Changes in arterial, mixed venous and intraerythrocytic ion concentrations during prolonged exercise



Reasons for performing study: Prolonged equine exercise can cause hypochloraemic alkalosis and hypokalaemia secondary to the loss of hypertonic sweat. Movement of ions in and out of erythrocytes during exercise may help regulate acid-base balance and changes in plasma ion concentrations. The extent to which this happens during prolonged equine exercise has not been reported.

Objectives: To measure changes in blood gases and major plasma and intraerythrocytic (iRBC) ion concentrations of horses undergoing prolonged submaximal exercise.

Methods: Six horses were trotted at ∼30% V̇O2max on a treadmill for 105 min. Arterial (a) and mixed venous (v) blood samples were collected every 15 min, and pre- and post exercise. Blood gases and plasma (pl) concentrations of sodium, potassium, chloride and protein were measured and their iRBC concentrations calculated and compared (P<0.05).

Results: PaCO2 decreased in all horses. pl[Cl-]v decreased and [HCO3-]v increased. Due to the exhalation of CO2 and chloride shifting, [HCO3-]a<[HCO3-]v, pl[Cl-]a>pl[Cl-]v and iRBC[Cl-]a<iRBC[Cl-]v. pl[K+]a and pl[K+]v both initially increased then decreased and horses were hypokalaemic post exercise. Both iRBC[Cl-]a and iRBC[Cl-]v decreased over the course of exercise but there was no change in the arteriovenous difference between them. There was no arteriovenous difference in pl[K+]. iRBC[K +]a>iRBC[K+]v. Conversely, iRBC[Na+]a<iRBC[Na+]v. pl[Na+]a<pl[Na+]v and [TP]a<[TP]v.

Conclusions: Significant arteriovenous differences in iRBC and plasma concentrations of chloride, potassium and sodium reflect the role that movement of ions across erythrocyte cell membranes play in regulating acid-base balance and plasma concentrations of these ions. Exhalation of CO2 has a major influence on this ion flux.


Evaporative cooling of sweat is the principal means by which horses dissipate heat during exercise (Carlson 1983; Hodgson et al. 1993). Endurance horses experience a substantial loss of fluid, as well as sodium, potassium and chloride, due to the hypertonic nature of their sweat (Rose et al. 1980; Snow et al. 1982; McCutcheon et al. 1995; Kingston et al. 1999). The concentration of sodium [Na+], potassium [K+] and chloride [Cl-] ions in equine sweat varies, depending largely on the exercise intensity, temperature, relative humidity and previous heat acclimation. Horses undergoing moderate intensity exercise in cool conditions can have sweating rates of 25 ml/m2/min. (McCutcheon and Geor 1998). If humidity is 45–55%, the sweating rate increases by approximately 30–35%, which corresponds to fluid losses of 10–15 l/h (Geor and McCutcheon 1998; McCutcheon and Geor 1998). McCutcheon et al. (1995) demonstrated that sealed pouches for sweat collection decreased evaporative losses and allowed for a more accurate determination of sweat losses for a given area of skin. The concentration of sodium, potassium and chloride in sweat collected from pouches at the end of 30 min of low intensity exercise were found to be 143 ± 9, 28.2 ± 2.1 and 158 ± 7.1 mmol/l, respectively. A significant result of the disproportionate ion loss observed in endurance horses is the development of a metabolic (hypochloraemic) alkalosis (Rose et al. 1979, 1991; McKeever et al. 1991) due to increases in strong ion difference (SID: SID =[Na++ K+]–[Cl-+ lactate]) (Stewart 1983). Concentrations of bicarbonate [HCO3-], hydrogenions [H+], carbonate [CO32−], base [A-], acid [HA] and hydroxide [OH-] ions and pH are all dependent variables. Thus, changes in the concentrations of the dependent variables can only occur if the concentrations of the independent variables (strong ions, weak acids and PCO2) increase or decrease (Stewart 1983). The hypochloraemia increases the SID, which in turn increases the pH.

Previous studies have demonstrated that horses undergoing low intensity exercise have only minor increases in lactate, while bicarbonate generally increases over the course of the run (Grosskopf and Van Rensburg 1983; Rose et al. 1991; Barton et al. 2003). Electroneutrality is maintained through the release of bicarbonate from erythrocytes in exchange for chloride (chloride shift), which leads to an increase in pH and thus contributes to the observed metabolic alkalosis. In addition, endurance horses may hyperventilate during the run.

The decrease in PCO2 results in a respiratory alkalosis (Pan et al. 1983; Rose et al. 1991; Bayly et al. 1995; Hopkins et al. 1998). Consequently, a mixed respiratory and metabolic alkalosis can be observed in endurance horses.

The loss of electrolytes in sweat can lead to a variety of homeostatic disturbances in horses undergoing prolonged submaximal exercise although, despite electrolyte imbalances, horses are still able to finish 160 km endurance races at speeds thataverage up to 20 km/h (Schott et al. 2006). Previous studies have demonstrated that erythrocytes may play an important role in plasma ion homeostasis during exercise of varying intensities in horses and man (Lindinger et al. 1992; Ferrante et al. 1995; Väihkönen et al. 1999; Bayly et al. 2006). Although numerous reported studies have focused on electrolyte changes in horses running in endurance races, to our knowledge, none have included intraerythrocytic ion measurements. Most studies focusing on electrolyte changes in endurance horses have been performed in the field and have produced somewhat inconsistent results due to the many variables associated with field studies (e.g. terrain, temperature variation and owner compliance). Of the few treadmill studies focusing on electrolyte changes due to prolonged submaximal exercise, we were unable to find any that have reported intraerythrocytic ion changes.

The purpose of this study was to explore the many factors, including intraerythroctyic ion changes, that appear essential to understanding the effects of prolonged hypertonic sweating and its subsequent effects on acid-base balance and plasma homeostasis in endurance horses.

Materials and methods


The Washington State University Animal Care and Use Committee approved this study. Five Thoroughbreds and one Quarter Horse, previously acclimated to a high-speed treadmill, were utilised. Prior to training, all horses underwent surgery to relocate the left carotid artery to a subcutaneous position. The horses were fed alfalfa-grass hay ad libitum and a manufactured concentrate with an added commercial electrolyte supplement (Apple Elite Electrolyte)1 to meet their nutritional requirements, and had access to pasture and water at all times. Bodyweights ranged from 487.7–513.2 kg (mean ± s.e.m. 500 ± 36 kg) on the day of the trial. The horses were aged 9–14 years.


To ensure a high level of performance, the horses began training on the treadmill 7 weeks prior to commencement of the trials. Horses underwent a combination of interval and long distance exercise 4 days/week, with exercise intensity and duration increasing weekly. Initially, all training was done at a +4% slope, but then increased to a +6% slope one week prior to the study. At the end of the 7 week training period, the horses underwent 2 incremental exercise tests to measure V̇O2max on a +6% treadmill slope, using a previously described protocol (Rose et al. 1990). Each horse's speed at 30% V̇O2max was calculated using the regression equation from the linear portion of the inline image-speed curve. Thirty percent of V̇O2max speed is representative of submaximal treadmill exercise that can be sustained for up to 120 min (Naylor et al. 1993; Schott and Hinchcliff 1993; Bayly et al. 1995). Two weeks after determining V̇O2max, the horses ran at their calculated 30% V̇O2max speed for 105 min on a 6% slope. All exercise was conducted in a temperature-controlled room. Three large industrial fans were in place to assist with cooling. The room temperature was set to 21°C and increased a few degrees due to the heat produced by the horses. Humidity was not measured on the day of the run, but 2 horses were re-run to determine the change in humidity. The humidity increased for one horse from 34–50% and for the other from 41–54%. The horses were allowed to feed freely on the morning of the trial, but food and water were withheld from 07.00 h until the final post exercise sample was taken.


The areas for catheterisation were surgically prepared and locally anaesthetised with lidocaine prior to the start of the trial. One catheter was placed percutaneously into the translocated carotid artery segment for arterial blood collection. Another catheter was passed via the right jugular vein into the pulmonary artery for the collection of mixed venous blood. The position of the catheter tip was verified by monitoring blood pressure wave forms. To maintain patency, catheters were flushed with small volumes of heparinised saline every 5 min for the duration of the run. A heart rate monitor and rectal thermometer probe (∼20 cm length) were also placed prior to the exercise test. Rectal temperature rather than mixed venous temperature was measured in order to spare the horses from an additional invasive procedure and because it has been shown that there is no difference between these temperatures when horses exercise submaximally like this for 15 min or more (Hodgson et al. 1993).


Before the collection of each sample, blood was drawn from the catheter for one minute to ensure a fresh sample. At the time of sampling, 2 heparinised syringes of arterial and mixed venous blood were simultaneously drawn anaerobically. Ten samples were collected in total: one pre-exercise, every 15 min during the run, and 20 min and 60 min post exercise. One syringe from each pair was immediately analysed for blood gases, pH and HCO3- (HCO3- was calculated). The sample in the other syringe was divided equally and immediately transferred to a sodium fluoride potassium oxalate Vacutainer (for determination of plasma lactate concentration) and 3 lithium heparin Vacutainers: the first for measurement of packed cell volume (PCV), total protein concentration [TP] and concentrations of plasma ions (calcium [Ca2+], [K+], [Cl-] and [Na+]); the second for determination of iRBC concentrations of the same ions; and the third for storage in case further measurements were needed. The tube containing blood for plasma ion analysis was refrigerated at temperatures ≤10°C until the samples could be processed (≤120 min). Once the final sample was collected, all samples were centrifuged at 1400 gfor 10 min and analysed within 1 h. The vacutainer containing blood for measurement of iRBC concentrations was immediately put into a freezer at −20°C. Temperature and heart rate were also recorded at each sampling time. Microcentrifuge tubes were filled with blood and centrifuged immediately. PCV and [TP] were measured using a microhaematocrit tube reader and refractometer (VEE GEE CLX-1)2, respectively. Lactate concentration was measured using a lactate analyser (YSI 2300 STAT)3. Plasma ion concentrations (Na+, Cl- and K+), temperature-corrected blood gases and pH were measured using an ion specific electrode blood gas machine (IL Synthesis 1725 Blood Gas analyser)4. iRBC ion concentrations were measured after lysing the erythrocytes with three repeated freeze-thaw cycles. The samples were kept frozen at −20°C until the freeze-thaw cycles were conducted. All samples were placed at room temperature and allowed to thaw (approximately 2 h). The samples were then placed back into the freezer until all samples were deemed frozen (approximately 3 h) (Tran-Son-Tay et al. 1990; Buranakarl et al. 2009). Following the third thaw, the samples were centrifuged at 1400 gfor 10 min. The Na+ and Cl- ion concentrations in the supernatant (whole blood) were measured using an ion specific electrode blood gas machine. The remaining supernatant was diluted 1 part supernatant to 3 parts de-ionised water and K+ (whole blood) was measured using an ion specific electrode blood gas machine. The following equation was used to determine iRBC ion concentrations (Buono and Yeager 1986; Ferrante et al. 1995; Bayly et al. 2006):


Statistical analysis

Using a 2-way repeated measures ANOVA (P<0.05), data were analysed to determine the effects of prolonged submaximal exercise on the changes of the aforementioned ions, blood gases and metabolites in mixed venous and arterial blood as well as intraerythrocytic ion changes. Because the analysis was conducted by repeated measures analysis of variance, the Huyhn-Feldt corrected P values were used to interpret the main effects and interaction in the ANOVA. Data were expressed as mean ± s.e.m.


Exercise speed ranged from 4.2–5.1 m/s with a mean of 4.62 ± 0.05 m/s (16.63 km/h). The average distance run in the 105 min was 29.12 km. Heart rate did not exceed 160 beats/min. The mean weight loss over the 105 min run was 21.2 ± 2.2 kg. This loss could not be entirely attributed to sweat because some of the horses defaecated and urinated while on the treadmill. Maximum and minimum values for every variable were recorded in Thoroughbreds at each measurement time. Despite the breed difference there was no indication that the Quarter Horse responded any differently to the exercise test, with all its values lying within the range of the Thoroughbreds' results.

Blood gas, pH, and temperature

PaCO2 decreased over time and pHa was more alkaline than pHv, although both increased throughout the run (Table 1). PaCO2 and PvCO2 both increased post exercise (in comparison to 105 min). Both pHa and pHv decreased post exercise. pHa and pHv were 7.48 ± 0.01 and 7.45 ± 0.01, respectively, prior to the run and, after 105 min, the values were 7.57 ± 0.03 and 7.52 ± 0.02, respectively (Table 1). Temperature increased (P<0.001) from 37.3 ± 0.2°C before exercise to 40.2 ± 0.5°C after 105 min.

Table 1. Bicarbonate (HCO3-) concentration, carbon dioxide tension (PCO2), and pH in arterial (a) and mixed venous (v) blood and plasma (pl ), and intraerythrocytic (iRBC) ion concentrations of chloride (Cl-) in arterial (a) and mixed venous (v) blood. Samples were collected before the run, every 15 min during the run and 20 and 60 min following the run. Data are expressed as mean ± s.e.m.
Time (min)pHPCO2 (mmHg)HCO3- (mmol/l)Cl- (mmol/l)
  • Denotes values that were significantly different (P<0.05) at this sampling point as compared to pre-exercise values.

  • *

    *Denotes arterial values that were significantly different (P<0.05) from the venous value at that same sampling time.

  • ^

    ^Denotes values that were significantly different (P<0.05) from the value at 105 min (end of exercise)

  • ▪The ANOVA main effect for sampling time was significant; however, subsequent multiple comparison did not identify specific differences with respect to pre-exercise values.

Pre-7.48 ± 0.01*7.45 ± 0.0141 ± 0*47 ± 0.731 ± 0.71*33 ± 0.6297 ± 0.697 ± 0.687 ± 2*90 ± 2
157.54 ± 0.004*7.46 ± 0.0137 ± 0.8*53 ± 0.532 ± 0.81*37 ± 0.8198 ± 0.6*97 ± 0.689 ± 2*90 ± 2
307.55 ± 0.01*7.50 ± 0.0137 ± 1*50 ± 0.932 ± 0.8*37 ± 0.7497 ± 0.7*96 ± 0.786 ± 1*89 ± 2
457.57 ± 0.02*7.49 ± 0.0136 ± 2*49 ± 132 ± 0.97*37 ± 0.6897 ± 0.6*95 ± 0.686 ± 2*88 ± 2
607.58 ± 0.01*7.50 ± 0.0135 ± 2*48 ± 232 ± 1.34*37 ± 1.0896 ± 0.6*94 ± 0.785 ± 2*87 ± 2
757.59 ± 0.02*7.51 ± 0.0234 ± 2*48 ± 332 ± 1.15*38 ± 1.0796 ± 0.8*94 ± 0.784 ± 2*87 ± 2
907.58 ± 0.03*7.52 ± 0.0336 ± 3*47 ± 332 ± 1.14*37 ± 0.7895 ± 0.8*94 ± 0.786 ± 1*88 ± 1
1057.57 ± 0.03*7.52 ± 0.0237 ± 3*48 ± 233 ± 0.68*38 ± 1.1195 ± 0.8*94 ± 0.885 ± 2*88 ± 1
Post207.48 ± 0.01^7.46 ± 0.01^43 ± 1*^46 ± 132 ± 0.5733 ± 0.68^95 ± 0.795 ± 0.8^87 ± 2*89 ± 2
Post607.48 ± 0.004^7.46 ± 0.01^45 ± 0.7*^50 ± 0.8^33 ± 0.5635 ± 0.7196 ± 0.7^96 ± 0.7^86 ± 1*89 ± 3

Total protein and PCV

The [TP] and PCV increased throughout the duration of the run. [TP]v and PCVv were greater than [TP]a and PCVa although both [TP] and PCV's arterial and venous changes occurred in parallel with the result that there were no changes in the magnitude of the arteriovenous differences in either variable (Table 2). PCVv rose 26% from 0.42 ± 0.01 to 0.53 ± 0.01 (P<0.001), while PCVa increased 27% (P<0.001). [TP]v and [TP]a increased by 16% and 15% (P<0.001), respectively, over the course of the run.

Table 2. Plasma (pl) and intraerythrocytic (iRBC) ion concentrations of sodium (Na+) and potassium (K+) in arterial (a) and mixed venous (v) blood, and plasma concentrations of ionised calcium (Ca2+) and total protein (TP) in arterial (a) and mixed venous (v) blood. Samples were collected before the run, every 15 min during the run and 20 and 60 min following the run. Data are expressed as mean ± s.e.m.
Time (min)Na (mmol/l)K (mmol/l)Ca (mmol/l)TP (g/100ml)
  • Denotes values that were significantly different (P<0.05) at this sampling point as compared to pre-exercise values.

  • *

    *Denotes arterial value that were significantly different (P<0.05) from the venous value at that same sampling time.

  • ^

    ^Denotes values that were significantly different (P<0.05) from the value at 105 min (end of exercise).

  • •The ANOVA main effect was significant for the source of the blood sample (arterial or mixed venous); however, subsequent multiple comparison did not identify specific differences between arteriovenous pairs.

Pre-132 ± 0.6132 ± 0.441 ± 9*44 ± 74.4 ± 0.24.4 ± 0.271 ± 6*75 ± 51.3 ± 0.021.4 ± 0.026.0 ± 0.16.1 ± 0.1
15132 ± 0.6*133 ± 0.550 ± 7*50 ± 45.3 ± 0.15.4 ± 0.171 ± 5*72 ± 31.3 ± 0.021.3 ± 0.026.4 ± 0.16.5 ± 0.1
30132 ± 0.7*133 ± 0.653 ± 6*57 ± 65.3 ± 0.15.3 ± 0.167 ± 3*65 ± 31.3 ± 0.021.3 ± 0.026.2 ± 0.16.4 ± 0.1
45132 ± 0.7*133 ± 0.745 ± 6*50 ± 75.3 ± 0.15.3 ± 0.173 ± 3*66 ± 41.2 ± 0.051.3 ± 0.026.2 ± 0.16.4 ± 0.1
60132 ± 0.7133 ± 0.643 ± 4*55 ± 75.0 ± 0.15.1 ± 0.175 ± 2*69 ± 31.2 ± 0.041.2 ± 0.046.5 ± 0.16.5 ± 0.1
75133 ± 0.8134 ± 0.840 ± 6*57 ± 85.0 ± 0.15.0 ± 0.172 ± 3*65 ± 31.2 ± 0.051.2 ± 0.046.6 ± 0.16.7 ± 0.1
90133 ± 0.9*134 ± 0.949 ± 6*57 ± 84.6 ± 0.14.7 ± 0.169 ± 4*69 ± 41.2 ± 0.061.2 ± 0.056.7 ± 0.16.8 ± 0.2
105133 ± 0.8*134 ± 0.948 ± 5*64 ± 74.4 ± 0.14.5 ± 0.171 ± 3*64 ± 41.2 ± 0.061.2 ± 0.056.9 ± 0.27.1 ± 0.2
Post20134 ± 0.8^135 ± 0.742 ± 4*61 ± 93.4 ± 0.1^3.4 ± 0.1^75 ± 3*69 ± 41.3 ± 0.05^1.3 ± 0.05^6.7 ± 0.26.8 ± 0.2
Post60134 ± 0.5^135 ± 0.7^37 ± 5*54 ± 83.5 ± 0.03^3.4 ± 0.03^80 ± 4*70 ± 51.3 ± 0.03^1.3 ± 0.03^6.6 ± 0.2^6.6 ± 0.2^

Plasma and intraerythrocytic ions

Exercise had no significant effect on any of the iRBC ion concentrations except iRBC[Cl-]a and iRBC[Cl-]v, although there were consistent arteriovenous differences (Table 1). pl[Na+]a did not change during exercise, but increased post exercise, while pl[Na+]v was increased after 105 min, and increased further in the post exercise period (Table 2). pl[Na+]a and iRBC[Na+]a were both less than pl[Na+]v and iRBC[Na+]v, respectively, for the duration of the run (Table 2). pl[K+]a and pl[K+]v both increased during the first 15 min of exercise, then decreased from 15–105 min and continued to decrease post exercise (Table 2). There were no arteriovenous differences in pl[K+]. iRBC[K+]a was greater than iRBC[K+]v (Table 2).

Both pl[Cl-]a and pl[Cl-]v decreased during exercise (Table 1). pl[Cl-]a was greater than pl[Cl-]v. There was significant change in iRBC[Cl-] over the course of the collection period (P = 0.018). However, the arterial and venous changes occurred in parallel, with the result that there was no change in the arteriovenous iRBC[Cl-] (Table 1). [HCO3-]v increased with exercise and was greatest after 105 min, then decreased post exercise (Table 1). [HCO3-]a did not change during the run although it was less than [HCO3-]v for the duration (Table 1). Ionized [Ca2+]a and [Ca2+]v both decreased from the pre-exercise values to 105 min and increased during the recovery period (Table 2). pl[La-]a and pl[La-]v both increased with exercise.


This is the first report documenting both plasma and intraerythrocytic concentrations of ions and associated movements of ions between plasma and erythrocytes in arterial and mixed venous blood over the course of prolonged submaximal exercise in horses. Resting iRBC values reported here are similar to those previously observed in horses prior to commencing short-term (<2 min) high intensity exercise to fatigue (Bayly et al. 2006).

All of the horses hyperventilated during the 105 min run. The decrease in PaCO2 reflected increased ventilation, which, when combined with the loss of Cl- in sweat and subsequent increase in SID, led to the increase in pH and HCO3-. CO2 is predominately carried to the lungs in the form of HCO3-. After CO2 is produced in active tissue, it diffuses into the erythrocyte where its conversion to HCO3- and H+ is catalysed by carbonic anhydrase. To ensure that this reaction continues to produce HCO3-, HCO3- moves out of the erythrocyte and Cl- moves in to maintain electroneutrality. Exhalation of CO2 drives this reaction in the reverse direction with the result that HCO3- moves into the erythrocyte and is converted to CO2 and water. Consequently, subsequent to the exhalation of CO2 and the associated increase in pH despite the decrease in [HCO3-] from pulmonary to carotid artery, Cl- moves out of the erythrocytes to maintain electroneutrality. Collectively, this explains why [HCO3-]a was less than [HCO3-]v, pl[Cl-]a was greater than pl[Cl-]v and iRBC[Cl-]a was less than iRBC[Cl-]v. As PvCO2 increased (due to increased production), the chemical equilibrium shifted toward the production of HCO3- and H+, explaining both the [HCO3-] findings and why pHv was less than pHa.

In keeping with the pH findings, it was expected that pl[Na+]a would be greater than pl[Na+]v and iRBC [Na+]a would be less than iRBC[Na+]v. While the latter was the case, pl[Na+]a was also less than pl[Na+]v. A decrease in [TP] from mixed venous to arterial blood was also observed despite the fact that it is generally accepted that protein remains within the intravascular space, particularly when traversing the pulmonary capillaries. The arteriovenous difference in pl[Na+] has been previously observed in horses undergoing strenuous exercise (Schott et al. 2002; Bayly et al. 2006), which indicates either that sodium was lost somewhere between the pulmonary artery and carotid artery or that water or another low or nonprotein fluid such as lymph was added to the circulation. It could be argued that, if this was the case, similar results would be seen for the other principal ions, K+ and Cl-. However, Cl- is already moving into the plasma from the erythrocyte secondary to the movement of HCO3- into the erythrocyte and K+ is an extremely labile ion that is actively pumped into the cell in exchange for both H+ and Na+ in order to keep the resting membrane potential close to its optimal value. The thoracic duct is the major lymphatic vessel that returns lymph to the circulation. Connection of this vessel to the vasculature usually occurs near the vena cava and thus could not contribute to the observed decrease in [Na+] and [TP] between the pulmonary and systemic arterial circulations.

Vengust et al. (2006) reported that horses undergoing intense exercise experience a loss of fluid from the pulmonary capillaries into the lung interstitium, which is probably due to increased pulmonary microvascular pressures and a decreased colloid osmotic gradient in response to fluid release from erythrocytes secondary to high O2 tension in the lungs. Lung water accumulation in response to strenuous exercise has also been reported in man (Anholm et al. 1999; McKenzie et al. 2005). This occurrence seems unlikely in our study as the lost fluid would have to be hypertonic with respect to sodium and protein in order to explain our results. However, it is possible that during the course of exercise, movement of water or other nonprotein containing fluid from erythrocytes (Vengust et al. 2006) and/or surrounding tissues into the pulmonary microvasculature contributed to the small, but significant, observed differences between pl[Na+]a and pl[Na+]v and [TP]a and [TP]v.

That iRBC[K+]a was greater than iRBC[K+]v can be associated with the fact that pHv was more acidic than pHa. Increased production of [H+] in venous blood (due to increased CO2 and thus HCO3-) causes K+ ions to leave the erythrocyte. It is well known that contracting skeletal muscle also releases K+ and this phenomenon may also have contributed to the increase in pl[K+]. The H+/ K+ exchange is less prominent in arterial blood due to the decreases in CO2 and subsequent decrease in [H+]. The finding that iRBC[K+]a was greater than iRBC[K+]v is also consistent with the observation that iRBC[Na+]a was less than iRBC[Na+]v.

Packed cell volume increased by 27% and 26% in arterial and venous blood, respectively, while [TP] only increased by 15% in arterial blood and 16% in venous blood. The overall increase for PCV and [TP] from the pre-exercise sample to 105 min is likely to be indicative of dehydration, but the greater increase in PCV is probably due to the additional effect of some splenic contraction, although the aforementioned possible loss of albumin into the lung interstitium cannot be totally discounted.

Lactate did not increase above 2.0 mmol/l during the run or the recovery period. This correlates with numerous other studies that have been performed on endurance horses (Grosskopf and Van Rensburg 1983; Rose et al. 1991; Barton et al. 2003). The overall decrease in ionized [Ca2+] is consistent with the hypertonic nature of equine sweat and has been previously reported (Barton et al. 2003; Schott et al. 2006).

In conclusion, significant arteriovenous differences in iRBC and plasma concentrations of Na+, K+, Ca2+ and Cl- indicate that there is regular and dynamic flux of ions between plasma and erythrocytes, and that these shifts play a role in regulating acid-base balance and plasma concentrations of key ions during prolonged submaximal exercise. Finally, exercise-induced changes in plasma concentrations of these ions are not necessarily associated with corresponding changes in iRBC concentration of the same ion. Another significant finding of this study was that the ion flux observed between plasma and erythrocytes appears to be strongly influenced by the exhalation of CO2. Perturbation of homeostasis, caused by ion losses, is believed to be a major factor leading to fatigue in endurance horses (Carlson 1985; Smith and Wagner 1985). Further understanding of these homeostatic alterations could lead to improved understanding of the origins of fatigue in endurance horses and more effective electrolyte supplementation programmes during endurance races.


This work was supported in part by the Washington State University College of Veterinary Medicine Summer Research Fellowship Program and the Washington Equine Research Program. The technical assistance of Eric Renner, Ben Crosland, Belinda Buchholz, Ashley Culp, Heidi Talbott, Jack Fillerup and Mike Clayton is greatly appreciated.

Conflicts of interest

The authors have no potential conflicts.

Manufacturers’ addresses

1 Farnam, Phoenix, Arizona, USA.

2 VEE GEE Scientific Inc, Kirkland, Washington, USA.

3 YSI Inc, Yellow Springs, Ohio, USA.

4 Instrumentation Laboratory, Orangeburg, New York, USA.