Ultrarunners sometimes use a training strategy that involves running in a fasted or glycogen-depleted state. The premise behind this strategy is that withholding carbohydrates before and/or during prolonged training sessions will result in beneficial metabolic adaptations that help the body better metabolize fat and spare glycogen (stored carbohydrate) during aerobic exercise, resulting in improved race performances.
Endurance training in and of itself results in an increased ability to better metabolize fat and spare glycogen as a long duration energy source. But can training in a fasted or glycogen-depleted state lead to further beneficial adaptations and performance gains in trained ultrarunners?
This article explores the research literature to weigh the pros and cons of what has been called the “train low, compete high” (Hansen et al. 2005; Hawley & Burke 2010; Bartlett et al. 2015) method where athletes intentionally withhold carbohydrates before and/or during training sessions (“train low”) to make the availability of carbohydrates during competition (“compete high”) stretch further.
Based on the research evidence, I recommend that ultrarunners avoid training with low carbohydrate availability, especially for higher intensity training sessions. The quality and duration of training sessions become compromised without adequate carbohydrate availability and the research literature has not demonstrated any real-world improvements in endurance performance by engaging in “train low” strategies even where studies demonstrate some positive metabolic adaptations. Ultrarunners are better served by proper fueling strategies that support the training volume required to be successful in the sport.

Training with Low Carbohydrate Availability
Given the widely documented importance of carbohydrates for prolonged endurance performance (e.g., Hawley et al. 1997; Cermak & van Loon 2013), training sessions with low carbohydrate availability are not only difficult for athletes but impede the effectiveness of the sessions — particularly for sessions that involve higher intensity workloads. However, research has demonstrated that training with low carbohydrate availability can induce adaptations in skeletal muscle that regulate energy metabolism, such as the increased production of mitochondria (the sites within cells where energy production takes place) and an increased utilization of fatty acids in energy production (Aird et al. 2018, Keller et al. 2001, Philp et al. 2012, Pilegaard et al. 2002, Steinberg et al. 2006, Wojtaszewski 2003, Yeo et al. 2008a).
In recent years, coaches and ultrarunners have implemented training sessions that attempt to take advantage of the potential benefits of a “train low, compete high” strategy. Training with low carbohydrate availability is typically achieved in one of two ways:
- Reduced endogenous carbohydrate availability, or training with low muscle and liver glycogen stores (i.e., low endogenous sources of carbohydrates). This means depleting the body’s glycogen stores, for example, through one workout early in the day followed by another workout later in the day with little carbohydrate replacement in between; or, through an evening workout that depletes glycogen levels followed by a run the next morning with limited carbohydrate intake overnight.
- Reduced exogenous carbohydrate availability, or training without ingesting any supplemental carbohydrates before (e.g., pre-workout meal) or during (e.g., sports drinks, gels) a workout to avoid supplying supplemental glucose to the working muscles. This means starting a training session in a fasted state and/or doing long training sessions with limited carbohydrate intake, such as doing a long run (e.g., 2+ hours) without consuming any carbohydrates during the run.
The use of fasted or glycogen-depletion training sessions is based on the hypothesis that training with low carbohydrate availability vs. high carbohydrate availability results in greater endurance adaptations that lead to performance gains.
If training with low carbohydrate availability results in endurance adaptations that lead to performance gains; then the use of fasted or glycogen-depletion runs may be warranted in an ultrarunner’s training program. However, if training with low carbohydrate availability does not result in endurance adaptations that lead to performance gains; then ultrarunners would be better served by avoiding fasted or glycogen-depletion runs.
The rest of this article examines the research evidence and discusses the implications of the evidence with practical recommendations for coaches and runners.
Research Evidence
Studies have variously examined the role of reduced endogenous carbohydrate availability and reduced exogenous carbohydrate availability on endurance adaptations and performance gains.
Reduced endogenous carbohydrate availability
One way to train with low carbohydrate availability is to exercise in a glycogen-depleted state with reduced endogenous carbohydrate availability — that is, with low muscle and liver glycogen stores.
A key initial study to examine reduced endogenous carbohydrate availability was conducted by Hansen and colleagues (2005). Their participants consisted of untrained individuals who engaged in a 10-week knee-leg extension (“kicking”) exercise program. Each participant trained one leg under low carbohydrate availability and the other leg under high carbohydrate availability. The “low” state was achieved by training that leg twice per day every other day so that the leg began the second workout on those training days with low muscle glycogen. The “high” state was achieved by training that leg once per day every day, ensuring the leg began each workout with high muscle glycogen. At the end of the study, the leg trained under low carbohydrate availability showed greater improvement in time to fatigue over the leg trained under high carbohydrate availability. Muscle biopsies also found that the leg trained under low carbohydrate availability showed greater metabolic adaptations. At least for untrained individuals, the study demonstrated some beneficial endurance adaptations from training with low carbohydrate availability.
To understand whether the same benefits would hold for well-trained individuals, Yeo and colleagues (2008b) selected participants with more than three years of endurance training and put them through a three-week cycling program. Half of the participants (the “high” group) trained once per day under high carbohydrate availability, alternating days of aerobic training and high-intensity interval training. The other half (the “low” group) trained twice per day every other day, with the aerobic workout done first followed by two hours of rest without any carbohydrate intake, and then the high-intensity interval training. Similar to the study on untrained individuals (Hansen et al. 2005), the “low” state meant participants began their second workout with low muscle glycogen, while the “high” state meant participants began each workout with high muscle glycogen. The results showed positive metabolic improvements while training with low carbohydrate availability — this time on trained individuals — including enhanced fat oxidation during aerobic activity. However, there was no significant difference in time-trial performance across the two groups.
Hulston and colleagues (2010) studied the impact of low muscle glycogen on fat metabolism in well-trained cyclists over a 3-week period. Half of the study participants were assigned to a “low” group that trained twice per day every other day so that they entered the second of those workouts with low glycogen availability. The other half were assigned to a “high” group that trained once per day so that they entered their workouts with high glycogen availability. Although results showed that high-intensity power output was better in the “high” group, fat oxidation improved in the “low” group. But time trial performance improved similarly in both groups. As the authors conclude, “Training with low muscle glycogen reduced training intensity and, in performance, was no more effective than training with high muscle glycogen. However, fat oxidation was increased after training with low muscle glycogen, which may have been due to the enhanced metabolic adaptations in skeletal muscle” (Hulston et al. 2010: 2046).
Reduced exogenous carbohydrate availability
Another way to achieve low carbohydrate availability is to withhold carbohydrate supplementation before and/or during exercise (reduced exogenous supply). This means avoiding pre-training meals and/or not using sports drinks or gels during training so that supplemental glucose is not supplied to the working muscles. “Glucose supplementation during exercise inhibits whole-body fat oxidation by suppressing plasma free fatty acid (FFA) levels while concomitantly reducing the entry of long-chain fatty acids into the mitochondrion, an effect that persists for several hours after ingestion” (Hawley & Burke 2010: 155). The question is whether training without supplemental carbohydrates results in adaptations that lead to improved endurance performance.
De Bock and colleagues (2008) studied whether training in a fasted state promoted “adaptations of fat metabolic pathways” (1051). Participants in their study engaged in a 6-week training program. Half of the participants began morning workouts in a fasted state — the “low” group. The other half ate a carbohydrate-rich breakfast 90 minutes prior to the workout and then ingested additional carbohydrates during exercise — the “high” group. Although training in a fasted state (the “low” group) led to “a decrease in exercise-induced glycogen breakdown and an increase in proteins involved in fat handling,” rates of fat oxidation were similar across both groups (1045). The adaptations gained by training in a fasted state did not exceed “overall training adaptations” in this study (1053).
Cox and colleagues (2010) studied endurance-trained cyclists and triathletes engaged in a 28-day cycling program. While maintaining a standard diet for all participants outside of training, half of the subjects were given supplemental carbohydrates during each hour of training (the “high” group) and the other half were given water (the “low” group). The “high” group showed increased activity of one metabolic marker (citrate synthase) important for energy production in mitochondria, demonstrating a different outcome than other studies that showed beneficial metabolic adaptations from training “low.” However, these metabolic changes did not translate into improvement in endurance performance as measured in a time trial. The authors conclude that “these metabolic changes do not alter the training-induced magnitude of increase in exercise performance” (126).
Van Proeyen and colleagues (2011), found beneficial metabolic adaptations in cyclists who trained in a fasted state. Participants engaged in a 6-week endurance program consisting of cycling. The “low” group trained in a fasted state while the “high” group ate a carbohydrate-rich breakfast 90 minutes prior to exercise and then ingested additional carbohydrates during exercise. Performance in a 60-minute time trial similarly increased across both groups, but the “low” group showed a greater increase in “the exercise intensity corresponding to the maximal rate of fat oxidation,” as well as upregulation of several metabolic markers and the prevention of an “exercise-induced drop in blood glucose concentration.” The authors conclude that training in a fasted state is more effective at increasing “muscular oxidative capacity” among other benefits (Van Proeyen et al. 2011: 236). However, there was no performance difference between the two groups in the time trial at the end of the study.
Reduced endogenous and exogenous carbohydrate availability
In many training situations, ultrarunners may use a combination of the two approaches discussed earlier where they start a run in a glycogen-depleted state (low endogenous carbohydrate availability) and then limit carbohydrate intake during the training session (low exogenous carbohydrate availability).
Morton and colleagues (2009) explored how these different types of carbohydrate availability interact with recreationally active participants in a 6-week training program. Participants were assigned to one of three groups with all groups completing the same amount of training. Groups 1 and 2 trained twice per day on two days each week, which meant participants in both groups entered their second session on those days with reduced glycogen levels. However, group 1 ingested supplemental carbohydrates before and during those second sessions (group 1 = low endogenous + high exogenous), while group 2 consumed a placebo (group 2 = low endogenous + low exogenous). Group 3 trained once per day on four days each week, which meant they entered each workout with high glycogen levels, and they did not ingest any carbohydrates during their training (group 3 = high endogenous + low exogenous). Endurance performance as measured by time to exhaustion was similar across all three groups, as were several cellular adaptations, but group 2 did exhibit “an enhanced stimulus for inducing oxidative enzyme adaptations of skeletal muscle,” suggesting that “training under conditions of reduced carbohydrate availability from both endogenous and exogenous sources” could be beneficial for low-intensity exercise, “although this does not translate to improved performance during high-intensity exercise” (Morton et al. 2009).
Discussion of Implications
Research evidence generally demonstrates that training with low carbohydrate availability leads to metabolic adaptations positive for endurance activities, in both untrained and trained individuals. In theory, this should be beneficial to ultrarunners. “However, despite creating conditions that should, in theory, enhance exercise capacity, the effects of this train-low strategy on a range of performance measures are equivocal” (Hawley & Burke 2010: 155). The cellular adaptations found in studies do not necessarily translate into actual performance gains. Even where evidence exists demonstrating that training with low carbohydrate availability leads to increases in beneficial metabolic markers, the same studies often do not show any real-world performance gains as measured in time-trial performance. This is true whether low carbohydrate availability is achieved by limiting endogenous or exogenous sources — that is, by training in a glycogen-depleted state or limiting the ingestion of carbohydrates before and/or during workouts.
Marquet and colleagues (2016) note, “Part of the reason for this ‘disconnect’ between ‘mechanistic’ and performance outcomes is that the dietary-training strategies that successfully augment markers of training adaptation simultaneously reduce the intensity at which athletes can train during key high-intensity interval training (HIT) sessions” (664). Given the various questions surrounding the optimal “train low” strategies (Bartlett et al. 2015), Marquet and colleagues (2016) recommend a periodized strategy that manipulates carbohydrate availability around various training goals. In their study, they divided trained triathletes into two groups for a 3-week training program. Group 1 (“sleep low”) performed evening interval sessions with high carbohydrate availability (i.e., high glycogen stores) followed by morning sessions that ensured both low endogenous carbohydrate availability (due to the restriction of carbohydrates overnight) and low exogenous carbohydrate availability (no carbohydrate intake during the workout). Group 2 performed the same training with high carbohydrate availability throughout. In contrast to many of the studies reviewed earlier, the “sleep low” group that periodized their carbohydrate restriction showed “significant improvements in submaximal cycling economy, as well as supramaximal cycling capacity and 10-km running time” (Marquet et al. 2016: 663).
Marquet et al. (2016) was the first study to document both metabolic improvements and practical performance gains, which suggests that a periodized approach to training with low carbohydrate availability may be warranted. Impey and colleagues (2016) similarly support the recommendation of “fueling for the work required.” That is, “when the goals of the training session are to complete the highest workload possible over more prolonged durations, then adequate CHO should be provided in the 24 h period prior to and during the specific training session” (Impey et al. 2016: 13; see also, Jeukendrup 2017). This mitigates the problem of not being able to train hard enough when in a glycogen-depleted and/or fasted state to attain the training effects targeted by higher intensity training sessions.
However, more recent review articles have found that long-term performance gains are not always achieved with a periodized “train low” strategy (Gejl & Nybo 2021; Podlogar & Wallis 2022). In their systematic review and meta-analysis of performance effects of periodized carbohydrate restriction, Gejl & Nybo (2021) found no overall effect of periodized “train low” strategies on endurance performance. Podlogar & Wallis (2022) explain why performance gains are elusive: “Even if a training session is initiated with adequate muscle glycogen stores, they will be markedly reduced by the end of it, creating a suitable environment for activation of crucial molecular signaling pathways thought to be responsible for positive adaptations” (S13). Endurance athletes — especially ultrarunners accustomed to large training volumes — need not intentionally seek out opportunities to train with low carbohydrate availability because endurance training naturally leads to the conditions for positive adaptations. Furthermore, “the capacity to oxidize fat should naturally come together with improved training status [i.e., increased training volume]” (Podlogar & Wallis 2022: S13). As Hetlelid and colleagues (2015) have shown, well-trained runners have higher fat oxidation rates at higher intensities than recreational runners even without taking into account dietary manipulations. Endurance training, in and of itself, is associated with general metabolic adaptations and performance improvements. Given that training with low carbohydrate availability negatively impacts the quality and/or duration of those training sessions and that research generally does not demonstrate performance gains as a result, “training with low carbohydrate availability should likely be at best viewed as a more time efficient way to train rather than the optimal way” (Podlogar & Wallis 2022: S13).
Conclusion & Recommendations
A review of the research literature does not support the hypothesis that training with low carbohydrate availability vs. high carbohydrate availability results in greater endurance adaptations that lead to performance gains.
In recent years, coaches and ultrarunners have used glycogen-depletion and fasted runs in an attempt to stimulate positive metabolic adaptations that include mitochondrial biogenesis and increased fat oxidation. Although research on training strategies that restrict carbohydrate availability have demonstrated some positive metabolic adaptations, they have not demonstrated clear real-world improvements in endurance performance.
Since going into training sessions in a glycogen-depleted state and/or restricting carbohydrates during prolonged training sessions negatively impacts the quality and duration of those workouts, ultrarunners are better off focusing on adequate carbohydrate fueling to support the training volume required to be successful in the sport.
There may be situations where runners have low carbohydrate availability before or during training sessions. For example, due to time constraints, a morning run may need to be done before eating breakfast. Or, due to logistical constraints, it may be difficult to carry or access the carbohydrates needed to adequately fuel during a long run. But even if all training cannot take place under optimal carbohydrate availability, runners should ensure they are properly fueled for higher intensity training sessions to get the most out of those workouts.
Another factor mitigating against training with low carbohydrate availability is the need to avoid a negative energy imbalance in the athlete’s overall diet. Ultrarunners require substantial calories to support high training volumes. Restricting calories during training can lead to an overall caloric deficit and the condition known as RED-S, or “relative energy deficiency in sport.” RED-S can lead to “short and long term consequences on health and sport performance” (Vardardottir et al. 2020). Ultimately, endurance athletes need adequate calories and nutrients to excel. With that in mind, plus the lack of documented improvements to endurance performance associated with training strategies that restrict carbohydrate availability, ultrarunners are better served by dietary strategies that support adequate carbohydrate availability during training.
References
Aird, T.P.; R.W. Davies; and B.P. Carson. 2018. “Effects of fasted vs fed-state exercise on performance and post-exercise metabolism: A systematic review and meta-analysis.” Scandinavian Journal of Medicine & Science in Sports 28: 1476-1493. DOI: 10.1111/sms.13054
Bartlett, Jonathan D.; John A. Hawley; and James P. Morton. 2015. “Carbohydrate availability and exercise training adaptation: Too much of a good thing?” European Journal of Sport Science 15(1): 3-12. DOI: 10.1080/17461391.2014.920926
Cermak, N.M. and L.J.C. Van Loon. 2013. “The use of carbohydrates during exercise as an ergogenic aid. Sports Medicine 43: 1139–1155. DOI: 10.1007/s40279-013-0079-0
Cox, Gregory R; Sally A Clark; Amanda J Cox; Shona L Halson; Mark Hargreaves; John A Hawley; Nikki Jeacocke; Rodney J Snow; Wee Kian Yeo; and Louise M Burke. 2010. “Daily training with high carbohydrate availability increases exogenous carbohydrate oxidation during endurance cycling.” Journal of Applied Physiology 109(1): 126-134. DOI: 10.1152/japplphysiol.00950.2009
De Bock, K; W Derave; BO Eijnde; MK Hesselink; E Koninckx; AJ Rose; P Schrauwen; A Bonen; EA Richter; and P Hespel. 2008. “Effect of training in the fasted state on metabolic responses during exercise with carbohydrate intake.” Journal of Applied Physiology 104(4): 1045-1055. DOI: 10.1152/japplphysiol.01195.2007
Gejl, KD and L. Nybo. 2021. “Performance effects of periodized carbohydrate restriction in endurance trained athletes — a systematic review and meta-analysis.” Journal of the International Society of Sports Nutrition 18: 37. DOI: 10.1186/s12970-021-00435-3
Hansen, Anne K; Christian P Fischer; Peter Plomgaard; Jesper Løvind Andersen; Bengt Saltin; and Bente Klarlund Pedersen. 2005. “Skeletal muscle adaptation: training twice every second day vs. training once daily.” Journal of Applied Physiology 98(1): 93-99. DOI: 10.1152/japplphysiol.00163.2004
Hawley, JA; EJ Schabort; TD Noakes; and SC Dennis. 1997. “Carbohydrate-loading and exercise performance. An update.” Sports Medicine 24(2): 73–81. DOI: 10.2165/00007256-199724020-00001
Hulston, Carl J.; Michelle C. Venables; Chris H. Mann; Cara Martin; Andrew Philp; Keith Baar; and Asker E. Jeukendrup. 2010. “Training with Low Muscle Glycogen Enhances Fat Metabolism in Well-Trained Cyclists. Medicine & Science in Sports & Exercise 42(11): 2046-2055. DOI: 10.1249/MSS.0b013e3181dd5070
Hawley, J. A., & Burke, L. M. 2010. Carbohydrate Availability and Training Adaptation: Effects on Cell Metabolism. Exercise and Sport Sciences Reviews 38(4): 152-160. DOI: 10.1097/JES.0b013e3181f44dd9
Impey, Susan G.; Kelly m. Hammond; Sam O. Shepherd; Adam P. Sharples; Claire Stewart; Marie Limb; Kenneth Smith; Andrew Philp; Stewart Jermson; D. Lee Hamilton; Graeme L. Close; and James P. Morton. 2016. “Fuel for the work required: a practical approach to amalgamating train-low paradigms for endurance athletes.” Physiological Reports 4(10): e12803. DOI: 10.14814/phy2.12803
Jeukendrup, Asker E. 2017. “Periodized nutrition for athletes.” Sports Medicine 47: 51–63. DOI: 10.1007/s40279-017-0694-2
Keller, C; A Steensberg; H Pilegaard; T Osada; B Saltin; BK Pedersen; and PD Neufer. 2001. “Transcriptional activation of the IL-6 gene in human contracting skeletal muscle: influence of muscle glycogen content.” FASEB Journal. 15(14): 2748-2750. DOI: 10.1096/fj.01-0507fje
Marquet, Larie-Anne; Jeanick Brisswalter; Julien Louis; Eve Tiollier; Louise M. Burke; John A. Hawley; and Christophe Hausswirth. 2016. “Enhanced Endurance Performance by Periodization of Carbohydrate Intake: ‘Sleep Low’ Strategy.” Medicine & Science in Sports & Exercise 48(4): 663-672. DOI: 10.1249/MSS.0000000000000823
Morton, James P; Louise Croft; Jonathan D Bartlett; Don PM Maclaren; Thomas Reilly; Louise Evans; Anne McArdle; and Barry Drust. 2009. “Reduced carbohydrate availability does not modulate training-induced heat shock protein adaptations but does upregulate oxidative enzyme activity in human skeletal muscle.” Journal of Applied Physiology 106(5): 1513Y-1521. DOI: 10.1152/japplphysiol.00003.2009
Philp, Andrew; Mark Hargreaves; and Keith Baar. 2012. “More than a store: Regulatory roles for glycogen in skeletal muscle adaptation to exercise.” American Journal of Physiology-Endocrinology and Metabolism 302(11): E1343–E1351. DOI: 10.1152/ajpendo.00004.2012
Pilegaard, Henriette; Charlotte Keller; Adam Steensberg; Jørn Wulff Helge; Bente Klarlund Pedersen; Bengt Saltin; and P Darrell Neufer. 2002. “Influence of pre-exercise muscle glycogen content on exercise-induced transcriptional regulation of metabolic genes.” Journal of Physiology 541: 261-271. DOI: 10.1113/jphysiol.2002.016832
Podlogar, Tim and Gareth A. Wallis. 2022. “New Horizons in Carbohydrate Research and Application for Endurance Athletes.” Sports Medicine 52(Suppl 1): S5–S23. DOI: 10.1007/s40279-022-01757-1
Steinberg, Gregory R; Matthew J Watt; Sean L McGee; Stanley Chan; Mark Hargreaves; Mark A Febbraio; David Stapleton; and Bruce E Kemp. 2006. “Reduced glycogen availability is associated with increased AMPKalpha2 activity, nuclear AMPKalpha2 protein abundance, and GLUT4 mRNA expression in contracting human skeletal muscle.” Applied Physiology Nutrition and Metabolism 31(3): 302-312. DOI: 10.1139/h06-003
Van Proeyen, Karen; Karolina Szlufcik; Henri Nielens; Monique Ramaekers; and Peter Hespel. 2010. “Beneficial metabolic adaptations due to endurance exercise training in the fasted state.” Journal of Applied Physiology 110: 236–245. DOI: 10.1152/japplphysiol.00907.2010.
Vardardottir, Birna; Sigridur Lara Gudmundsdottir; and Anna Sigridur Olafsdottir. 2020. “Health and performance consequences of Relative Energy Deficiency in Sport (RED-s).” Laeknabladid 106(9): 406-413. DOI: 10.17992/lbl.2020.09.596
Wojtaszewski, Jorgen; Christopher MacDonald; Jakob N Nielsen; Ylva Hellsten; D Grahame Hardie; Bruce E Kemp; Bente Kiens; and Erik A Richter. 2003. “Regulation of 5’AMP-activated protein kinase activity and substrate utilization in exercising human skeletal muscle.” American Journal of Physiology-Endocrinology and Metabolism 284(4): E813-E822. DOI: 10.1152/ajpendo.00436.2002
Yeo, Wee Kian; Sarah J. Lessard; Zhi-Ping Chen; Andrew P. Garnham; Louise M. Burke; Donato A. Rivas; Bruce E. Kemp; and John A. Hawley. 2008a. “Fat adaptation followed by carbohydrate restoration increases AMPK activity in skeletal muscle from trained humans.” Journal of Applied Physiology 105: 1519Y26. DOI: 10.1152/japplphysiol.90540.2008
Yeo, Wee Kian; Carl D Paton; Andrew P Garnham; Louise M Burke; Andrew L Carey; and John A Hawley. 2008b. “Skeletal muscle adaptation and performance responses to once a day versus twice every second day endurance training regimens.” Journal of Applied Physiology 105(5): 1462-1470. DOI: 10.1152/japplphysiol.90882.2008