Effect of a Ketogenic Diet on Submaximal Exercise Capacity and Efficiency in Runners

David M. Shaw; Fabrice Merien; Andrea Braakhuis; Ed Maunder; Deborah K. Dulson


Med Sci Sports Exerc. 2019;51(10):2135-2146. 

In This Article


The present study investigated the effect of a 31-d KD on submaximal exercise capacity in endurance-trained runners. We provide novel findings regarding exercise efficiency related to oxygen uptake and EE. We found that 1) a 31-d KD preserved mean submaximal exercise capacity without the requirement of CHO restoration or supplementation, 2) exercise efficiency was impaired at intensities above 70% V̇O2max as evidenced by oxygen uptake that could not be explained by shifts in RER alone and increases energy expenditure shifts in RER alone, 3) RER at V̇O2max may have implications as a performance surrogate after keto-adaptation, and 4) keto-adaptation increased submaximal endurance variability as evidenced by a twofold increase in the 90% CI for TTE and DTE in the post- compared with pre-KD trial, which coincided with reduced endurance capacity in five of eight runners, indicating a meaningful risk of an endurance decrement at an individual level.

The aim of the dietary interventions was to polarize metabolic states, as it has been suggested that increased fat oxidation rates and blood D-βHB concentrations after keto-adaptation can overcome the necessity of acute CHO fuelling strategies during continuous exercise lasting for several hours.[7] Because we could not replicate previous studies by providing meals, snacks, and fluid to participants during the adaptation period,[9,11] we implemented a novel approach to measure and verify prescribed dietary intakes. This included an image-assisted (alongside a fiducial marker) weighed dietary record for 40% of the dietary adaptation days to reduce the potential of underreporting and misreporting,[28] coding errors, and daily dietary variation.[29] When combined with urinary AcAc (daily) and blood D-βHB (days 3, 7, 14, 21, and 28 and post-KD tests) measures, we can confirm that all participants were extremely compliant to the KD. Although the HD was lower in CHO compared with recommendations for moderate training loads (4.6 g·kg·d−1 vs 5–7 g·kg·d−1),[5] this difference is negligible and did not seem to affect training adaptation or exercise capacity. Moreover, studies using intensified training protocols during dietary adaptation periods, such as the aforementioned investigation of keto-adaptation in elite race walkers,[11] may limit direct comparison to the present study because of interferences from a training response.

Acute fuelling strategies before and during the run-to-exhaustion trials would have, in part, compensated for suboptimal CHO intakes in the days prior. This required participants to ingest either a high- or low-CHO meal 2 h before the run-to-exhaustion trials and either CHO or fat while running according to the trial allocation. Arguably, the rate of CHO intake in our study (~55 g·h−1) was below recommendations for optimal performance of these durations (~90 g·h−1).[14] Nevertheless, 55 g·h−1 is comparable to recent publications of real-world dietary behaviors in ultraendurance athletes,[30] with even small rates of CHO supplementation (i.e., 10 g·h−1) providing a performance benefit.[14] In turn, fat oxidation was ~4-fold lower and CHO oxidation ~2.5-fold higher in the high-CHO trials compared with keto-adapted trial. This was likely underpinned by a combination of 1) elevated muscle[31] and hepatic glycogen content,[32] 2) elevated blood glucose uptake into the muscle,[33] 3) maintenance of blood glucose concentration,[3] and 4) reduction in hepatic glycogen utilization.[2] However, if CHO was acutely ingested in the post-KD trial, this would oppose adaptations to the KD and suppress hepatic ketogenesis, thus compromising rates of fat oxidation and ketolysis.[8]

To investigate exercise efficiency after keto-adaptation, we estimated oxygen uptake that could not be explained by shifts in RER from pre- to post-KD conditions, assuming no change to EE. At lower intensities (particularly <60% V̇O2max), the shift in RER could fully explain the increase in oxygen uptake. This coincides with a previous study demonstrating similar oxygen uptake at 62%–64% V̇O2max,[9] which suggested that keto-adaptation improved exercise efficiency at lower intensities. Nevertheless, we found that RER could only account for 14% of the increase in oxygen uptake at 13.5 km·h−1 during the metabolic test (~77% V̇O2max when keto-adapted) and 55% of the increase in oxygen uptake during the first 2 h of the run-to-exhaustion trial (~72% V̇O2max when keto-adapted). These effects likely underpinned the increase in EE observed between the pre- and post-KD trials and have previously been demonstrated in sedentary individuals,[16] but were thought to be abrogated by endurance training.[17] However, we demonstrate that this effect persists with keto-adaptation regardless of training status. The unaccounted oxygen uptake may be due to elevated fatty acid–activated transcription factor peroxisome proliferator–activated receptor α. In addition to peroxisome proliferator–activated receptor α's upregulation of fat oxidative genes, it also regulates the expression of mitochondrial uncoupling proteins.[34] Nevertheless, the effect of a ketogenic or LCHF diet on mitochondrial uncoupling in human skeletal muscle is unclear;[16,17,35] therefore, this remains speculative.

Our findings also indicate that keto-adaptation impairs high-intensity, endurance performance. Similar to previous studies,[9,11] CHO oxidation was truncated at near maximal exercise intensities after keto-adaptation, which manifested as a throttling of RER at V̇O2max. This may be due to an attenuation of glycogenolysis and pyruvate dehydrogenase activity[36] and could have underpinned the 1-km·h−1 reduction in velocity at V̇O2max. Furthermore, RPE was nonsignificantly higher at >65% V̇O2max during the metabolic test after keto-adaptation, which is in line with an earlier study investigating a 3-wk KD intervention in elite race walkers.[11] However, during the submaximal run-to-exhaustion, RPE did not differ, despite HR being 7–9 bpm higher. This could be due to increased sympathetic nervous system activity,[37] with potential neural effects after keto-adaptation compensating for the increase in metabolic stress. However, a decrement in high-intensity exercise performance does not necessarily negate ultraendurance performance. Therefore, with anecdotal reports of (ultra-) endurance athletes successfully employing LCHF diets and improved performance[18,19] or exercise capacity[9] in select individuals, including in the present study, further research remains warranted to understand the individual response to LCHF and KD for athletes competing at submaximal exercise intensities.

A potential surrogate to identify the individual response to keto-adaptation is RER at V̇O2max. In the present study, when RER was >1.0 compared with <1.0 (n = 4) at V̇O2max (n = 4) after keto-adaptation, between-diet comparisons of mean change from prediet to postdiet demonstrated a higher endurance capacity and lower lactate concentration at exhaustion. This response occurred despite no difference in exercise intensity relative to VT2, V̇O2max and vV̇O2max. It is important to note, however, that 3- to 5-min stages during a graded exercise test may not be reflective of steady-state pulmonary V̇CO2, particularly at high intensities, which can overestimate RER.[38] Although corrective models have been proposed, these have not been validated but do highlight a potential limitation of the current analysis. Nonetheless, a reduction in RER at V̇O2max is associated with impaired high-intensity performance because of a reduction in maximal CHO utilization;[9,11] with the present findings also suggesting a relationship with endurance capacity at submaximal intensities.

Differences in endurance capacity between groups with RER <1.0 compared with >1.0 at V̇O2max were also unrelated to rates of fat oxidation and serum glucose and blood D-βHB concentrations. Because MFO can increase to >1 g·min−1 within 3 d of adaptation to an LCHF diet[39] and blood KB concentrations increase rapidly (hours to days) in response to low-CHO availability (e.g., starvation and exhaustive exercise and a KD),[8] they are unreliable markers for optimal keto-adaptation. Potentially, differences in endurance capacity were due to lower rates of lactate oxidation[40] and gluconeogenesis via lactate[12] in the group with RER <1.0, thus resulting in the accumulation of blood lactate. Lactate production may also be higher because of increased shuttling of pyruvate to lactate dehydrogenase because of the inhibition in pyruvate dehydrogenase activity.[36] Elevated blood lactate concentrations at submaximal exercise intensities also appear in endurance athletes self-reporting chronic (>8 months) keto-adaptation;[10,12] therefore, it is uncertain whether this impairment in CHO metabolism resolves or persists with long-term adherence to a KD.

In conclusion, these findings demonstrate that 31 d of keto-adaptation can preserve mean submaximal exercise capacity. However, exercise efficiency was impaired, endurance variability increased, and there was a greater risk of an endurance decrement at an individual level after keto-adaptation. Moreover, the suggestion that longer adaptation periods to a KD are necessary to enhance endurance performance is currently unsubstantiated, although shifts in RER at V̇O2max may provide a time-course adaptation or performance surrogate to monitor such changes.