Fuel metabolism during ultra-endurance exercise

Pflügers Arch – Eur J Physiol (1998) 436:211–219

© Springer-Verlag 1998

O R I G I NA L A RT I C L E

&roles:H.G. Laurie Rauch · John A. Hawley Timothy D. Noakes · Steven C. Dennis

Fuel metabolism during ultra-endurance exercise

&misc:Received: 19 September 1997 / Received after revision: 15 December 1997 / Accepted: 22 January 1998

&p.1:Abstract Cyclists either ingested 300 ml 100 g/l U[14C] glucose solution every 30 min during 6 h rides at 55% of VO2max (n=6) or they consumed unlabelled glucose and were infused with U-[14C] lactate (n=5). Maintenance of euglycaemia limited rises in circulating free fatty acids, noradrenaline and adrenaline concentrations to 0.9±0.1 mM, 27±4 nM and 2.0±0.5 nM, respectively, and sustained the oxidation of glucose and lactate. As muscle glycogen oxidation declined from 100±13 to 71±9 µmol/min/kg in the last 3 h of exercise, glucose and lactate oxidation and interconversion rates remained at approximately 60 and 50 and at about 4 and 5 µmol/min/kg, respectively. Continued high rates of carbohydrate oxidation led to a total oxidation of around 270 g glucose, 130 g plasma lactate and 530 g muscle glycogen. Oxidation of some 530 g of muscle glycogen far exceeded the predicted (about 250 g) initial glycogen content of the active muscles and suggested that there must have been a considerable diffusion of unlabelled lactate from glycogen breakdown in inactive muscle fibres to adjacent active muscle fibres via the interstitial fluid that did not equilibrate with 14C lactate in the circulation. &kwd:Key words Glucose and lactate oxidation and interconversion · Carbohydrate utilisation · Fat oxidation&bdy:

Introduction It is now thought that carbohydrate (CHO) loading before and CHO ingestion during moderate-intensity . exercise at 70–75% of maximum oxygen uptake (VO2,max) both increase time to exhaustion. Depending on whether . the subject has eaten recently, fatigue at ≥70% of VO2,max coincides with either a reduced conversion of liver glyH.G.L. Rauch · J.A. Hawley · T.D. Noakes · S.C. Dennis ( ✉) MRC/UCT Bioenergetics of Exercise Research Unit, University of Cape Town Medical School, Sports Science Institute of South Africa, PO Box 115, Newlands, 7725, South Africa&/fn-block:

cogen to plasma glucose, leading to hypoglycaemia [11] or, in non-fasted subjects, a critically low (22±4 mmol/ kg wet wt.) working muscle glycogen content [ 3]. CHO ingestion may also be advisable during pro. longed low-intensity exercise at 55% VO2,max. We recently . found that two out of six subjects who cycled at 55% VO2,max for 3 h without CHO ingestion fatigued after 2.5 h [26]. Both those subjects stopped cycling when their rates of fat oxidation no longer increased to compensate for their declining rates of CHO oxidation. These data suggest that not all athletes may be able to sustain low-intensity exercise by an increasing oxidation of fat [15]. There is also the question of whether athletes who perform ultra-endurance events at competition pace for 4 h or more become increasingly reliant on fat oxidation towards the end of exercise [10, 15–17, 32, 36]. To our knowledge, there have been few studies of fuel oxidation in ultra-endurance exercise at the relatively high (>200 W) absolute work rates that top athletes are capable of sustaining for such durations. Most studies of very prolonged exercise have used protocols in which the subjects were either walking [36] or just total energy intake and expenditure was recorded [28]. Only Stein et al. [32] examined substrate utilisation in subjects. who cycled for 3 h and then ran for 5 h at some 53% of VO2,max. However, in those subjects, muscle glycogen oxidation was negligible after about 4 h of exercise and their approximately 0.4 g/min rate of CHO oxidation was not sufficient to sustain maximum, about 1 g/min rates of plasma glucose oxidation [19]. Accordingly, one aim of this study was to examine the oxidation of ingested U-[14C] glucose during 6 h of ultra-endurance exercise at competition pace, when sufficient CHO was ingested to maintain euglycaemia and limit rises in circulating catecholamine and free fatty acid concentrations. An additional purpose of this study was to investigate the influence of rising plasma catecholamine concentrations on the turnover and oxidation of intravenously infused U-[14C] lactate. O’Brian et al. [25] have noted that the oxidation of 600–700 g CHO

212

during marathon runs without CHO ingestion exceeded the estimated stores of some 380 g of glycogen in the approximately 15 kg of muscle estimated to be active in running and the some 100 g of glycogen in liver. They suspected that there might have been progressive redistribution of CHO, via lactate, from less active to more active muscle fibres.

Materials and methods Eleven competitive male endurance cyclists were recruited for this study, which was approved by the Research and Ethics Committee of the Faculty of Medicine, University of Cape Town. As radio-labelled tracers were administered and “arterialised” venous blood was sampled, the procedures and risks were carefully explained to the subjects before they signed written consent forms. Subjects and preliminary testing To minimise each cyclist’s exposure to radio-labelled tracers, the cyclists were divided into two groups, which were matched for . age, mass and VO2,max. One group (n=6) ingested U-[14C] glucose. The other group (n=5) ingested unlabelled glucose and received a continuous intravenous infusion of U-[ 14C] lactate. Subjects in the U-[14C] glucose and the U-[14C] lactate groups were exposed to total radiation doses of about 5 mrem. Radiation doses of 500 mrem/year or 130 mrem/13 week are regarded as safe in South Africa [3,5]. The matching of the subjects in the U-[ 14C] glucose and U[14C] lactate groups is described in Table 1. Percentage body fats were estimated from the sum of biceps, triceps, subscapular and supra-iliac skin-fold measurements using standard formulae [13]. Peak sustained power outputs (Wpeak) were measured on an electrically braked cycle ergometer (Lode, Groningen, Netherlands), as described by Hawley and Noakes [18]. During the incremental cycle exercise tests, subjects wore a nose-clip and breathed through a mouthpiece connected to an Oxycon Alpha automated gas analyser (Mijnhardt, The Netherlands). Before each test, the analyser was calibrated with a Hans Rudolph 3-l syringe (Vacumed, Ventura, USA), room air and a 5% CO2: 95% N2 gas mixture. Analyser outputs. were processed by a computer which calculated ventila. tion (V. E), oxygen consumption (VO2) and carbon dioxide produc. tion (VCO2) values for each breath (in litres/min). V O values 2,max . were the average of the highest VO2 values measured over 30 s in . the final work rate. Measurements of VO2,max were used to adjust . the work rates to 55% of VO2,max in the 6 h experimental rides that were conducted a week later. Experimental trial Two days before a trial, the subjects were asked to refrain from strenuous exercise and to consume their habitual diets. The habitual diets of the subjects were similar in both groups. Dietary recalls over 2 week days and one weekend day suggested that the subjects consumed about 50% of their energy as CHO (476±35 g), about 30% as fat (112±10 g) and about 13% as protein (108±99). Energy intakes were some 14 mJ/day. On the day of the trial, subjects came to the laboratory 2–3 h after eating a standardised breakfast containing 100–120 g CHO. A reasonably substantial pre-trial breakfast was designed to mimic the practices of most ultra-endurance athletes before prolonged exercise. Shortly after the subject’s arrival in the laboratory, an 18-G teflon cannula (Jelco, Johnson and Johnson, Halfway House, South Africa) was inserted into the left forearm antecubital vein and connected to a 3-way stopcock (Uniflex, Mallinckrodt, Hennef-Seig, Germany). The 18-G cannula was covered with a heating pad to dilate the vein and was used for the periodic sampling of “arterialised” venous blood [23]. At

the same time, a 20-G cannula was inserted into the right forearm antecubital vein of five of the subjects and connected to a calibrated auto syringe (Travenol, Hookset, N.H.,USA). The 20-G cannula was used for a continuous infusion of a sterile, pyrogen-free saline solution containing U-[14C] lactate (5.6 GBq/mmol) tracer (Amersham, Bucks, UK), throughout the 6-h ride. U-[14C]-lactate was infused at a rate of approximately 300 kBq/h (or some 8.5 µCi/h), and enabled us to estimate rates of plasma lactate turnover and oxidation, as described later. During the rides, the cyclists drank 300 ml flavoured 100 g/l glucose solution every 30 min. This amount of glucose was designed to match the peak (about 1 g/min) rates of ingested CHO oxidation during prolonged exercise [19]. In those trials where the subjects were not infused with U-[ 14C]-lactate, the glucose solution ingested over the last 4 h of exercise contained some 1.5 MBq (or about 40 µCi) of U-[14C]-glucose (11 GBq/mmol) tracer (Amersham). This tracer was used to calculate the rates of plasma glucose and ingested glucose oxidation from measurements of specific radioactivities and gas exchange. Assays of circulating metabolite and hormone concentrations Before the start of exercise and at 30 min intervals during the rides, “arterialised” venous blood samples (11 ml) were drawn for measurements of circulating metabolite and hormone concentrations. Aliquots of 3, 3, 1 and 4 ml of each blood sample were placed into four successive ice cold tubes containing either (i) potassium oxalate and sodium fluoride, (ii) ethylenediamine tetraacetate (EDTA), (iii) EDTA plus 100 µl aprotinin (Midrain, Novo Nordisk, Johannesburg, South Africa) or (iv) lithium heparin. The tubes were then immediately centrifuged at 2,500 g for 10 min at 4°C and the plasma stored at –80°C for later measurements of (i) glucose and lactate concentrations and specific radioactivities, (ii) free (non-esterified) fatty acid (FFA) and insulin concentrations, (iii) glucagon concentrations and (iv) noradrenaline and adrenaline concentrations, respectively. Plasma glucose concentrations were determined in a glucose analyser (Glucose Analyser 2, Beckman, Fullerton, Calif., USA). Circulating lactate and FFA concentrations were measured with spectrophotometric (Model 35, Beckman) enzymatic assays using commercially available kits (Lactate PAP, bioMérieux, Lyon, France; FFA half-micro test; Boehringer Mannheim, Mannheim, Germany). Plasma insulin and glucagon concentrations were determined by Coat-A-Count Insulin and Double Antibody Glucagon radioimmunoassays (both Diagnostic Products, Los Angeles, Calif., USA). Plasma adrenaline and noradrenaline concentrations were measured by high-pressure liquid chromatography with a Waters 460 electrochemical detector (Waters, Mass., USA). First, a dehydrobenzylamine internal standard in HCl (0.1 mM) was added to the plasma sample (1.5 ml) to give a final dehydrobenzylamine concentration of 10 pg/ml and, then, 0.4 ml tris(hydroxymethyl)aminomethane (TRIS)-HCl buffer, pH 8.7 (2.0 M) and 10 mg acid-washed aluminium hydroxide were added to the mixture. The mixture was then shaken for 15 min, after which the supernatant was aspirated from the precipitate. The precipitate was then washed with 3×1 ml Tris-HCl buffer, pH 8.1 (20 mM) before the catecholamines were eluted into 0.1 ml of glacial acetic acid (1% v/v ). Glacial acetic acid extracts were centrifuged at 2,000 g for 15 min and the supernatants stored at –20°C for ≤2 days prior to chromatography. Aliquots (15 µl) of the supernatants were introduced with a Waters 712 Wisp injector onto a 3.9×150 mm (particle size 5 µl) Resolve C18-Novapak column (Waters) and eluted at a flow rate of 1.0 ml/min with a mixture of sodium acetate (50 mM), citric acid (20 mM), sodium-1-octane-sulphonate (3.75 mM), di-n-butylamine (1.0 mM) and NaEDTA (0.135 mM) in methanol (95% v/v), pH 4.3, using a Waters 510 pump. . Measurements of expired V14CO2 specific radioactivities . . VO2 and VCO2 were also measured at 30-min intervals, over 5-min periods, throughout the trial. During those periods, a sample of ex-

213 pired air was passed through a solution containing 1 ml hyamine hydroxide in methanol, 1 ml 96% ethanol and a drop of 1% phenolphthalein indicator. Expired air was bubbled through this mixture for 2–3 min until the phenolphthalein turned from pink to clear, at which point 1 mM of CO2 had been absorbed [3, 5]. Liquid scintillation cocktail (10 ml, Ready Gel, Beckmann) was then added to the solution and 14CO2 disintegrations per min (dpm) per millimole specific radioactivity was measured in a liquid scintillation counter (Packard 1500 Tri-Carb, Downers Grove III., USA). All counts were corrected for differences in quench and background radioactivity. Measurements of plasma glucose and lactate specific radioactivities Corresponding plasma glucose and lactate specific radioactivities were measured in aliquots of the plasma collected for determinations of circulating glucose and lactate concentrations. To deproteinise the sample and to drive off H14CO3– as 14CO2, 1 ml plasma was acidified with 70 µl HClO4 (3.5 M). The sample was then centrifuged at 2,500 g for 20 min at 4°C and the protein-free supernatant removed and stored on ice. The precipitate was then resuspended in 2×0.76 ml of HClO4 (0.13 M), re-centrifuged on two further occasions and the supernatants were added to those previously saved. The pH of the combined supernatants was then returned to 7–8 by the addition of about 200 µl K2CO3 (3 M) in TRIS-HCl, pH 8 buffer (10 mM). The sample was then centrifuged as before to remove the precipitated KClO 4 and the supernatant passed through an anion exchange column (Extra-Sep RC SAX, Chromatography Research Supplies, Addison, III., USA), that had been conditioned with 20 ml ethanol followed by 20 ml distilled water adjusted to pH of about 8 by traces of NaOH. Glucose was eluted in the void volume with 3×1 ml of distilled water adjusted to a pH of about 8 and lactate was subsequently eluted with 2×1 ml CaCl2 (1 M) adjusted to a pH of about 2 with HCl. Subsequently, the eluates were evaporated to near dryness (about 0.3 ml) at 60°C over some 20 h before first distilled water (1 ml) and then Ready Gel (Beckman) liquid scintillation cocktail (15 ml) was added for radioactive counting, as described previously. Typically, radiolabelled glucose and lactate recoveries were both 85–95%. Estimates of plasma lactate turnover Changes in plasma U-[14C] lactate specific radioactivities were used to estimate the rates of plasma lactate turnover, using the equation of Steele [31]: Lac Ra=[I–(V×Lac×dSA/dt)]/SA In this equation, Lac Ra is the rate of appearance of un-labelled lactate in the circulation (in micromoles per minute per kilogram body mass; I is the infusion rate of U-[14C] lactate tracer in disintegrations per minute per kilogram); V is the predicted 100 ml/kg non-steady state lactate distribution volume [29]; Lac is the mean lactate concentration in consecutive plasma samples (in micromoles per millilitre) dSA/dt is the change in plasma 14C lactate specific (radio) activity over time (in disintegrations per minute per micromole per minute) and SA is the mean 14C lactate specific radio-activity (in disintegrations per minute per micromole) in successive plasma samples. Because U-[14C] lactate infusions were delayed until the start of the 6-h cycle rides to limit the subjects’ exposure to radioactivity, Lac Ra values were only measured after 2 h of exercise, when constant plasma 14C lactate SA values indicated that tracer equilibration was complete. Potential errors in measurements of lactate turnover arising from sites of infusion and sampling and from possible isotope equilibration between lactate, pyruvate and alanine are discussed later.

Measurements of plasma glucose and lactate oxidation Measurements of 14C glucose and 14C lactate oxidation. had to be corrected for the approximately equal contributions to VCO2 from the oxidation of secondarily labelled lactate and glucose. When U[14C] glucose was ingested, plasma lactate SA increased to about 12% of that of glucose and, when U-[14C] lactate was infused, plasma glucose SA rose to about. 4% of that of lactate. Corrections for the likely contributions to V14CO2 from the oxidation of secondarily labelled lactate and glucose were achieved by including their SA values in the estimates of directly 14C-labelled glucose or lactate oxidation, as shown below: . Rox=SA CO2/(SA glu+SA lac)×VCO2 In this equation, Rox is the rate of oxidation of glucose or lactate (in micromoles per minute .per kilogram; SA CO 2, SA glu and SA 14C lactate (in lac are the SA values for V14CO2, 14C glucose and . disintegrations per minute per millimole) and VCO2 is in micromoles per minute per kilogram. Since the conversion of one molecule of U-[14C] glucose to six molecules of 14CO2 decreases the . SA in disintegrations per minute per millimole by 6, the VCO2 values did not need to be divided by 6 to allow for the six CO 2 molecules arising from the oxidation of one glucose molecule. For the same reason, no allowance was needed for the three CO2 molecules arising from the oxidation of one lactate molecule. Before 14C glucose and 14C lactate oxidation were measured, the equilibration of 14CO2 with the body HCO3–/CO2 pool was presumed to be essentially complete (see Discussion). Interconversions of glucose and lactate Rates of plasma glucose to lactate (glu→2 lac) and plasma lactate to glucose (2 lac→glu) conversions (in micromoles per minute per kilogram) were estimated using the equations of Kreisberg et al. [21]: glu→2 lac=lac Ra×0.5×SA lac/SA glu 2 lac→glu=(glu Rox×2.0×SA glu/SA lac)×1.8 where lac Ra is the rate of lactate appearance in the plasma (in micromoles per minute per kilogram body mass), 0.5 is the glucoseto-lactate interconversion ratio and SA lac and SA glu are the plasma lactate and glucose 14C SA values in disintegrations per minute per millimole. In the second equation, the rate of glucose oxidation (glu Rox) is presumed to be similar to the rate of glucose turnover (in micromoles per minute per kilogram) [3, 5], 2.0 is the lactate-to-glucose interconversion ratio and SA glu and Sa lac are the plasma glucose and lactate 14C SA values (disintegrations per minute per millimole). The second equation is multiplied by 1.8 to correct for an observed approximately 45% loss of 14C from oxaloacetate to the tricarboxylate cycle during gluconeogenesis at rest [9]. Calculations of total CHO and fat oxidation Rates of CHO. and fat oxidation (CHO Rox and fat Rox) were cal. culated from VO2 and VCO2 (in litres per minute) values using the following equations of Frayn [14]: . . CHO Rox=4.55 V.CO2–3.21 VO2 . Fat Rox=1.67 (VO2–VCO2) In these equations, respiratory exchange ratios (RER) were assumed to be from predominantly non-protein sources. CHO Rox and fat Rox in grams per minute were converted to micromoles C 6 per minute per kilogram and micromoles C16:0 per minute per kilogram by dividing the values by 180 and 256 (molecular weights of glucose and palmitate), respectively.

U-14 C glucose (n=6)

U-14 C lactate (n=5)

(A) Characteristics Mass (kg) Body fat (%) Wpeak (W) HR . max (beats/min) VO2,max (l/min)

77±4 13±1 411±13 191±2 5.1±0.2

79±4 14±2 425±15 194±6 5.3±0.2

(B) Trial work rates % of Wpeak % of HR . max % of VO2,max

48.4±0.3 73±1 55±1

48.7±0.4 73±1 54±1

&/tbl.:

3.02±0.07 0.86±0.01 154±4 15.8±0.7 2.99±0.09 0.87±0.01 151±4 15.3±0.5 3.05±0.09 0.87±0.01 148±3 14.8±0.6 3.01±0.09 0.87±0.01 147±3 13.8±0.4 3.00±0.09 0.88±0.01 144±4 13.5±0.3

360 330 300 270 240 210

2.93±0.08 0.88±0.01 143±3 13.0±0.3 2.95±0.08 0.88±0.01 142±2 12.4±0.4 2.89±0.09 0.90±0.01 138±2 11.9±0.4 2.89±0.08 0.91±0.01 135±2 11.4±0.4 &/tbl.:

Table 1 Subject characteristics and work rates during the trials. Values are means ± SEM (W.peak is peak sustained power output, HRmax maximum heart rate, VO2,max maximum oxygen consumption)&/tbl.c:&

2.85 ±0.08 0.92±0.01 134±3 11.3±0.4

Repeated ingestion of CHO maintained plasma glucose and lactate concentrations at about 5.5 and 1.0 mM re-

2.87±0.09 2.88 ±0.09 0.94 ±0.01 0.93±0.01 129 ±2 132±2 10.7 ±0.4 10.8±0.4

Circulating metabolite and hormone concentrations

. VO2 (I/min) RER HR(beats/min) RPE

During the 6-h trials, the cyclists rode at about 50% of Wpeak, which corresponds to approximately 73% of max. imum HR (HRmax) and about. 55% of VO2,max (Table 1). From 0.5 to 6 h of exercise, VO2 increased slightly from 2.9±0.1 to 3.0±0.1 l/min and RER declined steadily from 0.94±0.01 to 0.86±0.01 (P