High‐Fat and Low‐Carbohydrate Dietary Environments Are Linked to Reduced Idiopathic Epilepsy Incidence and Prevalence

Abstract

Dietary manipulations like ketogenic diets are established interventions for recalcitrant epilepsy. However, it remains unknown whether specific macronutrient exposure through dietary environments could possibly extend to primary preventive qualities, associated with changes in epilepsy disease burden (prevalence and incidence). Here, macronutrient supply, GDP, and idiopathic epilepsy disease burden data were collated from more than 150 countries from 1990 to 2018. Nutritional geometry generalized additive mixed models (GAMMs) modeling unraveled that dietary environments with high‐fat and low‐carbohydrate supplies were linked to lower epilepsy incidence and prevalence. Our analyses suggested a plausible primary preventive role of dietary manipulations for epilepsy.

Full text

Annals of Clinical and Translational Neurology, 2025; 0:1–5
https://doi.org/10.1002/acn3.70017
1 of 5
Annals of Clinical and Translational Neurology
BRIEF COMMUNICATION OPEN ACCESS
High- Fat and Low- Carbohydrate Dietary Environments
Are Linked to Reduced Idiopathic Epilepsy Incidence
and Prevalence
DuanNi1,2, 3 | AlistairSenior3,4,5 | DavidRaubenheimer3,4 | StephenJ.Simpson3,4 | RalphNanan1,2,3
1Sydney Medical School Nepean, The University of Sydney, Sydney, New South Wales, Australia | 2Nepean Hospital, Nepean Blue Mountains Local Health
District , Sydney, New South Wales, Australia | 3Charles Perkins Centre, The University of Sydney, Sydney, New South Wales, Australia | 4School of Life
and Environmental Sciences, The University of Sydney, Sydney, New South Wales, Australia | 5Sydney Precision Data Science Centre, The University of
Sydney, Sydney, New South Wales,Australia
Correspondence: Ralph Nanan (ralph.nanan@sydney.edu.au)
Received: 21 January 2025 | Revised: 30 January 2025 | Accepted: 30 January 2025
Funding: This work was supported by the Norman Ernest Bequest Fund.
Keywords: dietary environment| high- fat low- carbohydrate| idiopathic epilepsy| macronutrient
ABSTR ACT
Dietary manipulations like ketogenic diets are established interventions for recalcitrant epilepsy. However, it remains unknown
whether specific macronutrient exposure through dietary environments could possibly extend to primary preventive qualities, associ-
ated with changes in epilepsy disease burden (prevalence and incidence). Here, macronutrient supply, GDP, and idiopathic epilepsy
disease burden data were collated from more than 150 countries from 1990 to 2018. Nutritional geometry generalized additive mixed
models (GAMMs) modeling unraveled that dietary environments with high- fat and low- carbohydrate supplies were linked to lower
epilepsy incidence and prevalence. Our analyses suggested a plausible primary preventive role of dietary manipulations for epilepsy.
1 | Introduction
Epilepsy is one of the leading neurological disorders globally
[1]. Epilepsy pathogenesis is multifactorial, and environmental
factors, like alcohol use [2], represent critical risk factors in its
development. Other aspects, like diet and nutrients, might also
be important, as reflected by the therapeutic effects of ketogenic
diets on recalcitrant epilepsy [3–6]. Previous diet- related studies
mostly concentrated on the effects of individual nutrient on ep-
ilepsy treatment [3, 4, 6, 7]. In contrast, less is known regarding
the primary prevention of epilepsy via nutritional and dietar y
modifications, as studies of this kind are intrinsically difficult to
carry out at a population level across a time scale.
Here, we leveraged nutrient supplies, a proxy for diet and nutri-
ent environments, to interrogate the potential associations be-
tween nutrient exposures and epilepsy incidence and prevalence,
harnessing the powerful tool, nutritional geometry generalized
additive mixed models (GAMMs), we previously conceptualized
[8, 9]. We found that dietary environments, high in fat and low
in carbohydrate supplies, are associated with reduced idiopathic
epilepsy prevalence and incidence, indicative of its protective ef-
fects. This might extend the clinical applications of dietary ma-
nipulation for primary prevention of epilepsy.
2 | Materials and Methods
2.1 | Data Collection
Disease burden data (incidence and prevalence) of idiopathic ep-
ilepsy were from the Global Burden of Disease (GBD) database
[10], which is defined as recurrent, unprovoked seizures without
identified underlying disease and is presumed to have genetic
This is a n open access ar ticle under the terms of t he Creative Commons Attr ibution License, which p ermits use, dis tribution and repro duction in any medium, p rovided the orig inal work is
properly cited.
© 2025 T he Author(s). Annals of Clini cal and Trans lational Neuro logy publishe d by Wiley Periodic als LLC on behal f of American Neur ological Assoc iation.

2 of 5 Annals of Clinical and Translational Neurology, 2025
bases. As reported in our previous studies [8, 9], nutrient supply
data (supply of kcal of nutrient/capita/day) were obtained from
the Food and Agriculture Organization Corporate Statistical
Database (www. fao. org/ faost at/ en/# home), reflecting nutrients
available for human consumption, as a close proxy of their ac-
tual dietary intake, but excluding the ones for other uses like ag-
ricultural utilization. Global gross domestic product (GDP) data
(US$/capita), as an indicator of socioeconomic wealth, was from
the Maddison project [11]. Countries or time points with no re-
cord were excluded, and the resulting data spanning from 1990
to 2018 covering more than 150 countries, across all continents,
were further analyzed.
2.2 | Generalized Additive Mixed Models
Details of the analyses are described in Supporting
Information S1 and our previous works [8, 9, 12]. In brief, we
leveraged the state- of- the- art nutritional geometry GAMM for
analysis. GAMM is a form of multiple regression, sharing simi-
lar assumptions to generalized linear models. GAMMs account
for the nonlinear terms as nonparametric smoothed functions,
often in the form of splines, and provide a flexible manner to
estimate the nonlinear associations. These nonlinear and inter-
active effects are gaining more and more interest in nutritional
research, with growing ev idence highlighting their implications
in multiple aspects of health and diseases [8, 9, 13], emphasiz-
ing the significance of multidimensional thinking in nutritional
research.
Here, an array of GAMMs was utilized to analyze the impacts of
nutrient supplies on epilepsy disease burden, and the individual
country from which the data originated was modeled with a ran-
dom effect, accounting for some confounders like ethnic and ge-
netic differences among countries. Importantly, GAMMs could
also adjust for potential confounders like GDP as a proxy for so-
cioeconomic status, and time, which was achieved in a similar
way to standard linear regression used in classic epidemiological
study [13]. In this context, GAMMs with multiple variables con-
sider all combinations of the individual, additive, and interactive
effects for parameters like macronutrient supply, year, and GDP
data. Modeling outcomes were evaluated using Akaike informa-
tion criterions [14].
GAMMs' results are interpreted via visualization onto the mul-
tidimensional nutritional space, a common platform for nutri-
tional geometry analysis, which deciphers the impacts from
multi- parameters (e.g., different macronutrients) simultane-
ously, rather than the traditional simple single regression/correla-
tion. Details for their interpretation are described in Supporting
InformationS1 and in our previous work [8, 9, 12, 13, 15].
3 | Results
As in Figure 1A, globally, epilepsy prevalence and incidence
steadily increased from 1990 to 2018. This is accompanied by
concurrent increases in global GDP and correlated changes in
nutritional supply landscapes, with the most striking increase
in fat supply (Figure1B, FigureS1). Epilepsy incidence exhibits
a heterogeneous global distribution, affecting most countries. In
2018, there were 159 countries with data records, and the high-
est incidence was reported in Saudi Arabia, while the lowest in
Kiribati (Figure1C).
To interrogate the potential associations between epilepsy dis-
ease burden and nutrient supplies, a series of GAMMs were
explored, which included combinations of factors including nu-
trient–nutrient interactions, as well as their changes over time
and impacts from socioeconomic status. A model considering
the interactions between macronutrient supplies and GDP, with
an additive effect of time, was favored (TablesS1 and S3), illus-
trating the interactive effects of macronutrient supplies and so-
cioeconomic status on epilepsy disease burden.
The results from 2018 are presented in Figure1D as an exam-
ple. This was the most recent year with relatively comprehensive
data coverage. The modeled associations between macronutri-
ent supplies and epilepsy incidence were presented as response
surfaces mapped onto macronutrient supply plots. We focused
on the effects of fat (x- axis) and carbohydrate (y- axis) supplies.
Protein supply was held at 25% (low), 50% (median), and 75%
(high) quantiles of global supply. Within the modeling sur-
faces, red represented higher, while blue denoted lower epilepsy
incidences.
Our modeling found that carbohydrate supply was strongly cor-
related with an increased epilepsy incidence, while fat supply
had the opposite association after accounting for total energy
supply (Figure1D, TablesS1 and S2). This is visualized via the
purple isocaloric line. Along this vector, the total nutrient energy
supply was held constant, and carbohydrate was isocalorically
replaced with fat. A higher fat:carbohydrate ratio was linked to
lower epilepsy burden, with the lowest epilepsy incidence found
in environments with high- fat but low- carbohydrate supplies
(bottom right). Across increasing quantiles of protein supplies,
there was only a mild change in epilepsy incidence, suggest-
ing an insignificant impact from protein. Epilepsy incidence
appeared similar with changes in the total energy supply from
macronutrients because along the red radial, altering the total
energy supply but keeping the fat:carbohydrate ratio unchanged
minimally impacted epilepsy disease burden.
Our modeling took into consideration the time effects, and the
associations between carbohydrate supply and higher epilepsy
burden were consistent across various time points (Figure1D,E).
We have also performed sensitivity tests by excluding countries
with the highest (United Arab Emirates) or lowest (Democratic
People's Republic of Korea) epilepsy incidences, which showed
consistent results (FigureS2), highlighting the robustness of our
analyses. Similarly, carbohydrate supply was also associated with
increased epilepsy prevalence, while fat supply had the opposite
effect (Figure2, TablesS3 and S4). These associations were in-
dependent of sex (FigureS3A,B). Notably, further analyses sub-
dividing plant- and animal- based fats revealed that they were
similarly linked to epilepsy incidence (FigureS3C).
4 | Discussion
Our work represents the first study to correlate idiopathic epi-
lepsy disease burden with global nutritional environments, as

3 of 5
FIGUR E  | Global associations of macronutrient supplies a nd idiopathic epilepsy disease burden. (A) Global age- standa rdized prevalence (black)
and incidence (red) of idiopathic epilepsy of both sexes as functions of years. (B) Global gross domestic product (GDP) per capita (in US dollars,
black), and macronutrient supplies (carbohydrate: green, protein: blue, fat: cyan) as functions of years. (C) A global overview of the age- standardized
idiopathic epilepsy incidence of both sexes in 2018. (D) Modeled effects of macronutrient supplies on age- standardized idiopathic epilepsy incidence
rate of both sexes, with 2018 results shown as representative (see Supporting InformationS1 for statistics and interpretations). (E) Modeled effects of
macronutrient supplies on age- standardized idiopathic epilepsy incidence rate of both sexes in 1990 (left), 2000 (middle), and 2010 (right).
1990 20002010 2020
0
5000
10000
15000
0
500
1000
1500
2000
GDP/capita (US$)
Supply kcal/capita/day
19902000 20102020
0
150
0
15
200
250
300
350
400
450
20
30
40
50
60
Prevalence, rate/100k
Incidence, rate/100k
●Epilepsyprevalencerate
●Epilepsyincidence rate
●GDP ●Carbohydrate
●Protein ●Fat
AB
Fat(kcal/capita/day)
)yad/atipac/lack(etardyhobraC
Protein(kcal/capita/day) =263 Protein(kcal/capita/day) =331 Protein(kcal/capita/day)=390
Epilepsy incidencerate (bothsexes,age-standardized, 2018)
D
40
2500
2000
1500
2500
2000
1500
2500
2000
1500
400 8001200 1600400 800 1200 1600400 800 1200 1600
41
42
43
44
45
40
41
42
43
44
46
41
42
43
44
45
Epilepsy new cases per100,000 (bothsexes,age-standardized, 2018)
C
1990
Protein(kcal/capita/day) =210
2000
Protein(kcal/capita/day) =226
2010
Protein(kcal/capita/day) =250
40.4
40 39.6 39.2
43.4
42.7
42
41.3
40.6
39.9
39.6
39.3
39
38.7
EEpilepsy incidence rate (both sexes, age-standardized)
500 1000 1500400 800 1200 400 800 1200 1600
1500
2000
2500
1500
2000
2500
1500
2000
2500
Fat(kcal/capita/day)
)yad/atipac/lack(etardyhobraC

4 of 5 Annals of Clinical and Translational Neurology, 2025
reflected by nutrient supplies. Our analyses unveil a potential
beneficial role of fat for epilepsy prevention. This has been ad-
justed for its plausible interactions with other macronutrients,
total energy supply, and socioeconomic status, using the cutting-
edge GAMM approach [8, 9, 16]. Such systematic analysis rep-
resents the most comprehensive population- level study in this
regard to our knowledge, showcasing that increased fat supply
is linked to reduced epilepsy incidence and prevalence, partic-
ularly coupled with decreased carbohydrate supply, suggesting
that dietary environments high in fat but low in carbohydrate
might exhibit protection against epilepsy development.
Our analyses were based on nutrient supplies, a proxy of the di-
etary and nutrient environment. Validation with more detailed
dietary and nutritional data would be warranted. Also, it remains
to be interrogated at which developmental stage across the lifes-
pan exposure to a high- fat, low- carbohydrate environment will
be the most beneficial for epilepsy prevention. Future cohort-
or population- based diet and nutrient studies might shed more
light on these aspects and also evaluate the potential adverse ef-
fects [17].
Of note, despite the strong associations between dietary envi-
ronment and epilepsy disease burden, the exact causality re-
mains to be elucidated. There is a previous report on children
with epilepsy exhibiting a preference for fat over carbohydrate
in their diets. This might also partly explain our modeling find-
ings [18].
Mechanistically, a recent murine study reported that ketone
body β- hydroxybutyric acid is the main contributor to the ther-
apeutic effect of high- fat, low- carbohydrate ketogenic diets in
epilepsy [5]. This is also supported by a clinical study reporting
the negative association between circulating ketone body levels
and seizure frequency in children with medically intractable ep-
ilepsy receiving a ketogenic diet [19]. Therefore, it would be of
interest to inspect which threshold levels of ketosis are required
for epilepsy treatment or prevention. In addition, a high- fat,
low- carbohydrate diet is likely to influence the gut–brain axis
[20] and immunometabolism [21], which are also critical for the
physiology and pathology of the CNS system.
One of the limitations of our study lies in the complexities in id-
iopathic epilepsy, with its multifactorial pathogenesis nature and
changing diagnosis and reporting criteria across countries over
time. As defined by GBD, idiopathic epilepsy refers to cases with-
out identified underlying disease and is presumed to have some
genetic basis. This at least excludes epilepsy cases of other causes
like infection or trauma [10]. Our modeling partially accounted
for the differences across countries regarding ancestral and ge-
netic backgrounds, as well as correcting for GDP as a reflection of
socioeconomic status, a critical factor impacting the affordability
of nutrient fat, the key component in ketogenic diets. Still, further
in- depth studies are needed to thoroughly decipher other con-
founders for idiopathic epilepsy. This would include investigation
into the interplay between genetic and environmental factors,
specifically diets and nutrients. Additionally, our modeling has
accounted for the time factor, showing a consistent pattern across
various time points, highlighting the robustness of our results,
despite the possible changes in disease reporting criteria. On the
other hand, albeit comprehensive data for food processing and
diet quality are still lacking on a global scale. Such data would
add value to dissect the plausible impacts from nutrient quality,
determined by the degree of food processing.
Together, our ecological analyses have revealed a robust asso-
ciation between a nutrient environment with high- fat and low-
carbohydrate supplies and decreased epilepsy incidence and
prevalence. This might inform future nutrient- based epilepsy
prevention.
Author Contributions
Concept and design: Duan Ni and Ralph Nanan. Acquisition, anal-
ysis, and interpretation of data: Duan Ni, Alistair Senior, David
Raubenheimer, Stephen J. Simpson, and Ralph Nanan. Drafting of the
manuscript: Duan Ni and Ralph Nanan. Critical rev ision of the manu-
script for important intellectual content: All authors.
FIGUR E  | Modeled effects of macronutrient supplies on age- standardized idiopathic epilepsy prevalence rates of both sexes, with 2018 results
shown as representative.
Protein (kcal/capita/day) = 263Protein (kcal/capita/day) = 331Protein (kcal/capita/day) = 390
Fat (kcal/capita/day)
)yad/atipac/lack( etardyhobraC
Epilepsy prevalence rate (both sexes, age-standardized, 2018)
440
400
360
420
340
400
380
360
320
340
320
340
380
400
2500
2000
1500
2500
2000
1500
2500
2000
1500
400 800 1200 1600 400 80012001600 400 80012001600
380
360

5 of 5
Acknowledgments
This project is supported by the Norman Ernest Bequest Fund.
Ethics Statement
The authors have nothing to report.
Conflicts of Interest
The authors declare no conflicts of interest.
Data Availability Statement
All data used in the present study are publicly available as described in
Materials and Methods in Supporting Information.
References
1. O. Devinsky, A. Vezzani, T. J. O'Brien, etal., “Epilepsy,” Nature Reviews.
Disease Primers 4 (2018): 18024, https:// doi. org/ 10. 1038/ nrdp. 2018. 24.
2. K. N. Woo, K. Kim, D. S. Ko, H. W. Kim, and Y. H. Kim, “Alcohol
Consumption on Unprovoked Seizure and Epilepsy: An Updated Meta-
Analysis,” Dr ug and Alcohol Dependence 232 (2022): 109305, https:// doi.
org/ 10. 1016/j. druga lcdep. 2 022. 109305.
3. I. D'Andrea Meira, T. T. Romão, H. J. Pires do Prado, L. T. Krüger, M.
E. P. Pires, and P. O. da Conceição, “Ketogenic Diet and Epilepsy: What
we Know So Far,” Frontiers in Neuroscience 13 (2019): 5, https:// doi. org/
10. 3389/ fnins. 2019. 00005 .
4. G. R. Lum, S. M. Ha, C. A. Olson, et al., “Ketogenic Diet Therapy
for Pediatric Epilepsy Is Associated With Alterations in the Human Gut
Microbiome That Confer Seizure Resistance in Mice,” Cell Reports 42
(2023): 113521, https:// doi. org/ 10. 1016/j. celrep. 2023. 113521.
5. Y. N. Qiao, L. Li, S. H. Hu, et al., “Ketogenic Diet- Produced Beta-
Hydroxybutyric Acid Accumulates Brain GA BA and Increases GABA/
Glutamate Ratio to Inhibit Epilepsy,” Cell Discovery 10 (2024): 17,
https:// doi. org/ 10. 1038/ s4142 1- 023- 00636 - x.
6. A. Verrotti, G. Iapadre, L. Di Francesco, L. Zagaroli, and G. Farello,
“Diet in the Treatment of Epilepsy: What we Know So Far,” Nutrients
12, no. 9 (2020): 2645, https:// doi. org/ 10. 3390/ nu120 926 45 .
7. J. Wells, A. Swaminathan, J. Paseka, and C. Hanson, “Efficacy and
Safety of a Ketogenic Diet in Children and Adolescents With Refractory
Epilepsy- A Review,” Nutrients 12, no. 6 (2020): 1809, https:// doi. org/ 10.
3390/ nu120 61809 .
8. D. Ni, A. M. Senior, D. Raubenheimer, S. J. Simpson, L. Macia, and
R. Nanan, “Global Associations of Macronutrient Supply and Asthma
Disease Burden,” Allergy 79, no. 7 (2024): 1989–1991, https:// doi. org/ 10.
1111/ all. 16067 .
9. A. M. Senior, S. Nakagawa, D. Raubenheimer, and S. J. Simpson,
“Global Associations Between Macronutrient Supply and Age- Specific
Mortality,” Proceedings of the National Academy of Sciences of the United
States of America 117 (2020): 30824–30835, https:// doi. org/ 10. 1073/ pnas.
20150 58117 .
10. Global Burden of Disease Collaborative Network (Institute for Health
Metrics and Evaluation (IHME), 2021).
11. R. I. Jutta Bolt, Herman de Jong, and Jan Luiten van Zanden, vol.
GD- 174 (Groningen Growth and Development Center, GGDC Research
Memorandum, 2018).
12. D. Ni, A. Senior, D. Raubenheimer, S. J. Simpson, and R. Nanan,
“Plant- Based Fat Supply Is Associated With Reduced ADHD Disease
Burden,” medRxiv (2024), https:// doi. org/ 10. 1101/ 2024. 01. 30. 24301956.
13. A. M. Senior, V. Legault, F. B. Lavoie, etal., “Multidimensional Asso-
ciations Between Nutrient Intake and Healthy Ageing in Humans,” BMC
Biology 20 (2022): 196, https:// doi. org/ 10. 1186/ s1291 5- 022- 01395 - z.
14. H. Akaike, “Second International Symposium on Information The-
or y,” 19 73.
15. J. A. Wali, D. Ni, H. J. W. Facey, etal., “Determining the Metabolic
Effects of Dietary Fat, Sugars and Fat- Sugar Interaction Using Nutri-
tional Geometr y in a Dietary Challenge Study With Male Mice,” Nature
Communications 14 (2023): 4409, https:// doi. org/ 10. 1038/ s4146 7- 023-
40039 - w.
16. S. N. Wood, Generalized Additive Models: An Introduction With R
(Chapman and Hall/CRC, 2017).
17. M. Ulamek- Koziol, S. J. Czuczwar, S. Januszewski, and R. Pluta,
“Ketogenic Diet and Epilepsy,” Nutrients 11, no. 10 (2019): 2510, ht tp s://
doi. org/ 10. 3390/ nu111 02510 .
18. A. Amari, L. Dahlquist, E. H. Kossoff, E. P. G. Vining, W. H.
Trescher, and K. J. Slifer, “Children With Seizures Exhibit Preferences
for Foods Compatible With the Ketogenic Diet,” Epilepsy & Behavior 11
(2007): 98–104, https:// doi. org/ 10. 1016/j. yebeh. 2007. 04. 005.
19. J. R. Buchhalter, S. D'Alfonso, M. Connolly, etal., “The Relationship
Between d- Beta- Hydroxybutyrate Blood Concentrations and Seizure
Control in Children Treated With the Ketogenic Diet for Medically In-
tractable Epilepsy,” Epilepsia Open 2 (2017): 317–321, https:// doi. org/ 10.
1002/ epi4. 12058 .
20. R. Li, Y. Liu, J. Wu, etal., “Adaptive Metabolic Responses Facili-
tate Blood- Brain Barrier Repair in Ischemic Stroke via BHB- Mediated
Epigenetic Modification of ZO- 1 Expression,” Advanced Science (Wein-
heim, Baden- Wurttemberg, Germany) 11 (2024): e2400426, https:// doi.
org/ 10. 1002/ advs. 20240 0426.
21. D. Ni, J. Tan, J. Reyes, etal., “High Fat Low Carbohydrate Diet is
Linked to Protection Against CNS Autoimmunity,” biorxiv (2024),
https:// doi. org/ 10. 1101/ 202 4. 07. 23. 604865.
Supporting Information
Additional supporting information can be found online in the
Supporting Information section.