Factors contributing to plasma TCO2 and acid-base state in Ontario Standardbred racehorses

Authors


  • Supported by the Ontario Horse Racing Industry Association and the Natural Sciences and Engineering Research Council of Canada.

email: waller@pharmacy.ohio-state.edu

Summary

Reasons for performing study: Standardbred and Thoroughbred racehorses around the world are tested for performance enhancing substances. Among these are blood alkalising substances that raise plasma pH and total carbon dioxide (TCO2) concentration. However, many horses have an elevated TCO2 due to dietary, environmental and health concerns without having been administered an alkalising substance.

Objectives: The purposes of this study were to determine the acid-base profile of a cross section of Standardbred horses in racing/race training in Ontario and the main independent variables that contributed to acid-base state.

Materials and methods: On nonracing days, blood from 211 horses at rest, from 9 training facilities, was analysed within 30 min for plasma pH (7.406 ± 0.039; mean ± s.e.), PCO2 (50.0 ± 3.4 mmHg), from which [HCO3-] (31.2 ± 2.8 mmol/l) and [TCO2] (33.1 ± 2.9 mmol/l; range 25.66–42.9) were calculated. From these, a subset of 161 horses had full data sets for plasma protein and strong ion concentrations. These data were further analysed by facility and level of TCO2. Data on nutrition, training, racing and medications were also collected.

Results: There were significant differences amongst facilities with respect to plasma pH, TCO2, strong ion difference ([SID]), PCO2 and total weak acid concentration ([Atot]). Horses having the highest TCO2 (37.0–42.9 mmol/l, n = 16) had significantly higher [SID] (52.9 ± 0.8 mEq/l) and PCO2 (52.5 ± 0.7 mmHg) and relatively low [Atot] (14.9 ± 0.7 mEq/l) compared to average TCO2 (32.1.0–34.9 mmol/l) horses (n = 75). In horses with the lowest TCO2 (n = 11) the greatest contributor was elevated [Atot] (21.0 ± 0.7 mEq/l) and unmeasured (acetate, citrate, proprionate, butyrate) weak acids (7.0 ± 0.2 mEq/l) while [SID] (49.6 ± 0.8 mEq/l) and PCO2 (47.8 ± 1.0 mmHg) were similar to average TCO2 horses. Thirty-two horses had a TCO2 ranging from 35.0–36.9 mmol/l).

Conclusions: There is a wide range of acid-base state and factors contributing to acid-base state amongst Standardbred race horses in Ontario. Dietary, environmental and handling practices and health concerns, that elevate plasma [SID], lower [Atot] and lower the concentration of unmeasured weak acids are the primary contributors to alkalosis and elevated TCO2.

Introduction

Measurement and interpretation of acid-base status is particularly important amongst racing jurisdictions to determine if horses have been administered alkalinising substances for the purpose of enhancing performance. Currently, a serum total carbon dioxide concentration ([TCO2]) threshold of 37 mmol/l is used by many racing jurisdictions to identify horses administered a prohibited alkalinising agent (Auer et al. 1993). However, unlike some other tested-for drugs and banned substances, TCO2 occurs naturally in the horse and is affected by a host of factors including exercise, feeding, electrolyte administration and diurnal variation (Lindinger and Waller 2008). Therefore, it is important to determine the major specific factors that generate or contribute to disturbances in TCO2 and acid-base state in racehorses. Identification of the origins of acid-base disturbances is facilitated by use of the physicochemical approach to acid-base status, which uses both dependent and independent variables to describe acid-base state. According to the physicochemical approach, the independent variables that determine plasma acid-base status are the concentration of strong ions in solution defined as the strong ion difference ([SID]), the partial pressure of carbon dioxide (PCO2) and concentration of weak acids in solution defined as the total weak acid concentration ([Atot]). Therefore, the dependent acid-base variables [H+], bicarbonate concentration ([HCO3-]) and total carbon dioxide concentration ([TCO2]) only change when one or more of the independent variables are altered (Stewart 1983; Lindinger and Waller 2008).

In the testing threshold-defining study by Auer et al. (1993), a TCO2 range from 26.6–36 mmol/l was reported in 61 normal Standardbred racehorses from Australia, while a study of 205 New Zealand racehorses reported a TCO2 range of 26–36.5 mmol/l (Irvine 1992). A large study of 2349 Thoroughbreds racing in California reported a range of TCO2 from 25–37.5 mmol/l (Cohen et al. 2006), while a [HCO3-] (which comprises ∼95% of TCO2) ranged from 23–38 mmol/l in 30 Standardbreds from the northeastern US (Soma et al. 2000). Clearly, there exists a wide range of normal TCO2 values in racehorses. However, of the limited studies to date reporting TCO2 or [HCO3-], none have measured and reported the major plasma variables that determine acid-base status. Therefore, the factors that contribute to TCO2 and other dependent acid-base variables have been incompletely characterised. The purpose of the present study was to determinethe acid-base profile of a cross section of Standardbred horses in racing/race training in Ontario and determine the main independent variables that contributed to acid-base state. It was hypothesised that there would be a wide variation in plasma acid-base and electrolyte variables in a subpopulation of Standardbred racehorses.

Materials and methods

Horses

Jugular venous blood from 211 Standardbred horses in race training was sampled at rest and analysed for plasma gases and ions within 30 min using a NovaStat Profile 9+ Analyser1. Horses were stabled at 9 different facilities in Southern Ontario, Canada and trained by 62 different trainers. They were typically fed between 06.00 and 08.00 h and blood sampled between 09.30 and 11.00 h over a 5 month period (Sept–Jan). Only horses that had not raced for the previous 2 days or that would not be racing in the coming 2 days were sampled. This avoided altered acid-base and electrolyte state associated with pre and post race medications and supplements, many of which directly or indirectly affect acid-base state. Blood was typically sampled before the training workout of the day; in some cases, blood was sampled at least 70 min after the conclusion of a training workout. Owners of facilities and trainers of horses did not have prior knowledge that blood samples were to be obtained from their horses. We obtained information on feeding practices, special routines and medications, if any, for each horse. Animal use procedures were approved by the University of Guelph Animal Care Committee, in accordance with the Guidelines of the Canadian Council on Animal Care.

Sample analysis

Seven or 10 ml lithium heparin vacutainer tubes2 were fully filled with blood via a 22 gauge needle inserted into the jugular vein by the researchers. Samples were kept on ice water until analysed (within 30 min). Prior to analysis, the tube was repeatedly and gently inverted for ∼60 s in order to fully mix the contents. The stopper was removed and the tube advanced fully over the aspirator tube so that blood was drawn 2–3 cm below the surface. Blood was aspirated into a NovaStat Profile 9+ analyser and electrode measurements obtained for plasma pH, PCO2, PO2, Na+, K+, Ca2+, Cl- and lactate-. Collected samples appeared to be anaerobic due to the consistently low PO2 of the samples. The instrument calculated [HCO3-] and TCO2 (CV 0.21–0.61%) and [H+] was calculated from pH.

The instrument calculated plasma [HCO3-] using the Henderson-Hasselbalch equation:

image

and plasma TCO2 was calculated as:

image

Where pK1′= 6.091 and S = 0.0307, the solubility coefficient of CO2 in plasma at 37°C.

It is important to note that if the sample was not maintained anaerobic by sampling technique or by the character of the vacutainer tube, then the TCO2 at the time of measurement of pH and PCO2 would be lower than when the blood was in the horse. Vacutainers were used for direct collection of blood from the jugular vein in order to minimise introducing air into the sample, ease and rapidity of sample collection and ability to achieve minimal dead space within the vacutainer tube.

Remaining blood was centrifuged, plasma removed and stored at -80°C for later analysis of total protein concentration using refractometry (CV 0.83%)3. A subset of data from 161 horses, for which complete data for plasma protein ([PP]) and strong ion concentrations were obtained, was used for a detailed analysis. Calculation (Acid-Basics II Software)4 of dependent acid-base variables was computed using values of plasma [Atot] ([plasma protein]× 2.04; Constable 1997) and yielded pH values that were consistently too high. In order to determine the possible contribution of unmeasured weak acids (UWA), an [UWA] was solved for each plasma sample by inputting values for plasma [SID], PCO2 and pH using software written by Dr John M. Kowalchuk (University of Western Ontario). This resulted in an average [Atot] of 13 mmol/l and an average [UWA] of 8.5 mmol/l and near-perfect agreement between measured and calculated dependent variables. The constants used in the calculations were 1.74 × 10−7 for the Ka for [Atot]; 6.0 × 10−11 Eq/l for the K3; 2.46 × 10−11 (Eq/l)2/mmHg for the KC and 4.4 × 10−14 (Eq/l)2 for the K′w (Stewart 1981). Lindinger (2004) and Lindinger and Waller (2008) provide a detailed description of the acid-base concepts and equations.

The contribution of each of the independent variables ([SID], PCO2, [Atot], [UWA]) to the concentrations of dependent acid-base variables (H+, pH, HCO3- and TCO2) was calculated using Acid-Basics II Software (P.D. Watson, University of South Carolina). For example, to determine the contribution of [SID], [SID] was allowed to change while PCO2, [Atot] and [UWA] were held constant at their normative values of 40 mmHg, 13 mmol/l and 8.5 mmol/l, respectively.

Statistics

The data were analysed with respect to training facility and TCO2. Plasma TCO2 was binned into the following 5 categories: 25.6–28.9, 29.0–30.9, 31.0–34.9, 35.0–36.9 and 37.0–42.9 mmol/l. Comparisons amongst facilities and TCO2 category were performed using two-way and one-way analysis of variance. When a significant F ratio was obtained, the Holm-Sidak post hoc test was used to test for differences between means. The TCO2 data were analysed for normality using the Shapiro-Wilk test.

Results

In addition to hay, the horses were typically fed a diet of sweet feed or some sort of similar, often pelleted, commercial concentrate. A total of 20 different supplements and 32 medications were reported. There was no detectable effect of medications and supplements as horses that received these did not have an acid-base profile that differed significantly from the mean of category 3 TCO2 horses.

Figure 1 provides histogram representation of the dependent acid-base variables pH, [H+] and [TCO2] for the 211 horses for which measures of pH and PCO2 were obtained. These data show a broad distribution and a reasonably normal distribution for the sample size.

Figure 1.

Frequency distribution for 211 Standardbred racehorses in training.

Table 1 provides a descriptive summary of the measured and calculated acid-base variables for the 161 horses for which complete acid-base and electrolyte data were obtained. For some variables ([PP], Hct, BE, [HCO3-] and [TCO2]) the mean values are greater than what is typically reported in tables of normative values (Lindinger and Waller 2008) and the [SID] is also higher than the normal values of 42 mEq/l. Also evident is that there is wide range in these variables that generally exceeds the ranges provided in tables of normative values.

Table 1. Measured and calculated plasma and blood variables of 161 Standardbred racehorses
VariableMeans.d.s.e.RangeMaxMin
  1. nCa2+ is the ionised calcium concentration normalised to a plasma pH of 7.4. [PP]= plasma protein concentration. [Atot]= weak acid concentration based on measured plasma protein concentration. [UWA]= concentration of unmeasured weak acids. Hct = haematocrit. Hb = haemoglobin concentration. BE = excess.

[H+] (nmol/l)38.93.50.417.548.931.4
pH7.4110.03790.002980.1927.5037.311
PO2 (mmHg)45.110.80.982.711330.7
PCO2 (mmHg)49.83.40.32462.138.1
Na+ (mEq/l)1441.90.212148136
K+ (mEq/l)3.80.440.032.54.92.4
Cl- (mEq/l)972.40.21110291
nCa2+ (mEq/l)2.880.120.010.923.142.22
[SID] (mEq/l)49.63.50.2818.159.441.3
PP [Atot] (mmol/l16.03.40.316.824.57.7
[UWA] (mmol/l)5.31.10.095.68.22.6
[PP] (g/100 ml)7.81.70.18.212.03.8
Glucose (mmol/l)5.80.80.076.911.34.4
Lactate (mEq/l)0.50.40.032.63.00.4
Hct (%)393.20.3154833
Hb (g/100 ml)12.91.10.095.216.110.9
BE (mmol/l)7.03.20.320.018.0-2.0
HCO3- (mmol/l)31.92.80.216.841.324.5
TCO2 (mmol/l)33.42.80.216.942.926.0

The frequency distribution for independent acid-base variables for the subgroup of 161 horses (Fig 2) was similar in appearance to those shown in Figure 1. The TCO2 data conformed to a normal distribution as determined using the Shapiro-Wilk test (W-Statistic = 0.984; P = 0.070; Passed).

Figure 2.

Frequency distribution of independent plasma acid-base variables for 161 Standardbred racehorses. [SID]= strong ion difference, [Atot]= plasma weak acid concentration based on measured plasma protein concentration. [UWA]= concentration of unmeasured weak acids.

The data were further analysed with respect to concentration of TCO2, such that data were binned into 5 categories with [TCO2]: 1) 25.6–28.9, n = 11; 2) 29.0–30.9, n = 24; 3) 31.0–34.9, n = 78; 4) 35.0–36.9, n = 32 and 5) 37.0–42.9, n = 16; mmol/l. The characteristics of each category are shown in Figure 3 (dependent variables) and Figure 4 (independent variables). It is evident from Figure 4 that increased [SID] and PCO2 contributed to the increased pH, [TCO2], [HCO3-] and decreased [H+] (categories 4 and 5 compared to category 3). In contrast, increased [Atot] and [UWA] were main contributors to the decreased pH, [TCO2], [HCO3-] and increased [H+] (categories 1 and 2 compared to category 3).

Figure 3.

The dependent variables by TCO2 category. Refer to text for category ranges and n. The number(s) above a bar indicate which categories it is significantly different from (P<0.05; one-way ANOVA). * indicates each mean is significantly different from every other mean.

Figure 4.

The independent variables within each TCO2 category. Refer to text for category ranges and n. The number(s) above a bar indicate which categories it is significantly different from (P<0.05; one-way ANOVA).

A more detailed breakdown of the contributions of the independent variables to only [H+] and TCO2 (since pH and [HCO3-] are represented by [H+] and [TCO2]) are provided for categories 5, 3 and 1. There were no notable differences between high, median and low [TCO2] horses with respect to general health status and medications.

High TCO2 horses (category 5): Horses having a [TCO2] above 37.0 mmol/l were represented by 8 trainers, with one trainer having 6 such horses (2 on lasix), another trainer 3, and 3 trainers 2 horses each. Thirteen of the horses were actively racing, one was on prolonged rest (highest [TCO2]), one was at 10 days of post race recovery and the remainder training in preparation to qualify. Three horses received lasix, one was on penicillin, one received herbal medicines and biotin, one was on a bronchodilator and another received iodine and MSM. Grain feeding varied from hay with oats or grain (n = 3) to those receiving high performance concentrate feeds (n = 8).

Increased [SID] was a greater contributor to the increased [TCO2] for category 5 (9.6 ± 0.8 mmol/l) compared to increased PCO2 (1.2 ± 0.1 mmol/l), while the low [Atot] and [UWA] contributed 1.2 ± 0.5 and 0.4 ± 0.2 mmol/l, respectively, to the increased TCO2. With respect to the decreased [H+], the contribution of increased [SID] (−10.9 ± 0.7 nmol/l) was matched by the effect of increased PCO2 (10.7 ± 0.6 mmol/l), while lesser contributions of −1.4 ± 0.8 and −0.5 ± 0.3 nmol/l are attributed to [Atot] and [UWA], respectively. A very similar pattern of contribution was evident for category 4 horses.

Median TCO2 horses (category 3): The median category was also characterised by a somewhat elevated [TCO2] and decreased [H+]. Increased [SID] was the major contributor (6.2 ± 0.3 mmol/l) to the elevated TCO2, while PCO2, [Atot] and [UWA] had only minor effects (0.9 ± 0.03, 0.6 ± 0.2 and 0.2 ± 0.1 mmol/l, respectively). With respect to normal [H+] of category 3, the effects of increased [SID] (−7.7 ± 0.3 nmol/l) were offset by the effect of increased PCO2 (8.0 ± 0.3 nmol/l) and there were only minor contributions from [Atot] (−0.6 ± 0.4 nmol/l) and [UWA] (−0.2 ± 0.1 nmol/l).

Low TCO2 horses (category 1): There were 11 horses that had [TCO2] less than 29.0 mmol/l. Three of these horses regularly received lasix, one horse received lasix, inhaled steroids, ventopulmin and MSM. Three horses received ventipulmin, one horse was on NSAIDs and another on Regumate.

Increased [SID] (above 42 mEq/l) had a contribution to increasing [TCO2] by 6.8 ± 0.7 mmol/l. PCO2 had minimal effect (0.8 ± 0.1 mmol/l) while increased [Atot] and [UWA] contributed to a lowering of [TCO2] (−3.2 ± 0.5 and −1.0 ± 0.2 mmol/l, respectively). With respect to [H+] the lowering effect of [SID] (−8.4 ± 0.8 nmol/l) was offset by the raising effects of PCO2, [Atot] and [UWA] (6.8 ± 0.8, 6.9 ± 1.1 and 2.3 ± 0.4 nmol/l, respectively).

Lasix horses: There were 39 horses that received lasix (frusemide) on a regular basis for racing and/or heavy training. There were no differences in the means or ranges of plasma acid-base and other plasma and blood variables in these horses (Table 2) compared to the remainder of entire group (Table 1).

Table 2. Measured and calculated plasma and blood variables of horses that received lasix
VariableMeans.d.s.e.RangeMaxMin
  1. Data are from 39 horses. nCa2+ is the ionised calcium concentration normalised to a plasma pH of 7.4. [PP]= plasma protein concentration. [Atot]= weak acid concentration based on measured plasma protein concentration. [UWA]= concentration of unmeasured weak acids. Hct = haematocrit. Hb = haemoglobin concentration. BE = base excess.

[H+] (nmol/l)38.43.280.5314.948.633.7
pH7.4170.0360.0060.1597.4727.313
PO2 (mmHg)43.88.81.441.172.431.3
PCO2 (mmHg)49.02.50.411.554.442.9
Na+ (mEq/l)1442.20.38.7148.7140
K+ (mEq/l)3.90.320.051.314.463.15
Cl- (mEq/l)992.20.49.410191.6
nCa2+ (mEq/l)2.940.120.020.443.202.76
[SID] (mEq/l)51.03.90.614.358.544.3
PP [Atot] (mmol/l12.41.70.36.616.49.8
[UWA] (mmol/l)5.60.80.14.53.47.9
[PP] (g/100 ml)6.11. 40.123.28.04.8
Glucose (mmol/l)5.60.50.12.67.14.5
Lactate (mEq/l)0.50.40.11.11.50.4
Hct (%)393.60.6144733
Hb (g/100 ml)12.91.20.24.615.611
BE (mmol/l)7.33.60.614.413.6−0.8
HCO3- (mmol/l)32.03.00.511.937.525.6
TCO2 (mmol/l)33.53.10.511.939.027.1

Gender: A total of 69 mares, 127 geldings and 15 stallions were sampled. Mean [TCO2] was 33.2 ± 0.36 and 33.1 ± 0.24 mmol/l for the female and male horses respectively, and there were no difference in any variables between genders (P = 0.830)

Training facility: The database was also analysed with respect to training facility and by trainer. No effects by trainer were evident, as most trainers had horses that spanned the range of [TCO2]. The facilities at a race track (60.3% of horses) had significantly lower [SID] and PCO2 than 6 out of 7 off-track training facilities. Therefore, the majority of off-track training facilities were characterised, on average, by horses with higher plasma pH and [TCO2] than horses stabled and trained at the track. It is noteworthy that the effect on [SID] was not from [Na+] or [K+], as there were no differences in these amongst facility. The main effect on [SID] was from [Cl-], such that off-track facilities had about a 2 mEq/l lower [Cl-]. Off track facilities were also represented by higher [PP], and therefore [Atot], than at the track. There were also no differences amongst facility for glucose, lactate, haematocrit and haemoglobin.

Discussion

The present study appears to be the first to determine all the dependent and independent variables that contribute to acid-base state in a large subpopulation of Standardbred racehorses. The main findings are that there is a wide variation of key plasma acid-base and electrolyte variables in Standardbred racehorses in racing/race training. There were also key differences in the contributions to pH and TCO2 associated with the state of alkalosis and TCO2, as well as with training facility. It is important to note that blood was sampled from horses a minimum of 2 days before or after a race day. This avoided altered acid-base and electrolyte state associated with pre- and post race medications and supplements, many of which directly or indirectly affect acid-base state. It is also important that the horsemen did not know we were coming to their facility to sample blood from their horses and is therefore highly likely that normal feeding and medication practices were followed. Thus, we had a very good representation of the acid-base state of a cross section of Standardbred racehorses representing many facilities and trainers.

Regarding the values for TCO2, it should be noted that these were calculated from measures of pH and PCO2 within the extracellular phase of blood samples aspirated by the Nova StatProfile blood analyser. Values of TCO2 determined in this way typically underestimate the total amount of CO2 that can be liberated from the plasma sample upon acidification to pH 2 or lower (Burnett et al. 2001), although one study reported that calculated TCO2 exceeded ‘measured’ TCO2 by ∼1 mmol/l between 20 and 40 mmol/l (Chittamma and Vanavanan 2008). Values for TCO2 thus vary from instrument to instrument depending on the constants used, the state of calibration and accuracy of the pH and PCO2 measures. The reference standard for measuring TCO2 in blood requires acidification of the sample and titration of the acidified sample with NaOH to a predetermined pH and the TCO2 is then calculated using the normality and volume of added NaOH (Burnett et al. 2001). The Beckman synchron EL-ISE uses a modification of this technique using a pH electrode with the tip covered by a silicone membrane. When CO2 is released from an acidified plasma sample, CO2 diffuses through the silicone membrane and lowers the pH of a bicarbonate solution bathing the electrode tip.

Horses having the highest [TCO2] (37.0–42.9 mmol/l) had significantly higher [SID] and PCO2 and relatively low [Atot] compared to horses with average [TCO2] (32.1.0–34.9 mmol/l). The elevated [SID] was entirely due to a lower plasma [Cl-] in these horses, as there were no differences in [Na+] or [K+] compared to the horses with average [TCO2]. It is possible that diet or feeding practices were responsible for the lower [Cl-], as 10 of these horses were receiving a commercial pelleted feed as opposed to oats or grain. These results suggest that some horses may naturally demonstrate TCO2 levels close to, or in excess of, the 37 mmol/l testing threshold, even when no alkalinising substances have been given. Certainly, normal management and feeding-induced causes for elevated TCO2 should be taken into consideration when using TCO2 threshold tests.

In horses with the lowest TCO2 (<29 mmol/l), the greatest contributor was elevated [Atot] and [UWA], while [SID] and PCO2 were similar to average TCO2 horses. Of the 11 horses with the lowest [TCO2] there was nothing remarkable about medications or feeding regimen, although 3 horses were on lasix and 3 on ventipulmin. Ventipulmin, a β2 adrenergic agonist, could contribute to lowering TCO2 by lowering pCO2 (Rose et al. 1983); however, β2 adrenergic agonists have been shown to increase plasma volume expansion (Ewaldsson et al. 2006) and also potentially induce hypoalbuminaemia (Sawaya and Lunn 1998), both of which would act to decrease [Atot]. Therefore, the ventipulmin is not likely responsible for the lowered TCO2 and increased [H+] in these horses. Instead, dietary, environmental or handling practices that dehydrate plasma volume and elevate the UWAs appear to be the cause of lowered [TCO2] in these horses.

The inclusion of a [UWA] component in the present study was necessary in order to obtain agreement between measured and calculated dependent variables using the physicochemical approach. Indeed, a limitation of the calculation of [Atot] proposed for use in horses previously (Constable 1997) is that it was determined from a relatively small subset of ‘normal’ horses and assumes normal and constant concentrations of unmeasured plasma acids such as phosphate, citrate, acetate, proprionate and butyrate (the physico-chemical approach classifies these as weak acids while equilibrium theory classifies these as strong acids [Nguyen et al. 2009]). However, many factors can contribute to alterations in the UWAs and potentially result in inaccurate calculation of [Atot]. For example, horses fed a high grain diet (such as the racehorses in the present study) produce less acetate and tend to have lower plasma VFA concentrations than those fed mainly forage (Hintz et al. 1971). Certain medications and/or supplements may also affect concentrations of plasma weak acids. Lasix administration results in increased renal excretion of both citrate (Kaufman et al. 1985) and phosphate (Haas et al. 1977). In addition, any discrepancies between measured and calculated dependent variables may also represent differences between horses in the weak acid constant (KA) used in the calculations. Certainly the cross-section of high performance racehorses from the present study would be likely to experience feeding and supplementation practices that could result in significant variation in [UWA] and deviation from ‘normal’ weak acid calculations.

The majority of off-track training facilities were characterised by horses with higher plasma pH and TCO2 than horses stabled and trained at the track, concomitant with increased [PP] and decreased [Cl-] with no change in [Na+] or [K+]. The combination of lower [Cl-] and increased [PP] suggests a dehydration and possible chloride depletion alkalosis characteristic of horses with acute or chronic dehydration (Carlson 1987). Exercise-induced dehydration in horses is associated with the development of a mild systemic alkalosis that results from extracellular Cl- loss in excess of Na+ loss as a result of thermoregulatory sweating (McCutcheon et al. 1995). Indeed, we have previously found that Standardbreds exhibit increased plasma [TCO2] during recovery from prolonged submaximal exercise due to decreased plasma [Cl-], and concluded that sweat-induced losses of Cl- can significantly increase [TCO2] by as much as 2 mmol/l with an 8 l dehydration (Waller et al. 2007). The chloride depletion alkalosis is also an important concern with dehydration due to equine transport (Friend 2000), excitement (McConaghy et al. 1995), certain medications such as lasix (Sosa Leon et al. 1998) and high ambient temperatures (Kerr and Snow 1983). Importantly, supplementation with electrolytes according to estimated sweat losses attenuates decreases in plasma [Cl-], restores plasma hydration, and results in decreased [TCO2] compared to when no electrolytes are given (Waller et al. 2007). However, of the 211 horses sampled in the present study, only 7 (6 off-track, 1 at-track) were reported to be receiving regular electrolyte supplementation. Therefore, it is certainly possible that dehydration and depletion of electrolytes contributed to elevated [TCO2] in these horses.

In contrast to previous studies that showed higher prerace [TCO2] in horses administered frusemide (Soma et al. 2000; Cohen et al. 2006), in the present study there were no differences in the means or ranges of plasma acid-base variables or other blood variables in the horses that received lasix on a regular basis for racing and/or heavy training compared to the horses that did not. Lasix, a loop diuretic, is known to have an alkalinising effect in sedentary and exercising horses (Freestone et al. 1989; Carlson and Jones 1990) as a result of substantial renal Cl- excretion without concomitant increases in renal cation excretion. However in contrast to the previous studies that obtained blood samples 1–2 h prior to racing (∼time of peak effect for frusemide), sampling in the present study occurred on nonrace days, likely when the effects of prior lasix administration would be absent. There were also no gender differences in [TCO2] in the present study. This is in contrast to a larger study of Thoroughbred racehorses (n = 2349) which found that females had lower serum [TCO2] than males (Cohen et al. 2006); however, the magnitude of difference was physiologically very small (31.1 and 30.7 mmol/l for males and females, respectively) and the clinical significance is unknown.

Overall, there was a wide range of medications and feed supplements reported in these horses, with 47% of horses receiving medications and 22% receiving at least one feed supplement. The most common feed supplements were vitamins, iron supplements, various herbals and electrolytes. The most common medications used in on-track horses was frusemide>anti-inflammatories>bronchodilators>herbal medicines>antibiotics>other (including hormone therapies, EPM medications and anthelmintics)>anabolic steroid, while the most common medications used in off-track horses was anti-inflammatories>frusemide>herbal medicines>bronchodilators>antibiotics = anabolic steroids>omeprazole>muscle relaxants>others (including omeprazole, hormone therapies, EPM medications and anthelmintics). A complete description of the veterinary drugs and herbal medications used in these horses is provided by Pearson (2009).

Conclusions

It is concluded that the normal range of [TCO2] in a subpopulation of Standardbred racehorses from Ontario is 17 mmol/l, ranging from 25.6–42.9 mmol/l and that some horses normally have a plasma [TCO2] that closely approaches or exceeds the testing threshold of 37 mmol/l. Dietary, environmental and handling practices that elevate plasma [SID], lower [Atot] and lower the concentration of unmeasured weak acids are the primary contributors to alkalosis and elevated [TCO2]. The main advantage of the physicochemical approach is the ability to identify and quantify the origins of an acid-base disturbance. The knowledge of why changes in [H+] and [TCO2] occur enhances the understanding of acid-base physiology and may be useful in developing alternative testing strategies to determine whether illegal alkalinising agents have been given.

Acknowledgements

The authors would like to thank the owners, trainers and grooms of the horses for providing access and assistance. This research was supported by the Ontario Horse Racing Industry Association and the Natural Sciences and Engineering Research Council of Canada.

Conflicts of interest

The authors declare no potential conflicts.

Manufacturers' addresses

1 Nova Biomedical, Waltham, Massachusetts, USA.

2 Becton-Dickenson, Mississauga, Ontario, Canada.

3 Atago Clinical Refractometer, Atago, Japan.

4 P.D. Watson, Columbia, South Carolina, USA.

Ancillary