1 INTRODUCTION

Ketogenic dietary therapies (KDTs) are high-fat, adequate-protein, restricted carbohydrate regimens leading to ketosis, a metabolic state in which ketone bodies, made from the breakdown of fatty acids in the liver, represent the primary source of energy (Lambrechts et al., 2017). Ketogenic dietary therapies are now considered safe and well-established treatments for adult and paediatric refractory epilepsy, with increasing randomised controlled trials (RCT) demonstrating their efficacy (Kim et al., 2016; Lambrechts et al., 2017; Neal et al., 2008; Raju et al., 2011; Sharma et al., 2013). The hypothesised therapeutic mechanisms of ketogenic dietary therapies action are many and involve ATP-sensitive potassium channel modulation, enhanced purinergic and GABAergic neurotransmission, attenuation of neuroinflammation, as well as an expansion in bioenergetic reserves and stabilisation of the neuronal membrane potential through improved mitochondrial function (Zhu et al., 2022). Ketogenic dietary therapies can also modulate the reactive oxygen species production, enhancement of synthesis, and release of the inhibitory modulator adenosine (Zhu et al., 2022). Moreover, recent studies have analysed the effect of the ketogenic dietary therapies on the gut microbiome in both mice and human studies, highlighting the neuromodulator effect through faecal microbial profiles (Dahlin et al., 2022; Olson et al., 2018). Beyond their utility as an anticonvulsant and anti-epileptogenic treatment, ketogenic dietary therapies have been shown to have neuroprotective properties, admitting the potential of ketogenic dietary therapies as a disease-modifying intervention in neurological disorders other than epilepsy, such as migraine, autism spectrum disorder (ASD), and neurodegenerative disorders (Zhu et al., 2022). Typically, seizure reduction or seizure freedom have been considered primary outcomes in clinical trials on drug-resistant epilepsy (DRE) populations (El-Rashidy et al., 2013; Kim et al., 2016; Lambrechts et al., 2017; Neal et al., 2008; Raju et al., 2011; Sharma et al., 2013; Sharma et al., 2016; Sondhi et al., 2020) but more recently other “beyond seizure effects”, such as cognitive function, behaviour, and quality of life (QoL) have been implemented as secondary outcomes to be objectively measured (Carroll et al., 2022; IJff et al., 2016; van Berkel, & IJff DM, Verkuyl JM., 2018). The effects of ketogenic dietary therapies on sleep have been poorly questioned and explored in patients with epilepsy or other neurological disorders undergoing ketogenic dietary therapies. Nevertheless, sleep quality and daytime somnolence represent important domains highly related to disease symptoms manifestation and quality of life. Carroll and colleagues (Carroll et al., 2022) recently searched for a parental opinion on the outcomes of importance in patients with epilepsy undergoing ketogenic dietary therapies, and “time spent asleep” and “daytime sleepiness” have been included among relevant “physical functioning” outcomes in patients undergoing ketogenic dietary therapies.

Importantly, sleep represents a prominent aspect of life and patient care (Zucconi & Bruni, 2008). Sleep problems have been found to result in a worsening of disease symptoms and additional learning and behaviour disturbances, affecting the whole family's health and well-being (Angriman et al., 2015). Indeed, sleep is crucial for restorative and cognitive effects, being required for neural plasticity and memory consolidation (Angriman et al., 2015; Gorgoni et al., 2013). Since sleep problems are common in neurological conditions for which ketogenic dietary therapies are recognised as an effective intervention – namely drug-resistant epilepsy, autism spectrum disorder, and migraine – the present scoping review aimed to: (a) determine the coverage of the body of literature about this topic; (b) identify and map available evidence of ketogenic dietary therapies effects on sleep quality and structure in these populations (patients with epilepsy, autism spectrum disorder, or migraine); and (c) inform research and clinical practice in the field.

2 METHODS

2.2 Study selection

The main cores of this review were the inclusion of any ketogenic dietary therapy type and the effect this treatment could have on sleep and sleep disorders in these patients, both subjectively and objectively measured. Inclusion criteria of basic diagnoses grouped epilepsy, autism spectrum disorder, and migraine as conditions that could benefit from ketogenic dietary therapy (all age groups and subtypes).

Studies that investigated the impact of the ketogenic dietary therapies on sleep objective (e.g., polysomnographic records) and subjective (e.g., standardised questionnaires, parental reporting) outcomes in patients with epilepsy, autism spectrum disorder, or migraine were included. The exclusion criterion was the absence of information on sleep outcome after implementation of ketogenic dietary therapies.

The studies identified through the literature search were assessed for inclusion through two stages by two independent authors (CAQ and LP). All records were initially considered based on title and abstract to evaluate their consistency with the aims of these reviews. Articles were selected for further review if the abstracts did not include enough information to determine eligibility. Then, the full text of each paper was assessed against inclusion and exclusion criteria. The review protocol has been published and is available on OSF (https://osf.io/s2bny). The flow diagram (Figure 1) maps out the number of records identified, included, and excluded during this process.

3 RESULTS

3.1 Characteristics of included studies

A total of 20 articles published between 2001 and 2022 were included: nine prospective studies divided between eight single-centre (Ebus et al., 2014; Guzel et al., 2019; Hallböök, Köhler, et al., 2007; Hallböök, Lundgren, & Rosén, 2007; Li et al., 2023; Merlino et al., 2023; Wu et al., 2018; Zhu et al., 2016) and one multi-centre (Miranda et al., 2011), eight retrospective studies, divided between three single-centre (Nordli et al., 2001; Peng et al., 2019; Reyes et al., 2015) and five multi-centre, one of which longitudinal study (Sofou et al., 2017) and four cross-sectional (Alqahtani & Mahmoud, 2016; Caraballo et al., 2011; Frye et al., 2011; Ünalp et al., 2021), two case reports (Anurat et al., 2022; Melikishvili et al., 2022), one two-arm open RCT (Kverneland et al., 2018). Of these 20 articles, only one study is related to migraine while 18 studies have epilepsy as a starting diagnosis and one study is related to autism spectrum disorder.

The sample size of the included studies ranged from 2 to 733 patients and the age of the participants ranged from 3 months to 45 years with the majority of the studies focussed on paediatric population (see Figure 2). The most utilised type of ketogenic dietary therapy is the classic ketogenic diet (cKD) (15/18 studies) followed by the modified Atkins diet (MAD) (6/18 studies). The timing of sleep outcome evaluation ranged from 3 days (Melikishvili et al., 2022) to 4 years (Alqahtani & Mahmoud, 2016) from the initiation of ketogenic dietary therapy.

The general characteristics of the selected studies are reported in Table 1.

TABLE 1. General design, demographics, and clinical data of included studies.

Study	Country	Study design	Diagnosis	Sample size	Mean age (year or month)	KDT
Alqahtani & Mahmoud, 2016	Saudi Arabia	Qualitative	Epilepsy	30	N/A	CKD
Anurat et al., 2022	Thailand	Case report	GLUT1DS	2	7 year	MAD
Caraballo et al., 2011	Argentina	Retrospective multicentre	DRE, GLUT1DS, RETT, and cryptogenetic epilepsy	216	5 year	Classic and medium-chain triglyceride
Ebus et al., 2014	Netherlands	Prospective follow up	DRE	34	16 year	Classic KD, MCT, Atkins, Mixed
Frye et al., 2011	USA	Cross sectional	Clinical seizures, subclinical epileptiform discharges in comorbidity (patients with ASD)	733	12 year	MAD
Guzel et al., 2019	Turkey	Prospective follow up single centre	DRE	389	4 year	CKD
Hallböök, Lundgren, & Rosén, 2007	Sweden	Prospective follow up	DRE	18	7.5 year	CKD
Hallböök, Köhler, et al., 2007	Sweden	Prospective follow up	DRE	18	7.5 year	CKD
Kverneland et al., 2018	Norway	RCT	DRE	75	Diet group 36 year, Control group 37 year	MAD
Melikishvili et al., 2022	Georgia	Case report	Non convulsive status epilepticus	2	4.5 year	CKD
Miranda et al., 2011	Denmark	Prospective multicentre follow up	DRE	83	8.1 year	MAD and CKD
Nordli et al., 2001	USA	Retrospective	DRE	32	1.2 year	CKD
Peng et al., 2019	China	Retrospective single centre	FIRES	7	8 year	CKD
Reyes et al., 2015	Argentina	Retrospective single centre	ESES	12	8.5 year	CKD
Sofou et al., 2017	Sweden	Longitudinal cohort	PDC	19	6 year	CKD and MKD
Ünalp et al., 2021	Turkey	Cross sectional	DRE	14	4.4 year	CKD
Wu et al., 2018	China	Prospective single centre follow up	DRE	52	3 month to 7 year	CKD
Zhu et al., 2016	China	Prospective single centre follow up	DRE	42	25.2 month	CKD
Li et al., 2023	China	Prospective single centre follow up	DRE	181	N/A	CKD
Merlino et al., 2023	Italy	Prospective single centre follow up	Migraine	70	46.1 ± 14.4 year	CKD

4 DISCUSSION

This scoping review provides evidence surrounding the impact of ketogenic dietary therapies on sleep in patients with neurological disorders undergoing ketogenic dietary therapies. Only a small number of studies considered the effects of ketogenic dietary therapies on sleep and almost only in epileptic populations. This is unsurprising, since to date ketogenic dietary therapies have been much more implemented in patients with epilepsy than other neurological disorders such as autism spectrum disorder and migraine.

Indeed, the application of ketogenic dietary therapies in autism spectrum disorder and migraine is more recent, and available studies show heterogeneous design and outcome data are strictly related to core disease symptoms (Varesio, 2021; Tereshko et al., 2023).

However, sleep problems and sleep disruption are very often present in the neurological conditions for which ketogenic dietary therapies represent a treatment option, and this comorbidity should be considered relevant and taken into account once ketogenic dietary therapies are implemented. The relationship between sleep and epilepsy has long been studied and found to be bidirectional, whereby factors such as sleep deprivation, daytime sleepiness, and sleep disordered breathing may be a trigger for seizures, but may in turn be caused by epilepsy itself (Manni & Terzaghi, 2010). Patients with frequent or medically intractable seizures are recognised to have multiple sleep abnormalities including increased latency to sleep onset and increased number of stage shifts, leading to inadequate sleep quality (Nobili et al., 2021). This can, in turn, exacerbate daytime drowsiness and behavioural problems, thus substantially affecting the quality of life (Méndez & Radtke, 2001). This complex relationship between sleep and epilepsy is further increased by the effects of ASMs, which may improve sleep, directly (by seizure reduction) or indirectly (improvements in QoL), or worsen sleep via adverse events (Jain & Glauser, 2014).

The literature findings (Jain & Glauser, 2014; Liguori et al., 2021) highlight the opportunity of considering sleep alterations in the comprehensive evaluation of ASMs and non-drug treatments (among which KDT), particularly for patients with comorbid sleep disorders. Unfortunately, sleep assessment before and after treatment initiation is often lacking in research studies as well as in the clinical practice.

Indeed, the present literature search documented a wide heterogeneity in study design and substantial paucity of data, as the inclusion of subjective or objective measures of sleep as an outcome is lacking.

The parents-reported measures of subjective sleep improvements were defined as general sleep quality improvement, improvement of alertness and/or daytime somnolence, and daily activity. Only rarely standardised sleep questionnaires or wearable devices such as an actigraph have been implemented in the follow-up. In the only study reporting an objective measurement of sleep changes using polysomnography under ketogenic dietary therapy during follow-up (Hallböök, Lundgren, & Rosén, 2007), a significant reduction of total daily sleep time, an increase in REM sleep, and a reduction in NREM sleep stage 2 were observed and these sleep changes correlated with an improvement of quality of life and of attentional behaviour.

All in all, we found that ketogenic dietary therapies might modify both sleep and wake states in treated patients. This effect could be related to an improvement of seizure occurrence, but a seizure-independent effect, as already described for other domains such as the cognitive and behavioural effects of ketogenic dietary therapies (Chinna-Meyyappan et al., 2022), can be deserved. Unfortunately, the design of the available studies does not allow this evaluation. Therefore, strategies should be prioritised to evaluate the direct effect of ketogenic dietary therapies on sleep.

The neurophysiopathological basis underlying the possible effect of ketogenic dietary therapies on sleep is essentially unknown at present. Sleep being such a complex physiological process regulated both locally and globally by different cellular and molecular mechanisms, and considering the potential multiplicity of experimental and clinical evidence on the mechanisms of action of ketogenic dietary therapy, the possible effects of ketogenic dietary therapy on sleep could result from biochemical changes of various types and magnitude. Moreover, the influence of ketogenic dietary therapy on sleep might also depend on the diet characteristics (caloric intake from food, distribution of food intake, ketosis levels) as direct and/or indirect effects, which need to be clarified.

Neurotransmitter regulation, immune modulation, metabolism changes and activation or inhibition of specific cellular pathways through ketogenic dietary therapy might rely on shared processes of sleep and slow wave activity regulation. For instance, it has been documented that ketogenic dietary therapy might increase the levels of brain-derived neurotrophic factor protein, which in turn improves sleep complaints (Colucci-D'Amato et al., 2020). Modification of energy substrates might also provide changes in initiation and regulation of vigilance states and, importantly, ketogenic dietary therapy have been proved to increase the extracellular and brain tissue content of adenosine (Ruskin et al., 2020) and to reduce plasma lactate levels (Cox et al., 2016). Furthermore, since numerous humoral factors including cytokines result in sleep regulation (Zielinski et al., 2016), the described anti-inflammatory effects of ketogenic dietary therapy decreasing pro-inflammatory cytokines such ascTNF-α and IL-1β (Cullingford, 2004; Rahman et al., 2014), could also play a role in sleep and slow wave activity regulation. Studies on mice treated with ketogenic dietary therapy have shown that prolonged fasting in free-running conditions advances the phase of body temperature rhythms as well as wheel-running, reflecting a phase-advance in the endogenous circadian clock (Oishi et al., 2013). Even if the neuronal and molecular mechanisms of these circadian clock regulations were not identified, several studies have indicated that calorie restriction or fasting affect the circadian clock (Challet, 2010; Oishi et al., 2013).

Macrostructure changes (increase in REM sleep and a reduction in NREM sleep stage 2) focus attention on anatomical structures such as hypothalamus and ventrolateral preoptic nucleus (VLPO), which are regions controlling REM and non-REM via the locus coeruleus, peduncle-pontine tegmental, and laterodorsal tegmental nuclei. It has been speculated that the increase in REM sleep and decrease in NREM sleep is induced by changes in GABAergic and galaninenergic functions (Hallböök, Lundgren, & Rosén, 2007) in VLPO. Indeed, the study of Lu and colleagues (Lu et al., 2000) showed that selective activation of VLPO galanin neurons increases NREM sleep and their inhibition decreases NREM sleep, thus confirming a pivotal role for VLPO galaninergic regulation of sleep. Data on the influence of ketogenic dietary therapy on galanin levels are not univocal (Tabb et al., 2004; Weinshenker, 2008), however, VLPO galanin neurons contain other neuropeptides besides galanin that can potentially take part in the regulation of sleep structure.

In epilepsy, few studies have considered the effect of reduced interictal spiking during sleep on sleep patterns. The recurrence of interictal epileptic discharges has been recognised to affect cognition and sleep continuity (Manni et al., 2005), even in the absence of clinical or subclinical seizures (Ebus et al., 2012; van Bogaert et al., 2012). Alterations of sleep patterns could account for learning impairment and sustained frequency of interictal epileptic discharges may impair sleep-related learning consolidation processes. Neurophysiological and functional neuroimaging evidence suggests that interictal epileptic discharges may impact cognition through either transient effect on brain processing mechanisms, or more long-lasting effects leading to prolonged inhibition of brain areas distant from but connected with the epileptic focus (van Bogaert et al., 2012). Therefore, a reduction of interictal epileptic discharges during sleep as result of ketogenic dietary therapy could help normalise cognitive processing and restore sleep continuity.

Effects of ketogenic dietary therapies on sleep are reported also in non-epileptic patients. Ketogenic dietary therapies may improve sleep disturbances in migraine patients independently of migraine improvements and anthropometric changes (Merlino et al., 2023). People with migraine might have significantly poorer subjective sleep quality and altered sleep architecture compared with healthy individuals: less total sleep time, lower REM sleep, and shorter sleep onset latency (Stanyer et al., 2021). Poor sleep quality and higher rates of insomnia can act as a trigger or be a consequence of migraine (Merlino et al., 2023).

Ketogenic dietary therapies can be a non-pharmacological treatment for epilepsy in subjects with autism spectrum disorder. Knowledge of sleep changes induced by ketogenic dietary therapies represents a potential new therapeutic approach in these patients, considering on the one hand that these patients commonly present with sleep problems, such as insomnia and circadian rhythm disturbances (Carmassi et al., 2019), and on the other hand that the consequences of sleep problems on patients’ global functioning are not negligible (Han et al., 2022).

In these neurological conditions, an assessment of sleep quality and changes should be implemented before and after the introduction of ketogenic dietary therapies (see Figure 3 for the proposed sleep assessment during follow-up).

5 CONCLUSIONS

Several changes in sleep quality and sleep structure under ketogenic dietary therapies have been outlined in the included literature, namely overall sleep quality, improvement of difficulty falling asleep, nighttime awakenings, increase of REM sleep, and improvement of daytime sleepiness.

These sleep changes can be relevant in the light of their connections with quality of life, cognition, behaviour, and core disease symptoms.

The results of our review suffer from the lack of studies that have sleep assessment as their primary objective, the difference in methodology adopted among the included studies, and the paucity of research aimed at non-epileptic populations. Moreover, no specific research has been carried out on the possible biochemical changes induced by ketogenic dietary therapies leading to changes in sleep structure, in a scenario where the mechanisms of action of ketogenic dietary therapies are not yet clarified.

Looking at future research, since knowledge about sleep effects of ketogenic dietary therapies still represents an unmet need in clinical practice, we suggest a comprehensive screening and characterisation of sleep quality and sleep problems at baseline in patient candidates for ketogenic dietary therapy (drug-resistant epilepsy, migraine, and autism spectrum disorder, etc.), to be monitored during the whole follow-up with standardised sleep questionnaires, clinical observations, and objective (actigraphy/polysomnography) examinations whether a sleep disorder is suspected or documented. In addition, since sleep related findings might deserve a correlation to quality of life and neuropsychological changes, we suggest always including quality of life and neuropsychological assessment in these patients.

Future research should be oriented on secondary outcomes such as the effects of ketogenic dietary therapy on sleep and their correlation with core disease symptoms, cognition and behaviour, and quality of life. Moreover, neurophysiological studies on the objective evaluation of sleep macrostructure and microstructure associated with the consideration of possible biochemical changes during dietary therapies might shed light on possible mechanisms leading to direct and indirect ketogenic dietary therapy-driven changes in sleep.

AUTHOR CONTRIBUTIONS

Ludovica Pasca: Conceptualization; writing – original draft; methodology; writing – review and editing; project administration; investigation. Carlo Alberto Quaranta: Writing – original draft; writing – review and editing; visualization; data curation; investigation. Serena Grumi: Formal analysis; methodology; conceptualization; investigation. Martina Paola Zanaboni: Visualization; validation. Anna Tagliabue: Validation; visualization. Monica Guglielmetti: Validation; visualization. Helene Vitali: Writing – review and editing; visualization. Elena Capriglia: Writing – review and editing; visualization. Costanza Varesio: Visualization; validation. Federico Toni: Validation; data curation. Lino Nobili: Writing – review and editing; validation; supervision. Michele Terzaghi: Supervision; writing – review and editing; methodology; writing – original draft. Valentina De Giorgis: Data curation; supervision; writing – review and editing; writing – original draft.

CONFLICT OF INTEREST STATEMENT

The authors certify that they have NO affiliations with or involvement in any organization or entity with any financial interest, or non-financial interest in the subject matter or materials discussed in this manuscript. We thank the Italian Ministry of Health, “Ricerca Corrente 2022-2024” granted to IRCCS Mondino Foundation.