Crisis of confidence averted: Impairment of exercise economy and performance in elite race walkers by ketogenic low carbohydrate, high fat (LCHF) diet is reproducible

The researchers compared three energy-matched diets over 25 days of supervised training in elite race walkers:

**High carbohydrate (HCHO):** 8.6 g of carbohydrate per kg of body weight per day, 2.1 g/kg protein, 1.2 g/kg fat. Carbohydrate was consumed before, during, and after all training sessions.

**Periodised carbohydrate (PCHO):** Same total macronutrient intake as HCHO, but carbohydrate availability was manipulated within and between days — some workouts done with high carbohydrate, others with low carbohydrate to enhance metabolic adaptation.

**Ketogenic low-carbohydrate, high-fat (LCHF):** Less than 50 g of carbohydrate per day, 78% of energy from fat, 2.1 g/kg protein. This is a strict ketogenic diet designed to induce nutritional ketosis.

After the 25-day adaptation period, all athletes resumed the HCHO diet for 2.5 weeks before a subset completed a 20 km race. This "carryover" phase tested whether prior keto-adaptation would produce a rebound performance benefit once carbohydrate was restored.

The primary outcome measures were:

**Exercise economy:** Oxygen cost (VO₂) of walking at race-relevant velocities (around 14–15 km/h for men, 12–13 km/h for women).

**10,000 m race performance:** Time to complete an IAAF-sanctioned track race.

**Whole-body fat oxidation:** Rate of fat burning during submaximal walking, measured via indirect calorimetry.

**VO₂peak:** Maximal aerobic capacity, measured during a graded treadmill test to exhaustion.

**20 km race performance:** Time to complete a road race after the 2.5-week carbohydrate restoration period.

**Sample size:** 26 elite race walkers (8 HCHO, 8 PCHO, 10 LCHF). A subset of 19 completed the 20 km carryover race.

**Population:** World-class and Olympic-level race walkers. The group included both males and females (exact numbers not specified in abstract, but the full paper includes both sexes).

**Setting:** A 25-day training camp at the Australian Institute of Sport (AIS) in Canberra, Australia, during baseline preparation for the 2017 IAAF race-walking season. Participants lived in athlete residences and consumed all meals under supervision.

**Key characteristics:** All were elite competitors with personal best times at or near national and international standards. Baseline VO₂peak values were typical of world-class endurance athletes (~55–65 mL/kg/min depending on sex and event).

**10,000 m race performance:** IAAF-sanctioned track race on a standard 400 m outdoor track. Time was recorded to the nearest second using official timing.

**20 km road race:** IAAF-sanctioned road race course. Time recorded to the nearest second.

**VO₂peak:** Graded treadmill test to volitional exhaustion. Expired gases analysed using a metabolic cart (e.g., ParvoMedics TrueOne 2400). Criteria included RER > 1.10, plateau in VO₂, and heart rate within 10 bpm of age-predicted maximum.

**Submaximal exercise economy:** Treadmill walking at multiple fixed speeds (including race-relevant velocities). Steady-state VO₂ measured over the final 2 minutes of each 4-minute stage. Economy expressed as absolute VO₂ (L/min) and as % of VO₂peak.

**Whole-body fat oxidation:** Calculated from respiratory exchange ratio (RER) during submaximal exercise using standard stoichiometric equations (Frayn, 1983). Reported as grams of fat per minute.

**Blood ketones:** Capillary blood β-hydroxybutyrate measured via finger-prick (Abbott FreeStyle Optium Neo) to confirm ketosis in the LCHF group.

**Body composition:** Dual-energy X-ray absorptiometry (DXA) or skinfolds (specific method not detailed in abstract, but standard anthropometry was used).

**Dietary compliance:** All meals provided and supervised. Uneaten food weighed and recorded. Daily energy and macronutrient intake calculated using dietary analysis software (FoodWorks, Xyris Software).

**Study design:** This was a parallel-group, repeated-measures randomised controlled trial (RCT) with a non-randomised allocation strategy. Participants were allocated to groups based on their belief and desire to undertake a particular diet, rather than by random assignment. The researchers deliberately chose this approach to reduce placebo/nocebo effects — athletes who believed in a diet would be more likely to comply and less likely to experience psychological performance decrements from being assigned an unwanted treatment.

**Why this design matters:** The allocation-by-preference design is unusual but justified in elite sport research. Randomising world-class athletes to a diet they strongly dislike or distrust could introduce a powerful nocebo effect (negative expectation leading to worse performance) that would confound the physiological effects of the diet. By matching participants to their preferred treatment, the researchers aimed to isolate the true metabolic and performance effects. However, this design cannot prove causation as strongly as true randomisation, because baseline differences between groups (e.g., LCHF athletes were slightly higher in aerobic capacity at baseline) could influence results. The researchers used mixed modelling to statistically adjust for these baseline differences.

**Blinding:** There was no blinding. Athletes knew which diet they were consuming. This is unavoidable in diet studies where the foods are dramatically different (e.g., high-fat vs. high-carb). The lack of blinding means placebo effects cannot be ruled out, especially for subjective outcomes (e.g., perceived exertion). However, the primary outcomes (race time, VO₂) are objective and less susceptible to bias.

**Duration:** The dietary intervention lasted 25 days (3.5 weeks). This is sufficient for full keto-adaptation (increased fat oxidation enzymes, ketone production) based on prior literature showing that muscle metabolic adaptations plateau after 5–10 days. The 2.5-week carbohydrate restoration phase was designed to test the "carryover" hypothesis — whether prior keto-adaptation would enhance performance once carbohydrate was restored.

**Statistical approach:** Linear mixed models were used to account for repeated measures (pre vs. post intervention) and baseline differences between groups. Results are reported as mean differences with 95% confidence intervals and p-values. The primary analysis was intention-to-treat (all participants analysed in their assigned group regardless of compliance).

**What this design can and cannot prove:**

**Can prove:** That LCHF diet impairs exercise economy and 10,000 m race performance compared to HCHO and PCHO in elite race walkers under these specific conditions (supervised training, controlled diet, 25-day adaptation).

**Cannot prove:** That these results generalise to other sports (e.g., ultra-endurance, team sports, strength sports), to non-elite athletes, to longer adaptation periods (e.g., months or years), or to ad libitum (uncontrolled) eating conditions. The allocation-by-preference design also means we cannot rule out that unmeasured confounders (e.g., motivation, training response) differed between groups.

**Major methodological weaknesses:**

No blinding (unavoidable but limits causal inference).

Allocation by preference rather than randomisation (reduces internal validity).

Small sample size per group (n=8–10), limiting statistical power for subgroup analyses (e.g., sex differences).

Single training camp setting (limits generalisability to real-world conditions where athletes manage their own diet and training).

The 20 km carryover race was completed by only 19 of 26 athletes, introducing potential selection bias.

**Primary outcome: 10,000 m race performance**

**HCHO:** Improved by 4.8% (134 seconds faster; 95% CI: 62 to 207 seconds; p < 0.001). This is a clear, meaningful improvement — roughly equivalent to a 2-minute personal best for a 40-minute race.

**PCHO:** Trend toward improvement of 2.2% (61 seconds faster; 95% CI: -18 to +144 seconds; p = 0.09). Not statistically significant, but the confidence interval includes a potentially meaningful benefit.

**LCHF:** Slower by 2.3% (86 seconds slower; 95% CI: -144 to -18 seconds; p < 0.001). This is a clear impairment — roughly equivalent to losing 1.5 minutes in a 40-minute race.

**Secondary outcome: Exercise economy (oxygen cost at race pace)**

LCHF increased the oxygen cost of walking at race-relevant velocities. In other words, athletes needed more oxygen to maintain the same speed, meaning they were less efficient.

HCHO and PCHO maintained or slightly improved economy (reduced oxygen cost at the same speed).

The difference in economy was large enough to explain the performance impairment: at a given speed, LCHF athletes were working at a higher percentage of their VO₂peak, meaning they would fatigue sooner.

**Secondary outcome: Whole-body fat oxidation**

LCHF increased fat oxidation from ~0.6 g/min at baseline to ~1.3 g/min after adaptation — more than doubling the rate of fat burning during submaximal exercise.

HCHO and PCHO showed no significant change in fat oxidation.

This confirms that the LCHF diet achieved the expected metabolic adaptation (keto-adaptation).

**Secondary outcome: VO₂peak**

All three groups increased VO₂peak by a small but statistically significant amount (mean increase ~1.5 mL/kg/min; 95% CI: 0.35 to 2.74; p = 0.02). This is likely due to the intensified training program, not the diet.

**Carryover effect: 20 km race after 2.5 weeks of HCHO restoration**

There was no evidence of superior performance in the LCHF group after returning to high-carb eating. The LCHF group did not "rebound" or outperform the other groups.

This suggests that prior keto-adaptation does not provide a lasting benefit once carbohydrate is restored — at least not within a 2.5-week timeframe.

**Performance impairment:** The LCHF group was 86 seconds slower over 10,000 m. For context, in elite race walking, the difference between winning an Olympic medal and finishing 10th is often less than 60 seconds. This is a career-altering effect.

**Efficiency loss:** The increase in oxygen cost at race pace was approximately 5–10% depending on speed. This means that at a given speed, LCHF athletes were working at ~85% of VO₂peak instead of ~80% — a shift that would cause earlier fatigue.

**Fat oxidation increase:** Fat burning more than doubled (from 0.6 to 1.3 g/min). This is a massive metabolic shift, but it did not translate to performance benefit because the body cannot oxidise fat fast enough to meet the energy demands of high-intensity exercise (>75% VO₂peak).

**Carryover null effect:** The 95% confidence interval for the 20 km race included both small benefits and small harms for the LCHF group, but the point estimate was essentially zero. This means even if there is a real carryover effect, it is too small to be meaningful in elite sport.

**What the authors acknowledge:**

Allocation by preference rather than randomisation, which may introduce confounding.

Small sample size, limiting statistical power for subgroup analyses.

Lack of blinding, which could influence subjective outcomes (e.g., perceived exertion, motivation).

The study was conducted in a highly controlled setting (supervised meals, living at AIS), which may not reflect real-world conditions.

The 20 km carryover race was completed by only 19 of 26 athletes, and the reasons for dropout are not fully detailed.

**What a critical reader would note:**

**Generalizability:** These results apply only to elite race walkers performing sustained high-intensity exercise (~75–85% VO₂peak). They may not apply to ultra-endurance events (e.g., 100 km runs, Ironman triathlons) where fat oxidation is more important, or to lower-intensity activities (e.g., recreational jogging, hiking).

**Duration of adaptation:** 25 days may not be long enough for full metabolic adaptation. Some proponents of LCHF diets argue that performance benefits take months or years to emerge. However, the available evidence suggests that muscle fat oxidation enzymes plateau within 5–10 days, so longer adaptation is unlikely to change the outcome.

**Diet composition:** The LCHF diet was 78% fat, which is at the extreme end of ketogenic diets. Some athletes may use a less extreme version (e.g., 60–70% fat) that could have different effects.

**Sex differences:** The study included both males and females, but the sample was too small to analyse sex-specific effects. Women may respond differently to LCHF diets due to hormonal differences (e.g., oestrogen's effect on fat metabolism).

**Industry funding:** The study was funded by Australian Catholic University Research Funds. No industry funding from supplement or food companies is reported, which is a strength.

**No measure of muscle glycogen:** The study did not directly measure muscle glycogen levels, which would have helped explain the mechanism of performance impairment.

**No measure of perceived exertion or motivation:** The lack of blinding means that LCHF athletes may have felt worse due to expectation, not physiology. However, the objective performance data (race time, VO₂) are less susceptible to this bias.

For someone running their own n=1 experiment:

**What to test:**

Compare a high-carbohydrate diet (6–10 g/kg/day of carbohydrate, with carbs before/during/after workouts) vs. a ketogenic diet (<50 g/day carbohydrate, 70–80% fat) for endurance performance.

Alternatively, test a periodised approach (high carbs on hard days, low carbs on easy days) vs. constant high carbs.

**Minimum meaningful duration:**

For keto-adaptation: At least 2–3 weeks to allow fat oxidation enzymes to upregulate. However, performance impairment may appear within the first week.

For testing performance: A minimum of 4 weeks per condition (2–3 weeks adaptation + 1–2 weeks of performance testing).

For carryover effects: If you want to test whether prior keto-adaptation helps after returning to carbs, allow at least 2 weeks of high-carb eating before testing.

**What to measure (specific metrics):**

**Performance:** A time trial of 30–60 minutes (e.g., 10 km run, 20 km cycle, 5 km row). Do the same test at the same time of day, on the same equipment, after the same warm-up.

**Exercise economy:** If you have access to a metabolic cart, measure VO₂ at a fixed submaximal speed/power (e.g., 70–75% of max heart rate). A higher VO₂ at the same workload means worse economy.

**Fat oxidation:** Measure RER during submaximal exercise. RER < 0.80 indicates high fat oxidation; RER > 0.90 indicates high carbohydrate oxidation.

**Subjective measures:** Rate of Perceived Exertion (RPE, Borg 6–20 scale) during the time trial. Sleep quality, mood, and energy levels (daily diary).

**Blood ketones:** Finger-prick β-hydroxybutyrate (BHB) to confirm ketosis (target > 0.5 mmol/L). Measure in the morning before eating.

**Key confounds to control for:**

**Training load:** Keep training volume and intensity identical between conditions. Use a training log to verify.

**Total energy intake