Popular enthusiasm for the Paleo diet, including a relatively high proportion of fat and protein presumed characteristic of the diet of our hunter/gatherer ancestors, has re-ignited the long standing debate about the nutritional merits of fat and carbohydrates, especially for athletes. In recent posts I have compared the effects of a high-fat diet with those of a high-carbohydrate diet on metabolic processes that have the potential to effect endurance performance, health and life expectancy. We have examined the evidence of these different diets on the development of preferential use of fat rather than carbohydrates for fuel during exercise; on risk of sustained elevation of the stress hormone, cortisol; on insulin resistance and inflammation; and on weight control. The evidence shows that a high fat diet does promote the use of fats as fuel during exercise, potentially beneficial in warding off disabling glycogen depletion during prolonged exercise. However both types of diet are associated with risks of sustained elevation of cortisol, insulin resistance and chronic inflammation. Particular components of each type of diet, specifically high glycaemic index (GI) carbohydrates which produce a rapid rise in insulin after ingestion, and omega-6 fatty acids, which are pro-inflammatory, are associated with high risk. With regard to weight control, the evidence indicates that low fat and low carbohydrate diets are equally effective. This post examines the evidence for effects on the ultimate outcomes: race performance, health and longevity.
When it comes to evidence regarding the effect of nutrition on performance, there are conflicting findings. In a meta-analysis of 20 studies comparing the effects of high fat with high carbohydrate diet on endurance exercise performance, Erlenbusch and colleagues found that averaged across all studies, subjects consuming a high-carbohydrate diet exercised significantly longer until exhaustion, but there were substantial differences between the findings of different studies, probably reflecting differences in the subjects studied and the design of the study. The benefit of the high carbohydrate diet was relatively large in studies of untrained subjects, but there was very little difference between the two types of diet in studies of trained athletes. In light of the fact that endurance training itself increases capacity for utilization of fats as fuel, it is plausible that in hitherto untrained subjects, a relatively brief period of high fat consumption is inadequate to produce a substantial capacity for fat utilization, so maximizing efficiency of glucose utilisation might be of greater value in such subjects.
There are some noteworthy studies that have reported greater benefit for a high fat diet in trained athletes. An early study from Tim Noakes lab in Capetown compared the effects of 2 weeks of high fat (70%) low carbohydrate (7%) diet with a high carbohydrate (74%), low fat (12%) diet in trained cyclists. The high fat diet led to higher fat utilization and improved performance at moderate exercise intensity, without deterioration of performance at high intensity. The importance of starting an endurance event with well stocked glycogen stores suggests that greater benefit might be obtained for a periodized nutritional strategy in which high fat diet is followed by a brief period of carbohydrate loading. A subsequent study from the Capetown lab using the nutritional periodization strategy found that high-fat consumption for 10 days prior to carbohydrate loading was associated with an increased utilization of fat, a decreased reliance on muscle glycogen, and improved time trial performance in a 20 Km time-trial following 150 minutes of medium intensity cycling.
Other studies of trained athletes reported equivocal results. Carey and colleagues tested the effect of fat adaptation using a nutritional periodization strategy, on performance during a one hour time trial following 4 hours of aerobic cycling. As expected the fat adaptation resulted in increased fat utilization. Power output was 11% higher during the time trial and distance covered was 4% greater, but this effect was not statistically significant. Nonetheless, in 5 of the 7 cyclists, the improvement in performance after fat adaptation was substantial, raising the possibility that the number of subjects was too small to provide adequate statistical power to test for a performance benefit.
Yet other studies indicated no benefit and perhaps even harm from the fat adaptation strategy. A further study from the Capetown lab by Haverman and colleagues compared 100km cycling time trial performance and also 1 Km sprint performance following 6 days of high fat consumption and 1 day of carbohydrate loading with performance following 6 days of high carbohydrate consumption and 1 day of carbohydrate loading. The anticipated enhancement of fat utilization was observed, but there was no significant difference between diets in 100-km time-trial performance, while 1-km sprint power output was significantly worse after the high fat diet. The investigators concluded that despite increasing fat utilization, the strategy of high fat diet followed by carbohydrate loading compromised high intensity sprint performance. This raises the possibility that the increased fat utilization might reflect impaired ability to use carbohydrates rather than an enhanced ability to utilize fats.
Thus, the tide of evidence has turned against the hope that fat adaptation produced by a period of one or two weeks of high fat consumption might be a worthwhile strategy for improving endurance performance. In contrast, this strategy might actually impair high intensity performance – an issue that is potentially of some importance even in events lasting several hours in which surges or hills might play a part in race outcome. The evidence does not rule out the possibility that some individuals might enjoy an improvement in endurance performance, but at this stage, the evidence does not justify a general recommendation of this strategy.
Perhaps improvement in performance from rather drastic dietary adjustments over a period of a few weeks is not the issue of greatest importance to the endurance athlete, for whom training is an undertaking extending over many months or years. Rather, the question of greater importance is the effect of long term nutrition on long term health. Although no studies have examined the long term effects of long term nutrition in endurance athletes, recent evidence has provided increasing clarity regarding the optimum diet for long term health in the general population.
Long term health and life expectancy
We will focus on evidence related to heart health because heart disease is the greatest cause of mortality in the general population and in addition there is some evidence that extensive endurance training and racing might in fact increase the risk of cardiovascular disease in athletes. Furthermore, most evidence suggests that a healthy diet for the heart minimises risk of cancer, though there are instances where foods that appear healthy for the heart have been linked to increased risk of cancer. Although depression is associated with only a modest risk of premature death, it is the illness causing the greatest degree of disability world-wide (according to the World Health Organization). Furthermore, mental state is of substantial importance in athletic performance. Therefore, I will also briefly address the evidence regarding the association between diet and depression.
In a recent comprehensive review of nutritional recommendations for cardiovascular disease prevention Eilat-Adar and colleagues found that both low fat and low carbohydrate diets are a healthy alternative to the typical Western diet. They note that low carbohydrate diets are associated with lower levels of potentially harmful tryglycerides and with higher levels of beneficial cholesterol in high density lipoprotein (HDL). Low-carbohydrate diets, which include 30%–40% of calories from carbohydrates and are low in saturated fat but high in mono-unsaturated fat, were found to be safe in healthy and overweight individuals at follow-up for up to 4 years. We will return to the controversial issue of saturated fat later. Eliat-Adar also found good evidence that Mediterranean diets, which include high consumption of fruit, vegetables and legumes, together with moderatley large amounts of fish but less red meat, may improve quality and life expectancy in healthy people, as well as in patients with diabetes, and heart disease. Mediterranean diets are preferable to a low-fat diet in reducing triglyceride levels, increasing HDL cholesterol, and improving insulin sensitivity.
A meta-analysis of trials by the Cochrane Collaboration – an organization which does extremely rigorous and conservative reviews of medical treatments – also concluded that evidence suggests favourable effects of the Mediterranean diet on cardiovascular risk factors, though with their usual caution, they stated that more trials are needed.
One trial that warrants special mention is the Spanish Prevención con Dieta Mediterránea (PREDIMED) trial, in which 7,216 men and women aged 55 to 80 years were randomized to 1 of 3 interventions: Mediterranean diets supplemented with nuts or olive oil and control diet. During a follow-up period of near to five years, nut consumption was associated with a significantly reduced risk of all-cause mortality. Subjects consuming more than 3 servings/week of nuts had a 39% lower mortality risk. A similar protective effect against cardiovascular and cancer mortality was observed.
With regard to the issue of saturated versus unsaturated fats, a recent re-analysis of the large and well conducted West Sydney Heart study found that replacing dietary saturated fat with omega- 6 linoleic acid, for subjects with known cardiovascular disease, actually led higher all-cause death rate, and higher death rate from coronary heart disease and cardiovascular disease. The authors also performed a new meta-analysis of previous studies and found that the pooled data also provided a strong trend towards higher death rate when saturated fat was replaced by omega-6 linoleic acid. This finding is contrary to the prominent advice to substitute polyunsaturated fats for saturated fats in worldwide dietary guidelines for reducing risk of coronary heart disease. The most plausible explanation is that the increased death rate is due to the pro-inflammatory effects of omega-6 fatty acids.
The frequent reports in both popular press and the medical literature linking various foodstuffs to cancers of various types makes this topic a mine-field. In part this situation reflects the heterogeneity of cancer and the multiplicity of different factors that might contribute to the cause in different cases. Nonetheless, in general, the evidence indicates that diets that are healthy with regard to weight control and cardiovascular outcome tend to be associated with lower risk of cancer. For example, a recent large review found that adherence to the Mediterranean diet was associated with lower risk of certain cancers, especially cancers of the digestive tract, consistent with the finding from the PREDIMED study mentioned above. However, in light of the fact that a key difference between typical Western diets and the Mediterranean diet is the larger relative amount of omega-3 fats in the Mediterranean diet, it is noteworthy that some studies have reported that omega-3 fats are associated with increased rate of prostate cancer, while others have reported a decreased rate. This should encourage caution against simplistic conclusions that a food item is invariably healthy in all amounts and all circumstances.
Many studies using relatively low quality methodology to assess diet and/or mental state have reported an association between adherence to a ‘healthy’ diet and decreased risk of depression. More recently, several studies have addressed this issue using more rigorous methodology. A meta-analysis by Psaltopolou and colleagues of studies examining the association between Mediterranean diet and risk of various neurological and mental disorders found that the Mediterranean diet was associated with a decrease in risk of depression of approximately 30%. This reduction was very similar in magnitude to the reduction of risk of stroke and for cognitive impairment. However, association cannot establish cause, and it is possible that other life-style factors associated with adherence to a healthy diet account for the better physical and mental health. The most conclusive evidence comes for randomized controlled trials in which individuals are randomly allocated to different diets. In the PREDIMED trial, the group who were allocated to the Mediterranean diet augmented with extra nuts experienced a 20% lower rate of depression over a period of 3 years, compared with those on a low fat diet. This was not a statistically significant reduction. However, in those who had type 2 diabetes, the Mediterranean diet with extra nuts produced a 40% reduction in occurrence of depression which was significant. Thus the balance of evidence does suggest that a Mediterranean diet augmented by nuts produces a reduction which is significant at least in those who already show other evidence of adverse metabolic effects.
There is overwhelming evidence that diet plays a large role in health and longevity, and after many years of confusing debate, there is emerging clarity about the type of diet that is healthiest. This is neither a high fat/low carbohydrate Paleo diet nor a low fat/high carbohydrate diet. Rather, a substantial body of evidence suggests a Mediterranean diet is preferable, especially when augmented with extra nuts.
There is some variability between studies in what is taken to be the Mediterranean diet, but the consistent features include high consumption of fruits, vegetables and legumes (beans, nuts, peas, lentils); low consumption of red meat and meat products but substantial consumption of fish; near equal proportions of omega-3 and omega-6 fats; moderate consumption of milk and dairy products; and low to moderate red wine consumption. The status of grains and cereals is ambiguous. The Mediterranean diet adopted in PREDIMED included a high consumption of grain and cereals. In general, whole grains and cereals appear healthy though gluten sensitivity is an issue for at least some individuals.
While the evidence for the Mediterranean diet is largely based on studies of the general population with emphasis on heart health, rather than being focused on athletes, the disconcerting evidence that male athletes who have run numerous marathons over a period of many years are at risk of atherosclerosis (as discussed in detail in my post of 30th May, 2012) suggests that a ‘heart-healthy’ diet should be a high priority for endurance athletes.
When it comes to endurance performance, there is no clear evidence in favour of any particular diet. However the consistent evidence that a high fat/low carbohydrate diet promotes preferential utilization of fats during exercise appeared promising at first. It is disappointing that this apparently beneficial adaptation is not reflected in enhanced performance, even in ultra-endurance events. On contrast, there is actually evidence that it can harm high intensity performance, such as 1 Km cycling time trial performance. However, the fact that at least some individuals do appear to show an endurance performance benefit from a high fat diet (followed by brief duration carbohydrate loading) as observed in the study by Carey and colleagues, makes me reluctant to dismiss the potential value of at least moderately high fat consumption. One crucial issue is to identify why the clear evidence of improved fat utilization does not generally lead to enhanced performance. It appears that the fat adaptation strategy, at least in the form of a rapid increase in proportion of fat to a quite high level over a periods of a few weeks, is in some way harming the utilization of carbohydrates as much as it might be improving the utilization of fats.
In my opinion, one candidate mechanism by which high fat consumption might harm carbohydrate metabolism in muscle is the elevation of cortisol associated with the fat adaptation strategy used in the studies. One immediate effect of high cortisol is the decrease in accessibility of the glut4 transporter molecules that transport glucose into muscle. Furthermore, sustained elevation of cortisol can produce a decrease in sensitivity of glucocorticoid receptors that mediate the various effects of cortisol, including its anti-inflammatory effects, thereby possibly leading somewhat paradoxically to chronic inflammation. This is speculation on the basis of what is known about mechanisms rather than direct evidence of beneficial or harmful effects in practice. Nonetheless, it appears to me plausible that a gradual introduction of a higher proportion of fats, at least up to the modest levels in the Mediterranean diet, over a more sustained period might produce promote preferential utilization of fat during exercise in a manner that translates into improved endurance performance.
In light of the evidence that glycogen depletion during training can enhance training effects, I consider that during normal training, consumption of carbohydrates is potentially counter-productive, in most instances. Exceptions might include high intensity sessions; very prolonged sessions; or for the purpose of testing the planned strategy for race day in the final few long runs of marathon/ultra-marathon preparation. However, the need to start an endurance event with glycogen stores well stocked suggests that at least a brief period of carbohydrate loading, and ingestion of carbohydrates during long events, is highly desirable.
In summary, I consider that the emerging evidence provides strong support for the proposal that the optimum nutrition for most endurance athletes is a Mediterranean diet, but with carbohydrate loading immediately prior to long races.