INTRODUCTION

Alzheimer’s Disease (AD), which is estimated to contribute to 60–70% of late-onset dementia seen in 50 million individuals worldwide, is a complex neurodegenerative disease that includes behavioral changes in addition to progressive decline in cognitive functions such as memory, language, perception, and judgment [32, 47, 67]. The decline in cognitive and functional capacity creates the main causes of dependency and disability in older adults, increasing the global disease burden significantly [41]. Although the pathogenesis of AD has not yet been fully elucidated, the widely acknowledged mechanisms are based on neuronal damage in the hippocampus as a result of the accumulation of amyloid β (Aβ) plaques and advanced phosphorylation of tau proteins. Mitochondrial dysfunction, changes in the blood-brain barrier (BBB), and glucose hypometabolism are thought to contribute to this process [37, 60].

Age is known to be a strong risk factor for late-onset AD, and the prevalence of AD, which is 3% over the age of 65, is reported as 32% the age of 85 and over [1, 22]. Other important risk factors are sex and genetic predisposition. Also, some comorbidities (brain damage, diabetes mellitus, cardiovascular diseases, etc.) and it is reported that modifiable risk factors related to lifestyle, such as physical activity, nutritional status, tobacco, or alcohol habits, can also be effective by 40% [13, 35].

Currently, approved pharmacological treatments of AD have not yet gone beyond being able to alleviate behavioral symptoms by regulating neurotransmitters [48]. Although there are several drugs approved by the Food and Drug Administration (FDA) currently (e.g. memantine, acetylcholinesterase inhibitors), pharmacological advancements have many obstacles [21]. Also, one study reported successful results of only 0.4% of 413 clinical pharmacology studies conducted between 2002 and 2012 [9]. Other treatment approaches are active or passive immunization, γ or β secretase inhibitors, and antiaggregatory drugs. However, none of the aforementioned treatment approaches ameliorate neurodegeneration, stop or alleviate AD progression [56].

Despite being defined in the early 1900s [40], the fact that no effective treatment has yet been found for AD has increased the interest in alternative or adjuvant treatment options based on diet and lifestyle changes [3, 43]. The ketogenic diet is also among the issues that draw attention in this context, due to its potential neuroprotective effects. The purpose of this review is to provide a summary of current evidence on the potential neuroprotective roles of ketogenic diets in AD.

ETIOPATHOGENESIS OF ALZHEIMER’S DISEASE

Although the etiology of AD has not been clarified yet, genetic, and environmental factors are thought to be effective. The pathogenesis differs according to the most common type of the disease, sporadic (late-onset) form or familial (early-onset) AD [33]. According to genome-wide association studies, many genetic loci are associated with AD risk [55]. The primary gene associated with AD risk is Apolipoprotein E (APOE). The epsilon 4 allele (APOE ε4) of this gene is the strongest genetic risk factor for AD, while the APOE ε2 allele is known as the strongest protective genetic factor. Although it is thought that APOE ε4 contributes to AD pathology mainly as a result of its negative effects on Aβ plaque accumulation and clearance, it has been suggested that it may also cause neurofibrillary degeneration, increase in BBB permeability, and changes in astrocyte functions (including neuroinflammation) in recent years [54]. It is also known that the response to AD treatment is mainly related to the APOE ε4 allele [65]. Another important risk factor for AD is insulin resistance or type 2 diabetes associated with reduced glucose metabolism. It is thought that the decrease in glucose uptake to the brain directly affects neuronal activity and brain metabolism, and this process is one of the early markers of AD [12]. Moreover, it has been reported that glucose hypometabolism in the brain may play a key role in the pathophysiology of dementia [31].

The pathogenesis of sporadic form of AD is mainly attributed to cerebral atrophy, amyloid plaques, and neurofibrillary tangles (NFTs). Although many hypotheses have been proposed for the pathophysiology of sporadic AD, the most emphasized hypotheses in clinical studies to date are the amyloid hypothesis, neurotransmitters hypothesis, mitochondrial cascade hypothesis, and tau hyperphosphorylation hypothesis [33]. The amyloid hypothesis, first described in 1991, begins with changes in the processing of amyloid precursor protein (APP) that is involved in the formation of insoluble Aβ fibrils through the control of γ and β secretases [20]. Plaque deposition of Aβ fibrils not only directly inhibits synaptic signals but also causes excessive phosphorylation of tau proteins through activation of “tau protein kinase-1,” leading to an increase in insoluble NFTs. The increase in these plaques and tangles causes a local pro-inflammatory response because of microglial activation, thus neurotoxicity develops, and neurodegeneration progresses, finally, cognitive functions such as learning and memory are impaired [14, 15, 60]. The target of the current therapeutic practices is mainly to prevent the accumulation of Aβ plaques or to increase the clearance of these plaques [21, 24].

The cholinergic hypothesis, first proposed in 1976 [11], is based on an imbalance in the level of neurotransmitters such as acetylcholine, dopamine, noradrenaline as a result of cholinergic system failure in AD [19]. Although acetylcholinesterase inhibitor drugs (e.g. donepezil and rivastigmine) that are FDA-approved are used based on this mechanism, a recent meta-analysis evaluated their effects on dementia as moderate [28]. Another hypothesis, the mitochondrial cascade hypothesis, is based on age-related and oxidative damage-related mitochondrial dysfunction and the associated disruption of epigenetic neuronal regulation. This hypothesis refers to a mechanism of action intertwined with the cholinergic hypothesis, the amyloid hypothesis, and tau hyperphosphorylation [33].

Tau protein is a phosphoprotein found mainly in neuronal axons in the brain and is critical for the maintenance of synaptic functions and neuronal signaling. Phosphorylation of tau protein is controlled by protein kinase/phosphatase activities, so it can be triggered by the increase of Aβ plaques. However, it is estimated that tau hyperphosphorylation has a complex mechanism that has not been fully elucidated. In this context, a relatively clear mechanism is that the increase in NFT because of tau hyperphosphorylation may cause a decreased synapses, cellular dysfunction, and neurotoxicity [14, 56, 60]. Moreover, tau protein is thought to be more closely associated with cognitive functions than Aβ plaque [7]. In recent years, tau hyperphosphorylation has started to draw more attention in the therapeutic focus, as a result of the inefficacy of therapeutic approaches focused mainly on Aβ to date [14].

KETOGENIC DIET

The ketogenic diet (KD), which has been used in modern medicine since the 1920s, was developed mainly for the treatment of refractory epilepsy [64]. Other practice areas include neurological diseases including AD, cancer, diabetes, and psychological disorders [36, 66]. The principle of the ketogenic diet is that it consists of low carbohydrates (≤10% total energy) and high fat. This macronutrient composition causes a systematic shift from glucose metabolism to fatty acid, thereby increasing the use of ketone bodies (KBs) such as acetoacetate (AcAc) and β-hydroxybutyrate (β-OHB) as energy substrates. The physiological effect of KD, which is acknowledged as a kind of biochemical starvation model, is based on promoting the use of KBs instead of glucose, which is the main energy substrate, especially in the central nervous system [39, 57, 59]. The potential of the ketone bodies as using energy substrates to stabilize the mitochondrial metabolism makes they are suitable therapeutic agents [57].

Whereas carbohydrates constitute approximately 45–60% of the daily energy in the traditional diet, with 25–30% of fat and 12–15% of protein, these ratios in classic KD are up to 2–10% for carbohydrates, 80–90% for fat, and approximately 10–15% for protein [51, 52]. Fundamentally, it is carried out with the principle of replacing carbohydrate sources in the diet with fat, and most ketogenic diets mainly contain long-chain fatty acids. Achieving adequate ketosis in these conditions requires radical changes in dietary habits and strict adherence to diet in the long term. Therefore, medium-chain triglyceride (MCT) KD is recommended as an alternative form of classic KD. In this KD model, carbohydrate restriction is more moderate (10–20% E), while total fat content may be lower (70% E) than in the classic ketogenic diet. MCTs are superior to long-chain fatty acids for faster and more effective ketosis, as they are easily absorbed in the intestinal lumen after being broken down into medium-chain fatty acids and transported to the liver and undergo rapid β-oxidation [2, 52]. MCT-KD generally contains 60% octanoic acid (8-carbon fatty acid), 40% decanoic acid (10-carbon fatty acid) [2]. Another type of KD is the modified Atkins diet (MAD), in which fat is around 70% of the daily energy and carbohydrate is limited to 5%, but there is no change in the energy or protein [66].

POTENTIAL NEUROPROTECTIVE EFFECTS OF THE KETOGENIC DIET

The existing hypotheses regarding the mechanism of the ketogenic diet can be examined under two headings: the effects of direct carbohydrate restriction in the diet and the use of fatty acids (and therefore KBs metabolism) for energy [6, 53].

The possible effects of carbohydrate restriction are related to the fact that—as already mentioned—impaired glucose metabolism is observed as one of the early markers of AD. Also, down-regulation of GLUT1 and GLUT3, which are major glucose transporters in the brain, correlates with tau hyperphosphorylation and decreased glucose uptake in the brain, increasing the risk of cognitive dysfunction and AD [29, 34]. On the other hand, the fact that KBs uptake to the brain does not change in AD, unlike glucose, has also supported the prominence of KBs as an alternative energy source [10]. Thus, a metabolic intervention such as KD may be a strategy to prevent or alleviate the progression of AD, due to by suppressing the glycolytic pathway. In ketogenic metabolism, acetyl-CoA is synthesized from AcAc and β-OHB and enters the Krebs cycle, thus skipping the glycolysis step in conditions such as metabolic stress. KBs are faster energy sources and suppress glycolytic ATP production, while increasing mitochondrial oxidation and ATP production, and thus may alleviate mitochondrial dysfunction, which has an important role in neurodegeneration [63]. It has also been reported that KD can upregulate hippocampal genes encoding mitochondrial and energy metabolism enzymes [4]. All these physiological effects, beyond the potential benefits for energy metabolism or brain and metabolic regulation, may also mitigate the adverse effects of impaired glucose metabolism in AD [6, 25, 53].

Ketone bodies can affect cellular signaling pathways indirectly through AMP-activated protein kinase (AMPK) and mechanistic target of rapamycin (mTOR) mechanisms and Peroxisome proliferator-activated receptor-alpha (PPARα). Although multiple mechanisms have been identified in this context, the fact that PPARα is a key transcriptional factor in lipid metabolism and related ketosis regulation forms the basis of these interactions [18, 53].

In addition to its cellular metabolic effects, KBs have also been reported to prevent the accumulation and neurotoxicity of Aβ plaques [27, 62]. One of the related mechanisms may be the ameliorating effect of KD on mitochondrial dysfunction. Also, it is thought that mitochondrial dysfunction may cause Aβ increase and neurotoxicity by triggering oxidative damage and causing changes in APP production [26].

Oxidative damage and inflammation, which play an important role in Alzheimer’s neurotoxicity, are other potential target areas of KD [46]. Ketosis suppresses ROS production by improving mitochondrial respiration. Fundamentally, β-OHB can increase the efficiency of the electron transport chain by activating Nrf2 (nuclear factor erythroid-derived 2-related factor 2), stimulating the stimulation of the cellular antioxidant system, modulation of the ratio of oxidized and reduced nicotinamide adenine dinucleotide (NAD+/NADH), and synthesis of uncoupling proteins (UCPs) [46]. The anti-inflammatory activity is attributed to the following mechanisms: (1) direct suppression of systemic inflammation through NF-kB (nuclear factor kappa-light-chain-enhancer of activated B cells) inhibition and PPARγ activation. (2) β-OHB passing through the BBB activates the HCA2 (hydroxyl-carboxylic acid receptor 2) receptor synthesized in macrophages, and microglia, reducing neuroinflammation. (3) β-OHB down-regulates the synthesis of pro-inflammatory cytokines (IL-1, IL-8) by suppressing the NLRP3 (NOD-like receptor 3) inflammasome [46, 68].

CLINICAL STUDIES

The APOEε4 allele, which is an important risk factor for AD, is a factor that directly determines the response to ketosis intervention. According to a study about this scope, while ketogenic agent (carpylidene) intervention increased cerebral blood flow in the lack of the ε4 allele (APOEε4–), it has not any significant effect in individuals with ε4 (APOEε4+) [61].

The first randomized controlled study on the cognitive effects of ketogenesis in older adults was conducted by Reger et al. In this study, the acute (90 min) effect of oral MCT (40 mL) supplementation was evaluated in 20 individuals with a diagnosis of mild cognitive impairment (MCI) or AD. It has been reported that increasing β-OHB concentration correlates with improvement in cognitive functions, while Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-cog) and paragraph recall are improved only in APOEε4(–) individuals [50]. Another study on the effect of MCT examined the acute (2 h) and long-term (90 day) effects in 152 individuals with mild to moderate AD. Similarly, increased β-OHB concentration was associated with higher ADAS-cog score only in APOEε4(–) individuals at days 45 and 90 [23].

Studies using different forms of MCT (encapsulation or with carriers such as ketogenic formula, milk, or yogurt) have used doses varying from 20–30 g, and only one study used 56 g per day [49]. The administration period generally varied between 3 and 6 months. The results of these studies are consistent with the point that MCT supplementation improves verbal memory, attention, or total cognitive score (ADAS-cog) through increased circulating KBs. However, studies have also reported that a significant response is generally observed in the lack of the APOEε4 allele [16, 17, 42, 44, 49, 61]. The reviewed studies are summarized in Table 1.

Table 1.   Summary for human studies with evidence of Alzheimer’s Disease or mild cognitive impairments
Full size table

Many studies have examined the effects of ketosis with MCT supplementation without changing the routine diet. However, studies on the effect of carbohydrate-restricted or ketogenic diets are still limited. In one of these studies, in individuals with MCI, a 6‑week low-carb diet (5–10% E; dietary ketosis) resulted in improved verbal memory, correlated with increased circulating KBs, compared to a high-carb diet (50% E) [30]. In another study, the effectiveness of a ketogenic diet (<15 g/d carbohydrate) versus a low-fat diet (<10 g/d fat) for 6 weeks (and a 6-week washout period) in pre-diabetic MCI individuals was investigated. While there was no change in the low-fat diet, KD increased ketosis in the brain and increased total tau protein in the cerebrospinal fluid (CSF) [8]. These results are particularly striking in the context of delaying AD in individuals at higher risk (MCI/prediabetic).

Two studies were found that evaluated the 12-week effect of parallel dietary interventions. In the first of these studies, the effectiveness of MAD against a healthy diet model recommended by the National Institute of Aging (NIA) for the elderly in the MCI/early AD period was examined. According to MAD criteria, daily net carbohydrate was limited to a maximum of 20 g, while fat was high to provide satiety, the protein was kept at the required level. The diet of the control group (NIA) was applied with prescriptions suitable for a diet that is rich in vegetables-fruits, seafood, legumes, oilseeds, while limited in saturated fat and processed foods. Although the memory composite score did not differ between the groups, it was observed that episodic memory and mood improved in the MAD group compared to baseline [5]. In another study, in addition to the cognitive effects of the ketogenic diet in AD, its effects on daily activities and quality of life were examined. According to the results of the crossover study evaluating the modified KD (fat, 58% E) or low-fat (11% E) control diet, the KD intervention improved daily activity and quality of life, although it did not improve cognitive functions [45].

In a pilot study conducted on 15 individuals with mild to moderate AD, a high-fat (70% E) diet with a KD ratio of at least 1 : 1 was applied for 3 months, and then mental tests were re-evaluated after a 1-month washout period. At the end of the 3-month period, a significant increase in ADAS-cog scores (mean 4.1) was reported if the target ketosis was achieved. However, after the washout period, cognitive scores did not differ from the baseline values [58]. This finding is critical in drawing attention to the uncertainty of the persistence of the beneficial effects of KD and the optimal period of administration.

CONSIDERATIONS AND SIDE EFFECTS

Studies on the use of the KD in AD generally include interventions of 3–6 months, no serious side effects have been reported, and generally well tolerated. However, it should also be considered that KD may have some risks for older adults, especially in the long term. Short-term side effects are associated with a high-fat content by the nature of KD, may include nausea, constipation, diarrhea (especially MCT), anorexia, transient dyslipidemia, hypoglycemia, dehydration, and weight loss. Gastrointestinal side effects can usually be managed with dietary interventions (such as smaller portions and frequent meals or modulation of fiber, fluid, and sodium intake). For this reason, MCT is used gradually increasing. As a matter of fact, the tolerance and safety of the applications were evaluated at an “acceptable” level in many clinical studies [5, 17, 45]. Relatively longer-term risks include atherosclerosis, nephrolithiasis, liver dysfunction, protein/micronutrient deficiencies, decreased bone mineral density, neuropathy in the optic nerves [38, 66].

Unintentional weight loss is an important issue that needs to be addressed especially in older adults. Although it is a desired effect of KD to provide weight loss in overweight/obese individuals, it should be followed carefully as it may pose a risk for general health and functional capacity in older adults. It was also reported that KD provided a significant reduction in body weight and BMI in a 3-month period [45]. It should be noted that weight loss is an important potential KD side effect in this population, especially considering that ketosis (fat, 58% E) in this study was not that strict compared to other KD studies.

Ketogenic diets may induce nutrient deficiencies (especially micronutrients), due to their macronutrient composition and the nature of ketosis that suppresses appetite. In this context, according to a study by Taylor et al., a KD rich in MCT oil, eggs, olive oil, oilseeds, and especially non-starchy vegetables and avocado can provide the targeted macronutrient composition (fat 70% E, carb <%10 E). If there is a “healthy ketogenic diet” planned by a registered dietitian, it has also been shown in this study that the recommended dietary allowances (RDA) can be met in older adults [59].

CONCLUSIONS

In the absence of effective treatment approaches for slow or stop AD, evidence for KD applications holds promise for more than symptomatic treatment, especially in MCI/early AD. However, the current evidence varies widely in study design, participants, and outcome variables, and clinical effects over 6 months are not yet known. Therefore, it can be said that there is still a long way to go before the application of KD in older adults can be accepted as a therapeutic approach. The fact that it is not an easy-to-apply diet model and that it requires strict compliance are difficulties that may complicate longer-term studies. On the other hand, the optimal KD method and duration of application, how long the beneficial effects are permanent are still uncertain issues. Therefore, further clinical studies focusing on the neuroprotective effect of ketogenic diet approaches along with the general health and nutritional status of older adults in the long term are critical.