%) 2 (10%)
Kang et al. [10] 2007 14 Prospective 2–14 5 (43%) 5 (43%) 3 (36%)
Ito et al. [14] 2008 1 Retrospective 7 1 (100%) 1 (100%) 1 (100%)
Kossoff et al. [15] 2008 30 Prospective 18–53 10 (33%) 3 (10%) 1 (3%)
Kossoff et al. [17] 2008 1 Retrospective 2 1 (100%) 1 (100%) 0 (0%)
Carrette et al. [18] 2008 8 Prospective 30–54 1 (13%) 0 (0%) 0 (0%)
Chapman and Cardenas [19] 2008 1 Retrospective 5 1 (100%) 0 (0%) 0 (0%)
Harris et al. [20] 2008 1 Retrospective 5 1 (100%) 1 (100%) 1 (100%)
Weber et al. [16] 2009 15 Prospective 2–17 6 (40%) 2 (13%) 0 (0%)
Porta et al. [21] 2009 10 Prospective 0.4–16 2 (20%) 2 (20%) 0 (0%)
Slaughter et al. [22] 2009 1 Retrospective 10 1 (100%) 0 (0%) 0 (0%)
Kumada et al. [23] 2010 2 Retrospective 4–5 2 (100%) 2 (100%) 2 (100%)
Miranda et al. [24] 2010 33 Prospective 1–16 13 (39%) 6 (18%) 0 (0%)
Kossoff et al. [31] 2010 5 Prospective 4–18 3 (60%) 2 (40%) 1 (20%)
Tonekaboni et al. [25] 2010 51 Prospective 1–16 16 (31%) 6 (12%) 5 (10%)
Kossoff et al. [8] 2011 30 Prospective 3–16 21 (70%) 13 (43%) 3 (10%)
Smith et al. [26] 2011 18 Prospective 18–55 4 (22%) 1 (6%) 0 (0%)
Levy et al. [27] 2011 1 Retrospective 6 1 (100%) 1 (100%) 1 (100%)
Takahashi et al. [28] 2011 1 Retrospective 1 1 (100%) 1 (100%) 1 (100%)
Ito et al. [29] 2011 6 Retrospective 7–17 6 (100%) 6 (100%) 2 (33%)
Groomes et al. [32] 2011 13 Retrospective 4–20 12 (92%) 6 (46%) 2 (15%)
Kumada et al. [30] 2012 10 Prospective 1–17 4 (40%) 3 (30%) 3 (30%)
Sharma et al. [12] 2012 15 Prospective 0.5–3 9 (60%) 8 (53%) 6 (40%)
Cervenka et al. [33] 2012 25 Prospective 18–66 6 (27%) 3 (14%) 1 (5%)
Kim et al. [11] 2012 20 Retrospective 2–17 5 (42%) 3 (25%) 2 (17%)
Kossoff et al. [34] 2013 8 Retrospective 15–44 6 (75%) 4 (50%) 2 (25%)
Sharma et al. [13] 2013 50 Prospective 2–14 26 (52%) 6 (11.5%) 5 (10%)
Ramm-Petersen et al. [35] 2013 2 Retrospective 16–49 2 (100%) 2 (100%) 1 (50%)
El-Rashidy et al. [36] 2013 15 Prospective 2–4 8 (62%) 0 (0%) 0 (0%)
Haberlandt et al. [37] 2013 1 Retrospective 6 1 (17%) 1 (17%) 1 (17%)

Adult patients combined 92 28 (30%) 11 (12%) 3 (3%)

Children combined 342 180 (53%) 100 (29%) 52 (15%)

All patients combined 434 208 (48%) 111 (26%) 55 (13%)

of ketosis. Similar to absence epilepsy in children, the MAD may have a
unique role for another idiopathic generalized epilepsy with seizures
that can persist into adulthood, JME (juvenile myoclonic epilepsy)
[34]. As a direct result of the success of the MAD for adults, our hospital
opened an Adult Epilepsy Diet Center in August 2010, with now
over 100 patients treated to date. Several clinical trials are underway
including dietary treatment of status epilepticus and examining if
KetoCal® shakes are similarly effective and well-tolerated in adults as
they are in children. Other adult epilepsy diet centers are actively
being formed, including one in London [41].
There are some families in which the KD may be preferable to the
MAD. As stated previously, infants under age 2 years are nutritionally
at risk, and the strict supervision and calculation of calories by dieti-
tians with the KD may provide a level of safety not present with
the MAD. This is valid as well for other children who are nutritionally
at risk such as those who are underweight, with poor oral intake,
or with other complex systemic illnesses. In addition, we have no-
ticed that some families become anxious regarding the inherent flex-
ibility and self-creation of meal plans with the MAD and are more
comfortable with the set recipes with the KD. Lastly, for emergent
situations such as new-onset infantile spasms or status epilepticus,
the KD (with a fasting period) may potentially work more quickly
and effectively; therefore, if time is critical, the MAD may not be
adequate.
3.4. What clues does the modified Atkins diet give us into the mechanisms of
action of all dietary therapies?
One of the most scientifically fascinating aspects of research into
alternative diets such as the MAD and LGIT relates to the mechanisms
of action of ketogenic therapies. For decades, and still today, the KD
is hypothesized to work by breaking down fats into ketone bodies,
which can be used for energy [1]. This was believed to be the prima-
ry, if not sole, mechanism of action of the KD. The MAD does not
always achieve long-term high levels of urinary ketosis, and many
children and adults can lose ketosis over time yet have preserved sei-
zure control [5]. It is possible that an initial period of high ketosis
is important or even that low levels of stable ketosis may improve
seizures, but the MAD has raised questions about the classic belief
that persistent, high ketosis is critical. Further study on this issue is
warranted.
The MAD trials also suggest that the first month is the most impor-
tant to achieve success. Data from KD studies have shown this as well,
with improvement tending to occur within 2–4 weeks of onset and
4:1 ketogenic ratios being more effective than lower ratios [42,43].
What occurs during the first month that metabolically “shocks” the
system and improves seizures? This change could be attributed to in-
creased utilization of fat, the effect of elevated ketones, or other meta-
bolic changes that have yet to be identified. If better elucidated, then,
alternative methods to metabolically alter the brain could be devised,
perhaps needing to be used only for short periods of time. Considering
the potential adverse effects of long-term dietary therapy use, effective
short-term dietary therapy may have profound implications on health
and nutrition [44].
In addition, as the MAD becomes more widely used, we suspect that
it will continue to be used for epilepsy indications not traditionally
attempted, such as in idiopathic generalized epilepsies. If determined
to be effective, animal models of these epilepsies could then be tested
with dietary therapy, and the potential molecular targets could be
identified.
 E.H. Kossoff et al. / Epilepsy & Behavior 29 (2013) 437–442
3.5. What is the future of the modified Atkins diet 10 years from now?
Ten years since its introduction, some trends have occurred that
might allow for prediction as to the future of the MAD in 2023. There
is little doubt that the MAD will continue to be utilized in adults with re-
fractory epilepsy, and this age group may even be the primary age group
for which the MAD is used. Adolescents may also be universally started
on this instead of the classic KD.
Early case reports have already suggested that the MAD may have
a unique role in regions of the world, including developing countries,
where the resources of a full ketogenic diet center (including dieti-
tians) may not be available. A 2-year-old boy in Honduras with
Lennox–Gastaut syndrome was successfully treated with the MAD
after materials were translated into Spanish, and his mother was
educated by a local child neurologist [17]. There was also a cost
savings overall for this family; the higher costs of meats and dairy
in this country were offset by the reduction in costs for anticonvul-
sants. Similar projects are underway in India, China, and other coun-
tries. A prospective clinical trial for adults with epilepsy using solely
email correspondence with a neurologist (no dietitian) demonstrated
feasibility and preliminary efficacy [33]. Telemedicine may improve
the efficacy of “remote” dietary treatment and provide access to
this treatment for patients unable to travel long distances to a KD
center.
As further trials of the MAD continue, we suspect that the incidence
of adverse effects, especially the long-term ones, may be less than with
the KD because of the inherent higher protein content. This has not been
shown to date, but there is evidence in one study [40]. If established,
children receiving the KD beyond 1–2 years may be counseled to switch
to the MAD in cases where prolonged therapy is indicated. For children
with GLUT-1 deficiency who need long-term dietary treatment, the
MAD could prove very helpful [29].
In addition, if the MAD continues to be demonstrated as similarly
effective, yet less restrictive and safer than the KD, it would be a log-
ical choice for trials of new-onset epilepsies. Families and patients
are often averse to perceived risks of anticonvulsants, and although
the MAD is not free of adverse effects, it may be seen as a more
“natural” approach by many. Trials would perhaps specify a time
limit (e.g., 1–2 months) after which the MAD would be discontinued
and an anticonvulsant started if seizure control is not achieved. In
addition, nonepilepsy uses under investigation (e.g., brain tumors,
Alzheimer's disease, ALS, and migraines) where a strict KD may be
beneficial, yet a hospital admission and restrictive meal plans would
negatively impact quality of life, could be tested with the MAD instead
[45–48].
4. Summary
Now, 10 years after its development, it is clear that the MAD
is effective and offers a comparable alternative for those consid-
ering dietary therapy but reluctant to implement the classic
KD. Research provides clues to the underlying mechanisms of di-
etary therapies and suggests that an intense metabolic alteration
through increased fat metabolism during the first month may
be critical. Studies also suggest that this less restrictive diet can pre-
serve efficacy and improve tolerability and health of the children
and adults receiving dietary treatment. Finally, the MAD provides
exciting new applications for dietary therapies, including those in
adults, in other neurologic diseases, and in regions with limited
resources.
Disclosures
Dr. Kossoff is on the Scientific Advisory Board for Atkins Nutritionals,
Inc. Dr. Kossoff, Dr. Cervenka, and Ms. Henry receive grant support for a
clinical trial involving KetoCal© by Nutricia, Inc.
441
Key questions (answered)
1. How is the modified Atkins diet different from the traditional
classic ketogenic diet?
The MAD composition is also carbohydrate-restricted but
requires slightly less fat. The primary difference is the lack
of restrictions on protein, calories, and fluids, as well as
the outpatient, nonfasting initiation period.
2. Is the modified Atkins diet effective?
Yes, with 10 years of use, there are 423 children and adults
reported in 31 studies from multiple centers who have un-
dergone MAD. When added together, 187 (47%) patients
overall have had a N50% seizure reduction, which is compa-
rable to the results found for the ketogenic diet.
3. Which patients should receive the modified Atkins diet
instead of the ketogenic diet?
For children under the age of 2 years and those receiving
formula-only nutrition, the ketogenic diet is preferable.
For adolescents and adults, the MAD is probably the better
option. In children in other age groups, diet prescription is
at the discretion of their neurologist and dietitian, as well
as their parents.
4. What clues does the modified Atkins diet give us into the
mechanisms of action of all dietary therapies?
Data suggest that a strict, highly ketotic, high-fat initiation
period (probably 1 month) may be critical. Long-term high
levels of ketosis may not be required. Perhaps a short-term,
intense, metabolic alteration is one mechanism by which
dietary therapy is effective.
5. What is the future of the modified Atkins diet 10 years from
now?
We suspect that the MAD will be used primarily for adults and
adolescents and that children receiving long-term KD will be
transitioned to it. There may be additional utility in regions
of the world with limited resources, as a first-line therapy for
some forms of epilepsy, and even for nonepilepsy indications.
References
[1] Kossoff EH, Hartman AL. Ketogenic diets: new advances for metabolism-based
therapies. Curr Opin Neurol 2012;25:173–8.
[2] Wilder RM. The effect of ketonemia on the course of epilepsy. Mayo Clin Bull
1921;2:307–8.
[3] Kossoff EH, Krauss GL, McGrogan JR, Freeman JM. Efficacy of the Atkins diet as
therapy for intractable epilepsy. Neurology 2003;61:1789–91.
[4] Pfeifer HH, Thiele EA. Low-glycemic-index treatment: a liberalized ketogenic diet for
treatment of intractable epilepsy. Neurology 2005;65:1810–2.
[5] Kossoff EH, McGrogan JR, Bluml RM, Pillas DJ, Rubenstein JE, Vining EP. A modified
Atkins diet is effective for the treatment of intractable pediatric epilepsy. Epilepsia
2006;47:421–4.
[6] Atkins RC. Dr. Atkins' new diet revolution. New York: M. Evans and Company; 1992.
[7] Kossoff EH, Dorward JL. The modified Atkins diet. Epilepsia 2008;49(Suppl. 8):37–41.
[8] Kossoff EH, Dorward JL, Turner Z, Pyzik PL. Prospective study of the modified Atkins
diet in combination with a ketogenic liquid supplement during the initial month.
J Child Neurol 2011;26:147–51.
[9] Kossoff EH, Turner Z, Bluml RM, Pyzik PL, Vining EPG. A randomized, crossover com-
parison of daily carbohydrate limits using the modified Atkins diet. Epilepsy Behav
2007;10:432–6.
[10] Kang HC, Lee HS, You SJ, Kang DC, Ko TS, Kim HD. Use of a modified Atkins diet in
intractable childhood epilepsy. Epilepsia 2007;48:182–6.
[11] Kim YM, Vaidya VV, Khusainov T, Kim HD, Kim SH, Lee EJ, et al. Various indications for
a modified Atkins diet in intractable childhood epilepsy. Brain Dev 2012;34:570–5.
[12] Sharma S, Sankhyan N, Gulati S, Agarwala A. Use of the modified Atkins diet in infan-
tile spasms refractory to first-line treatment. Seizure 2012;21:45–8.
[13] Sharma S, Sankhyan N, Gulati S, Agarwala A. Use of the modified Atkins diet for
treatment of refractory childhood epilepsy: a randomized controlled trial. Epilepsia
2013;54:481–6.
[14] Ito S, Oguni H, Ito Y, Ishigaki K, Ohinata J, Osawa M. Modified Atkins diet therapy for
a case with glucose transporter 1 deficiency syndrome. Brain Dev 2008;30:226–8.
[15] Kossoff EH, Rowley H, Sinha SR, Vining EP. A prospective study of the modified
Atkins diet for intractable epilepsy in adults. Epilepsia 2008;49:316–9.
442 E.H. Kossoff et al. / Epilepsy & Behavior 29 (2013) 437–442
[16] Weber S, Molgaard C, Taudorf K, Uldall P. Modified Atkins diet to children and ado-
lescents with medical intractable epilepsy. Seizure 2009;18:237–40.
[17] Kossoff EH, Dorward JL, Molinero MR, Holden KR. The modified Atkins diet: a poten-
tial treatment for developing countries. Epilepsia 2008;49:1646–7.
[18] Carrette E, Vonck K, de Herdt V, Dewaele I, Raedt R, Goossens L, et al. A pilot trial
with modified Atkins' diet in adult patients with refractory epilepsy. Clin Neurol
Neurosurg 2008;110:797–803.
[19] Chapman K, Cardenas JF. Hemimegalencephaly in a patient with a neurocutaneous
syndrome. Semin Pediatr Neurol 2008;15:190–3.
[20] Harris ML, Patel H, Garg BP. Intractable seizures, developmental delay, and the keto-
genic diet. Semin Pediatr Neurol 2008;15:209–11.
[21] Porta N, Vallee L, Boutry E, Fontaine M, Dessein AF, Joriot S, et al. Comparison of sei-
zure reduction and serum fatty acid levels after receiving the ketogenic and modi-
fied Atkins diet. Seizure 2009;18:359–64.
[22] Slaughter L, Vartzelis G, Arthur T. New GLUT-1 mutation in a child with treatment-
resistant epilepsy. Epilepsy Res 2009;84:254–6.
[23] Kumada T, Miyajima T, Kimura N, Saito K, Shimomura H, Oda N, et al. Modified
Atkins diet for the treatment of nonconvulsive status epilepticus in children. J
Child Neurol 2010;25:485–9.
[24] Miranda MJ, Mortensen M, Povlsen JH, Nielsen H, Beniczky S. Danish study of a
modified Atkins diet for medically intractable epilepsy in children: can we achieve
the same results as with the classical ketogenic diet? Seizure 2011;20:151–5.
[25] Tonekaboni SH, Mostaghimi P, Mirmiran P, Abbaskhanian A, Abdollah Gorji F,
Ghofrani M, et al. Efficacy of the Atkins diet as therapy for intractable epilepsy in
children. Arch Iran Med 2010;13:492–7.
[26] Smith M, Politzer N, Macgarvie D, McAndrews MP, Del CM. Efficacy and tolerability
of the modified Atkins diet in adults with pharmacoresistant epilepsy: a prospective
observational study. Epilepsia 2011;52:775–80.
[27] Levy Y, Peleg-Weiss L, Goldberg-Stern H. The modified Atkins diet for intractable
epilepsy may be associated with late-onset egg-induced anaphylactic reaction:
a case report. Nutrition 2011;27:380–2.
[28] Takahashi H, Imai K, Takayama R, Mine J, Ohtani S, Ikeda H, et al. Remarkable effect
of a modified ketogenic diet in a boy with focal seizures followed by epileptic
spasms in a cluster. No To Hattatsu 2011;43:305–8.
[29] Ito Y, Oguni H, Ito S, Oguni M, Osawa M. A modified Atkins diet is promising as a
treatment for glucose transporter type 1 deficiency syndrome. Dev Med Child
Neurol 2011;53:658–63.
[30] Kumada T, Miyajima T, Oda N, Shimomura H, Saito K, Fujii T. Efficacy and tolerability
of modified Atkins diet in Japanese children with medication-resistant epilepsy.
Brain Dev 2012;34:32–8.
[31] Kossoff EH, Bosarge JL, Comi AM. A pilot study of the modified Atkins diet for
Sturge–Weber syndrome. Epilepsy Res 2010;92:240–3.
[32] Groomes LB, Pyzik PL, Turner Z, Dorward JL, Goode VH, Kossoff EH. Do patients with
absence epilepsy respond to ketogenic diets? J Child Neurol 2011;26:160–5.
[33] Cervenka MC, Terao NN, Bosarge JL, Henry BJ, Klees AA, Morrison PF, et al. Email
management of the modified Atkins diet for adults with epilepsy is feasible and ef-
fective. Epilepsia 2012;53:728–32.
[34] Kossoff EH, Henry BJ, Cervenka MC. Efficacy of dietary therapy for juvenile myoclonic
epilepsy. Epilepsy Behav 2013;26:162–4.
[35] Ramm-Pettersen A, Nakken KO, Skogseid IM, Randby H, Skei EB, Bindoff LA, et al.
Good outcome in patients with early dietary treatment of GLUT-1 deficiency syn-
drome: results from a retrospective Norwegian study. Dev Med Child Neurol
2013;55:440–7.
[36] El-Rashidy OF, Nassar MF, Abdel-Hamid IA, Shatla RH, Abdel-Hamid MH, Gabr SS,
et al. Modified Atkins diet vs classic ketogenic formula in intractable epilepsy. Acta
Neurol Scand 2013. http://dx.doi.org/10.1111/ane.12137.
[37] Haberlandt E, Karall D, Jud V, Baumgartner SS, Zotter S, Rostasy K, et al. Glucose
transporter type 1 deficiency syndrome effectively treated with modified Atkins
diet. Neuropediatrics 2013. http://dx.doi.org/10.1055/s-0033-1349225 [in press].
[38] Neal E, Chaffe H, Schwartz R, Lawson MS, Edwards N, Fitzsimmons G, et al. The ke-
togenic diet for the treatment of childhood epilepsy: a randomised controlled trial.
Lancet Neurol 2008;7:500–6.
[39] Kossoff EH, Bosarge JL, Miranda MJ, Wiemer-Kruel A, Kang HC, Kim HD. Will seizure
control improve by switching from the modified Atkins diet to the traditional keto-
genic diet? Epilepsia 2010;51:2496–9.
[40] Chen W, Kossoff EH. Long-term follow-up of children treated with the modified
Atkins diet. J Child Neurol 2012;27:754–8.
[41] Cervenka MC, Henry B, Nathan J, Wood S, Volek JS. Worldwide dietary therapies for
adults with epilepsy and other disorders. J Child Neurol 2013;28:1034–40.
[42] Kossoff EH, Laux LC, Blackford R, Morrison PF, Pyzik PL, Hamdy RM, et al. When do
seizures improve with the ketogenic diet? Epilepsia 2008;49:329–33.
[43] Seo JH, Lee YM, Lee JS, Kang HC, Kim HD. Efficacy and tolerability of the ketogenic
diet according to lipid:nonlipid ratios—comparison of 3:1 with 4:1 diet. Epilepsia
2007;48:801–5.
[44] Groesbeck DK, Bluml RM, Kossoff EH. Long-term use of the ketogenic diet in the
treatment of epilepsy. Dev Med Child Neurol 2006;48:978–81.
[45] Maroon J, Bost J, Amos A, Zuccoli G. Restricted calorie ketogenic diet for the treat-
ment of glioblastoma multiforme. J Child Neurol 2013;28:1002–8.
[46] Henderson ST. Ketone bodies as a therapeutic for Alzheimer's disease.
Neurotherapeutics 2008;5:470–80.
[47] Paganoni S, Wills AM. High-fat and ketogenic diets in amyotrophic lateral sclerosis.
J Child Neurol 2013;28:989–92.
[48] Kossoff EH, Huffman J, Turner Z, Gladstein J. Use of the modified Atkins diet for
adolescents with chronic daily headache. Cephalalgia 2010;30:1014–6.