Effects of Ketone Bodies on Endurance Exercise

Abstract

Priorities for every athlete include improving endurance performance, optimizing training, nutrition, and recovery. Nutritional strategies are crucial to support athletes to perform at the highest level, and considering that muscular and hepatic glycogen stores are limited, alternative strategies to maximize fat metabolism have been suggested. A ketogenic diet has been proposed as a possible method of providing metabolic fuel during prolonged periods of exercise. However, clinical trials and empirical experience have produced contrasting results regarding the ergogenic value of a ketogenic diet. For this reason, using ketone esters and/or salts have been proposed to obtain nutritional ketosis without limiting carbohydrate intake. Exogenous ketones should not only represent an alternative metabolic fuel source, sparing carbohydrates, but they also may increase postexercise glycogen replenishment, decrease proteolysis, and act as metabolic modulators and signaling metabolites. While there are some encouraging results showing an increase in endurance performance, contrasting evidence regarding the efficacy of exogenous ketones for endurance performance is present and further studies should be performed to make a definitive statement.

Introduction

In the pursuit of excellence, both amateur and professional athletes are looking for innovative strategies and/or techniques to improve their performance. In this framework, any gain, both large and small, may represent a priority considering that the gap separating winners and losers is seconds. Nutrition, training, recovery, and ergogenic aids represent the typical areas where athletes and their staff may find new ways to improve their performance; in particular, nutrition and supplementation has recently gained the attention of both athletes and sport scientists alike (^1–3). It is well known that in endurance exercise performed at medium and high intensity, carbohydrates represent the main metabolic substrate (⁴). Metabolism of fat is often insufficient to sustain prolonged, high-intensity exercise. For example, maximum fat oxidation rate (FatMax) in endurance athletes ranges from 0.4 to 0.6 g·min⁻¹ reached at 40% to 60% of maximum oxygen uptake (V˙O_2max) (⁵), whereas in fat or ketone adapted ultraendurance athletes performing specific training sessions, FatMax may reach 1.2 to 1.3 g·min⁻¹ (^6,7). A FatMax of 0.4 to 0.6 g·min⁻¹ corresponds to ~3.64–5.46 kcal·min⁻¹, whereas 1.2–1.3 g·min⁻¹ corresponds to ~10.92 to 11.83 kcal·min⁻¹. Running a marathon at 2:06:36 (h:min:s) requires an average speed of 20 km·h⁻¹ which corresponds to ~22–23 kcal·min⁻¹ considering a gross efficiency of ~200 mL O₂·kg⁻¹·km⁻¹, whereas cycling at 300 W requires ~18 kcal·min⁻¹ considering a gross efficiency of ~22% (^8–10). It is clear that even the highest FatMax is not enough to sustain the mechanical power required by some endurance events. Fat stores are virtually limitless considering that an a 70-kg endurance athlete with 4% of fat mass may store roughly 3 kg of fat mass corresponding to ~27000 kcal, but the inability of fat metabolism to support high intensity efforts impairs high-end performance representing a significant limit of fat diets. At the opposite, the oxidation of glucose deriving from glycogenolysis is able to sustain high mechanical power but it also should be considered that muscular and hepatic glycogen stores are limited and an adapted athlete following a strenuous protocol of carbohydrate loading may store 200–220 mmol of glycogen·kg⁻¹ d.w. muscle which corresponds to ~500 to 700 g of glycogen (2000–2800 kcal) (^11,12). This quantity is not enough to end a marathon at 20 km·h⁻¹ (i.e., 2:06:36 h:min:s) which costs roughly 2800 kcal, Indeed, the low muscle mass of the best marathon runners is able to store ~2000 kcal in muscular mass involved in running. Aside from running, a 1-d stage in cycling may require ~5 or 6 h of effort performed at average power output of 270 to 320 W not considering sudden changes of pace (corresponding to 4000–6000 kcal) or an Ironman triathlon (i.e., 3.8 km swim, 180-km time trial and a 42-km run) that elite athletes complete in ~8 h considering 50 min for the swim leg, 4:30 (h:min) for the bike leg and 2:50 (h:min) for the run leg corresponding to ~8000 kcal. It should be considered that exogenous carbohydrates supplementation may increase carbohydrates availability, but exogenous carbohydrates oxidation rates is limited to 1.8 g·min⁻¹ if the correct blend of sucrose and fructose is introduced. Furthermore, this high amount may easily cause gut distress and nausea limiting the ability of introducing carbohydrates during training and/or racing (¹³).

Ketogenic diets (KD), ketone salts (KS), and/or ketone esters (KE) have gained the interest of many sports scientists who suggest that ketone bodies (KB) may represent a near-endless metabolic substrate able to sustain even the highest metabolic power required by some endurance events.

Physiology

The word ketone comes from the German word Aketon, for “acetone,” which is the simplest ketone (¹⁴). Ketone bodies are organic, lipid-derived compounds produced in the liver during low levels or absence of carbohydrate intake as seen in starvation, KD, or prolonged fasting to provide a biochemical fuel source for peripheral tissue (¹⁵). The term “ketone bodies” refers to three molecules, acetoacetate (AcAc), 3- β-hydroxybutyrate (β-OHB) and acetone. However, only AcAc and acetone are “ketones” containing a carbonyl group with two hydrocarbon atoms. β-OHB is a KB, but is not a ketone since one of its hydrocarbon atoms is replaced by a hydroxyl group. The two most abundant KB are AcAc and β-OHB, but only AcAc is produced in the liver and can be subsequently enzymatically converted to β-OHB or spontaneously degraded to acetone which is the less abundant KB (¹⁶).

Ketogenesis represents the process in which fatty acids are converted into AcAc in mitochondria of perivenous hepatocytes in the presence of different factors. Low levels of insulin, high levels of glucagon and epinephrine stimulate the release of fatty acids from adipose tissue which are converted to acetyl CoA by β-oxidation in the liver. During periods of low carbohydrate intake, glycolysis produces small amounts of pyruvate and consequently low quantities of oxaloacetate are produced in the mitochondria via pyruvate carboxylase (¹⁷). Even a low-medium protein intake below 1.7 g·kg⁻¹·d⁻¹ is necessary to induce ketogenesis; indeed high amount of protein stimulates gluconeogenesis increasing glucose availability and suppressing ketogenesis (¹⁸). On average, 1.6 g of amino acids is required to synthesize 1 g of glucose (¹⁹). Moreover, low intracellular glucose causes oxaloacetate to be preferentially utilized in the process of gluconeogenesis, instead of condensing with acetyl CoA. The high ratio acetyl CoA/oxaloacetate diverts acetyl CoA to KB formation (²⁰). Subsequently, acetyl CoA is converted to acetoacetyl CoA. Acetoacetyl CoA is transformed into hydroxymethylglutaryl (HMG) CoA by mitochondrial HMG CoA synthase. HMG CoA is then cleaved to liberate AcAc in a step mediated by HMG CoA lyase. The reduction of AcAc to β-OHB is catalyzed by 3-HB dehydrogenase (HBD), and acetone is formed by the spontaneous decarboxylation of acetoacetate (²¹). In healthy adults, the liver is capable of producing up to 150–185 g·d⁻¹ of KB (^16,21,22). Once produced, KB cross mitochondrial and cellular membranes and are transported to peripheral tissue via the bloodstream (i.e., brain, heart, muscle) into the cytosol and mitochondria via the monocarboxylate transporters (MCTs). In mitochondria they are converted back to acetyl-CoA via β-ketoacyl-CoA transferase, not present in liver, which uses succinyl-CoA as the CoA donor, forming succinate and acetoacetyl-CoA (²³). Ketone bodies can be completely oxidized as a fuel source by skeletal muscle and they have a similar RER to glucose, 1.0 and 0.89 for AcAc and for β-OHB, respectively (¹⁶).

After overnight fasting ketonemia ranges from 0.1 to 0.5 mmol·L⁻¹, whereas after 5 d or more of fasting ketonemia may reach 7 to 10 mmol·L⁻¹. Four days of KD are able to raise ketonemia to 1 to 2 mmol·L⁻¹; however, plasma KB concentrations are strictly dependent on carbohydrate intake and physical exercise and they may reach 7 to 8 mmol·L⁻¹. Physical exercise, in particular if performed in the fasted state and for a prolonged period, is able to raise ketonemia to 0.5 to 1 mmol·L⁻¹ and in the postexercise period ketonemia may raise to 1 to 4 mmol·L⁻¹ if no carbohydrates are ingested (¹⁶). Ketone ester ingestion may raise plasmatic KB concentrations from 1 to 5 mmol·L⁻¹ (¹⁵). Veech also suggests an increased efficiency of KB in generating metabolic energy compared with glucose and fatty acids. Indeed, KB could increase the hydraulic efficiency of heart by 28% compared with glucose in animal models as observed by Sato et al. (^16,22,24,25). A possible explanation may be represented by the ability of KB metabolism to induce changes in mitochondrial electron transport system increasing increasing the energy of the transport chain redox span between site I and site II. This results in an increased energy release by the electron traveling across that span by means of the reduction of nicotinamide adenine dinucleotide (NAD⁺) and the oxidation of the Coenzyme Q (CoQ) couple and in an increase in redox energy of respiratory chain (²⁶). As a consequence, the energy of the protons ejected from the mitochondria at the energy-conserving sites increases causing a corresponding increase in the energy of adenosine triphosphate (ATP) hydrolysis (²⁷). Such increase in ATP hydrolysis may be related to the intrinsec higher heat of β-OHB combustion than pyruvate; indeed the combustion in a bomb calorimeter of β-OHB is able to produce 31% more calories per C₂ unit than pyruvate. The greater enthalpy generated by β-OHB derives from the fact that β-OHB is more reduced than pyruvate showing a higher ratio of hydrogen:carbon present in each molecule (²²). Furthermore, KB metabolism may generate energy more efficiently even if compared with palmitate which is more reduced than KB. Indeed, three aspects should be considered: 1) the redox potential of the CoQ couple is not oxidized as it is during KB metabolism, but rather reduced, decreasing energy available for ATP synthesis; 2) during β oxidation, there is a significant loss of efficiency in producing ATP because half of the reducing equivalents enter at a flavoprotein site with an lower potential compared with NAD couple reducing the redox span between NAD⁺ couple and CoQ couple; 3) elevation of free fatty acids increases uncoupling proteins generating heat production (²²). Translating these aspects into skeletal muscle metabolism, the better efficieny of KB should be defined by the ability to generate more power for the same oxygen consumption requiring less oxygen per mole of carbon during their oxidation (²⁸). Indeed, endurance performance may be considered as a product of three physiological variables: V˙O_2max, lactate threshold and efficiency. V˙O_2max is determined by cardiac output, muscular blood flow, oxygen carrying blood capacity and mitochondrial volume density; therefore, peripheral oxygen availability potentially represents a limiting factor for endurance performance (^29,30). The better efficiency showed by KB in energy production compared with glucose should facilitate a higher-power output at the same V˙O₂, increasing maximal performance; at the same time, this efficiency should decrease V˙O₂ at the same power output and therefore increasing submaximal performance (i.e., time to exhaustion). Nevertheless, this aspect deserves to be evaluated in depth with further studies. It also should be considered that KB represent an alternative fuel source to glucose allowing the athlete to spare glucose and preserve glycogen stores for high intensity efforts, such as the final climb of a cycling stage or the last 10 km of a marathon. A decreased reliance on carbohydrates also may be helpful to support the high volume of training considering both the limited human glycogen stores and the slow process of muscular glycogen replenishment.

Ketogenic Diet

Ketogenic diet represents the most intuitive way to obtain biological ketosis forcing the body to produce KB by greatly reducing carbohydrate intake. Ketogenic diet is generally characterized by a total carbohydrate intake of less than 50 g·d⁻¹ and a moderate protein intake of approximately 1.5 g·kg⁻¹·d⁻¹ to induce ketogenesis (^31,32). Ketogenic diet has been representing an effective treatment for refractory epilepsy in children since the early 20th century (³³). The exact mechanism responsible for anticonvulsant properties is unclear; postulated hypotheses include reduced glucose utilization/glycolysis, reprogrammed transport, indirect impact on ATP-sensitive potassium channel or adenosine A1 receptor, alteration of sodium channel isoform expression, or effects on circulating hormones including leptin (^34,35). An interesting hypothesis suggests that raising resting membrane potential may inhibit the synchronous neuronal discharge characteristic of epilepsy (²²).

A number of variations of the Hopkins KD also has been considered as a treatment for weight loss, cancer chemotherapy adjuvant, Acyl CoA dehydrogenase deficiency, neurodegenerative diseases, and muscle wasting (²²). A well-formulated KD also may increase omega-3 daily amount if correct fat sources are chosen, with consequent decrease of inflammation and of insulin resistance and increase of FatMax and of muscular anabolic response to stimuli representing a therapeutic aid for diabetes and sarcopenia (^36,37). A well-known method to induce hyperketonemia also is represented by repeated ingestion of medium-chain triglyceride (MCTG); 20 to 30 g·d⁻¹ of MCTG may be a proper quantity to increase ketonemia to 0.29 mmol·L⁻¹ (³⁸). The first studies regarding the effects of KD on endurance performance were performed almost forty years ago by Phynney et al. considering the typical diet of Inuit observed by western scientists during expeditions in Arctic (^39–41). In particular, Phynney et al. observed that KD preserved time to exhaustion performed at 65% of V˙O_2max in highly trained athletes while respiratory quotient decreased to 0.72 indicating that fat metabolism was the sole substrate used by the athletes. It is worth noting that increasing reliance on fat oxidation and, consequently, on oxidative processes, may promote an adaptive response to oxidative stress induced by mitochondrial reactive oxygen species; this response is called mitohormesis, and it may enhance mitochondrial function (³¹). Other studies have investigated the effects of KD on different populations and also the impact of KD in athletes of different sports have been studied (^42–48). Nevertheless, when evaluating the outcome of these studies it should be considered that the metabolic demands of top level endurance exercise are complicated because they also include high intensity efforts performed at or above anaerobic threshold, and the evidence observed in sedentary and/or active subjects cannot be translated to elite athletes performing endurance exercise including high-intensity efforts (¹⁶). Indeed, endurance exercises requiring short bursts of high intensity efforts above anaerobic threshold and/or prolonged periods of time spent at intensity of anaerobic threshold may be impaired by KD. High levels of ketonemia may impair glycolysis limiting the functionality of fast twich fibers. Otherwise, ultra endurance performance and other endurance activities performed at low-medium intensity may be positvely affected by KD.

The main concern of KD may be represented by the need to train and race in the presence of scarce levels of carbohydrate. This aspect requires a keto-adaptation of at least 2 to 4 wk during which athletes may show signs of weakness, asthenia, lethargy, drowsiness and poor performance (^18,49,50). Despite no standard definition exists to define the ideal range of carbohydrate intake to achieve ketosis because of individual variability, a carbohydrate daily intake of 0.5 g·kg⁻¹·d⁻¹ or <50 g·d⁻¹ may represent an accurate assessment (¹⁴). Daily protein intake should range from 1.76 to 2.2 g·kg⁻¹ lean mass per day or 1.2–1.7 g·kg⁻¹ body weight per day, and the remaining calories should come from fat, covering 70% to 80% of daily energy intake (^31,32). It should be taken into account that a high daily intake of proteins may inhibit ketogenesis by stimulating gluconeogenesis (⁴⁹). Furthermore, very low calorie KD may require to be supplemented by sodium and potassium to maintain daily intakes for sodium at 3 to 5 g·d⁻¹ and for potassium at 2 to 3 g·d⁻¹ (¹⁸). Indeed, blood insulin levels tend to decrease in KD causing an increase in renal sodium excretion (^45,51). Consequently, it can be supposed that the adrenal cortex increases the compensatory secretion of aldosterone augmenting sodium reabsorption and potassium excretion. A study by McSwiney et al. showed promising results demonstrating that after 12 wk of KD (%carbohydrate:protein:fat = 6:17:77, 41.1 ± 13.3 g carbohydrates·d⁻¹, 3022.3 ± 911.1 kcal·d⁻¹), amateur athletes lost significant fat mass compared with athletes following an high carbohydrate diet (HCD) (%carbohydrate:protein:fat = 65:14:20, 2643.6 ± 358.0 kcal·d⁻¹). Body fat mass significantly decreased in the KD group, with a loss of 4.6 kg compared with 0.5 kg in the HCD group and they showed significant improvements in relative critical power in 3 min all-out cycling test (pre, 8.3 ± 2.2 w·kg⁻¹ vs post, 9.7 ± 2.3 w·kg⁻¹) and relative peak power during single 6-s sprint cycling test (pre, 13.7 ± 1.4 w·kg⁻¹ vs post, 14.5 ± 1.1 w·kg⁻¹) compared with baseline (⁴⁹). Also, relative critical power in 3 min all-out cycling test (HCD, 8.4 ± 2.2 w·kg⁻¹ vs KD, 9.7 ± 2.3 w·kg⁻¹) and relative peak power during single 6-s sprint cycling test (HCD, 13.8 ± 2.2 w·kg⁻¹ vs KD, 14.5 ± 1.1 w·kg⁻¹) were significantly higher in the KD group compared with the Hc group at the end of the treatment.

Absolute critical power in 3-min all-out cycling test and absolute power during single 6-s sprint cycling test did not differ significantly between the groups (⁴⁹). Moreover, the KD group showed a marked but not significant decrease in time required to complete 100-km cycling test (HC, 168.44 ± 9.14 min·s⁻¹vs KD, 161.53 ± 8.44 min·s⁻¹), and a significantly inferior RER, indicating a greater reliance of fat metabolism compared with the high carbohydrate group (⁴⁹). In evaluating this marked but not significant improvement in 100-km cycling test in the KD group, it should be considered the response of two subjects suggests a large individual variation in response to a KD. Indeed, these two subjects showed an improvement of ~12 min compared with an average improvement of 5 min in the KD group. Similarly, Zinn et al. showed a significant higher FatMax (pre, 0.6 ± 0.1 g·min⁻¹ vs post, 0.8 ± 0.1 g·min⁻¹), a significant increase in exercise intensity relative to V˙O_2max at which peak absolute FatMax occurred (pre, 48.2% ± 8.7% vs post, 63.2% ± 5.7% V˙O_2max) and a significant reduction in body weight (−4.0 ± 3.1 kg) following KD. Nevertheless, the incremental cycle test showed a significant decrease in time to exhaustion (−2 ± 0.7 min), and a marked but not significant reduction in peak power (−18 ± 16.4 W), in V˙O_2max (−1.69 ± 3.4 mLO₂·kg⁻¹·min⁻¹), and ventilatory threshold (−6 ± 44.5 W) following KD. However, it should be considered that only five amateur athletes were involved in this study and they were provided with a daily macronutrient prescription of <50 g total carbohydrate, 1.5 g·kg⁻¹ protein and ad libitum fat (⁵⁰). Similar results were observed in 42 healthy, nontrained subjects (V˙O_2max: 36.7 ± 8.5 mL·O₂⁻¹·kg⁻¹·min⁻¹) following KD for 6 wk. Subjects were asked to follow a KD according to their personal preferences but were advised to eat ad libitum but limit their carbohydrate intake to a maximum of 20 to 40 g·d⁻¹ to derive at least 75%, 15% to 20%, and 5% to 10% of total energy from fats, protein, and carbohydrates, respectively. 7-dy food records showed that the subjects consumed on average a KD 2224 ± 584 kcal with 71.6%, 20.9%, and 7.7% of total energy intake from fat, protein, and carbohydrate, respectively. A significant reduction in body weight (−2.0 ± 1.9 kg) and RER at rest (pre, 0.86 vs post, 0.79) was observed, but, at the same time, absolute power (pre, 241 ± 57 vs post, 231 ± 57 W) and V˙O_2max (pre, 2.55 ± 0.68 L·min⁻¹ vs post, 2.49 ± 0.69 L·min⁻¹) decreased during maximal incremental cycling test (⁵²). Zajac et al. reported similar outcomes in eight off-road cyclists with a training experience of at least 5 years and a minimal V˙O_2max of 55 mL·kg⁻¹·min⁻¹ after KD for 4 wk and HCD for other weeks in crossover design. The HCD included 50% carbohydrates, 30% fats, and 20% protein, while the KD was composed of 70% fat, 15% protein, and 15% carbohydrates. In these subjects a significant reduction in fat mass (HCD, 14.88 ± 3.78 kg vs KD, 11.02 ± 3.66 kg) and RER at rest (HCD, 0.88 ± 0.04 vs KD, 0.76 ± 0.01) and during submaximal cycling test (HCD, 0.84 ± 0.03 vs KD, 0.79 ± 0.02, at 90 min) was observed. Nevertheless, incremental cycling test showed that power at V˙O_2max (HCD 362 ± 16.09 vs KD 350 ± 14.60 W) and power expressed at lactate threshold (257 ± 10.60 vs KD 246 ± 9.50 W) decreased significantly in the KD group compared with the HCD group. Otherwise, relative V˙O_2max (HCD 56.02 ± 3.50 vs KD 59.40 ± 3.10 mL·kg⁻¹·min⁻¹, and relative V˙O₂ at lactate threshold (43.5 ± 1.8 mL·kg⁻¹·min⁻¹ vs 47.8 ± 2.1 mL·kg⁻¹·min⁻¹) significantly increased in the KD group compared with HCD group. This evidence may be explained by the significant weight loss observed in the KD group (⁵³). It is worth mentioning the study performed by Burke et al. for 3 wk on 21 world class walkers divided into three groups according to three different dietary approaches: HCD (8.6 g·kg⁻¹ carbohydrates, 2.1 g·kg⁻¹ protein, 1.2 g·kg⁻¹ fat), periodized carbohydrates diet (PHCD) (8.3 g·kg⁻¹ carbohydrates, 2.2 g·kg⁻¹ protein, 1.2 g·kg⁻¹ fat), low carbohydrated high fat diet (LCHF) (<50 g·d⁻¹ carbohydrates, 2.2 g·kg⁻¹ protein, 4.7 g·kg⁻¹ fat). Athletes’ performance was evalauted through an incremental walking test, a 10-km race and a 25-km standardized long walk. In the KD group FatMax improved significantly from the pretreatmen value of 0.62 ± 0.32 g·m⁻¹ to posttreatment of 1.57 ± 0.37 g·m⁻¹, whereas in the HCD and PHCD groups no change was observed in FatMax. Moreover, oxygen cost of race walking increased significantly in the KD group indicating an impairment in walking economy. The HCD and PHCD groups improved significantly 10 km race performance, showing 6.6% and 5.3% improvements in performance following the 3-wk diet and training intervention, respectively. In the KD group, 10-km race performance showed an impairment by 1.6%.

These outcomes provide insufficient evidence supporting ergogenic properties of KD on endurance performance; however, KD may be beneficial for athletes representing an aid for weight loss and for supporting training sessions performed at low-moderate intensity for long periods of time. High-volume low-intensity trainings are typical of the early season in the classical periodization of annual training schedule (^54,55). Indeed, stressing fat metabolism represents a priority of this period and KD may be helpful to increase FatMax in presence of low-moderate intensity with a synergic effect. Moreover, high volume of training requires an adequate caloric intake that may be difficult to obtain with high carbohydrate diet and it may detrimental for increasing fat oxidation considering the hyperinsulinism induced by high intake of carbohydrates. KD may represent a good option to support high energetic demand and, at the same time, to maximize fat oxidation. The ability of KD to support low-medium intensity efforts in ultra-endurance events without stressful carbohydrate-loading prior the event and exogenous carbohydrates supplementation during the event also may be helpful to avoid gastrointestinal upset and logistic problems.

In some specific cases and sports, KD may be used to facilitate weight loss in a short period of time without affecting performance. Nevertheless, KD may be a support in limiting weight gain in off-season, as often observed in athletes involved in endurance events.

Ketogenic Salts and/or Esters

Taking into account the clinical studies and the empirical experience of several athletes suggesting an ergogenic effect of KD in endurance events requiring medium and higher intensity, it appears that low muscular carbohydrate content during heavy sustained exercise impairs physical performance. Consequently, sport scientists proposed a mixed solution that matches both the benefits of a high carbohydrate diet and of the presence of KB. This proposal requires athletes to follow a high carbohydrate diet to guarantee adequate muscular and hepatic glycogen content, and, at the same time, to ingest a proper quantity of KS or KE before the efforts to ensure KB availability (Table) (⁵⁶). Ketone esters are compounds that are created through an ester linkage between a ketone bodies and an alcohol, whereas KS comprise of the free acid form of β-OHB buffered with sodium, potassium, and/or calcium salts (^26,57). Indeed, ingestion of KS or KE represent a fast and effective way to deliver nutritional ketosis for at least 2 h, with KE better tolerated and more effective in elevating blood KB levels compared with KS (⁵⁶). Ingestion of large quantities of KS is impractical due to resulting gastrointestinal distress, and potentially undesirable consequences of cation overload or acidosis (¹⁵). To be effective, KS and/or KE ingestion should not affect gastric emptying, carbohydrate intake or cause gut distress (¹⁶). The rate of KB uptake and oxidation by skeletal muscle during exercise at different intensities and duration also should be considered to provide an adequate quantity of KB to exercising muscles (^15,58). Ketone ester consumption rather than KS at fasted state with repetead administration may improve tolerability; moreover, ingestion of β-OHB monoester rather AcAc diester using flavored water may decrease gastrointestinal symptoms (^56,59).

Table: Summary of the effects of KE or KS administration on maximal cycling performance.

The first evidence regarding the ergogenic effects of KE supplement comes from data reported in a patent application filed in April 2013 where the administration of a solution containing 230 kcal of KB taken 60 min before exercise in a fasted state improved 30-min rowing performance (averaging 1% and up to 2%) in 22 elite and subelite heavy and lightweight male and female rowers. These improvements were not considered significant but in professional sports races are often won or lost with differences in performance less than 1% (¹⁶). Cox et al. confirmed an improvement of ~2% in the distance covered in 30-min cycling time trial in the group administered with KE (573 mg·kg⁻¹) + glucose (60% of total calories) compared with glucose alone. Indeed, the KE group cycled on average 411 ± 162 m further over the control group. Ketone ester administration resulted in plasma β-OHB concentration of 2 to 2.5 mmol·L⁻¹. Nevertheless, this study also elucidated many effects of KE on human metabolism during endurance exercise. First of all, circulating β-OHB levels decreased significantly when cycling intensity rose from 40% to 75% of V˙O_2max indicating an increasing β-OHB oxidation rate, which increase from 0.35 g·min⁻¹ at 40% V˙O_2max to 0.5 g·min⁻¹ at 75% of V˙O_2max. The estimated amount of β-OHB oxidation to total oxygen consumption ranges from 16% to 18% with a net sparing effect of glucose oxidation. Ingestion of KE 15 min before the start of exercise and 45 min into the cycling trial significantly lowered blood lactate by ~50% (~2–3 mM) compared with glucose or fat ingestion during 1 h of cycling at 75% V˙O_2max and suppressed the rise in plasma fatty free acids and in glycemia. Theoretically, a lower lactate production at given V˙O₂ may translate to a lower metabolic acidosis with potential ergogenic effects on high intensity performance. It seems likely that lower lactate levels were due to the suppression of glycolysis as indicated by skeletal muscle biopsy which revealed a higher content of muscular glycogen and a lower content of intramuscular tryacylglycerol after 2 h of cycling at 70% of V˙O_2max (⁵⁸). These promising results were not replicated by Leckey et al. who examined the effect of AcAc diester administration in a setting resembling the racing conditions of elite cycling and included top-level athletes. The athletes ate a high carbohydrate meal pre-race supplemented with caffeine and diet cola prior to trial, and they followed a typical incremental warm-up. Following on, the athletes performed an all-out test mimicking the 2017 world championship 31-km time trial course requiring maximum carbohydrate oxidation rate and, consequently, placing great metabolic demand on glycolysis. Ketone ester was administered ∼30 min before and immediately before commencing the warm up, and it was associated with a 2% reduction in overall performance and with a 3.7% reduction in power output. From a metabolic point of view, AcAc diester resulted in lower lactate levels at the end of time trial (−4.5 mmol·L⁻¹; −35%) and lower glycemia both after the ingestion of KS and after the end of 31 km time trial compared with control group. This evidence is consistent with the data of Cox et al. and suggests that KS ingestion may decrease glucose contribution to energy expenditure. An alternative reason explaining both the lower glycemia and the reduction in cycling performance may be represented by an impairment in glycosis promoted by KS ingestion. Indeed, the alteration of glycosis may decrease energy availability crucial in short-term endurance performance (⁶⁰). Nevertheless, all participants reported gastrointestinal discomfort associated with the intake of the KE and it can be suggested that these side effects may play a role in explaining performance impairments considering the different KE formula used by Cox et al. which did not result in gut distress. Finally, it should be considered that AcAc diester resulted in serum AcAc concentration of 0.5 mmol·L⁻¹ and in serum β-OHB concentration 0.4 mmol·L⁻¹, much lower than plasma β-OHB concentration of 2 to 4 mmol·L⁻¹ reported by Cox et al. (^59,61).

Achieving an adequate ketonemia represents an essential requisite to evaluate ergogenic properties of KB; KE generally provide a better gut absorption than KS, whose administration may cause gut distress and a not sufficient increase of ketonemia (¹⁵). Acute KS ingestion at both 60 and 15 min before submaximal cycling exercise providing 0.38 g·kg⁻¹ of β-OHB (i.e., ~18.5 g β-OHB) dissolved in 3.8 mL·kg⁻¹ water for each bolus resulted in a significant increase in ketonemia (⁵⁷). The graded cycling test consisted of six stages of 8 min in duration. Ketonemia appears to also be affected by the intensity and the length of exercise; indeed, the highest plasma β-OHB concentration after KS ingestion was observed during the last stage of exercise at 0.44 ± 0.15 mmol·L⁻¹. Ketone salt administration brought about a significant decrease in glycemia compared with the control group at all stages throughout exercise (i.e., 30% V˙O_2max, −0.19 ± 0.36 mM, 40% V˙O_2max, −0.21 ± 0.43 mM, 50% V˙O_2max, −0.27 ± 0.40 mM, 60% V˙O_2max, −0.21 ± 0.39 mM, 70% V˙O_2max, −0.17 ± 0.54 mM, and 80% V˙O_2max, −0.39 ± 1.24 mM) and an elevation in RER for intensity up to 60% V˙O_2max, being ∼0.03 higher in the KS group than control group. The effects of KS on RER at higher intensity were negligible and not significant, being ∼0.01 higher in the KS group compared with control group (⁵⁷). These outcomes suggest the contribution of KB to energy expenditure during exercise and KB sparing effect on glucose oxidation in well-trained cyclists (⁵⁷). Nevertheless, the majority of athletes (68%) reported gastrointestinal distress limiting, if confirmed, the potential effectiveness of KS ingestion (⁵⁷). Similarly, the ingestion of KS containing 0.3 g β-OHB·kg 30 min⁻¹ before the start of the exercise significantly decreased glycemia immediately postexercise and posttime trial (i.e., ~0.3–0.7 mmol·L⁻¹ lower than the placebo condition) and RER at 30% (i.e., ~0.85 in the placebo group vs ~0.83 in the ketone group) and 60% ventilatory threshold. Ketone salt administration alters substrate oxidation lowering total carbohydrates oxidation and augmenting total fat oxidation rates with a net glucose-sparing effect; however, KS administration impaired cycling time trial performance resulting in a significant decrease of power output by ~7%. Time to complete time trial was ~8% longer in KS than the control group (711 ± 137 s vs 665 ± 120 s; P = 0.03). It should be considered that racemic mixture of β-OHB resulted in lower concentrations of blood β-OHB (~1–1.2 mmol·L⁻¹) than KE supplements (~3 mmol·L⁻¹) (⁶⁰). In a study led by Rodger et al. administration of a supplement containing 11.7 g of β-OHB salt diluted with 100 mL of sugar-free lemonade increased average blood β-OHB concentrations from 0.20 mmol·L⁻¹ to 0.63 mmol·L⁻¹ far below the levels required for therapeutic ketosis (≥2 mmol·L⁻¹). The supplement was ingested 20 min before the trial and at the halfway point (45 min) during the 90-min submaximal cycling trial. Subjects were asked to cycle for 90 min at 80% of their second ventilatory threshold, prior to a 4-min maximal cycling test, separated by 2 min passive rest. It is likely that low ketonemia may explain the absence of significant change between placebo and treated group, apart from a not significant increase in RER (0.87 ± 0.05 vs 0.85 ± 0.03) during submaximal exercise and a significant increase in mean RER during the maximal test (1.01 ± 0.07 vs 0.96 ± 0.05) in the treated group compared with the control group. Also, a significant increase in power output was not observed in treated group (364 ±58 Wvs 355 ±46 W) (⁶²).

Finally, it has been observed that ingestion of KE beverage containing 0.573 mL·kg⁻¹ of the KE (R)-3-hydroxybutyl (R)-3-hydroxybutyrate in well-trained athletes after exhaustive exercise consisting in cycling at intermittent intensity for 2-minintervals, alternating 90% peak power efforts with 50% peak power recovery, is able to increase muscle glycogen content by 50% and glucose uptake by 32% during hyperglycemic clamp enhancing glycogen synthesis in skeletal muscle Indeed, total glucose uptake was 1.26 in the control group and 1.66 g·kg⁻¹ in KE group; simultaneously the increase in postexercise glycogen was 114 mmol·kg⁻¹·min⁻¹ in the KE group compared with 70 mmol·kg⁻¹·min⁻¹ in the control group These results seem to be associated with an increased insulin release, with insulin concentration twofold higher by the end of the clamp in the ketone group compared with control group (16 ± 3 mUI·L⁻¹ for control glucose and 31 ± 6 mUI·L⁻¹). Nevertheless, this augmentation of insulin release in response to KE administration requires the elevation of blood glucose levels. Furthermore, an increase in insulin level also may prevent muscular catabolism after exhaustive exercise (^16,63). Nevertheless, this effect on muscular glycogen replenishment were not confirmed by Vandoorne et al. In this study, ingestion of 1.5 g·kg⁻¹ of KE supplement and of a high dose mixture of carbohydrate and protein raised plasma level of β-OHB to ~ 5 mmol·L⁻¹ but it did not affect glycogen resynthesis after exhaustive exercise. Subjects were asked to warm up by unilateral cycling with the right leg on a cycle ergometer. The subjects then started an intermittent exercise protocol aimed to deplete muscle glycogen in the right leg by unilateral knee-extensions (70°–130° knee-angle) at a rate of 0.5 Hz. Subjects first performed an exercise bout, during which they produced a mean power output as high as possible for 5 min. Thereafter, they did 9 series of 30 knee-extensions at 30% of 1RM, followed by 5 series of 6 contractions at 70% of 1RM. The contraction series was interspersed by 30-s passive rest intervals (⁶⁴). Notwithstanding, KE administration markedly augmented mTORC1 signaling increasing phosphorylation of p-S6K1and 4E-BP1. This observation suggests that KE may improve anabolic response to exercise increasing protein synthesis in C2C12 myotubes, being mTORC1 a master regulator of protein synthesis in human skeletal muscle (⁶⁴).

KE supplementation also seems to be able to suppress appetite decreasing ghrelin and representing a potential support for athletes willing to lose weight in those sports where weight plays a pivotal role (⁶⁵).

Conclusions

It can be concluded that long-term KD may be favorable for endurance athletes, during the preparatory season, when a high volume and low to moderate intensity of training loads are performed during training. Indeed, KD may provide a high amount of calories avoiding the ingestion of a great quantity of carbohydrates which are able to cause gastrointestinal distress and hyperinsulinemia with deleterious effects on fat oxidation. The KD is able to maximize the increase of FatMax during exercise and the reduction of body mass and fat content promoted by high volume training. Furthermore, keeping in mind that KD represents a fat-rich diet, a well-formulated KD may provide a high amount of marine omega-3 coming from fatty fish (i.e., mackarel, salmon, tuna). Omega-3 fatty acids may decrease inflammation, insulin resistance, and postexercise muscle damage, and they also may increase the anabolic response to training stimuli and FatMax (^36,53). The increased reliance on fat oxidation promoted by KD seems to increase mitochondrial function, augmenting the adaptive response to oxidative stress generated in oxidative processes during mitochondrial energy production (³¹). However, KD seems not be able to sustain moderate-high exercise activity limiting exercise performance because of the low intake of carbohydrates in KD does not guarantee an adequate content of muscle and hepatic glycogen stores and reduces the capacity to use carbohydrate resulting from inhibition of pyruvate dehydrogenase (PDH) (⁶⁶). Reduced glycolytic flux sustained by decrease of PDH activity may impair performance during high intensity efforts that rely heavily on energy coming from glycolytic pathways (²⁸). Otherwise, elevated ketone concentrations may suppress glycolysis without causing any impairment in glycolytic flux but simply because ketones metabolism may hold hierarchical preference over carbohydrate metabolism. A decreased reliance on glycolysis may lower lactate production and attenuate accumulation of hydrogen ion and other metabolic by-products partly responsible to fatigue during high intensity efforts; moreover, a reduction of glycolytic metabolism may contribute to spare muscle glycogen stores (^28,58).

The potential ergogenic effects of exogenous intake of KE or KS may be related both to the suggested higher efficiency of KB in relatively short endurance activities where glycogen stores are not supposed to be depleted and to KB metabolic role as alternative fuel source in whose events where glycogen stores are likely to become depleted for a glucose-sparing effect. Theoretically, KB also should be able to support higher metabolic intensity compared with fat considering the higher RER. Ketone esters or KS also could be administered to increase postexercise muscle glycogen synthesis when co-ingested with carbohydrate after exhaustive exercise to enhance the recovery process (Fig.). Finally, suppression of appetite after KE ingestion may be helpful to decrease body fat (Fig.). However, it seems likely that a necessary prerequisite of any potential ergogenic effect of KB is reaching blood ketone concentrations >2 mmol·L⁻¹, and it also should be taken into consideration the metabolic differences and distinct pharmacokinetics properties between the ketone molecules delivered by exogenous KS and KE and this can impact their effects on exercise performance (^56,59). It is likely that optimal window of ketonemia ranges from 1 to 3 mmol·L⁻¹ for a proposed performance benefit (⁶¹). Indeed, KS administration generally failed to achieve ketonemia >2 mmol·L⁻¹ and this evidence may explain the lack of ergogenic effects. Therefore, it is necessary to evaluate carefully ketonemia when evaluating ergogenic properties of KB (Fig.). It is crucial to elucidate if KB capacity of sparing carbohydrate reserves is due to a reduction of glycolytic capacity via inhibition of PDH and phosphofructokinase-1 (PFK-1) by increases in NADH:NAD+, acetyl-CoA:CoA ratio or citrate (13). In this case carbohydrate utilization would be impaired by KB, limiting exercise performance at higher intensity in those physiological conditions that rely almost solely on anaerobic glycolysis or high glycolytic flux for ATP production (Fig.). Similarly, the suggested inhibition of lipolysis promoted by KB via inhibition of the nicotinic acid receptor (PUMA-G) should be carefully evaluated as it may impair endurance performance, as fatty acids are the preferred fuel source at low-moderate intensity (²⁶). Other studies are warranted to elucidate the effects of KB on substrate selection which represents the basic tenet to explain the interaction between KB and the human endurance performance (Fig.). So far, only the effects of acute ingestion of KE or KS have been investigated; it remains to be determined if chronic administration of KE or KS may cause different effects on human metabolism and performance. It can be hypothesized an adaptive metabolic process to KE or KS administration that may enhance responsiveness to KS or KE ingestion (Fig.). Finally, it should be noted that top-level athletes may respond differently to KS and/or KS ingestion compared with sedentary subjects considering the impact of genetics and of training on energy substrate metabolism and therefore further research is required to define these differences.

Figure: Possible aims of further research on ketone metabolism in endurance performance.