It has long been “common knowledge” that endurance exercise is fuel-limited. This is most familiar to general populations as a commonly observed limitation on foot races that says that you will run out of energy, “hit the wall,” or “bonk” at around 20 miles. For common race distances, the effect is most frequently observed in the marathon (26.2 miles). The standard explanation is that you can only store a limited amount of carbohydrate fuel in the form of muscle and liver glycogen, and that you will bonk when the supplies run out. Management of the “problem” is variously accomplished by “carb-loading” before the race to super-fill (“super-compensate”) the glycogen stores, and by various refueling strategies during the race (which support a large industry that provides special drinks, gels, bars, and the like).
At the same time, there is a growing body of evidence that there is no reason to depend on stored carbohydrate fuel which is limited in quantity if you can mobilize your fat stores as a source of fuel. The problem seems to be that anyone who is accustomed (adapted) to eating a high-carbohydrate diet tends to preferentially burn carbohydrates as fuel and does not utilize the energy available from fat in a way that spares carbohydrate use during endurance exercise. These are the so-called “sugar burners” as compared to “fat burners.”
There are a few possible strategies to enhance fat burning. The most commonly discussed strategy is to train at slower, more aerobic paces, to train the body to optimize using fat as fuel. This approach was discussed by Alan Couzens. Although carbohydrate was also restricted, the main emphasis was on low intensity training, with pretty amazing results from the combined approach.
A second strategy, perhaps even simpler, is to change to a low-carbohydrate diet. Maintaining glycogen supplies in a partially depleted state encourages the development of aerobic apparatus via the action of AMP Kinase (increased mitochondria and associated fat burning enzymes such as carnitine palmitoyltransferase, necessary to help transport fatty acids across the mitochondrial membrane). People with deficiencies of carnitine palmitoyltransferase are unable to utilize fat for fuel efficiently, and can even suffer damaged muscles and rhabdomyelosis resulting in kidney failure. There is also evidence that people suffering from obesity and diabetes may have impaired mitochondrial function, which inhibits their fat burning and at least partially explains the origins of their condition. Similarly, a deficiency of carnitine can result in impaired mitochondrial function and even lead to symptoms of metabolic syndrome.
There is an adaptation period of a few weeks to lowering carbohydrate intake during which the appropriate metabolic pathways and aerobic enzymes and cofactors are up-regulated to better utilize fat as a substrate, allowing blood glucose levels during exercise to be more easily maintained. As Alan Couzens explains, lower carb intake should result in a 20% increase in fat burning in a few weeks. Numerous scientific studies bear this out. A recent study, for example, concludes that a high fat diet increased the rate of whole-body and muscle fat oxidation while reducing the rate of muscle glycogenolysis during submaximal exercise, even after restoring high carbohydrate intake. In this study, a high fat diet was followed for up to two weeks (which we know is just long enough to see significant adaptation), during which usual training was followed. A high carbohydrate diet was restored for the last few days along with a taper before racing. The athletes were able to maintain the increased fat oxidation even after reinstating high carbohydrate intake.
There is at least one supplement (Vespa) which is alleged to improve fat-burning during exercise. There is also some evidence that, all else being equal, the athlete that has the most efficient fat-burning ability can consistently beat the athlete with the higher raw power output (higher VO2 max) in endurance competition.
An article in Running Times last year discussed the strategy of carb cutting to increase endurance, though the article suggests mainly forgoing carb supplementation on training runs, not going low carb in the training diet. The Ethiopian runners again are cited as evidence that a high carb diet is necessary to fuel runs, and there is speculation that they are somehow training themselves not to use glycogen as the predominant fuel by not consuming carbs while actually training. This conclusion seems unlikely to us, since it is well known that exercise causes an increase in blood sugar (due to action of epinephrine and glucagon), effectively mimicking the result of taking in supplemental carbohydrate while exercising. A marathoner does not “hit the wall” until they have run ~20 miles at race pace anyway, a condition not likely encountered during routine training runs for most of us.
Despite all that is known about “the wall,” people persist in believing that a high carbohydrate diet is necessary if you hope to be a good runner. A high carbohydrate diet may work well for Ethiopian runners, who are known for running prodigious mileage and training three times per day. At this training intensity, we seriously doubt their glycogen stores are ever completely full, and they may effectively be training in a perpetually glycogen depleted state, which would upregulate the same enzyme systems that occurs in the low carbohydrate condition. Perhaps what they are mainly training is the rapid replenishment of glycogen stores. However, for most of us amateur or older runners, we’re only fooling ourselves if we think we need to fuel exercise with a high carb diet. David does quite well on very minimal carbs (less than 50 g/day), while Cynthia prefers closer to 100 g/day. Since we generally only run once per day and even then usually less than 10 miles, there is no reason to consume any more carbs to fuel our exercise habits.
David has been a low-carb runner for more than two years now, keeping carbohydrate consumption low before, during, and after endurance exercise. While he won’t brag about his speed and has no plans to challenge any course records, he consistently finds that he has even energy levels and rapid post-event recovery with minimal muscle aches and pains. Fat fueling works for him! However, until very recently, while he refueled less often than is typical, he did routinely refuel using protein and fat for any event of more than 3–4 hours.
He was prompted to revisit the issue when he heard about a recent study (“Metabolic Factors Limiting Performance in Marathon Runners”) by Ben Rapoport when he had his fifteen minutes of fame on NPR in October 2010. Ben was attempting to model bonking in the context of marathon runners. David subsequently exchanged e-mail with Ben whose initial reaction was to interpret David’s personal experience as simply confirming Ben’s model, arguing that fat-fueling works, and that you don’t bonk as long as you run sufficiently slowly. David pointed out some flaws in Ben’s interpretation of the scientific data he analyzed and challenged him to define a performance that David might personally be able to achieve that he might consider sufficient to disprove his model, but he never responded to the challenge. We summarize here both a couple of the criticisms of Ben’s model and analysis and some more recent personal experience and experiments.
A key part of the experimental data that Ben used in building his model came from Romijn AJ, Coyle EF, Sidossis LS, Gastaldelli A, Horowitz JF, et al. (1993) “Regulation of endogenous fat and carbohydrate metabolism in relation to exercise intensity and duration,” Am J Physiol 265: E380–E391. That paper shows (in Fig. 8) the relative amount of particular substrates which are used in short-term exercise at three different intensities. The subject’s dietary habits were not identified, and any changes that might take place over more than a few minutes were ignored. The data show that more intense efforts use primarily muscle glycogen as fuel. They do not show the pathways that feed the muscle glycogen. Glycogen and glucose are rapidly interconverted, and over longer periods of exercise, new supplies of glucose are made by the liver. The liver, in turn, is capable of using any available fuel source via gluconeogenesis to keep blood glucose levels up. These fuel sources include ingested carbohydrates, protein, and fat as well as lactate, glycerol from liver and adipose tissue fat stores, and even body protein stores such as in muscle tissue. Which source predominates depends on availability. Some of the necessary up- and down-regulation of different metabolic pathways can take place on a short time scale to respond to instantaneous demand for fuel from the muscles and other bodily functions. However, there are also some key adjustments that respond more to average demand and supply to deal with changing diet, exercise demand, starvation, and the like, and these are events that tend to have time constants measured in days or weeks, not minutes or hours. The adaptation processes may be numerous and complex (or at least, we fully admit that we don’t understand all of them) and likely include adaptations in the skeletal muscles as well as in the liver. However, we think it is clear that the data extracted from Romijn do not actually show the underlying consumed fuel source for various levels of endurance exercise, and the data cannot reasonably be extrapolated to model and predict bonking.
One factor which Ben cites in favor of the importance of carbohydrate and fueling is the relative stoichiometry of oxygen consumption for each substrate. He states that carbohydrate oxidation typically generates approximately 120 kcal per mole of respired oxygen, whereas fatty acid oxidation typically generates only approximately 100 kcal per mole of oxygen. We would argue that this, while probably correct, is irrelevant in that the difference is relatively small, and the availability of oxygen is never a limiting factor at a typical aerobic marathon pace. There is always an excess of oxygen available. In any case, regardless of the underlying fuel used to generate blood glucose/glycogen, the immediate muscle activity (except for very short term bursts of power) is fueled by glucose.
Notwithstanding the possible flaws, Ben’s model does make predictions as to whether one is likely to bonk in a marathon. In particular, he has provided a calculator for the marathon. Putting in the best estimates David has for his personal situation, the calculator predicts that he should not bonk for finishing times over about 3 hrs. Since he can only sustain a 3-hour marathon pace for about 1 mile or so, he may never be able to really challenge the model experimentally using himself as a test subject.
However, David was inspired to test his need for refueling during running, and now has a significant number of test results that, if not sufficient to explicitly challenge Ben’s model, are at least startling to many people. For some time now, David routinely runs training runs of up to 4–5 hrs without refueling (water alone). This is still true when he runs with his son at faster speeds who typically chooses to refuel using at least a couple of energy bars or gels. David first pushed the envelope a bit on what was supposed to be a planned 18-mile run that stretched to about 26 miles when the route was blocked by an impassable rain-swollen river crossing. That run extended over about 7 hours including some significant rest breaks, so it was not at a strenuous pace, although it did include some significant elevation change. David consumed only water. The next experience was a 50K (31-mile), mostly flat run, again at a very modest pace, finishing in 6:50, again on water alone. Two weeks later, he completed a much more strenuous 50K on rough rocky trails with about 6000 ft of elevation gain (and loss) in 7:28 on water alone. In all three of these runs, David started with a breakfast of a two-egg cheese omelet and bacon about two hours beforehand. He felt some minor hunger around normal lunch time, but otherwise did not experience any significant fluctuations in energy level, and he certainly did not bonk. His overall weight was down about 3–4 lbs at the end of the day compared to the beginning, but looking at beginning-of-day weights across several days, any weight change was pretty much lost in the noise. Of course, it only takes about a pound of fat to fuel a 50K run, anyway. Another week later David ran yet another moderately hilly 50K event which was slower (8:04) due to snowy trail conditions. Due to a late start (10 am), he chose to eat a little at around the 6-hour mark, mostly because it was a long time since breakfast. However, the ability to do three 50K runs in four weekends, certainly confirms that post-event recovery was rapid.
So what’s the take-away message from all this? It’s interesting to compare our reaction and the experiments that David decided to try to others. Greg Crowther also blogged on Ben’s article. Greg’s big concern was inter- and intra-personal variability, and he questioned the validity of extrapolating from some sort of average measurement of performance to the needs of one individual on one particular day. He didn’t ask (as we did) if too many runners are needlessly afraid of bonking but rather, “if ingesting extra carbohydrates before and during a marathon might help you avoid ‘hitting the wall,’ why wouldn’t you do it? Especially after investing all of that time and effort in training, traveling to the race, etc.?” Hey, we’ll even go further: if ingesting fuel (whether carbohydrate, protein, or fat) during a race might help you maintain a faster pace for whatever reason and for whatever limited portion of the race, even it’s just a placebo effect (or tricking your Central Governor), then by all means go for it! But we also stand by our basic conclusions that (1) carbohydrates are simply not necessary before, during, or after endurance exercise, (2) they’re probably not even particularly beneficial, at least if you are suitably adapted to efficient fat-burning (and may be harmful to your long-term health) and not logging very high miles at high intensity, and (3) unless you’re able and motivated to run, say, an entire marathon at a high percentage of your VO2 max, consuming much of any food immediately before, during, or after the event is entirely optional. Moreover, there may be some benefit to having the stomach completely empty while racing, in order to avoid stomach upset, intestinal discomfort/diarrhea or nausea when running at high intensity. (David does continue to believe in significant protein consumption sometime in the 12 hours or so following strenuous exercise as an aid to muscle repair and recovery, but the timing of even that refueling does not appear to be nearly as critical as some authors suggest.) David will probably resume modest consumption of fat and protein on longer events, just because he’s not big on skipping meals entirely, anyway, but it’s reassuring to know that he can easily keep going without any fuel if for any reason he needs to.