Reasons for performing study: The quantity of glucose disposal during exercise (walk and trot) compared to rest by use of the hyperglycaemic clamp technique has not been reported previously and has relevance to nutritional requirements.
Hypothesis: Exercise (walk and trot) significantly increases glucose disposal compared to rest.
Methods: Seven healthy Dutch Warmblood mares, all in dioestrus, mean ± s.d. age 11.6 ± 2.4 years and weighing 569 ± 40 kg were fasted for 12 h prior to a hyperglycaemic clamp at rest (maintaining a steady state of the blood glucose concentration during 30 min), walk (10 min, 1.5 m/s), trot (20 min, 4.4 m/s), walk (10 min, 1.5 m/s) and rest again (maintaining a steady state during 30 min). Plasma glucose concentrations were measured every 5 min. The mean rate of glucose disposal was calculated by corrections for glucose loss via the glucose space and urine. A one-way ANOVA with post hoc Bonferroni was performed.
Results: The mean ± s.d. rate of glucose disposal was 15.0 ± 2.1 at first rest, 25.1 ± 6.2 at first walk, 37.4 ± 9.1 at trot, 33.0 ± 13.1 at second walk and 18.7 ± 4.6 µmol/kg bwt/min at second rest. Values at trot and at second walk differed significantly from values at first rest, whereas values at both rests were similar as well as at first rest and at first walk.
Conclusions: Mean rate of glucose disposal of Warmblood horses increased 2.5 times during trot compared to basal.
Potential relevance: The hyperglycaemic clamp technique is an attractive nonisotope method to assess the rate of glucose disposal in exercising horses.
Plasma glucose is an important source of energy consumed by muscles. Other sources are muscle glycogen and fatty acids. Research into glucose metabolism during exercise has been performed in different species. Quantification of endogenous glucose production and glucose uptake in dogs, man and rats during moderate exercise reveals a tight regulation of production and uptake (Jenkins et al. 1985; Vissing et al. 1988; Berger et al. 1994). Glucose uptake by muscle increases during physical exercise. Therefore, an increase in hepatic production (HGP) is required to meet this demand for blood glucose and to avoid hypoglycaemia (Coggan 1991). Berger et al. (1994), Jenkins et al. (1985) and Vissing et al. (1988) found in dogs, man and rats, respectively, that HGP increased as glucose uptake increased and, when infused i.v. with glucose, that HGP did not increase further. Only during strenuous exercise with glucose infusion did both HGP and plasma glucose increased, disrupting the balance (Jenkins et al. 1985; Vissing et al. 1988; Berger et al. 1994). In horses, however, there appears to be a mismatch in this tight regulation: even during moderate exercise plasma glucose concentrations may rise (Farris et al. 1995).
Geor et al. (2000a,b,c) were the first to report on glucose metabolism in horses during exercise in different circumstances. They utilised stable isotope labelled glucose in order to trace the uptake and found that during low intensity exercise plasma glucose concentration was almost unchanged and the increase of HGP almost equalled the glucose uptake (Geor et al. 2000b). After pre-exercise oral glucose administration, there was a more than 2-fold increase in whole body glucose uptake during exercise compared to 24 h fasted animals (Geor et al. 2000a). In horses infused with i.v. glucose during low intensity exercise the increase in plasma glucose only partially suppressed HGP and was associated with a 2-fold increase in glucose uptake and induced only a small rise in plasma glucose (Geor et al. 2000c). The difference between this and the results of Farris et al. (1995) might be because the exercise was of lower intensity.
The hyperglycaemic clamp technique is an alternative nonisotope based method to quantify the rate of glucose disposal. In typical glucose tolerance tests, the dose of glucose is fixed and the measure of tolerance is the plasma glucose concentration. In the clamp technique, the plasma glucose concentration is fixed and the glucose administered becomes the measure of tolerance. The hyperglycaemic clamp technique allows quantification of the sensitivity of beta cells to glucose and is used to increase the plasma glucose concentration acutely to a fixed hyperglycaemic plateau and maintain it at that concentration for approximately 2 h, thereby suppressing endogenous hepatic glucose production. A critical assumption made by using this technique is that during hyperglycaemia after the appropriate corrections for glucose loss via the glucose space and urine, the rate of glucose disposal is equal to the glucose infusion rate (Rijnen and van der Kolk 2003). This is due to the expected complete suppression of endogenous glucose production.
Knowledge of exercise physiology of horses depends heavily on the precise quantification of their glucose metabolism. Currently, horses are involved in very different kinds of exercise such as dressage, showjumping and endurance in which high performance is demanded of the muscles. Improvement of muscle strength, endurance or both would not only be of great interest to researchers, but also for trainers of horses. Furthermore, better refined advice for food composition and supplements could be provided.
The objective of the study reported here was to quantify glucose disposal during exercise (walk and trot) compared to rest by use of the hyperglycaemic clamp technique aiming at the availability of a nonisotope method to assess the rate of glucose disposal in exercising horses.
Materials and methods
This study was approved by the Committee on Animal Welfare of the University of Utrecht (DEC 2008III05050).
Seven Dutch Warmblood mares, aged 8–14 years (mean ± s.d.; 11.6 ± 2.4 years) were included in this study, all in dioestrus at the time of the experiment. The horses were used to frequent handling as they were used regularly in a student riding school. They were accustomed to treadmill exercise due to earlier research activities, but were otherwise untrained. Bodyweight ranged from 522–648 kg (569 ± 40 kg). The horses were housed on pasture but during the experiment were kept in individual stables for 3 days and fed hay ad libitum and 2 kg of pellets. Food was withheld for 12 h prior to the experiment to rule out glucose uptake from the small intestine. Water was supplied ad libitum. After the experiment the horses were fed hay and pellets, monitored for one night and returned to pasture.
Hyperglycaemic clamp technique
To become familiar with the method, a pilot of the experiment was performed 2 months prior to the definite experiments. Quantification of glucose disposal using this technique has been described previously (Rijnen and van der Kolk 2003). On the day before each experiment, two 13 cm MILA catheters were inserted in both jugular veins and covered with Vetrap and Leucoplast during the night. On the day of the experiment, the horses were weighed to calculate the priming dose of glucose. The horses were positioned on the treadmill (Kagra)1. An ECG device (Televet 100 version 4.0)2 was installed and heart rates were monitored during the whole experiment. Two blood samples were taken for determination of basal plasma glucose concentrations. Hereafter, the bodyweight-dependent-priming dose (BPD) of glucose (as a 50% solution) was given i.v. within 10 min. The priming dose varied from 125–156 ml (136.9 ± 9.7 ml) and was calculated by the following equation: BPD (ml) = 0.24 bwt.
After this, the continuous infusion started at a bodyweight-dependent rate (mean 0.321 ± 0.035 ml/h/kg bwt) with a calibrated infusion pump (Model Argus 414)3. Every 10 min, a venous blood sample (heparinised) was taken for immediate determination of glucose concentration in an automated analyser (ABL-605)4. Blood glucose concentrations were considered hyperglycaemic when ranging from 10–15 mmol/l. A steady-state concentration was defined as a plasma glucose concentration fluctuating within a narrow range (± 0.2 mmol/l) for at least 30 min (4 samples) and no earlier than 90 min after the start of the clamp. During the first steady-state hyperglycaemia, additional venous blood samples (heparin vacuettes) were taken for determination of plasma insulin concentration in order to allow quantification of the sensitivity of beta cells to glucose. These samples were centrifuged for 5 min at 6000 g. Plasma was separated and stored at −20°C until analysis. Insulin concentration was measured by means of a radioimmunoassay kit (Coat-A-Count TKIN2 836)5 validated for use in horses (van der Kolk et al. 1995). When this steady state (Rest 1) continued for ≥30 min, the exercise procedure was started. This comprised of 10 min walking (Walk 1), 20 min trotting (Trot) and another 10 min walking (Walk 2). Treadmill speed was chosen to obtain 30% HRmax during walking and 50% HRmax during trot. Average HRmax was determined in earlier research at the department.
During the exercise procedure we adjusted the infusion rate of glucose, in order to achieve steady concentrations of plasma glucose (10–15 mmol/l). In the limited amount of time per gait, a genuine steady state as described for resting situations could not be achieved, so minimal demand of steadiness during exercise was defined as: a plasma glucose concentration fluctuating within a narrow range (± 0.4 mmol/l) for 10 min, with a constant glucose infusion rate. During exercise, heparinised blood samples were taken every 5 min for immediate determination of glucose and lactate concentration in an automated analyser4. After the protocol, the horses rested again and the experiment ended when concentrations of blood glucose were steady for ≥30 min (at Rest 2). At the end of each experiment, urine samples were collected as a free catch sample without measuring urine volume to calculate the urinary loss of glucose.
In a separate setting under the same conditions performed 2 weeks after the experiment, all animals underwent the same exercise procedure without hyperglycaemic clamp to assess the glucose concentrations at Rest 1, after Trot and after Walk 2 for comparison. A venous blood sample (heparinised) was taken for immediate determination of glucose concentration in an automated analyser4.
During steady state of the blood glucose concentration, the glucose infusion rate must equal the glucose disposal rate, provided that endogenous glucose production is completely suppressed by hyperglycaemia. The rate of glucose disposal was computed as:
M (mmol/kg bwt/min) = INF (mmol/kg bwt/min) - UC (mmol/kg bwt/min) - SC (mmol/kg bwt/min), where M is glucose disposal rate, INF is glucose infusion rate, UC is rate of urinary glucose loss, and SC is the so-called space correction factor (DeFronzo et al. 1979). In the hyperglycaemic clamp test, the plasma glucose concentration is not maintained perfectly and a correction must be made. The space correction factor adjusts for glucose that has been added or removed from the glucose space i.e. extracellular volume. The plasma glucose concentrations at the beginning (G1) and end (G2) of each 10 min period are considered. The space correction is calculated as: SC (mmol/kg bwt/min) = (G2 − G1) × 0.019 (DeFronzo et al. 1979; Rijnen and van der Kolk 2003).
With reference to the hyperglycaemic clamp test, a small correction for urinary glucose loss has to be made. To calculate the amount of glucose loss via urine during this period, the urine glucose concentration was determined in urine collected within 15 min after ending the hyperglycaemic clamp test. It is assumed that, because a square wave of hyperglycaemia has been created, urinary glucose losses are distributed equally over the experimental period (DeFronzo et al. 1979). To calculate the rate of urinary glucose loss, the published model (DeFronzo et al. 1979; Rijnen and van der Kolk 2003) was adjusted by estimating urine production as 20 ml/kg bwt/24 h (approx 10 l/24 h). The formula for rate of urinary glucose loss was: UC (mmol/kg bwt/min) = urinary glucose concentration (mmol/l): (bwt × 144) (DeFronzo et al. 1979; Rijnen and van der Kolk 2003).
The plasma insulin concentration (pmol/l) was determined during the steady state of the blood glucose concentration at Rest 1 only. The mean plasma insulin concentrations during the steady state are called the I value; in the hyperglycaemic clamp test, this is a measure of the beta-cell response to glucose (DeFronzo et al. 1979). During hyperglycaemia, glucose uptake can be enhanced and as a consequence the glucose disposal rate can overestimate the amount of insulin-mediated glucose uptake. This can be corrected by computing the ratio of the glucose disposal rate to the plasma insulin concentration.
Unfortunately, insulin was only measured at the start of the experiment and during the first steady state due to financial restraints.
Assays and data analysis
Values for M, heart rate and lactate concentration as well as glucose concentrations from the separate setting were statistically analysed over time by means of a one-way ANOVA with post hoc Bonferroni testing performed with SPSS 16.0 for Windows6. Differences were considered significant at values of P<0.05.
The mean basal heparinised blood glucose concentrations ranged from 4.8–5.5 mmol/l (5.1 ± 0.27 mmol/l) after food was withheld for 12 h.
Treadmill speed was 1.5 ± 0.11 m/s during Walk 1, 4.4 ± 0.23 m/s during Trot and 1.5 ± 0.09 m/s during Walk 2.
Heart rates are shown in Figure 1. Heart rates at Rest 1 and Rest 2 were not significantly different, nor for Walk 1 and Walk 2. All other heart rates were significantly different (P<0.05).
Figure 2 shows that blood lactate concentrations were below the aerobic lactate threshold concentration of 4 mmol/l during the experiment without significant changes over time.
The mean rate of glucose disposal (M) was 15.0 ± 2.1 at first rest, 25.1 ± 6.2 at first walk, 37.4 ± 9.1 at trot, 33.0 ± 13.1 at second walk and 18.7 ± 4.6 µmol/kg bwt/min at second rest. Values at trot and at second walk differed significantly from values at first rest, whereas values at both rests were similar as well as at first rest and at first walk (see also Fig 3). During the steady state at rest, mean I value was 225.3 ± 41.8 pmol/l and mean M/I ratio was 0.100 ± 0.024.
At Rest 1, the average blood glucose concentration achieved during steady state was 13.3 ± 1.1 mmol/l. For Walk 1, Trot, Walk 2 and Rest 2 these values were 13.1 ± 1.2, 12.6 ± 1.3, 11.8 ± 1.1 and 11.3 ± 1.2 mmol/l, respectively. The mean time to reach a steady-state blood glucose concentration above hyperglycaemic threshold (10–15 mmol/l) was 116 ± 21 min (SC = −0.00054 ± 0.00085) at Rest 1 and 247 ± 33 min (SC = −0.00054 ± 0.00068) following the beginning of Rest 2. During walking and trotting exercise the ‘minimal demand of steadiness’ during exercise was met. The glucose concentrations were 10–15 mmol/l and revealed a s.d. of 0.18 ± 0.19 mmol/l. The glucose infusion rates ranged from 0.28–0.38 ml/h/kg bwt (0.32 ± 0.03 ml/h/kg bwt) during different gaits. Urinary glucose concentrations post exercise ranged from 8.1–67.9 mmol/l (24.2 ± 22.3 mmol/l).
In the separate sham setting, glucose concentration in Rest 1 was 4.90 ± 0.28 mmol/l, after Trot 4.77 ± 0.16 mmol/l and after Walk 2 4.78 ± 0.22 mmol/l without significant differences.
Glucose as a substrate for ATP production might come from various sources: glucose can be taken up in the small intestine, it can be produced by the liver via gluconeogenesis or by glycogenolysis in the liver and muscles. Muscle glycogen is the most important resource for glucose in the exercise phase. A relatively small amount of glucose is withdrawn from the blood (Hinchcliff and Geor 2004). In a state of hyperglycaemia, more glucose is taken from the blood circulation but the absolute amount of muscle glycogen consumed remains constant (Farris et al. 1995; Geor et al. 2000c). Under conditions of low intensity exercise and hyperglycaemia (and hyperinsulinaemia) it is unlikely that muscle glycogen is synthesised because in horses recovery of glycogen is typically slow, although hyperinsulinaemia does contribute to the activation of glycogen synthase as well as the transport of glucose into muscle by the insulin-activated glucose transporter GLUT-4 (Poso et al. 1995; van Dam et al. 2004). Unfortunately, insulin concentrations were not available during and after exercise.
The contribution of plasma glucose to total carbohydrate oxidation (plasma glucose, muscle glycogen and lactate) is approximately 26% during induced hyperglycaemia and 17% without exogenous glucose infusion (Geor et al. 2000c).
The assessed rate of glucose disposal (M) in the current study revealed differences between gaits. Mean rate of glucose disposal of Warmblood horses increased 2.5 times during Trot compared to the Rest 1 phase. In accordance with previous findings, trotting was associated with a significant increase in the rate of glucose disposal presumed predominantly due to muscular work load. In agreement with the current study, exercise in Standardbreds and Thoroughbreds resulted in an approximately 2-fold increase in whole body glucose uptake and the estimated contribution of plasma glucose to energy expenditure. (Geor et al. 2000c). In that study, rates of glucose uptake were compared between a group of Standardbreds and Thoroughbreds with glucose infusion and a control group during exercise. It had been assumed that there would be a mismatch between HGP and glucose uptake in the control group, i.e. a hyperglycaemia would occur during exercise. In contrast, a close match between HGP and glucose uptake during low-intensity exercise was found. In the control group, HGP was only partially suppressed during the whole exercise procedure. However, HGP increased mostly after 45 min of exercise. The assumption of HGP suppression by exogenous glucose infusion can be defended, because measurements in the current study were done during 20 min of low-intensity exercise. The separate sham setting without the hyperglycaemic clamp showed no significant differences in glucose concentration over time. This is in accordance with the study of Geor for horses in low-intensity exercise (Geor et al. 2000a).
Another important fact is that the glucose uptake is enhanced by hyperglycaemia. In a state of (induced) hyperglycaemia, muscles rely on glucose more than during normal blood glucose concentrations (Geor et al. 2000a). During hyperglycaemia, hyperinsulinaemia also occurs and this facilitates glucose uptake in the muscles and the whole body. In our study, the hyperglycaemia may be unnaturally promoting glucose disposal rather than fat metabolism. Therefore, it must be clear that the values reported in the current study are under these hyperglycaemic conditions and may be different under euglycaemic or natural conditions.
In conclusion, the hyperglycaemic clamp technique is an attractive nonisotope method to assess the rate of glucose disposal in exercising hyperglycaemic horses.
Further recommendations for research are quantification of glucose disposal by use of the hyperglycaemic clamp technique during extended moderate exercise and/or during high intensity anaerobic exercise.
Conflicts of interest
The authors have declared no potential conflicts.
1 Graber A, Fahrwagen, Switzerland.
2 Veterinary Telemetric ECG Systems, Frankfurt am Main, Germany.
3 Argus Medical AG, Heimberg, Switzerland.
4 Radiometer Copenhagen,Westlake, Ohio, USA.
5 Diagnostic Products Corp, Los Angeles, California, USA.