The Effects of Ketogenic Diets and Ketone Supplements on the Aerobic Performance of Endurance Runners: A Systematic Review

Abstract

Context:

Ketogenic diets and ketone supplements have gained popularity among endurance runners given their purported effects: potentially delaying the onset of fatigue by enabling the increased utilization of the body’s fat reserve or external ketone bodies during prolonged running.

Objective:

This systematic review was conducted to evaluate the effects of ketogenic diets (>60% fat and <10% carbohydrates/<50 g carbohydrates per day) or ketone supplements (ketone esters or ketone salts, medium-chain triglycerides or 1,3-butadiol) on the aerobic performance of endurance runners.

Data Sources:

A systematic search was conducted in PubMed, Web of Science, Pro Quest, and Science Direct for publications up to October 2023.

Study Selection:

Human studies on the effects of ketogenic diets or ketone supplements on the aerobic performance of adult endurance runners were included after independent screening by 2 reviewers.

Study Design:

Systematic review.

Level of Evidence:

Level 3.

Data Extraction:

Primary outcomes were markers of aerobic performance (maximal oxygen uptake [VO₂max], race time, time to exhaustion and rate of perceived exertion).

Results:

VO₂max was assessed by incremental test to exhaustion. Endurance performance was assessed by time trials, 180-minute running trials, or run-to-exhaustion trials; 5 studies on ketogenic diets and 7 studies on ketone supplements involving a total of 132 endurance runners were included. Despite the heterogeneity in study design and protocol, none reported benefits of ketogenic diets or ketone supplements on selected markers of aerobic performance compared with controls. Reduction in bodyweight and fat while preserving lean mass and improved glycemic control were reported in some included studies on ketogenic diets.

Conclusion:

This review did not identify any significant advantages or disadvantages of ketogenic diets or ketone supplements for the aerobic performance of endurance runners. Further trials with larger sample sizes, more gender-balanced participants, longer ketogenic diet interventions, and follow-up on metabolic health are warranted.

Dietary strategies can play an important role in exercise performance, ¹⁸ particularly in endurance exercise, where athletes compete at submaximal exercise intensities for 30 minutes or longer. ⁴⁹ Carbohydrates and fat are the 2 main energy sources for endurance athletes. It has long been believed that such athletes should adopt a carbohydrate-rich diet to improve and maintain their performance, and that the body depends more on carbohydrates as fuel with increasing exercise intensity and duration. Thus, high-carbohydrate diets are advised to maintain optimal concentrations of skeletal muscle glycogen and blood glucose and hence delay the onset of fatigue.^14,25 However, this prevailing concept that carbohydrates are the predominant fuel during high-intensity exercise has been challenged recently, ³⁶ in particular in a study by Prins et al, ⁴⁵ showing that athletes adapted to low-carbohydrate, high-fat diets could exercise at ~85% maximal oxygen uptake (VO₂max) using mainly fat as fuel.

Ketogenic diets that are rich in fat and low in carbohydrates have attracted the attention of endurance runners. Carbohydrate as a fuel suffers from biological constraints, providing up to approximately 1680 kcal of energy in the form of glucose in body fluids (~30 g, including 5 g blood glucose, ⁸ which is consumed rapidly during prolonged exercise), glycogen in skeletal muscle (~300 g), and glycogen in the liver (~90 g). ⁵⁵ Theoretically, adapting to a ketogenic diet could shift the body’s metabolism toward utilizing more fat instead of carbohydrates and thereby enabling the athlete to sustain exercise for longer. ³ Alternatively, consuming exogenous ketone supplements to increase the availability of ketone bodies and to reduce the dependence on carbohydrate fuelling has been suggested as a strategy to gain the benefits of a ketogenic diet without strict dietary adherence. ⁵² Ketone bodies may be more efficient in producing energy than glucose and fatty acids. ⁴³ Commercially available ketone supplements include medium-chain triglycerides (MCT); beta-hydroxybutyrate (βHB) salts; 1,3-butanediol; 1,3-butanediol βHB monoesters; and 1,3-butanediol acetoacetate (AcAc) mono- and di-esters.^43,50,52,53

The use of fat as a fuel for exercise has a long history. Until the mid-1970s, athletes at the Olympic Games were more likely to prefer high protein diets rather than high carbohydrate diets. ³⁸ Phinney and colleagues ³⁹ showed in 1983 that trained cyclists on a high-fat, low-carbohydrate diet with <20 g carbohydrates per day for 4 weeks demonstrated a metabolic shift from burning carbohydrates to burning fat, which caused a nonsignificant increase in mean endurance in a time-to-exhaustion (TTE) trial. Lambert et al ³⁰ showed in 1994 that trained cyclists consuming a high-fat diet (70% fat, 7% carbohydrates) for 2 weeks had a longer TTE than those on a high-carbohydrate diet. The authors repeated the study in 2001 and showed that a high-fat diet also improved cyclists’ performance in a 20-km time trial. ²⁹ They later showed that a 6-day high-fat diet (68% fat) followed by 24-hour carbohydrate loading did not make a difference in overall time to completion, but it impaired performance in some sprints during a 100 km cycling trial. ²⁴ This single negative finding raised scepticism about the effects of ketogenic diets on endurance performance in the sports science community, with an invited editorial in 2006 concluding “we were near to closing the door on its application in sports nutrition.” ¹⁰ In recent years, there have been renewed discussions in the academic and lay literature, as well as in social media, on using high-fat diets to enhance exercise performance. For instance, Zajac and colleagues reported favorable effects of a 4-week ketogenic diet on VO₂max, the gold standard for aerobic fitness, ⁴⁷ in off-road cyclists. ⁶⁵ In 2012, the British cycling team revealed their use of ketone supplements in the London Olympics. ⁴⁸ At the annual meeting of the American Medical Society for Sports Medicine in 2018, Volek presented data on how ketogenic diets can enhance the performance of both endurance and resistance exercise. ¹ An expert analysis in 2020 by the American College of Cardiology suggested that endurance athletes consuming a diet rich in unsaturated fats and plant-based proteins may consider adopting a ketogenic diet to meet performance goals. ⁵⁶ However, concerns have also been raised on how ketogenic diets and ketone supplements may impair exercise performance by reducing exercise economy or inducing side effects,^11,27 and a review published in 2015 concluded that there were no clear performance advantages of ketogenic diets. ⁹ Moreover, athletes playing different types of sports may have different needs to improve performance, and therefore the effects of ketogenic diets and ketone supplements on specific types of sports may be different.

Given the absence of specific reviews for endurance running, a popular sports activity as shown by increasing participation every year, ³² we set out to review the effects of ketogenic diets and ketone supplements on the aerobic performance of endurance runners.

Methods

A systematic search was conducted in 4 databases: PubMed, Web of Science, Pro Quest, and Science Direct for publications up to October 31, 2023. The following search terms were first used in PubMed: (“performance”[Title/Abstract] OR “effect”[Title/Abstract]) AND (“ketogenic diet”[Title/Abstract] OR “low-carbohydrate diet”[Title/Abstract] OR “high-fat diet”[Title/Abstract] OR “fat rich diet”[Title/Abstract] OR “Atkins diet”[Title/Abstract] OR “ketone supplement”[Title/Abstract] OR “ketone supplementation”[Title/Abstract] OR “ketone supplements”[Title/Abstract] OR “medium-chain triglyceride”[Title/Abstract] OR “1,3-butanediol”[Title/Abstract] OR “beta-hydroxybutyrate”[Title/Abstract]) AND (“long-distance runners”[Title/Abstract] OR “runners”[Title/Abstract] OR “running”[Title/Abstract] OR “cross-country runners”[Title/Abstract] OR “marathon runners”[Title/Abstract] OR “endurance runners”[Title/Abstract] OR “triathlon athletes”[Title/Abstract] OR “endurance athletes” [Title/Abstract]). Then, the same search terms were applied to Web of Science, Pro Quest, and Science Direct to retrieve additional studies. The literature search was repeated before manuscript submission to identify additional studies.

The included studies met the following criteria: (1) only adult endurance runners as participants; (2) reported the effects of ketogenic diets (defined as >60% fat and <10% carbohydrates or <50g carbohydrates per day) ¹⁸ compared with high-carbohydrate, low-fat diets [HCLF, defined as >40% of total energy from carbohydrates], or the effect of ketone supplements [ketone esters/salts, MCT or 1,3-butanediol]; (3) reported on measures of aerobic performance, including VO₂max, race time, TTE, and rate of perceived exertion (RPE). The studies were excluded if they were animal studies, reviews or letters, not on endurance runners specifically, did not test the effect of ketogenic diets or ketone supplements, or did not measure aerobic performance.

After identifying all eligible records, a data matrix was developed and data were abstracted on the following variables: study design, participant characteristics (including their habitual diet), types and duration of dietary interventions/types and dosage of ketone supplementation, comparison groups, ketosis testing, performance tests, markers of aerobic performance (VO₂ max and markers of endurance proxied by race time, TTE, and RPE), respiratory exchange ratio (RER) during exercise, changes in body composition, and side effects. A narrative synthesis was performed to assess the effects of ketogenic diets or ketone supplements on the aerobic performance of endurance runners.

Risk of bias assessment was conducted using the Cochrane risk of bias tool for randomized trials (RoB 2), ⁵⁹ nonrandomized studies of exposures (ROBINS-E), ²⁸ or nonrandomized studies of interventions (ROBINS-I), ⁵⁸ depending on the study design.

Two reviewers independently screened the title, and abstract of selected papers and full texts were retrieved when necessary. References of retrieved articles were assessed manually for potential additional studies. The 2 reviewers also independently extracted relevant information from included studies. A third reviewer resolved eventual disagreements on study selection and data extraction. One reviewer performed risk of bias assessment for included studies, the results of which was cross-checked by another reviewer. A third reviewer resolved disagreements on the assessment results.

This review used published data and therefore ethical approval was not needed. This study was not registered and a protocol was not prepared.

Results

The keyword search identified 264 articles. After screening by titles and abstract, 250 were excluded for the following reasons: animal studies (n = 197), reviews or letters (n = 13), did not test the effects of ketogenic diets or ketone supplements (n = 11), did not specifically recruit adult endurance runners (n = 21), or did not measure aerobic performance (n = 8). Among the 14 studies assessed in full text, 6 were further excluded because they did not test the effects of ketogenic diets or ketone supplements,^4,19 or did not specifically recruit adult endurance runners,^16,57,66,67 leaving 8 studies included.^{20,33,42-44,46,50,51} Two additional studies were obtained from Science Direct,^37,63 and another 2 were identified through hand search of the bibliography.^26,45 Searches performed in Web of Science and Pro-Quest did not identify any additional studies. Therefore, a total of 12 studies published between 2001 and 2023 were included (Figure 1). Details of each study are summarized in Table 1.^{20,26,33,37,42-46,50,51,63}

Figure 1.

Flowchart of study selection in this review.

Table 1.

Studies investigating the effects of ketogenic diets or ketone supplements on the performance of endurance runners

Author (Year)	Study Design and Sample Characteristics	Dietary Intervention		Performance Test (Test Condition)	Ketosis Testing	Main Results: VO2max	Main Results: Endurance Performance			Body Composition and RER	Side Effects	Risk of Bias
		Control	Intervention				TTE	Race Time	RPE (Overall)
Ketogenic diets
Heatherly et al (2018) 26	Nonrandomized crossover Recreationally competitive male runners habitually consuming moderate-to-high CHO diet (n = 8, 39.5 ± 9.9 years, baseline VO2max = 49.2 mL/kg/min)	Habitual ad libitum HCLF (first 7-10 days): 43% CHO, 17% protein, 38% fat, 2820 kcal/day	Ad libitum KD (next 3 weeks): 7% CHO, 29% protein, 64% fat, 1886 kcal/day (1000 kcal/day less energy than HCLF)	Endurance performance: 5 times 10 minutes on a treadmill at 5 km, 10 km, 21 km, 42 km and sub-42 km individual race paces in the heat, 5KTT on a road course (fasted)	Blood βHB before exercise: 0.7mmol/L	Postintervention VO2max not measured	Not measured	No significant difference in 5KTT, but KD displayed a trend of improved performance versus HCLF	No significant difference at any race pace (OMNI walk/run perceived exertion scale)	(1) RER: reduced after KD versus HCLF in all race paces (2) Body mass: ~2.5 kg (3%) decrease after KD versus HCLF (3) Body fat (BIA): lower after KD compared with HCLF with a significant decrease in skinfold thickness at multiple sites in the trunk region (4) Lean mass: no significant difference	Blood glucose: near-significant lower in KD pre-50 minute run; no significant difference post-50 minute run	Low (by ROBINS-I) a
Volek et al (2016) 63	Cross-sectional Elite male ultra-endurance runners habitually consuming either KD or HCLF (n = 20, KD: 34.1 ± 7.1 years, baseline VO2max = 64.3 mL/kg/min, HCLF: 32.9 ± 6.0 years, baseline VO2max = 64.7 mL/kg/min)	HCLF (habitual): 59.1% CHO, 14.4% protein, 25.0% fat, 3174 kcal/day	KD (habitual): 10.4% CHO, 19.4% protein, 69.5% fat, 2884 kcal/day	VO2max: graded treadmill test Endurance performance: 180-min running at 64% VO2max on a treadmill (Had CHO/KD drink 90 minutes before exercise)	Blood ketone before exercise blood βHB 0.3 mmol/L, total ketone 0.6mmol/L	No significant difference	Not measured	N/A	No significant difference between KD and HCLF (0-10 scale)	RER: Lower in KD than HCLF	(1) Muscle glycogen depletion and resynthesis: No significant difference after running and during recovery (2) Blood glucose: No significant difference at rest and during exercise (3) Blood lactate in the last hour of exercise: higher in KD versus HCLF	Low (by ROBINS-E) b
Prins et al (2019) 46	Randomized, counterbalanced, crossover Recreationally competitive distance male runners currently consuming a carbohydrate-based diet (>50% CHO) (n = 7, 35.6 ± 8.4 years, baseline VO2max = 61.6 ± 3.1 mL/kg/min in KD,baseline VO2max = 60.6 ± 8.4 mL/kg/min in HCLF)	HCLF (6 weeks): 56.4% CHO, 15.3% protein, 27.8% fat, 2837 kcal/day	Ad libitum KD (6 weeks): 6.0% CHO, 25.1% protein, 68.6% fat, 2947 kcal/day	VO2max: incremental test to exhaustion on treadmill Endurance performance: 5KTT at ~82% VO2max on a treadmill on days 4, 14, 28, 42 (fasted)	Blood βHB before exercise (mean): 0.5mmol/L	No significant difference	No significant difference	Impaired in the first 5KTT on day 4 in KD versus HCLF but no significant difference in the other time trials	No significant difference (OMNI walk/run perceived exertion scale)	(1) RER: peak and mean significantly lower in KD versus HCLF (2) No significant difference in body mass, lean mass, fat mass and body fat (BIA)	(1) Blood glucose: no significant difference at days 4, 14, 28 and 42 pre- and postexercise (2) Blood lactate: no significant difference at days 4, 14, 28 and 42 pre- and postexercise	Low (by RoB 2) c
Shaw et al (2019) 51	Randomized, counterbalanced, crossover Trained male endurance runners habitually consuming a mixed diet (n = 8, 29.6 ± 5.1 years, baseline VO2max = 59.4 ± 5.2 mL/kg/min)	HCLF (31 days): 42.9 ± 7.8% CHO, 18.6 ± 1.4% protein, 39 ± 7% fat, 3280 kcal/day	KD (31 days): 4.1 ± 0.8% CHO, 18.2 ± 3.5% protein, 78 ± 4% fat, 3122 kcal/day	VO2max: graded treadmill test to volitional exhaustion Endurance performance: run-to-exhaustion trial at 70% VO2max on a treadmill on days 0 and 31 (had standardized breakfast and CHO/KD drink every 20 minutes)	Urinary, Blood βHB before exercise 0.75 mmol/L	No significant difference	No significant difference between pre- and post-KD	No significant difference in running speed	No significant difference between KD and HCLF (6-20 Borg scale)	(1) RER: reduced post-KD versus pre-KD trial (2) Body mass: reduced in KD versus HCLF (3) Body fat (skinfold): lower post-KD versus pre-KD	(1) Blood glucose: no significant difference between pre- and post-KD during exercise; elevated from 1 hour exercise to exhaustion compared with baseline after HCLF but not after KD, which only showed an increase from pre-exercise to 1 hour exercise (2) Blood lactate: no significant difference between pre- and post-KD during exercise	Low (by RoB 2) c
Prins et al (2023) 45	Randomized, counterbalanced, crossover Highly trained competitive middle-aged male distance runners (n = 10, 39.9 ± 5.1 years, baseline VO2max = 58.7 ± 5.2 mL/kg/min)	HCLF (31 days): 60-65% CHO, 20% fat, 15-20% protein, 2596 kcal/day	KD (31 days): <50 g/day CHO, 75-80% fat, 15-20% protein, 2545 kcal/day, supplement added salt to taste at mealtime and 1-2 g/day sodium from broth	Endurance performance: self-paced 1-mile time trial on a treadmill (at least 3 hour postprandial)	Blood βHB ≥0.5 mmol/L throughout KD intervention	Not measured	Not measured	Not measured	No significant difference (6-20 Borg scale)	(1) RER: lower post-KD versus post-HCLF (2) BW and composition: no significant difference between KD and HCLF	(1) Blood glucose: no significant difference between HCLF and KD; elevated postexercise compared with baseline independent of diet (2) Blood lactate: no significant difference between HCLF and KD; increased postexercise compared with baseline independent of diet	Low (by RoB 2) c
Ketone supplements
Evans et al (2019) 20	Double-blind, placebo-controlled, randomized, crossover Endurance-trained runners (n = 8, 7M/1F, 33.5 ± 7.3 years, baseline VO2max = 62.0 ± 5.6 mL/kg/min)	CHO + PLA: 8% CHO -electrolyte solution + flavored placebo condition	CHO + KME: 8% CHO - electrolyte solution + 573 mg/kg 1,3-butanediol βHB monoesters	Endurance performance: 60-minute submaximal exercise at 65% VO2 max on a treadmill, a self-paced 10KTT (2 trials) (had standardized breakfast, prescribed drink before and during exercise)	Blood βHB before exercise in CHO + KME 0.99 mmol/L (30 minutes after ingestion)	Postintervention VO2max not measured	Not measured	No significant difference in 10KTT	No significant difference after submaximal exercise (6-20 Borg scale)	RER: No significant difference during submaximal exercise	(1) Blood glucose: lower in CHO + KME 30 minutes after ingestion (right before exercise); no significant difference at 20 minutes, 40minutes, 60 minutes into exercise and end of 10 KTT (2) Blood lactate after 10 KTT: no significant difference (3) GI discomfort: no significant difference (4) Cognitive performance: no significant difference before and after exercise	Low (by RoB 2) c
Misell et al (2001) 33	Double-blind, randomized, crossover Endurance-trained runners (n = 12, 30.5 ± 5.2 years)	56 g corn oil (LCT) supplement in 2 equal portions daily for 2 weeks	60 g MCT oil supplement in 2 equal portions daily for 2 weeks	VO2max: graded treadmill test Endurance performance: 85% VO2max for 30 minutes proceeded by 75% VO2max until exhaustion on a treadmill at the end of each phase (fasted)	Blood βHB in MCT group before exercise: 0.12 mmol/L	No significant difference	No significant difference	Not measured	Not measured	RER: higher in MCT group at 15 minutes but no differences at other timepoints	(1) Blood glucose: no significant difference pre-exercise and 45 minutes into exercise (2) Blood lactate after 45 minutes exercise: no significant difference (3) GI distress when consuming MCT but symptoms lessened over 2-week period	Low (by RoB 2) c
Oopik et al (2001) 37	Double-blind, randomized, crossover Endurance-trained runners (n = 7, 19.4 ± 1.7 years, baseline VO2max = 67.5 ± 4.8 mL/kg/min)	33.3 ± 7.7 g flavored commercial cooking oil in 2 equal portions daily for 7 days	34.1 ± 7.6 g MCT in 2 equal portions daily for 7 days	Endurance performance: run at 80.3% VO2max to exhaustion on a treadmill at the end of each phase (test condition not mentioned)	Blood βHB in MCT group before exercise: 0.4 mmol/L	Postintervention VO2max not measured	No significant difference but tended to be worse after MCT ingestion	Not measured	Not measured	Not measured	(1) Blood glucose: no significant difference pre- and postexercise (2) Blood lactate before or after exercise: no significant difference	Low (by RoB 2) c
Prins et al (2020) 43	Double-blind, placebo-controlled, randomized, crossover Recreational male distance runners consuming a Standard American Diet (44% CHO, 16.4% protein, 38.1% fat) (n = 13, 24.8 ± 9.6 years, baseline VO2max = 60.1 ± 5.4 mL/kg/min)	PLA (0 kcal orange tangerine liquid)	KS1: 22.1 g βHB salt + MCT KS2: 44.2 g βHB salt + MCT	Endurance performance: 5 KTT on a treadmill (3 trials) (had the prescribed drink 60 minutes before exercise)	Blood βHB before exercise: KS1: 0.6 mmol/L, KS2: 0.7 mmol/L	Postintervention VO2max not measured	Not measured	No significant difference between KS1, KS2 and PLA	No significant difference between KS1, KS2 and PLA (OMNI walk/run perceived exertion scale)	RER: No significant difference between KS1, KS2 and PLA	(1) Blood glucose: no significant difference between KS1, KS2 and PLA pre- and postexercise (2) Blood lactate: higher in KS1, nonsignificantly but trending higher in KS2 compared with PLA postexercise (3) Cognitive performance: near-significant faster reaction 30 minutes after KS2 (before exercise) compared with PLA; no significant difference after performance tests	Low (by RoB 2) c
Prins et al (2020) 44	Double-blind, placebo-controlled, randomized, crossover Recreational male endurance runners not consuming KD or ketone supplements (n = 10, 20.8 ± 1.0years)	Flavour matched PLA	300 mg/kg βHB salt + MCT	Endurance performance: 5KTT time trial on a treadmill (2 trials) (Had the prescribed drink 60 minutes before exercise, no other food/drinks 3 hours before exercise)	Blood βHB before exercise in βHB salt + MCT group: 0.6 mmol/L	Not measured	Not measured	βHB + MCT caused a nonsignificant 4.06% decrease on average	No significant difference between βHB + MCT and PLA (OMNI walk/run perceived exertion scale)	RER: No significant difference between βHB + MCT and PLA	(1) Blood glucose: no significant difference between βHB + MCT and PLA pre- and postexercise (3) Blood lactate: no significant difference between βHB + MCT and PLA during exercise	Low (by RoB 2) c
Scott et al (2019) 50	Double-blind, randomized, crossover Male runners (n = 11, 38 ± 12 years,VO2max = 64.2 ± 5.0 mL/kg/min)	CHO: 650 mL energy drink containing 110 g ± 5 g CHO	CHO + BD: 60 g CHO coingested with 500 mg/kg BD	Endurance performance: 60-min submaximal running at 75% VO2max, then a 5KTT on a treadmill (2 trials) (overnight fast; had the prescribed drink 75% 30 minutes before submaximal exercise, 25% right before submaximal exercise, 25% after submaximal exercise)	Blood ketone before exercise: βHB in CHO+BD 0.8 mmol/L	Postintervention VO2max not measured	Not measured	No significant difference in 5KTT performance between CHO and CHO + BD	No significant difference between CHO and CHO + BD during 5KTT (6-20 Borg scale)	RER: No significant difference between CHO and CHO + BD	(1) Blood glucose: higher in CHO + BD versus CHO after TT; no difference pre-exercise, 30 minutes and 60 minutes into exercise (2) Blood lactate: lower in CHO + BD versus CHO after 30 minutes submaximal exercise but similar after TT (3) GI discomfort: no significant difference	Low (by RoB 2) c
Poffe et al (2023) 42	Double-blind, randomized Moderately trained, recreational male ultrarunners (n = 18, 34.1 ± 8.7 years, VO2max = 53.4 ± 3.8 mL/kg/min)	Taste-matched noncaloric PLA containing 1 mmol/L bitter sucrose octaacetate dissolved in water	Pure ketone ester (βHB BD)	Endurance performance: self-paced 100 km time trial (n = 8) or run to exhaustion (60 km: n = 4; 80 km: n = 6) outdoors (had the prescribed drink 30 minutes before the test and every 30 minutes during the test)	Blood βHB in ketone ester: 2.0-2.5 mmol/L during the performance test	Not measured	No significance difference between PLA and ketone ester	No significant difference between PLA and ketone ester	No significant difference between PLA and ketone ester	Not measured	(1) Blood glucose: lower in ketone ester at 60 km, 80 km, and 100 km; no difference pre-exercise, at 20 km and 40 km (2) Blood lactate: no significant difference (3) GI discomfort: no significance difference (4) Cognitive performance: faster reaction and movement in reaction time task, and fewer false alarms in rapid visual information processing task in ketone ester compared with PLA after performance test	Low (by RoB 2) c

BD, 1,3-butanediol; βHB, β-hydroxybutyrate; BIA: body impedance analyser; BW, bodyweight; CHO, carbohydrate; F, female; GI, gastrointestinal; HCLF, high-carbohydrate low-fat diet; KD, ketogenic diet; 5KTT, 5 km running time trial; LCT, long-chain triglyceride; M, male; MCT, medium-chain triglycerides; PLA, placebo; RER, respiratory exchange ratio; RPE, rate of perceived exertion; TTE, time to exhaustion; VO₂max, maximal oxygen uptake.

^aROBINS-I considers bias due to confounding, participant selection, intervention classification, deviation from intended intervention, missing data, outcome measurement, and selection of reported results.

^bROBINS-E considers bias due to confounding, exposure measurement, participant selection, postexposure interventions, missing data, outcome measurement, and selection of reported results.

^cRoB 2 considers bias from the randomization process, deviation from intended interventions, missing outcome data, outcome measurement, and selection of the reported result.

Participant Characteristics

A total of 132 endurance runners (131 male, 1 female) participated in these 12 studies, with the female athlete participating in the study by Evans et al. ²⁰ Eight studies reported their participants were professional (n = 37),^37,45,63 or recreational,^26,42-44,46 runners (n = 54), whereas 4 studies did not specify training status.^20,33,50,51 Six studies specifically recruited subjects habitually consuming a moderate-to-high carbohydrate diet,^{26,43,45,46,51,63}, 1 of which also recruited those on long-term ketogenic diets for comparison. ⁶³ One study noted that subjects were not on ketogenic diets or ketone supplements before the intervention, ⁴⁴ and the remaining 5 studies on ketone supplements recruited participants regardless of usual dietary intake.^{20,33,37,42,50}

Intervention and Study Design

Five studies tested the effects of ketogenic diets,^{26,45,46,51,63} and 7 studies tested the effects of ketone supplements.^{20,33,37,42 -44,50}

Studies on Ketogenic Diets

The 5 studies employed randomized crossover,^45,46,51 nonrandomized crossover, ²⁶ or cross-sectional designs, ⁶³ and they all confirmed that participants achieved ketosis before performance tests by elevated blood and/or urinary ketone levels. The duration of consuming ketogenic diets ranged from 3 weeks to a mean of 20 months. Daily total energy intake was 1000 kcal less during ad libitum ketogenic diet intervention compared with HCLF in Heatherly et al, ²⁶ but was equivalent in the other included studies.^45,46,51,63 Apart from macronutrient manipulation, 2 studies also advised sodium supplementation.^45,46

Studies on Ketone Supplements

All 7 studies used a double-blind, randomized approach and measured blood ketone levels after intervention; of these 7 studies, 6 confirmed that participants achieved ketosis before performance tests,^{20,37,42-44,50} whereas participants from the remaining study did not achieve ketosis. ³³ Ketone supplements were 1,3-butanediol βHB monoesters (573 mg/kg bodyweight [BW], ²⁰ or 25 g before and 12.5 g per 30 minutes during the performance test), ⁴² 22.1 g or 44.2 g or 300 mg/kg BW of βHB salts plus MCT,^43,44 500 mg/kg BW of 1,3-butanediol, ⁵⁰ or 60 g or 34 g MCT.^33,37 In terms of dosage, 3 studies used a single dosage of ketone supplements,^20,44,50 1 randomly assigned participants to different dosages (1 serving or 2 servings) of βHB salts plus MCT, ⁴³ 2 gave daily supplements for 1 or 2 weeks,^33,37 and 1 gave 25 g before exercise and then 12.5 g every 30 minutes during the performance test to maintain ketosis. ⁴² Supplements were either given alone,^{33,37,42,43,51} or with carbohydrates.^20,50 The studies compared ketone supplements with placebo alone,^42-44 carbohydrates alone, ⁵⁰ cooking oils,^33,37 or compared carbohydrates plus ketone supplements with carbohydrates plus placebo. ²⁰

Performance Tests

Measures of aerobic performance are defined in Table 2. The majority of endurance performance tests were time trials,^{20,26,43-46,50} whereas 1 study used 180-minute running trials at 64% VO₂max, ⁶³ 3 used run-to-exhaustion trials at 70% to 85% VO₂max,^33,37,51 and 1 used both time trials and run-to-exhaustion trials as some participants failed to complete the 100 km time trial. ⁴² VO₂max was measured by an incremental test to exhaustion on a treadmill.^33,46,51,63 Race time, TTE, and RPE were assessed mainly by fixed-intensity or self-paced treadmill time trials,^{20,33,37,42-46,50,51,63} except that 1 study conducted a 5-km time trial on a road course with additional trials at different individual race paces. ²⁶ For studies on ketogenic diets, endurance performance tests were done under fasted conditions,^26,46 after a standardized breakfast with supplemental drinks in between,^51,63 or at least 3 hours postprandial. ⁴⁵ Participants of the studies on ketone supplements took performance tests under fasted conditions, ³³ shortly after having the prescribed drinks,^43,50 or had the prescribed drinks before and during exercise.^20,42,44 One study did not report the test condition. ³⁷

Table 2.

Measures of aerobic performance

Item	Description
VO2max	The highest rate at which oxygen can be taken up and utilized by the body during strenuous exercise; gold standard to assess aerobic fitness 47 ; sets the upper limit for endurance performance. 5 American College of Sports Medicine’s classification of exercise intensity based on percentage of VO2max 21 : <37% (very light), 37-45% (light), 46-63% (moderate), 64-90% (vigorous), ≥91% (near maximal to maximal).
Race time	The performance of timed running; may be affected by environmental humidity and temperature.26,47
TTE	Longest time a runner can sustain on a fixed workload. Measure of the endurance capacity. 34
RPE	A scale (eg, Borg 6-20 scale, OMNI walk/run perceived exertion 0-10 scale) to measure exercise intensity based on subjective physical sensations of effort during exercise: higher RPE indicates higher perceived intensity. 23

RPE, rate of perceived exertion; TTE, time to exhaustion; VO₂max, maximal oxygen uptake.

Effects of Ketogenic Diets/Ketone Supplements on Aerobic Performance Measures

VO₂max

Four studies reported the effects of ketogenic diets (compared with HCLF)^46,51,63 or ketone supplements (compared with corn oil), ³³ on VO₂max, and all showed no significant difference.

Markers of Endurance

Race time. Eight studies reported the effects on the race time of time trials,^{20,26,42-44,46,50} or the running speed of a run-to-exhaustion trial, ⁵¹ and all showed no benefits after ketogenic diets or ketone supplements.^{20,26,42-44,46,50,51} One study found impaired performance in the time trial within the first 4 days after starting ketogenic diets, whereas there was no significant difference in the subsequent time trials on days 14, 28, and 42. ⁴⁶

Time to exhaustion. Five studies assessed TTE, of which 3 reported no significant effects of ketogenic diets compared with HCLF,^46,51 ketone supplements compared with corn oil, ³³ or placebo. ⁴² The remaining study found a trending lower TTE after MCT ingestion compared with cooking oil. ³⁷

Rate of perceived exertion. Nine studies measured overall RPE, and none observed improvement or decline after consuming ketogenic diets or ketone supplements compared with controls.^{20,26,42-46,50,51,63}

Effects of Ketogenic Diets/Ketone Supplements on Body Composition and RER During Exercise

Some studies also assessed body composition and/or resting metabolism. Four studies on ketogenic diets measured body fat and lean mass using bioelectrical impedance or skinfold thickness.^26,45,46,51 Reduction in body fat was observed in 2 of these studies,^26,51 but not in the other 2.^45,46 Only 3 studies measured lean mass, and none showed any change with ketogenic diets.^26,45,46

Ten studies assessed RER during 1-mile or 5-km time trials,^{43 -46,50} 60-minute running at 65% VO₂max, ²⁰ 180-minute running at 64% VO₂max, ⁶³ run-to-exhaustion trials,^33,51 or at different race paces. ²⁶ Five studies demonstrated reduced RER with ketogenic diets compared with HCLF, aligning with increased peak and mean fat oxidation rates.^{26,45,46,51,63} However, most studies on ketone supplements found no significant change of RER after supplementation,^20,43,44,50 except that 1 study found higher RER after 15 minutes exercise but no differences at other timepoints. ³³

Side Effects

Some studies measured blood glucose and lactate levels and muscle glycogen dynamics. All studies on ketogenic diets found no differences in blood glucose before and after exercise compared with HCLF or pre-ketogenic diets,^{26,45,46,51,63} and 3 studies reported no difference in blood lactate after exercise,^45,46,51 while 1 found higher blood lactate in the last hour of exercise under a ketogenic diet. ⁶³ One study also measured muscle glycogen depletion and resynthesis and observed no difference between ketogenic diets and the control. ⁶³ All studies on ketone supplements measured blood glucose during exercise, 5 of which found no significant difference compared with placebo or cooking oils,^{20,33,37,43,44} 1 found higher level postexercise (60 minutes running followed by a 5 km time trial) due to a more profound elevation from 30 minutes into exercise compared with having carbohydrates alone, ⁵⁰ and 1 found lower level starting from 60 km into exercise until 100 km compared with placebo. ⁴² All 7 studies on ketone supplements also measured blood lactate after exercise, with 5 reporting no significant difference compared with placebo or cooking oils,^{20,33,37,43,44} 1 reporting higher levels compared with placebo, ⁴³ and 1 reporting lower levels compared with carbohydrates alone. ⁵⁰ Three studies reported no significant difference in gastrointestinal (GI) discomfort experienced after consuming ketone supplements,^20,42,50 whereas 1 study reported GI distress in all participants after consuming MCT with lessened symptoms over 2 weeks. ³³ Three studies assessed the effects of ketone supplements on cognitive performance before and after exercise by testing reaction time, as well as performance on a multitasking test—no significant difference at any timepoint was detected in 2 studies,^20,43 whereas ketone ester supplementation fully negated the psychocognitive decline induced by exercise in the third study compared with minimal effects under placebo. ⁴²

Risk of Bias Assessment

All studies included in this systematic review were considered to have low risk of bias as assessed by RoB 2,^{20,33,37,42-46,50,51} ROBINS-I, ²⁶ or ROBINS-E, ²⁸ depending on the study design (Table 1).

Discussion

This review of a total of 12 studies—5 on ketogenic diets and 7 on ketone supplements, involving mainly male endurance runners—found no advantages or disadvantages of ketogenic diets or ketone supplements on aerobic performance in terms of VO₂max, race time, TTE, and RPE, assessed mainly in time trials.

Our findings are consistent with other reviews of ketogenic diets on different types of athletes. Bailey and Hennessy ³ noted that ketogenic diets coupled with training programs seemed to increase VO₂max in race walkers compared with baseline similar to high carbohydrate controls.^11,12 A recent review investigating the efficacy of popular diets among athletes concluded that ketogenic diets may be effective in reducing body fat but they resulted in similar aerobic fitness and endurance capacity compared with high carbohydrate diets. ¹⁸ A meta-analysis also found ketogenic diets did not reduce the aerobic capacity and exercise performance of endurance athletes of mixed types of sports regarding their VO₂max, TTE, RPE, and maximal heart rate during graded exercise tests. ¹³ Shaw et al ⁵² reviewed ketogenic diets for endurance cycling, racewalking, and running and concluded that they did not seem to provide any advantage or disadvantage for high-intensity as well as prolonged moderate-intensity endurance exercise. These reviews on different endurance sports and our present review on endurance runners suggested that the macronutrient content of the pre-exercise diet may not make a significant difference to exercise performance in general.

However, it is important to note that many previous studies, including most of those in this review,^26,45,46,51 studied ketogenic diets for relatively short durations (3-6 weeks). Even though ketosis and utilization of fat can be achieved within days, it has been suggested that it can take months to achieve optimal “keto-adaptation” to enable fat and ketone bodies to provide an adequate and steady source of energy, although it is unclear how to determine whether sufficient keto-adaptation has occurred.^54,63 Supplementation of minerals is also the key to successful adaptation to ketogenic diets, ³⁹ but only 2 included studies advised sodium supplementation.^45,46 A lack of good adaptation to nutritional ketosis could possibly result in different degrees of side effects of ketogenic diets, such as hypoglycemia, ²² low muscle glycogen, ⁷ impaired muscle regeneration, ⁶ and low energy availability, that could adversely affect aerobic performance. As such, the previous findings of a reduction in exercise economy, i.e., more oxygen requirement at a given speed due to increased fat oxidation and thus more utilization of aerobic capacity, after 3 weeks of ketogenic diets in elite race walkers,^11,64 may need further controlled trials of longer ketogenic diet duration to verify.

Athletes are bound to develop hypoglycemia during prolonged exercise without supplementing carbohydrates throughout. Blood glucose concentration, rather than muscle glycogen, is the key to determining exercise performance and time to fatigue after 2 to 3 hours of exercise, and it is regulated by the central nervous system so as to prevent hypoglycemic brain damage, especially in those who are unable to utilize ketones adequately for brain function. ³⁵ Lambert et al ³⁰ showed that athletes randomized to a 2-week ketogenic diet were able to continue cycling at lower blood glucose concentrations before exhaustion than their HCLF counterparts. It remains unanswered by the trials so far whether ketogenic diets have benefits for prolonged exercise such as long-distance running among athletes who are better at using ketones to fuel their brain. Well-controlled studies on prolonged exercise are needed to test this hypothesis, eg, exercising to exhaustion at a high intensity after an overnight fast, followed by a short rest and then another exercise to exhaustion at a moderate intensity, with no carbohydrate ingestion during the whole period, as in the study by Lambert et al. ³⁰

Other than sports performance, ketogenic diets may also exert some health benefits on athletes compared with HCLF. Continuous glucose monitoring in the study by Prins et al ⁴⁵ demonstrated that, throughout the 31-day HCLF intervention, 30% of the athletes had mean and median fasting glucose >100 mg/dl, which indicated high risk of prediabetes. These subjects also had the greatest reduction of blood glucose when put under isocaloric ketogenic diets with equivalent exercise performance, ⁴⁵ suggesting that, compared with HCLF, ketogenic diets may help improve glycemic control while maintaining endurance capacity. These potential benefits, if confirmed, will have to be considered by the American College of Sports Medicine (ACSM) and the International Society of Sports Nutrition, which currently state, respectively, that a diet providing high carbohydrate availability remains the core advice for distance athletes, ⁶⁰ and ketogenic diets or ketone supplements should be used with caution due to insufficient evidence of benefit. ⁶¹

The 7 included studies on ketone supplements did not show any difference in VO₂max, race time, TTE, and RPE regardless of whether supplements were ingested chronically,^33,37 or acutely,^{20,42 -44,50} before,^43,44 or both before and during exercise,^20,42,50 and with carbohydrates,^20,50 or not.^42-44 The heterogenous ingestion protocols in the included studies, such as the types of ketone supplements, time and amount of ingestion, as well as involvement of carbohydrate supplements, limit analyses on how ketone supplements may affect aerobic performance. However, the dosage of ketone supplements in the included studies were sufficient to induce ketosis in most studies as indicated by pre-exercise blood ketone levels reaching concentrations expected when consuming ketogenic diets.^{26,42,46,51,63} Only 1 study did not achieve pre-exercise ketosis, but this may be because βHB had been cleared from circulation during overnight fasting. ³³ Studies on ketone supplements for endurance cyclists are scarce, with mixed findings including slight improvement in performance, ¹⁵ impaired performance, ³¹ and no difference in performance. ⁵³ Reviews of ketone supplements by Valenzuela et al ⁶² and Shaw et al ⁵² concluded no benefit of acute ketone supplements on performance in different cycling, running, and intermittent exercise protocols.

There were several unanswered questions as to how ketone supplements may affect aerobic performance. It was suggested that acute exogenous ketone supplementation may harm aerobic performance by suppressing lipolysis and increasing carbohydrate oxidation, therefore accelerating fatigue during prolonged events. ⁵² Although our review found unchanged or temporarily higher RER after consuming ketone supplements, RER may not accurately reflect carbohydrate and fat utilization without adjusting for βHB and AcAc because their respiratory quotients are 0.89 and 1.00, respectively. ⁴⁴ Blood lactate is a proxy for glycolytic activity, ⁵⁰ but the included studies had mixed results, with 5 reporting no difference,^{20,33,37,42,44} 1 showing lower, ⁵⁰ and 1 showing higher blood lactate after ketone supplements. ⁴³ Therefore, it remains uncertain whether ketone supplementation leads to altered fat and carbohydrate oxidation. ⁴⁴ Alternatively, it was suspected that a transient decrease in blood pH caused by acute ketone supplementation may cancel out the potential advantages of ketone supplements.^40,41 However, change in plasma pH was not measured in the studies giving acute supplementation.^20,42-44,50 Aerobic performance could also be affected negatively by GI discomfort after consuming ketone supplements. One study reported GI distress in all participants after MCT ingestion, with symptoms wearing off over 2 weeks, ³³ while 3 studies reported no significant discomfort compared with carbohydrate supplements with or without placebo.^20,42,50

Limitations

There are some limitations with this review. First, only 12 studies were included, and all had small sample sizes, with a total of only 132 participants with varied competing experience and training levels. Second, female athletes, who tend to rely less on carbohydrate fueling compared with male athletes during endurance exercise due to the effect of oestrogen, ¹⁷ were underrepresented. Third, the included studies were highly heterogenous in terms of population (elite and recreational runners), intervention (different study design and composition of ketogenic diets and ketone supplements), comparator (different control groups used in the included studies on ketone supplements), and outcome (different endurance performance tests with incomplete measures in some studies). Fourth, external factors, such as temperature and humidity, ²⁶ and internal factors, such as genetic differences in ketone and lipid metabolism, ² which may also affect aerobic performance with ketogenic diets or ketone supplements, were not considered. Finally, as discussed, the duration of ketogenic diets was relatively short (3-6 weeks) in most of the included trials,^26,45,46,51 so findings on exercise performance in the experimental groups could be attributed to a lack of keto-adaptation.

Conclusion

This review of 12 studies showed no advantages or disadvantages of ketogenic diets and ketone supplements on the aerobic performance of endurance runners. Further trials with larger sample sizes, including both male and female participants, with longer duration of ketogenic diet interventions and with longer-term follow-up on their metabolic health will improve our understanding on the overall impact of ketogenic diets on aerobic performance.