Dietary sugars

Common sugars in the human diet include the monosaccharides: glucose, fructose and galactose; and the disaccharides: sucrose (fructose–glucose), lactose (galactose–glucose) and maltose (glucose–glucose)⁽Reference Gonzalez, Fuchs and Betts⁹⁾. Dietary sugars can be consumed from a variety of food sources, which can influence resultant health effects. The WHO classifies sugars into those which are intrinsic (e.g. incorporated within the structure of intact fruit and vegetables or lactose/galactose from milk) v. free sugars⁽¹¹⁾. Free sugars are defined by the WHO as monosaccharides and disaccharides added to foods and beverages by the manufacturer, cook or consumer, along with sugars naturally present in honey, syrups, fruit juices and fruit juice concentrates. This classification system is useful for distinguishing between food sources of sugar that are energy dense (i.e. free sugars) and thus may contribute to weight gain^{(11, 12)}. However, this classification system does not specifically distinguish between ingestion of glucose-containing sugars and fructose- or galactose-containing sugars in relation to health, nor does it consider the physical activity status of the individual. This is interesting considering the fundamental differences in the intestinal absorption and hepatic metabolism of glucose, fructose and galactose, and how this metabolism is altered during exercise⁽Reference Gonzalez and Betts^8–Reference Tappy and Lê¹⁰⁾.

Before describing the hepatic metabolism of carbohydrates and sugars, it is important to clarify two points. First, the hydrolysis of glucose polymers such as maltodextrin and starch are typically not rate-limiting to intestinal absorption⁽Reference Hawley, Dennis and Laidler¹³⁾, and therefore (at least with regard to hepatic metabolism) free glucose, maltose, maltodextrin and starch can all be considered physiologically similar stimuli. Secondly, in typical human diets, free glucose is rarely consumed alone, but rather is usually consumed alongside fructose or lactose, or is consumed in polymer (non-sugar) form, such as maltodextrin and starch. Accordingly, whilst this review will refer to the specific types of sugars utilised in studies (e.g. glucose only v. fructose–glucose mixtures), it can be viewed that a non-fructose or non-galactose condition (such as glucose or maltodextrin ingestion) is physiologically representative of non-sugar intake (i.e. maltodextrin, starch, etc.), whereas fructose–glucose and galactose–glucose co-ingestion represent the physiological responses to typical sugar intake.

Hepatic metabolism of sugars at rest

Glucose and galactose are primarily absorbed across the intestinal lumen via the transport protein sodium-dependent GLUT2⁽Reference Daniel and Zietek¹⁴⁾, whereas fructose is primarily absorbed via GLUT5⁽Reference Daniel and Zietek¹⁴⁾. Once absorbed, these sugars are then metabolised very differently. Glucose is preferentially metabolised by extra-splanchnic tissues such as skeletal muscle, the brain and cardiac muscle⁽Reference Kelley, Mitrakou and Marsh^15, Reference Ferrannini, Bjorkman and Reichard¹⁶⁾, whilst fructose and galactose are primarily metabolised by the liver and to a lesser extent small bowel enterocytes and proximal renal tubules⁽Reference Tappy and Lê^10, Reference Williams¹⁷⁾. Peripheral tissues such as muscle are therefore only exposed to relatively small amounts of fructose and galactose⁽Reference Ercan, Nuttall and Gannon^18, Reference Chong, Fielding and Frayn¹⁹⁾.

Compared with galactose, the metabolic fate of fructose is relatively well characterised. At rest, the liver rapidly takes up and metabolises fructose into fructose-1-phosphate (P) via fructokinase (K _m for fructose: about 0·5 mm and V _max estimated at about 3 mm/min per g human liver)⁽Reference Heinz, Lamprecht and Kirsch^20–Reference Hers and Kusaka²²⁾. Fructose-1-P is then metabolised into triose-P (C₃ substrates) via aldolase B⁽Reference Mayes²³⁾. At rest, the majority of fructose-derived triose-P is converted via gluconeogenesis into glucose (about 50 %) and glycogen (about 15–25 %), but some of triose-P can be metabolised into pyruvate, then either oxidised within the liver or converted into lactate (about 25 %), which enters the systemic circulation and can increase blood lactate concentrations⁽Reference Tappy and Lê^10, Reference Björkman, Gunnarsson and Hagström²⁴⁾. One other fate of ingested fructose is the conversion into fatty acids via the process known as de novo lipogenesis⁽Reference Bar-On and Stein²⁵⁾. It has been suggested that lactate is the primary precursor to hepatic de novo lipogenesis with fructose intake⁽Reference Carmona and Freedland²⁶⁾, but the proportion of fructose that is ultimately converted into lipid is estimated at <1 % and therefore represents a quantitatively minor pathway of disposal⁽Reference Chong, Fielding and Frayn¹⁹⁾. Nevertheless, the effects of ingested fructose on de novo lipogenesis may still be important for metabolic health⁽Reference Sanders, Acharjee and Walker²⁷⁾.

Quantitative estimates of the metabolic fate of galactose in human subjects are scarce. It has been suggested that the primary pathway for human galactose metabolism is the Leloir pathway, the enzymes of which show highest activity in the liver⁽Reference Williams¹⁷⁾. This pathway involves four main steps: (1) phosphorylation of galactose by galactokinase (K _m for galactose: about 0·9 mm and V _max estimated at about 1·4 mm/min per g human liver⁽Reference Timson and Reece^28, Reference Sørensen, Mikkelsen and Frisch²⁹⁾) to yield galactose-1-P; (2) conversion of galactose-1-P and uridine diphosphate (UDP) glucose to UDP galactose and glucose-1-P by galactose-1-P uridyltransferase; (3) conversion of UDP-galactose to UDP-glucose by UDP-galactose-4-epimerase; and (4) conversion of UDP-glucose and diphosphate to glucose-1-P and uridine triphosphate by UDP-glucose pyrophosphorylase⁽Reference Williams¹⁷⁾. Of note, is an alternative pathway for step 2, known as the Isselbacher pathway, whereby galactose-1-P and uridine triphosphate are converted to UDP-galactose and diphosphate by the enzymes UDP-galactose pyrophosphorylase and UDP-glucose pyrophosphorylase⁽Reference Williams¹⁷⁾. Some tracer studies have attempted to determine the metabolic fate of oral galactose in human subjects, estimating that during ingestion of galactose at a rate of 33 µmol/kg per min (about 135 g over 360 min), the splanchnic uptake of galactose is saturable at about 15 µmol/kg per min⁽Reference Sunehag and Haymond^30, Reference Coss-Bu, Sunehag and Haymond³¹⁾. At this rate of ingestion, it is estimated that about 30–60 % of the ingested galactose is converted into glucose⁽Reference Sunehag and Haymond³⁰⁾, mostly via the direct conversation of hexose to glucose (about 67 %), with some converted via the indirect (hexose to C₃ substrates to glucose) pathway (about 33 %)⁽Reference Coss-Bu, Sunehag and Haymond³¹⁾. Ultimately, the metabolic fate of ingested galactose in human subjects therefore remains incompletely understood, although it has been speculated that liver glycogen synthesis is a major route⁽Reference Niewoehner, Neil and Martin^32, Reference Décombaz, Jentjens and Ith³³⁾ (Fig. 1).

Fig. 1. (Colour online) Major metabolic pathways of glucose, fructose and galactose in the human liver. TCA, tricarboxylic acid cycle; P, phosphate; UDP, uridine diphosphate. Based on^{(Reference Tappy and Lê10, Reference Williams17, Reference Mayes23–Reference Carmona and Freedland26, Reference Sunehag and Haymond30–Reference Décombaz, Jentjens and Ith33)}.

Hepatic metabolism of carbohydrates with exercise

Exercise increases energy expenditure, which is predominantly met during prolonged (>30 min) exercise by increases in both carbohydrate and fat oxidation compared with the resting state⁽Reference van Loon, Greenhaff and Constantin-Teodosiu³⁴⁾. The relative contributions of carbohydrate v. fat to exercise metabolism are influenced by the intensity and mode of exercise⁽Reference Achten, Venables and Jeukendrup³⁵⁾, preceding nutritional status⁽Reference Chen, Travers and Walhin^36–Reference Edinburgh, Hengist and Smith³⁸⁾, endurance training status⁽Reference van Loon, Jeukendrup and Saris³⁹⁾ and biological sex⁽Reference Fletcher, Eves and Glover⁴⁰⁾. Specifically, higher carbohydrate oxidation rates are seen with cycling v. running⁽Reference Achten, Venables and Jeukendrup³⁵⁾, higher v. lower exercise intensity⁽Reference van Loon, Greenhaff and Constantin-Teodosiu³⁴⁾, prior carbohydrate feeding v. fasting⁽Reference Gonzalez, Veasey and Rumbold³⁷⁾, in individuals who are less v. more endurance-trained⁽Reference van Loon, Jeukendrup and Saris³⁹⁾ and amongst men v. women⁽Reference Fletcher, Eves and Glover⁴⁰⁾. Of these predictive factors, the intensity of exercise seems to be the most potent in determining carbohydrate and fat utilisation⁽Reference van Loon, Greenhaff and Constantin-Teodosiu^34, Reference Romijn, Coyle and Sidossis⁴¹⁾. Even in highly trained athletes studied in the overnight fasted state, carbohydrates are the predominant fuel source during moderate-to-high intensity (>50 % peak oxygen consumption) exercise⁽Reference van Loon, Greenhaff and Constantin-Teodosiu³⁴⁾. Exercise is therefore a potent modulator of carbohydrate metabolism, with implications for the fate of ingested carbohydrate.

The primary sources of carbohydrate supporting exercise metabolism are muscle glycogen, and circulating glucose and lactate⁽Reference van Loon, Greenhaff and Constantin-Teodosiu³⁴⁾. In the fasted state, almost all the circulating glucose is derived from hepatic glycogenolysis and gluconeogenesis, with minor contributions from the kidneys and intestine⁽Reference Gonzalez, Fuchs and Betts⁵⁾. When compared with the capacity to store fat, the relatively limited capacity for human subjects to store carbohydrate has implications for the ability to sustain moderate-to-high-intensity exercise⁽Reference Gonzalez, Fuchs and Betts⁵⁾. Even amongst lean individuals (about 10 % body fat), sufficient energy is stored as fat in adipose tissue to theoretically sustain moderate-to-high-intensity exercise for many weeks. However, utilising fat as a fuel has a number of limitations in the context of exercise performance. Fat is a relatively ‘slow’ fuel; the rate of ATP resynthesis with fat is at least half that when utilising muscle glycogen⁽Reference Walter, Vandenborne and Elliott^42, Reference Hultman and Harris⁴³⁾. Fat is also an inefficient fuel on an oxygen basis, requiring about 10 % more oxygen consumption for an equivalent energy yield as glucose⁽Reference Jeukendrup and Wallis⁴⁴⁾. Consequently, during high-intensity exercise where rapid ATP resynthesis is required and/or muscle oxygen availability could be limiting, there are advantages to oxidising carbohydrates over fats. Finally, recent evidence implies that glycogen is more than just a fuel and is an important signalling molecule⁽Reference Ørtenblad, Westerblad and Nielsen⁴⁵⁾. Low glycogen concentrations in the intramyofibrillar region are associated with impaired sarcoplasmic reticulum calcium release rates and excitation contraction coupling⁽Reference Ørtenblad and Nielsen⁴⁶⁾. Therefore, specific depots of glycogen appear to play important roles in both fuelling and regulating skeletal muscle contractile function, hence achieving high carbohydrate availability before and during competition is a goal for athletes competing in almost all endurance sports⁽Reference Thomas, Erdman and Burke^47, Reference Burke, van Loon and Hawley⁴⁸⁾.

Low carbohydrate (glycogen) availability in muscle and liver is strongly associated with fatigue during prolonged exercise⁽Reference Bergström, Hermansen and Hultman^49, Reference Casey, Mann and Banister⁵⁰⁾. The amount of glycogen stored in muscle and liver glycogen prior to single, or repeated bouts of exercise positively correlate with subsequent exercise capacity⁽Reference Bergström, Hermansen and Hultman^49, Reference Casey, Mann and Banister⁵⁰⁾. A number of carbohydrate-related adaptations occur in response to regular endurance training that facilitate improvements in exercise performance. Endurance-trained athletes have a greater capacity to store muscle glycogen, and therefore display an increase in overnight-fasted muscle glycogen concentrations compared with people who are less endurance trained⁽Reference Gonzalez, Fuchs and Betts^5, Reference Areta and Hopkins⁵¹⁾. This increase in basal muscle glycogen concentrations with endurance training is also exaggerated on a high-carbohydrate diet⁽Reference Areta and Hopkins⁵¹⁾, suggesting that endurance-trained athletes can better tolerate high-carbohydrate diets by appropriately storing the excess carbohydrate as muscle glycogen. Interestingly, it seems that basal liver glycogen content does not adapt with endurance training, as endurance-trained athletes tend to exhibit similar liver glycogen concentrations to non-trained controls, when measured in the overnight fasted state⁽Reference Gonzalez, Fuchs and Betts⁵⁾. Whether this is also the case in the postprandial state remains to be established.

The increase in fasting muscle (but not liver) glycogen concentrations with endurance training provides trained athletes with a larger depot of glycogen to utilise during exercise and so postpones the point at which critically low muscle glycogen concentrations initiate fatigue. In addition to the greater storage capacity, trained athletes also utilise their muscle glycogen more conservatively during exercise⁽Reference Areta and Hopkins⁵¹⁾. This sparing of glycogen with endurance training is not specific to muscle, as the rate of liver glycogen utilisation is also attenuated in endurance-trained athletes compared with controls, particularly at moderate-to-high exercise intensities⁽Reference Gonzalez, Fuchs and Betts⁵⁾. Evidence regarding whether gluconeogenesis is altered with endurance training is currently equivocal, as some studies indicate endurance training is associated with an increase in absolute rates of hepatic gluconeogenesis⁽Reference Bergman, Horning and Casazza⁵²⁾, whereas others have shown reductions in hepatic gluconeogenesis after endurance training⁽Reference Bergman, Horning and Casazza⁵²⁾. When pooling all the currently published studies that have concomitantly estimated hepatic glycogenolysis and gluconeogenesis⁽Reference Bergman, Horning and Casazza^52–Reference Webster, Noakes and Chacko⁶⁶⁾, it is clear that endurance training is associated with lower rates of glycogenolysis (Fig. 2(a), whereas any difference in gluconeogenesis with training status and/or exercise intensity is relatively small and unlikely to be quantitatively important (Fig. 2(b). Furthermore, it is apparent that hepatic glycogenolysis is the predominant source of blood glucose during exercise in an overnight fasted state, and the increase in endogenous glucose appearance with increasing exercise intensity is almost entirely met by an increase in hepatic glycogenolysis, rather than gluconeogenesis (Fig. 2).

Fig. 2. (Colour online) Hepatic glycogenolysis (a) and gluconeogenesis (b) in endurance trained individuals and untrained individuals. Each dot represents a group of participants or exercise intensity from a study, and error bars represent 95 % CI (only calculated when published data were available to permit this). The shaded areas represent the 95 % CI of the trend lines. Data are from^{(Reference Bergman, Horning and Casazza52–Reference Webster, Noakes and Chacko66)}.

Interactions between carbohydrate ingestion and exercise occur on multiple levels and in both directions. Ingesting carbohydrates during exercise can increase total carbohydrate oxidation and suppress net liver glycogen utilisation and fat oxidation⁽Reference Gonzalez, Fuchs and Smith⁶⁷⁾. Whereas even modest exercise potently re-directs the metabolic fate of orally ingested sugars. For example, 60 min cycling at 100 W performed 90 min after fructose ingestion diverts more fructose away from storage (e.g. as glycogen) and increases fructose oxidation, without altering the conversion of fructose to glucose⁽Reference Egli, Lecoultre and Cros⁶⁸⁾. This may partly explain why daily exercise can completely prevent the increase in plasma TAG concentrations seen with high fructose intakes⁽Reference Egli, Lecoultre and Theytaz⁶⁹⁾. Remarkably, the protection offered from exercise against fructose-induced hypertriglyceridemia is seen independently from changes in net energy balance⁽Reference Egli, Lecoultre and Theytaz⁶⁹⁾, yet current recommendations for the health effects of dietary sugars rarely consider the context of physical activity status.

Since low carbohydrate availability is associated with impaired exercise tolerance, athletes engaging in competitive endurance events regularly consume carbohydrates during exercise⁽Reference Saris, van Erp-Baart and Brouns⁷⁰⁾. When ingesting glucose alone, the maximal rate at which human subjects can digest, absorb and metabolise glucose is about 1 g/min⁽Reference Gonzalez, Fuchs and Betts⁹⁾, which typically only represents about 44 % of total carbohydrate oxidation during exercise at moderate intensity (about 60 % peak oxygen system) and is therefore insufficient to fully meet the carbohydrate requirements of cycling-based exercise⁽Reference Jentjens, Achten and Jeukendrup⁷¹⁾. Consequently, oral ingestion of glucose is unable to prevent muscle glycogen depletion during prolonged exercise⁽Reference Gonzalez, Fuchs and Smith⁶⁷⁾. It is thought that the primary limitation to the metabolism of orally ingested glucose lies in the splanchnic region, and intestinal absorption of glucose appears to be saturated at about 1 g/min⁽Reference Gonzalez, Fuchs and Betts⁹⁾. Ingesting glucose at rates higher than 1 g/min during exercise is therefore likely to lead to accumulation of glucose in the gut and cause gastrointestinal distress. Interestingly, combining fructose with glucose appears to accelerate the digestion, absorption and utilisation of carbohydrate, such that exogenous carbohydrate oxidation rates can reach up to about 1·7 g/min, equating to about 70 % of total carbohydrate oxidation⁽Reference Gonzalez, Fuchs and Betts^9, Reference Jentjens, Achten and Jeukendrup⁷¹⁾. Under these conditions, the relative contribution from endogenous carbohydrate sources is therefore reduced from 100 % in the fasted state, to about 30 % with very high (2·5 g/min) carbohydrate ingestion rates⁽Reference Jentjens, Achten and Jeukendrup⁷¹⁾. The primary mechanism by which fructose–glucose mixtures can increase exogenous carbohydrate oxidation over glucose alone is thought to be that intestinal fructose transport utilises a separate pathway than glucose. Specifically, whilst glucose absorption via sodium dependent GLUT-1 is saturated at about 1 g/min, fructose is primarily transported via GLUT5, thereby taking advantage of this alternative pathway and delivering more total carbohydrate to the system⁽Reference Gonzalez, Fuchs and Betts⁹⁾.

Potential implications and applications

Exercise performance

The health and performance implications of carbohydrate intake can be dependent on the specific pathways through which different dietary sugars are absorbed and metabolised. In terms of endurance performance, the accelerated digestion, absorption and utilisation of fructose–glucose mixtures above glucose only, has potential benefits with regard to sparing glycogen stores whilst minimising gastrointestinal distress during exercise⁽Reference Gonzalez, Fuchs and Betts⁹⁾. Gastrointestinal complaints are relatively common in endurance events⁽Reference Pfeiffer, Stellingwerff and Hodgson⁷²⁾, which may directly impair performance, but also limit the ability to ingest adequate nutrition to fuel the demands of the exercise. It has recently been demonstrated that the ingestion of either glucose alone, or sucrose (glucose–fructose) can prevent liver glycogen depletion during prolonged (3 h) cycling at a moderate exercise intensity (55 % V_O2max)⁽Reference Gonzalez, Fuchs and Smith⁶⁷⁾. Whilst there was no further benefit of ingesting sucrose compared with glucose with respect to net liver glycogen depletion, the prevention of liver glycogen depletion with sucrose ingestion was attained with lower ratings of both gut discomfort and perceived exertion, compared with glucose ingestion⁽Reference Gonzalez, Fuchs and Smith⁶⁷⁾. Furthermore, when carbohydrates are ingested in large amounts during exercise (>1·4 g/min), the ingestion of glucose–fructose enhances endurance performance by about 1–9 % more than when glucose is ingested alone⁽Reference Rowlands, Houltham and Musa-Veloso⁷³⁾. Conversely, galactose appears to display relatively low rates of exogenous carbohydrate oxidation during exercise (about 0·4 g/min oxidised, when ingesting about 1·2 g/min), despite an apparent potential for faster intestinal absorption of galactose compared with glucose in perfusion studies⁽Reference Holdsworth and Dawson^74, Reference Cook⁷⁵⁾. Moreover, since galactose primarily shares a common intestinal absorption pathway to glucose, combining galactose and glucose ingestion is unlikely to provide the same benefits for exogenous carbohydrate availability and endurance performance as glucose–fructose co-ingestion.

In addition to manipulating carbohydrate availability during exercise, dietary sugars can also play an important role during post-exercise recovery, particularly in multi-stage events such as the Tour de France and the Marathon des Sables, where athletes are required to perform to the best of their ability with <24 h recovery. Within these scenarios, the primary limiting factor to recovery time is glycogen storage rate⁽Reference Burke, van Loon and Hawley^48, Reference Alghannam, Gonzalez and Betts⁷⁶⁾. Even with high carbohydrate intakes, it is thought to take between 20 and 24 h to fully restore muscle glycogen concentrations after exhaustive exercise⁽Reference Coyle⁷⁷⁾. Thus, on moderate carbohydrate diet, muscle glycogen repletion can take up to 46 h⁽Reference Piehl⁷⁸⁾. Therefore, intensive nutritional strategies can be applied to optimise muscle glycogen resynthesis post-exercise. Post-exercise muscle glycogen resynthesis rates display a biphasic response, with the most rapid net synthesis seen within the first 30 min following exercise in an insulin-independent phase⁽Reference Price, Rothman and Taylor⁷⁹⁾. Following this period, muscle glycogen synthesis rates become insulin-dependent and can fall to at least half the rate of that seen within the first 30 min post-exercise⁽Reference Price, Rothman and Taylor⁷⁹⁾. Muscle glycogen resynthesis rates are maximally stimulated with carbohydrate ingestion rates of ≥1 g/kg body mass per h⁽Reference Burke, van Loon and Hawley^48, Reference Alghannam, Gonzalez and Betts⁷⁶⁾, and this ingestion rate is also associated with optimal restoration of endurance capacity during short-term (4 h) recovery periods⁽Reference Alghannam, Gonzalez and Betts⁷⁶⁾. Therefore, athletes are advised to consume carbohydrate at a rate of 1–1·2 g carbohydrate/kg body mass per h during the early stages (4 h) of recovery⁽Reference Thomas, Erdman and Burke^47, Reference Burke, van Loon and Hawley⁴⁸⁾ and, when these ingestion rates are not achievable, the addition of certain (insulinotropic) proteins, such as milk proteins, to carbohydrate can potentially increase the efficiency of muscle glycogen resynthesis⁽Reference Betts and Williams⁸⁰⁾.

Current sports nutrition guidelines for recovery do not specify whether the carbohydrate should be from a particular source of sugar (e.g. glucose v. fructose v. galactose)⁽Reference Thomas, Erdman and Burke^47, Reference Burke, van Loon and Hawley⁴⁸⁾, which is understandable given that muscle glycogen resynthesis rates do not appear to differ whether glucose or glucose–fructose mixtures are ingested⁽Reference Gonzalez, Fuchs and Betts^9, Reference Fuchs, Gonzalez and Beelen⁸¹⁾, yet overlooks the clear potential for sugars to differentially affect liver glycogen resynthesis. Indeed, when pooling all currently published data that compare glucose with glucose–fructose mixtures in crossover designs⁽Reference Casey, Mann and Banister^50, Reference Fuchs, Gonzalez and Beelen^81–Reference Trommelen, Beelen and Pinckaers⁸⁵⁾, it is apparent that post-exercise muscle glycogen resynthesis rates do not differ between glucose ingested alone v. glucose–fructose (sucrose) mixtures (Fig. 3(a)). Extrapolating these data would suggest that 22 h are required to completely re-synthesise muscle glycogen from a fully depleted state, when following current sports nutrition guidelines, regardless of the type of carbohydrate ingested (Fig. 3(a)). This indicates that intestinal absorption of carbohydrate is not rate-limiting to post-exercise muscle glycogen resynthesis. Conversely, liver glycogen resynthesis appears to be potently accelerated by glucose–fructose co-ingestion compared with glucose (polymers) alone (Fig. 3(b))⁽Reference Décombaz, Jentjens and Ith^33, Reference Casey, Mann and Banister^50, Reference Fuchs, Gonzalez and Beelen⁸¹⁾, which may be in part due to greater exogenous carbohydrate availability and/or the specific hepatic metabolism of fructose.

Fig. 3. (Colour online) Studies (specified by reference citations) that have directly compared glucose ingestion alone, with either fructose–glucose or galactose–glucose mixtures, and measure rates of muscle (a) and liver (b) glycogen repletion post-exercise. Each circle represents a timepoint within a study. Error bars represent 95 % CI, and the shaded areas represent the 95 % CI of the trend line. For complete recovery of muscle glycogen stores, 600 mm/kgDM was chosen on the basis that muscle glycogen concentrations at exhaustion is typically about 115 mm/kgDM and the maximal muscle glycogen concentrations of relatively well-trained athletes (60–70/ml/kg per min) is between 600 and 800 mm/kgDM⁽Reference Areta and Hopkins⁵¹⁾. For complete recovery of liver glycogen stores, 80 g was chosen on the basis that liver glycogen concentrations in the overnight fasted state are about 280 mm/l. Assuming a liver volume of 1·8 litre and the molar mass of a glycosyl unit being 162 g/m, this equates to 80 g glycogen⁽Reference Gonzalez, Fuchs and Betts⁵⁾.

A further interesting observation is that liver glycogen resynthesis rates also appear to show a bi-phasic, time-dependent response, albeit in the opposite direction to skeletal muscle. Within the first 2 h post-exercise, net liver glycogen resynthesis rates are about 30–50 % slower than the 3–5 h period, independent of the type of carbohydrate ingested (2 (se 2) and 5 (se 2) g/h in the 0–2 h post-exercise v. 4 (se 2) and 8 (se 2) g/h in the 2–5 h post-exercise, with glucose and sucrose ingestion, respectively)⁽Reference Fuchs, Gonzalez and Beelen⁸¹⁾. Furthermore, with high rates of carbohydrate ingestion, fructose–glucose mixtures can reduce ratings of gut discomfort during recovery from exercise, compared with glucose ingestion alone⁽Reference Fuchs, Gonzalez and Beelen⁸¹⁾. Extrapolating these data (i.e. assuming that the first 6·5 h is representative of a full 24 h period) indicates that when only glucose is ingested, complete recovery of liver glycogen stores may take about 25 h (Fig. 3(b)). However, when glucose–fructose mixtures are ingested, then liver glycogen repletion could take as little at 11 h (Fig. 3(b)). When considering that the time between ending a stage and beginning the next stage in the Tour de France can be about 15 h, the accelerated recovery of liver glycogen stores with fructose–glucose mixtures is highly meaningful from a practical standpoint.

Interestingly, fructose is not the only sugar that more rapidly replenishes liver glycogen contents following exercise than glucose alone. The addition of galactose to glucose also accelerates post-exercise liver glycogen repletion when matched for total carbohydrate intake⁽Reference Décombaz, Jentjens and Ith³³⁾, and to a similar extent as fructose–glucose ingestion (Fig. 3(b)). Since intestinal galactose–glucose absorption should theoretically be slower than fructose–glucose absorption, it is tempting to speculate that the mechanisms by which fructose and galactose enhance liver glycogen resynthesis relate to hepatic metabolism, rather than intestinal absorption. These data also raise the following question: if the Leloir pathway (direct galactose–glucose conversion) is the primary pathway of human galactose metabolism, why is the liver glycogenic response to galactose ingestion more comparable to fructose than to glucose? With regard to generating useful data for applied practice, there is a need to establish the optimal dose and mixture of sugars for rapid liver glycogen resynthesis and whether this translates into improved endurance performance. Accordingly, dose–response studies and direct comparisons of combined galactose–fructose–glucose are warranted.

Whilst the effects of fructose–glucose and galactose–glucose ingestion on glycogen resynthesis are interesting and likely to be important for athletic performance, this will remain speculative in the absence of empirical data. Fortunately, a recent study compared the recovery of exercise capacity with glucose–maltodextrin ingestion v. fructose–maltodextrin ingestion⁽Reference Maunder, Podlogar and Wallis⁸⁶⁾. Since the maltodextrin is hydrolysed, absorbed and oxidised as quickly as free glucose⁽Reference Gonzalez, Fuchs and Betts⁹⁾, it can be considered the glucose–maltodextrin is physiologically almost identical to ingesting pure glucose. Athletes were first asked to run on a treadmill at 70 % $\dot{\rm V}{\rm O}_{2\max}$ to exhaustion. Following this, the athletes ingested 90 g carbohydrate/h during a 4 h recovery period as either a glucose–maltodextrin mixture, or fructose–maltodextrin mixture. After the recovery period, the athletes ran on the treadmill again at 70 % $\dot {\rm V}{\rm O}_{2\max}$ to exhaustion. During the glucose–maltodextrin trial, the second-bout capacity for these athletes to run was 61·4 (se 9·6) min, whereas, when fructose–maltodextrin was ingested in the recovery period, these athletes ran for 81·4 (se 22·3) min representing an improvement in second-bout endurance capacity of about 30 %⁽Reference Maunder, Podlogar and Wallis⁸⁶⁾. This provides the first evidence that fructose–glucose ingestion accelerates recovery of exercise capacity. When considered in light of the consistently reported acceleration of liver glycogen recovery, it may be sensible for athletes requiring rapid recovery during multi-stage events to consume fructose–glucose mixtures rather than glucose only. In terms of applying this in practice, it could mean the use of fruit smoothies to supplement carbohydrate intake rather than the commonly held view that pasta is a preferable choice for carbohydrate loading.