BREASTFEEDING MEDICINE

Volume XX, Number XX, 2019

ª Mary Ann Liebert, Inc.
DOI: 10.1089/bfm.2019.0190

The Ketogenic Diet Including Breast Milk for Treatment

of Infants with Severe Childhood Epilepsy:
Feasibility, Safety, and Effectiveness

Anastasia Dressler,1 Chiara Häfele,1 Vito Giordano,1 Franz Benninger,2 Petra Trimmel-Schwahofer,1
Gudrun Gröppel,1 Sharon Samueli,1 Martha Feucht,1 Christoph Male,1 and Andreas Repa1
Downloaded by Boston University package from www.liebertpub.com at 12/02/19. For personal use only.

Abstract

Objective: The ketogenic diet (KD) is a high-fat and restricted carbohydrate diet for treating severe childhood
epilepsy. In infants, breast milk is usually fully replaced by a ketogenic formula. At our center, mothers are
encouraged to include breastfeeding into the KD if still breastfeeding. This retrospective study describes
achievement and maintenance of ketosis with or without inclusion of breast milk.
Methods: Data were retrieved from a prospective longitudinal database of children treated with KD for
epilepsy analyzing infants <1 year of age. The time to achieve clinically relevant ketosis (‡2 mmol/L beta-
hydroxybutyrate) was compared with and without inclusion of breast milk into standard KD. Ketosis, nutritional
intakes, effectiveness, adverse effects, and successful continuation of breastfeeding were evaluated.
Results: A total of 79 infants were eligible for analysis. In 20% (16), breast milk was included. Infants with
breast milk included into the KD achieved relevant ketosis in 47 hours (interquartile range [IQR] 24–95)
compared with 41 hours (IQR 22–70; p = 0.779) in infants with standard KD. Beta-hydroxybutyrate at day 2
was 3.1 mmol/L (IQR 0.5–4.9) and 3.8 mmol/L (IQR 2.2–4.9). Infants with breast milk included received higher
amounts of carbohydrates at baseline and calories at 3 months. Seizure freedom and adverse effects showed no
relevant differences. No infections occurred in infants receiving breast milk. In two infants, KD was initiated
with breast-feds after bottle-feeding KD formula. In 31%, breastfeeding was continued after the KD, and in
25%, inclusion of breast milk and breastfeeding was maintained until complete weaning. Before discharge from
hospital, the amount of breast milk included was median 90 mL/day (IQR 53–203) equivalent to median 9%
(IQR 6–15).
Conclusions: Appropriate ketosis was achieved in most infants and maintained within 48 hours. Incorporation
of breast milk into KD is feasible, safe, and effective.

Keywords: ketogenic diet, breast milk, infants, epilepsy, beta-hydroxybutyrate

Introduction life or as central part of a mixed diet later on. To induce and
maintain ketosis in KD, only a restricted amount of carbo-

T he ketogenic diet (KD) is a high-fat, low-carbohydrate

diet1,2 used for the therapy of drug-resistant childhood
epilepsy.3 With increasing knowledge on the high efficiency of
the neonatal brain to metabolize ketone bodies4,5 and avail-
ability of specialized ketogenic infant formulas, the KD is
increasingly used for the management of intractable epilepsy
already in the first year of life.6–9
Exclusive breastfeeding is the optimal form of nutrition
during infancy, either exclusively during the first 6 months of
hydrates is tolerated and breast feeding is usually fully re-
placed by a specialized high-fat ketogenic formula.10 In view
of the considerable physical11–17 and mental18 health ad-
vantages of mother’s milk, it nevertheless seems desirable to
allow for incorporation of breastfeeding into the KD, the
more in a group of patients particularly vulnerable to physical
and mental comorbidities.19,20
Until now, there are three case series on the inclusion of
breast milk into KD.21–23 At our center, both KD with and

Departments of 1Pediatrics and Adolescent Medicine and 2Child and Adolescent Neuropsychiatry, Medical University Vienna,
Vienna, Austria.

1
 2 DRESSLER ET AL.
without inclusion of mother’s milk are well established.
Studies that directly compared the effectiveness of a KD with
and without breast milk are still lacking. The objective of our
study was to retrospectively analyze the feasibility, effec-
tiveness, and safety of a KD regimen that includes breast milk
compared to standard KD without breastfeeding in infants
younger than 1 year of age.
Methods
Study design
This study is a retrospective database analysis of infants
with childhood epilepsy that were treated with a KD at our
clinic (Department of Pediatrics, Medical University of
Vienna, Austria; September 1999 to December 2018). The
KD program at the study center was started in 1999 in tod-
dlers and older children. From 2006 onward, when a ready-
Downloaded by Boston University package from www.liebertpub.com at 12/02/19. For personal use only.
made ketogenic formula was available in Europe, infants
younger than 1 year of age were also included. The study
compared conventional KD without breastfeeding with a KD
that permitted ingestion of breast milk. Eligible participants
were infants <12 months of age fed with an exclusive or
mixed liquid diet, but without parenteral nutrition. The pri-
mary outcome was time to achieve clinically relevant ketosis.
Important secondary outcomes were treatment responder rate
and seizure freedom at 3 months and last follow-up visit and
the z-score of weight and height after 3 months. Moreover,
continuation of breastfeeding, nutritional intake, growth, and
adverse effects were analyzed.
Ketogenic diet
The KD was initiated without prior fasting or fluid re-
striction.24 Nursing mothers were invited to include their
breast milk into the KD. Before initiation of the diet, a
thorough neurologic and pediatric examination, video-
electroencephalogram, and fasting blood samples were ob-
tained. After consent, the KD was calculated by a trained
nutritionist using a computer-based algorithm (Supplemen-
tary Fig. S1). The KD was started at a fat/nonfat ratio of 1:1.24
A ketogenic formula (Ketocal; Nutricia Metabolics, Er-
langen, Germany; fat/nonfat ratio of 3:1 or 4:1) was therefore
mixed with either conventional infant formula (fat/nonfat
ratio of 0.4:1) or expressed breast milk. For mature breast
milk, an average amount of fat, proteins, and carbohydrates
was used for calculation (100 mL containing 75 calories,
4.03 g of fat, 1.13 g of proteins, and 7 g of carbohydrates,
equating to a fat/nonfat ratio of 0.4–0.6:1).25,26 To optimize
the fat content and to maximize the amount of breast milk,
mothers were encouraged to provide pumped hind milk. The
amounts of ketogenic formula were then increased stepwise
to reach a fat:nonfat ratio of 3:1 (maximum)—depending on
beta-hydroxybutyrate levels (2–5 mmol/L). When beta-
hydroxybutyrate levels were lower or within this range,
fat/nonfat ratio was increased on a daily basis in both groups.
After reaching stable ketosis, the fat:nonfat ratio was lowered
while maintaining ketosis to allow for a maximum amount of
ingested breast milk (or formula milk). In the breastfeeding
group, transition from pumped mother’s milk to breastfeed-
ing was encouraged. In this study, infants were breastfed after
bottle-feeding the calculated amount of ketogenic formula.
Plasma glucose, beta-hydroxybutyrate, and urine ketone levels
were measured three times a day while KD was established. At
home, urine ketone levels, seizures, nutrition, and adverse
effects were documented in a patient’s diary. At each outpa-
tient visit, a thorough pediatric, neurological, and nutritional
analysis was performed. Nutritional parameters (fluids, total
calories, protein, lipid, and carbohydrate intake) and growth
parameters (weight, height, and cranial circumference) were
assessed. At outpatient visits, blood analyses included fasting
plasma glucose and beta-hydroxybutyrates according to a
standardized protocol.3,9,27 Outpatient visits were performed
after 1 month and after 3, 6, and 12 months after KD start. For
infants not receiving the KD at 3 months, last follow-up on the
KD was taken for analysis.
The clinical data of infants with inclusion of breast milk
into their KD are displayed in Supplementary Table S1.
A detailed individual prescription of a KD that includes
breast milk is given in Supplementary Table S2.
Data collection
Data were retrieved from a database of pediatric patients
treated with KD. Patients were prospectively enrolled and
data entered from medical records and patient’s diaries
(collected at 1, 3, 6, and 12 months after KD start). The time
to clinically relevant ketosis was defined as time (hours) until
‡2 mmol/L blood beta-hydroxybutyrate27 was reached. The
data with statistical analysis are reported in the text (see
Results section) and the time course of beta-hydroxybutyrate
blood levels (first 5 days and at 1 and at 3 months of therapy)
was plotted and reported with the proportion of infants with
clinically relevant ketosis (‡2 mmol/L beta-hydroxybutyrate)
in Figure 1. The baseline characteristics were recorded as
shown in Table 1. Nutritional parameters and growth were
documented as shown in Table 3. The z-score differences
from initiation of KD to the end of study were calculated and
reported in the text (Results section). The effectiveness of
therapy is shown in Table 2. Treatment response was de-
fined as a seizure reduction >50% compared to baseline
(seizure count through 3 months before initiation of the
KD), seizure freedom was defined 100% seizure reduction
compared to baseline, if maintained through to the last
follow-up visit. Adverse effects were recorded using an
adverse effects diary and recorded as shown in Table 4. Low
fluid intake was defined as intake 30% below recommen-
dations,28 growth deficit as a z-score loss of body weight of >
-0.8 and excessive weight gain as z-score increase of >1.6.
A sample of a prescription of a typical patient’s KD that
included breast milk is shown with the amounts of breast
milk in Supplementary Table S2. For data entry into the
database parental consent was obtained. Data collection and
analysis were approved by the Ethics Committee of the
Medical University of Vienna (EK.-Nr. No. 1391/2013; EK-
Nr. 1591/2019).
Data analysis
Statistic data description was used as appropriate (fre-
quency, median, interquartile range [IQR], range). For com-
parisons between groups, medians, odds ratios, confidence
intervals, and Pearson’s chi-square, k-sample median test and
the median difference estimator (Hodge–Lehman) were used
as appropriate. For infants not receiving the KD at 3 months,
the last follow-up visit on the KD was used for analysis. Data
 BREAST MILK AND THE KETOGENIC DIET 3
Downloaded by Boston University package from www.liebertpub.com at 12/02/19. For personal use only.

FIG. 1. Ketone levels and relevant ketosis (mmol/L). The time course of ketosis is shown as the trajectories of the serum
beta-hydroxybutyrate levels (day 1 to 5 and at 1 and 3 months; median and standard deviation) and the percentage of infants
achieving a clinically relevant ketosis (>2 mmol/L beta-hydroxybutyrate) ( p = 0.008 at day 3). *p < 0.01, tested using the
chi-square test.

analysis was performed using the IBM Statistical Package Results

for Social Science (SPSS Statistics Version 25). The sig-
nificance level was set at p £ 0.05. The hypothesis to be in-
vestigated was that the time to achieve clinically relevant
ketosis was similar in infants in whom breast milk was
continued during treatment with a KD compared with infants
treated with KD alone.
Patient characteristics
A total of 201 children were screened and 83 patients
£1 year of age were identified. Four infants were excluded
(KD started while receiving parenteral nutrition). The whole
final cohort consisted of 79 participants. The age when the

Table 1. Baseline Characteristics

Parameter Breast milk (n = 16) Formula only (n = 63) p
Female 6 (38) 26 (41) 1.0
Etiology known 12 (75) 43 (68) 0.76
Age at KD start (months) 5.3 [1.1–8.2] 6.7 [4.8–9] 0.82
Age at epilepsy onset (months) 2.2 [0.6–6] 3.1 [1.1–5.4] 0.91
Epilepsy onset before KD (days) 24 [15–79] 69 [33–139] 0.18
Drug-naive 2 (13) 8 (13) 1.0
Number of AEDs 2 [1–2.75] 2 [1–3] 0.65
z-score weight 0.5 [-0.3 to 0.9] - 0.5 [-1.2 to 0.3] 0.04
z-score height 0.8 [-0.6 to 1.7] 0.2 [-1.1 to 0.6] 0.15
Categorical data are presented as numbers with percentages in parentheses (tested using the chi-square test). Continuous data are
presented as the median and interquartile range in squared brackets (tested using the k-sample median test).
AEDs, antiepileptic drugs; KD, ketogenic diet.
 4 DRESSLER ET AL.
Table 2. Effectiveness
Breast milk
Parameter (n = 16)
Treatment response (at 3 months) 10 (67)a
Treatment response (at final follow-up visit) 10 (67)a
Seizure freedom (at 3 months) 9 (40)a
Seizure freedom (at final follow-up visit) 8 (53)a
Categorical data are presented as numbers with percentages in parentheses (tested using the chi-square test), and the odds ratio between
groups with 95% CIs in parentheses.
a
Data of one patient missing.
b
Data of two patients missing.
CI, confidence interval.
KD was started was 6.2 months (median; IQR: 4.4–8.4
months; range: 14.6 days–12.0 months) with a majority of
infants (63/79; 80%) already receiving exclusively formula
Downloaded by Boston University package from www.liebertpub.com at 12/02/19. For personal use only.
and less infants being breastfed (16/79; 20%; exclusively: 10/
79; 12.7%; partially: 6/79; 7.6%). Upon invitation to include
breast milk into their infants KD, all mothers decided to do
so. Infants on formula received standard KD by default. More
infants were male (47/79; 69%). The etiology of epilepsy was
unknown in every third infant (31/79; 30%). The KD was the
first antiepileptic therapy in a small minority of infants (10/
79; 13%): most infants (69/79; 87%) received antiepileptic
drugs (AED) before KD initiation. In this study, the amount
of concomitant AEDs at initiation of KD was two (median;
range: IQR: 1–2, range: 0–4). The amount of previously used
AEDs until the KD was started was two (median; IQR 1–3,
range: 0–12). The median duration of KD was 10.7 months
(IQR 5.7–20.5 months; range: 0.4–74). Baseline character-
istics of participants by group are shown in Table 1. The z-
scores of body weight at baseline were significantly higher in
infants who received breast milk. Other parameters did not
differ significantly. The median duration of KD was 10.1
months (IQR: 7.0–25.8 months).
Ketosis and treatment effectiveness
The time to achieve clinically relevant ketosis did not differ
significantly between groups ( p = 0.78). Ketosis was reached in
infants with breast milk after 47 hours (median; IQR: 24–95
hours), in infants without breast milk after 41 hours (median;
IQR: 22–70 hours). The proportion of infants with a beta-
hydroxybutyrate level >2 mmol/L (defined as clinically rele-
vant) was significantly lower in infants with inclusion of breast
milk on the third day of the KD only (Fig. 1). Two infants in the
group without breast milk failed to achieve clinically rele-
vant ketosis (beta-hydroxybutyrate levels: 0.2–0.9 and 0.2–
1.8 mmol/L) and three infants achieved clinically relevant
ketosis as late as after 8, 9, and 15 days, because of medications
containing small amounts of carbohydrates, which was detected
and resolved. The time-course of beta-hydroxybutyrate levels
was not significantly different between both groups (Fig. 1). The
KD was equally effective in both groups. About two thirds of
infants showed seizure reduction by more than half, and in about
a third of infants, seizures ceased completely (Table 2).
Nutritional intake and growth
Infants who were breastfed at baseline received a lower
fat:nonfat ratio and significantly more carbohydrates (Table 3).
Formula only
(n = 63) Odds ratio and 95% CI p
44 (70) 0.9 (0.3–2.9) 1.00
44 (72)b 0.8 (0.2–2.4) 0.75
20 (27) 3.2 (1–10.3) 0.07
22 (36)b 2 (0.7–6.3) 0.25
After KD was implemented, infants who received breast milk
in their KD consumed significantly more energy (difference
between medians of 7 kcal, p = 0.044). Individual data of
infants who received a KD with breast milk are displayed in
Supplementary Table S1. The main type of epilepsy was
West Syndrome (14/16) of varying etiology. On the first day
of the KD, infants who also received breast milk received
144 mL of mothers’ milk (median; IQR 90–307 mL/day,
min. 34 mL–max. 600 mL), which was 30% (median; IQR 6–
50%) of liquid meals (fat/nonfat ratio: 1.83:1; median; IQR
1.5–2.5). On day 3, the amount of breast milk included was
90 mL (median; IQR 53–203 mL/day, min 24 mL–max.
260 mL), which was 9% (IQR 6–15%, range 3–38%) of
liquid meals. The fat/nonfat ratio of liquid meals was median
2.48:1 (IQR 2.01–2.6). The median duration of including
breast milk was 127 days (IQR: 17–166 days; range: 7–482
days). An example of a ketogenic prescription is given in
Supplementary Table S1.
Growth (z-score difference for weight and height from
baseline to last follow-up) was not significantly different
between the two groups: difference z-score of weight: 0.08
versus 0.25 (median, 95% CI: -0.28 to 0.76; p = 0.82), dif-
ference z-score of height: -0.22 versus 0.48 (median, 95%
CI: -0.41 to 1.22; p = 0.43).
Feasibility
Eight out of 16 (50%) mothers who started to include
breast milk into the KD continued to provide breast milk for
more than 3 months. Five mothers (31%) provided breast
milk less than a month (stopped after 7, 8, 10, and 10 and 19
days). Five mothers (31%) managed to nurse their infants on
the KD, with three mothers who switched to breastfeeding
after pumping milk at the start of KD and two mothers who
nursed from the first day of the KD. Four mothers (25%)
provided breast milk until weaning without switching to in-
fant formula.
Adverse effects
There was no significant difference in total adverse effects
(Table 4). Constipation and low fluid intake was as the most
frequent adverse effect in both groups. Intermittent high tri-
glycerides at 3 months were observed in up to 30% of infants.
Discussion
The KD is a well-established diet for severe early-onset
epilepsy syndromes.3,9 The present article reports for the first
 BREAST MILK AND THE KETOGENIC DIET 5

Table 3. Nutritional Parameters

Estimator median difference
Parameter Breast milk (n = 16) Formula only (n = 63) and 95% CI
At start
Ratio (fat:nonfat) 2.5 [2.1–2.5] 2.8 [2.5–3.0]a 0.4 (0–0.5)a
Fluids per kg per day (mL) 126 [118–152] 123 [99–144] -7.2 (-28 to 8.7)
Calories per kg per day (kcal) 98 [88–115] 92 [84–107] -6.5 (-18 to 3.4)
Proteins per kg per day (g) 2 [1.9–2.5] 2 [1.6–2.3] -0.1 (-0.4 to 0.1)
Lipids per kg per day (g) 9.3 [8.1–11] 8.8 [7.8–10] -0.6 (-1.6 to 0.5)
Carbohydrates per kg per day (g) 1.9 [1.5–2.2]b 1.1 [1–1.8]b -0.6 (-0.9 to -0.2)b
After 3 months
Ratio (fat/nonfat) 2.5 [2.0–2.9] 2.9 [2.0–3.0] 0.1 (0–0.5)
Fluids per kg per day (mL) 113 [97–135] 98 [82–115] -16 (-33 to 0.7)
Calories per kg per day (kcal) 91 [80–106]a 84 [74–93]a -8.8 (-19 to 0.1)a
Proteins per kg per day (g) 1.8 [1.6–2.4] 1.7 [1.6–2.0] -0.2 (-0.4 to 0.07)
Lipids per kg per day (g) 8.3 [7.4–9.4] 7.7 (6.8–8.8) -0.6 (-1.5 to 0.3)
Carbohydrates per kg per day (g) 1.5 [1.2–2.5] 1.2 (0.9–1.9) -0.3 (-0.8 to 0.03)
Downloaded by Boston University package from www.liebertpub.com at 12/02/19. For personal use only.

Z-score weight 0.6 [-1.3 to 0.9] -0.1 (-1.3 to 0.9) -0.3 (-1.2 to 0.6)
Z-score height 0.4 [-0.3 to 1.3] 0.2 (-1.4 to 1.1) -0.3 (-1.4 to 0.5)
Continuous data are presented as medians with interquartile range in squared brackets (tested using the k-sample median test) and median
estimator between groups (Hodges–Lehman) with 95% CIs in parentheses.
a
p < 0.05.
b
p < 0.01.
CI, confidence interval.

time the feasibility, effectiveness, and safety of a KD that intake after 3 months of therapy was observed in infants with
includes breast milk compared to the conventional exclu- KD where breast milk was included.
sively formula-based approach in infants. Clinically relevant Effectiveness of the KD at 3 months was high in both
ketosis was typically achieved within the first 48 hours of groups: infants with breast milk showed response in 67% and
therapy and maintained throughout the KD. A higher caloric seizure freedom in 40%; infants without breast milk showed
response in 70% and seizure freedom in 27%. The most
Table 4. Adverse Effects frequent adverse effects were low fluid intake and constipa-
tion in both groups. However, during the first 3 months no
Breast milk Formula infections occurred in infants receiving breast milk compared
Parameter (n = 16) only (n = 63)
to infants without breast milk included. The microbiome in
Any adverse effect 13 (81) 50 (79) mother’s breast milk and areolar skin shapes the infant’s gut
Tirednessa 1 (6) 5 (8) in early life, underscoring the importance of breastfeeding in
Beta-hydroxybutyrate 1 (6) 14 (22) the maturation the infants gut microbiome, which has im-
high (>5 mmol/L)b plications on infections and immunity.29,30
Hypoglycemia (<45 mg/dL)a 0 (0) 3 (5) In addition, the higher carbohydrate intake and lower
Glucose low (45–59 mg/dL)a 1 (6) 7 (11) fat/nonfat ratio was reflected by a lower percentage of hypo-
Low fluid intakeb 6 (38) 23 (37) glycemia and high lipids in infants with breast milk included.
Vomitingb 3 (19) 4 (6)
Also, a higher percentage of KD refusal and constipation seen
Intravenous liquids for 3 (19) 5 (8)
low fluid intakea in this group might reflect a flavor preference for breast milk
Refusal of KDb 2 (13) 3 (5) due to the mixed feeding of KD formula and breast milk and a
Solid food refusalb 1 (6) 17 (27) greater nutritional change.
Triglycerides high 2 (13) 20 (32) In two infants, KD was initiated with feedings at the breast
(>150 mg/dL)b after bottle-feeding the KD formula, and in five infants,
High cholesterol 0 (0) 3 (5) breastfeeding was successfully continued after the KD was
(>200 mg/dL)b stopped.
Constipationb 8 (50) 19 (30) Breastfeeding is the normative standard for infant nutrition,
Diarrheab 0 (0) 4 (6) providing healthy growth and development, recommended
Cholecystolithiasisb 1 (6) 1 (2)
exclusively during the first 6 months and advised to be
Infectionsb 0 (0) 9 (14)
Carnitine deficiencyb 0 (0) 2 (3) continued up to 1 year of age or beyond while gradually
Kidney stonesb 0 (0) 1 (2) introducing solid foods.31 Breast feeding as well as in-
Weight gainb 0 (0) 2 (3) cluding pumped breast milk does not only provide macro-
Growth deficitb 2 (13) 7 (11) and micronutrients but also promotes intestinal, immune,
and cognitive development.32 In infants with inborn errors
Data are presented as numbers with percentages in parentheses of metabolism such as phenylketonuria33,34 breast milk on
(tested using the chi square test).
a
During the first week of the KD. demand is frequently included after having given a small
b
During the first 3 months of the KD. amount of specialized formula when close clinical and
KD, ketogenic diet. metabolic monitoring is guaranteed.35
 6 DRESSLER ET AL.
The prescription of the KD allows only very restricted
amounts of carbohydrates, so that the inclusion of breast milk
is difficult, let alone the continuation of breastfeeding. In the
previously reported case series,21–23 feeding at the breast
while on KD was described in 3 cases,21 and expressed breast
milk was used in the other 14.22,23 Also in our cohort, two
infants were exclusively breastfed: the allowed amount of
breast milk was fed at the breast after the calculated amount
of ketogenic formula to guarantee the calculated fat/nonfat
ratio. However, most mothers in our cohort preferred to ex-
press breast milk. Regardless of the beneficial properties of
pumping human milk, infants fed with pumped milk have
been recently described to show a different and less diverse
microbiome than breastfed infants,36 so that breastfeeding
while on the KD should be encouraged.
As described in the study by Cole et al.,21 fat/nonfat ratios
were lower also in our patients who received breast milk,
which was also confirmed by a higher amount of calories per
Downloaded by Boston University package from www.liebertpub.com at 12/02/19. For personal use only.
day at 3 months. A gradual introduction and a modified
fat/nonfat ratio of 2:1 up to 2.5:1 allowed us to keep a high
amount of breast milk integrated in the KD while maintaining
ketosis.
Finally, including breast milk during KD in infants with
epilepsy allows for achieving stable ketosis and maybe to
maintain some the health benefits of breast milk as well as
mother–infant bonding.
Our data show the feasibility of including expressed breast
milk into the KD and to maintain breast-feeding—even until
weaning in 25% of patients. Treatment response (i.e., re-
duction in seizure frequency, as well as seizure freedom) is in
accordance with infants on standard KD.6,7,9,37,38 The higher
percentage of seizure freedom in infants with breast milk
included, although not significant, might reflect the docu-
mented epidemiological association of breastfeeding with a
decreased risk of epilepsy.39 On the contrary, infants with a
KD that included breast milk were slightly younger and re-
ceived less AEDs than infants without breast milk, which also
favors a better response to therapy.6 This potential additional
benefit of breast milk on the effectiveness of the KD still
remains to be studied, in particular with regard to develop-
mental and cognitive long-term outcome.
Strength and limitations
This is the first study that directly compared a KD with
breast milk included to the standard KD that is exclusively
based on formula. In addition, we analyzed the largest group
of infants (n = 16) and described in detail how to achieve
ketosis and maintain feeding at the breast, however, in a
cohort of 79 infants, this is a relatively small sample for
subgroup analysis. The fact that ketosis can be achieved in
most patients within the first 48 hours is promising: Data are,
however, still very limited, especially on mothers actually
nursing their children while on the KD. Since the team has
gained experience in including breast milk to the standard
KD during the study period, protocols changed insofar as in
the last year’s nursing while on the KD was accepted by
mothers earlier, that is, already from the first day of the KD.
Summary and Conclusions
Our results show that ketosis and hence seizure control is
feasible with inclusion of breast milk into the KD. We also
show that continuing nursing while implementing the KD is
complicated for both team and mothers and only achieved in
a minority. Effectiveness and safety are not different. How-
ever, infections do not occur when breast milk is included.
Based on our results, we suggest to aim at continuation of
breastfeeding at start of the KD by bottle-feeding the keto-
genic formula and to feed the remaining amount of tolerable
carbohydrates at the breast—as recommended also for inborn
errors of metabolism such as phenylketonuria.33–35
Future research should focus on the question how to further
optimize the composition of human breast milk, for example,
by using hind milk to allow for even higher amounts of breast
milk and booster the fat/nonfat ratio. Moreover, long-term
beneficial effects of breast milk with respect to growth,
cognitive development, and effects on the gut microbiome
are to be explored in future studies.
Disclosure Statement
A.D. has received travel reimbursement and speaker
honoraria from SHS, Nutricia, and Vitaflo; P.T. has received
travel reimbursement from SHS, Nutricia, and Vitaflo. M.F.
received travel reimbursement from SHS, Nutricia. None was
related to this article. The other authors have no conflicts of
interest to disclose.
Funding Information
No funding was received for this article.
Supplementary Material
Supplementary Figure S1
Supplementary Table S1
Supplementary Table S2
References
1. Kossoff EH, McGrogan JR. Worldwide use of the keto-
genic diet. Epilepsia 2005;46:280–289.
2. Freeman JM, Kossoff EH, Hartman AL. The ketogenic
diet: One decade later. Pediatrics 2007;119:535–543.
3. Kossoff EH, Zupec-Kania BA, Auvin S, et al. Optimal
clinical management of children receiving dietary therapies
for epilepsy: Updated recommendations of the International
Ketogenic Diet Study Group. Epilepsia Open 2018;3:175–
192.
4. Cremer JE, Braun LD, Oldendorf WH. Changes during
development in transport processes of the blood-brain
barrier. Biochim Biophys Acta 1976;448:633–637.
5. Hawkins RA, Willianson DH, Krebs HA. Ketone-body
utilization by adult and suckling rat brain in vivo. Biochem
J 1971;122:13–18.
6. Dressler A, Trimmel-Schwahofer P, Reithofer E, et al. The
ketogenic diet in infants—Advantages of early use. Epi-
lepsy Res 2015;116:53–58.
7. Hong AM, Turner Z, Hamdy RF, et al. Infantile spasms
treated with the ketogenic diet: Prospective single-center
experience in 104 consecutive infants. Epilepsia 2010;51:
1403–1407.
8. Prezioso G, Carlone G, Zaccara G, et al. Efficacy of ke-
togenic diet for infantile spasms: A systematic review. Acta
Neurol Scand 2018;137:4–11.
9. Dressler A, Benninger F, Trimmel-Schwahofer P, et al.
Efficacy and tolerability of the ketogenic diet versus high-
 BREAST MILK AND THE KETOGENIC DIET
dose ACTH for infantile spasms: A single center parallel-
cohort randomized controlled trial. Epilepsia 2019;60:441–
451.
10. Rubenstein JE. Use of the ketogenic diet in neonates and
infants. Epilepsia 2008;49 Suppl 8:30–32.
11. Kramer MS, Kakuma R. Optimal duration of exclusive
breastfeeding. Cochrane Database Syst Rev 2012:
CD003517.
12. Ip S, Chung M, Raman G, et al. Breastfeeding and maternal
and infant health outcomes in developed countries. Evid
Rep Technol Assess 2007:1–186.
13. OECD/EU. Health at a Glance: Europe 2016–State of
Health in the EU Cycle. Paris: OECD Publishing, 2016.
14. Gouveri E, Papanas N, Hatzitolios AI, et al. Breastfeeding
and diabetes. Curr Diabetes Rev 2011;7:135–142.
15. Paul IM, Bartok CJ, Downs DS, et al. Opportunities for the
primary prevention of obesity during infancy. Adv Pediatr
2009;56:107–133.
Downloaded by Boston University package from www.liebertpub.com at 12/02/19. For personal use only.
16. Zhou Y, Chen J, Li Q, et al. Association between breast-
feeding and breast cancer risk: Evidence from a meta-
analysis. Breastfeed Med 2015;10:175–182.
17. Sung HK, Ma SH, Choi JY, et al. The effect of breast-
feeding duration and parity on the risk of epithelial ovarian
cancer: A systematic review and meta-analysis. J Prev Med
Public Health 2016;49:349–366.
18. Girard LC, Farkas C. Breastfeeding and behavioural
problems: Propensity score matching with a national cohort
of infants in Chile. BMJ Open 2019;9:e025058.
19. Nickels KC, Wirrell EC. Cognitive and social outcomes of
epileptic encephalopathies. Semin Pediatr Neurol 2017;24:
264–275.
20. Weatherburn CJ, Heath CA, Mercer SW, et al. Physical and
mental health comorbidities of epilepsy: Population-based
cross-sectional analysis of 1.5 million people in Scotland.
Seizure 2017;45:125–131.
21. Cole NW, Pfeifer HH, Thiele EA. Initiating and main-
taining the ketogenic diet in breastfed infants. ICAN 2010;
2:177–180.
22. Fenton C, Randall R, Groveman SA, et al. Use of expressed
breast milk with the ketogenic diet. ICAN 2015;7:342–346.
23. Le Pichon JB, Thompson L, Gustafson M, et al. Initiating
the ketogenic diet in infants with treatment refractory epi-
lepsy while maintaining a breast milk diet. Seizure 2019;69:
41–43.
24. Dressler A, Stocklin B, Reithofer E, et al. Long-term out-
come and tolerability of the ketogenic diet in drug-resistant
childhood epilepsy—The Austrian experience. Seizure
2010;19:404–408.
25. Ballard O, Morrow AL. Human milk composition: Nu-
trients and bioactive factors. Pediatr Clin North Am 2013;
60:49–74.
26. Jenness R. The composition of human milk. Semin Peri-
natol 1979;3:225–239.
7
27. Kossoff EH, Zupec-Kania BA, Amark PE, et al. Optimal
clinical management of children receiving the ketogenic
diet: Recommendations of the International Ketogenic Diet
Study Group. Epilepsia 2009;50:304–317.
28. van der Louw E, van den Hurk D, Neal E, et al. Ketogenic
diet guidelines for infants with refractory epilepsy. Eur J
Paediatr Neurol 2016;20:798–809.
29. Pannaraj PS, Li F, Cerini C, et al. Association between
breast milk bacterial communities and establishment and
development of the infant gut microbiome. JAMA Pediatr
2017;171:647–654.
30. Hermansson H, Kumar H, Collado MC, et al. Breast milk
microbiota is shaped by mode of delivery and intrapartum
antibiotic exposure. Front Nutr 2019;6:4.
31. Section on B. Breastfeeding and the use of human milk.
Pediatrics 2012;129:e827–e841.
32. Neville MC, Anderson SM, McManaman JL, et al. Lacta-
tion and neonatal nutrition: Defining and refining the crit-
ical questions. J Mammary Gland Biol Neoplasia 2012;17:
167–188.
33. Banta-Wright SA, Kodadek SM, Houck GM, et al. Com-
mitment to breastfeeding in the context of phenylketonuria.
J Obstet Gynecol Neonatal Nurs 2015;44:726–736.
34. Banta-Wright SA, Kodadek SM, Steiner RD, et al. Chal-
lenges to breastfeeding infants with phenylketonuria.
J Pediatr Nurs 2015;30:219–226.
35. Pichler K, Michel M, Zlamy M, et al. Breast milk feeding
in infants with inherited metabolic disorders other than
phenylketonuria—A 10-year single-center experience.
J Perinat Med 2017;45:375–382.
36. Moossavi S, Sepehri S, Robertson B, et al. Composition
and variation of the human milk microbiota are influenced
by maternal and early-life factors. Cell Host Microbe 2019;
25:324–335.e4.
37. Kossoff EH, Hedderick EF, Turner Z, et al. A case-control
evaluation of the ketogenic diet versus ACTH for new-
onset infantile spasms. Epilepsia 2008;49:1504–1509.
38. Pires ME, Ilea A, Bourel E, et al. Ketogenic diet for in-
fantile spasms refractory to first-line treatments: An open
prospective study. Epilepsy Res 2013;105:189–194.
39. Sun Y, Vestergaard M, Christensen J, et al. Breastfeeding
and risk of epilepsy in childhood: A birth cohort study.
J Pediatr 2011;158:924–929.
Address correspondence to:
Anastasia Dressler, MD
Department of Pediatrics and Adolescent Medicine
Medical University Vienna
Waehringer Guertel 18–20
A-1090 Wien
Austria
E-mail: anastasia.dressler@meduniwien.ac.at