Key points are not available for this paper at this time.
We thank Professor Noakes for his insights on the contribution of hypoglycemia (or glycogen depletion) to exercise capacity (1). However, we feel that a holistic rather than reductionist approach is required to tackle the specific focus of this perspective: the effect of a ketogenic diet on athletic performance. We reiterate points from our original article (2): 1) that sports performance is explained by a complex interaction of factors, and 2) rather than claim a single truth to a superior dietary approach, sports scientists should identify nuances and context within the characteristics of the athlete and the event to determine the most suitable nutrition approach(es). We now present a sports-centric summary of the current literature on ketogenic diets and endurance sports performance, building a dashboard to highlight the nuances of each study rather than the traditional meta-analytical approach, which deliberately eradicates such important detail (see Figure 1). We examine each study for context (scenarios in which there are likely to be true differences between the ketogenic low-carbohydrate high-fat (LCHF) and high-carbohydrate availability (HCHO) approaches), but also caveats (issues with the study design that raise questions about interpretations). In keeping with the original theme, our analysis is limited to studies of ketogenic (extreme CHO restriction; i.e., <50 g·d−1) rather than generic LCHF diets, in humans (rather than species with profound differences in substrate utilization), in populations with habitual sports-specific training (at least tier 2 15), and involving protocols related to endurance sports, which have reasonable translation to sports performance. Performance outcomes were taken from the published reports of individual data that are fully transparent or through digitalization of figures using plotdigitizer.com. For time to exhaustion protocols, data were approximated to a change in time trial performance using the methods of Hopkins et al. (16). The difference between means was then calculated for each test by subtracting the mean difference in the LCHF condition from that of the HCHO/control group in the case of parallel group–designed investigations, or by direct comparison between treatments for crossover studies. We suggest a 2% change in performance as being of real-world significance, based on doubling a 1% within-athlete coefficient in variation in performance; although this is arbitrary and also specific to the athlete and the event, we propose that this is a generous but realistic representation of performance coefficient in variation in competitive athletes (17).FIGURE 1: Proposed dashboard of changes in endurance performance after LCHF diet interventions. - Datapoint (Study) Context Caveats LCHF adaptation alone A (6) Tier 3–5 M racewalkers (n = 10); 3 wk LCHF + base phase training; field conditions (real-life 10,000-m track races, with standardized −2 h prerace meal: HCHO meal for race 1 and LCHF for race 2). Parallel-group design: performance results = change from race 1 to race 2 in LCHF group compared with control HCHO group (n = 9). Fully supervised dietary intervention Nonrandomized (belief-chosen) allocation of treatments may amplify any positive effect of LCHF on performance (due to placebo effect and self-selection according to responsiveness) B (7) Tier 3–5 M + F racewalkers (9 M, 1 F); 3 wk LCHF + base phase training; field conditions (real-life 10,000-m track races, with standardized −2 h prerace meal: HCHO meal for race 1 and LCHF for race 2). Parallel-group design: performance results = change from race 1 to race 2 in LCHF group compared with control HCHO group (n = 5 M, 3 F). Fully supervised dietary intervention Nonrandomized (belief-chosen) allocation of treatments may amplify any positive effect of LCHF on performance as for A. No control of menstrual phase for female athletes spread across groups C (8) Tier 2 M middle-aged runners (n = 8); 3 wk LCHF; field conditions (fasted 5-km hilly run undertaken after 50-min preload run in the heat). Crossover design (no washout): performance results = change from run 1 to race 2. Ad libitum dietary intervention after education Order effect: LCHF undertaken after additional 3-wk training. HCHO trial undertaken under suboptimal conditions (overnight fasted, no CHO intake during preceding exercise or during run). Lack of dietary control led to mismatch (33% decrease) in LCHF energy intake: participants lost 2.1 kg BM in LCHF, which likely contributed to performance change on hilly course. D (9) Tier 2–3? M mixed endurance athletes (n = 9); 12 wk LCHF; laboratory conditions (100 km on cycling ergometer: baseline: HCHO meal and self-selected fueling in 2 h before ride, Intervention: HCHO = HCHO meal + 30–60 g·h−1 CHO during ride; LCHF = LCHF meal + water with electrolytes during ride). Parallel-group design: performance results = change from ride 1 to ride 2 in LCHF compared with control HCHO group (n = 11). Ad libitum dietary intervention after education LCHF group had higher BF at baseline and substantial BM (~6 kg) and BF loss (~5 kg) over intervention (<1 kg in HCHO group). Seven additional subjects failed to complete LCHF intervention and/or testing, suggesting greater opportunity for negative outcomes with this treatment. Other exercise capacity metrics collected in study (e.g., SS, CPT) measured relative to BM, exaggerating changes E (10) Tier 2–3? M cyclists (n = 5); 4 wk LCHF; laboratory conditions (fasted cycling to exhaustion at ~62%–64% V˙O2max).Crossover design (no washout): performance results = change between trial 1 (HCHO) and trial 2 (LCHF). Fully supervised dietary intervention Order effect: all LCHF trials undertaken after additional 4-wk training; suboptimal nutrition support in HCHO trial (overnight fasted, no CHO during exercise); large individual performance differences in response to LCHF; authors also noted "limitation of intensity of exercise that can be performed" with "throttling of function near V˙O2max" F (11) Tier 2? M runners (n = 7); 4 d LCHF; laboratory conditions (fasted 5-km TT on treadmill with simultaneous expired gas analysis). Randomized crossover design (2-wk washout): performance results = comparison between LCHF and HCHO trials. Ad libitum dietary intervention after education Performance validity: TT performed on motorized treadmill while simultaneously analyzing expired gas and RPE (every 500 m) likely reducing capacity to make small but valuable changes in pace as in real-life race G (11) As for F: 14 d LCHF As for F H (11) As for F: 28 d LCHF As for F I (11) As for F: 42 d LCHF As for F J (12) Tier 2–3? M middle-aged runners (n = 10); 31 d LCHF; laboratory conditions (treadmill 1-mile run with −3 h prerun meal). Crossover design (2–3 wk washout): performance results = changes between pretrial and posttrial for both LCHF and HCHO interventions. Ad libitum dietary intervention after education Performance validity: TT was performed while simultaneously analyzing expired gas and RPE (every 200 m) as for F K (13) Tier 2–3? M triathletes/runners (n = 8); 4 wk LCHF; laboratory conditions (treadmill running to exhaustion at 70% V˙O2max: baseline/HCHO trials = 2 g·kg−1 CHO pretrial meal + ~ 8 g·kg−1·h−1 CHO during run; LCHF = energy-matched LCHF pretrial meal and intake during run). Crossover design (2–3 wk washout): performance results = changes between pretrial and posttrial for both LCHF and HCHO interventions. Ad libitum dietary intervention after education Larger individual variability in performance changes with LCHF than HCHO treatment; 2 additional subjects failed to complete LCHF intervention, suggesting greater opportunity for negative outcomes with this treatment L (14) Tier 2? M endurance athletes (n = 7); 9 d LCHF; laboratory conditions (fasted cycling ergometer protocol: 90 min at 70% V˙O2max + incremental protocol to exhaustion − increase by 5% V˙O2max every 5 min. Parallel-group design: performance results = changes between pretrial and posttrial for LCHF and control HCHO group. Ad libitum dietary intervention with supplied CHO or fat-rich drinks daily Nonrandomized (voluntary) allocation of LCHF treatment may amplify any positive effect of LCHF on performance as for A. Suboptimal nutrition support for HCHO trial (overnight fasted and no CHO during exercise). Daily training implemented as a "race" involving same cycling protocol but allowing CHO intake in HCHO group; consistent reduction in endurance capacity also seen with LCHF LCHF + integration with other dietary strategies M (7) Tier 3–5 M + F racewalkers (n = 7 M, 1 F); LCHF/HCHO periodization: 3 wk LCHF + base phase training + 17 d HCHO + race taper; field conditions (real-life 10,000-m track race + 20-km road race: both following prerace meal according to diet; within 20-km race = ~50 g CHO according to individual choice). Parallel-group design: performance results = change from race 1 (time doubled) to race 2 compared with control HCHO group (n = 6 M, 5 F). Fully supervised intervention Nonrandomized (belief-chosen) allocation of treatments may amplify any positive effect of LCHF on performance as for A. No control of menstrual phase for female athletes spread across groups. Comparison of 10,000-m track race and 20-km race via normalization of race distance N (15) Tier 3–5 M racewalkers (n = 7); 6 d LCHF + 1 d HCHO diet (endogenous CHO) restoration); field conditions (real-life 10,000-m track races with standardized 2 g·kg−1 CHO meal −2 h prerace, preintervention and postintervention). Parallel-group design: performance results = change from race 1 to race 2 compared with control HCHO group (n = 6). Fully supervised intervention Nonrandomized (belief-chosen) allocation of treatments may amplify any positive effect of LCHF on performance as for A O (16) Tier 4 M triathlete (n = 1): 2 yr LCHF + 3-wk exogenous CHO training support (60 g·h−1 during 3-wk high-intensity sessions) + exogenous CHO performance support (10 g before trial); laboratory conditions (2 × 30 s cycling ergometer sprints). Case study involving documentation of 7-wk training: performance results = change power output from trial 1 (3 wk documented LCHF) to trial 2 (3 wk documented LCHF + exogenous CHO). Monitored dietary intervention Order effect: LCHF + CHO trials undertaken after additional 3-wk training. Unable to distinguish acute CHO performance effects from differences in adaptation to CHO-supported training P (16) As for O: laboratory conditions (10 g CHO before 4-min cycling ergometer sprint a ) As for O Q (16) As for O: laboratory conditions (30 g CHO during 20-km cycling ergometer TT). Performance results = change in performance time from trial 1 As for O R (16) As for O: laboratory conditions (60 g·h−1 CHO during 100-km cycling ergometer TT. Performance results = change in performance time from trial 1 As for O S (17) Tier 3–5 M + F racewalkers; 7 d LCHF + prerace exogenous ketone (ketone ester); (n = 8 M, 1 F); field conditions (real-life 10,000-m track race with −2 h prerace meal: Baseline/HCHO: 2 g·kg−1 CHO; LCHF = energy matched LCHF meal + 573 mg·kg−1 KE 30 min prerace). Parallel-group design: performance results = change from race 1 to race 2 compared with control HCHO (n = 7 M, 2 F). Fully supervised dietary intervention Nonrandomized (belief-chosen) allocation of treatments; may amplify any positive effect of LCHF on performance as for A. No control of menstrual phase for female athletes spread across groups Data are expressed as either mean ± SEM or mean for time trial (time difference) or average power (work rate) outcome.BF, body fat; BM, body mass; F, female; KE, ketone ester supplement; M, male; TT, time trial; V̇O2max, maximal aerobic capacity. Although we consider this to be a starting point for gaining a more nuanced picture of the performance effects of LCHF diets in their various versions, we note that 1) we are scratching the surface of the permutations of athlete, event, and LCHF strategy; 2) that greater transparency around individual data and experiences is needed to identify true responsiveness versus nonresponder status; and 3) (15) while noting the lack of investigations of true ultraendurance events in which moderate-intensity workloads and/or difficulty in supplying sufficient exogenous CHO to muscle and central nervous system substrate needs creates a different opportunity, the current literature does not support the proposal that LCHF strategies are beneficial for the performance in competitive endurance athletes. Furthermore, scenarios in which there might be equivalency to strategies that focus on high endogenous or exogenous CHO availability are matched by others in which there is detriment to performance.
BURKE et al. (Fri,) studied this question.