SEARCH

SEARCH BY CITATION

Keywords:

  • horses;
  • endurance;
  • exercise;
  • metabolic elimination;
  • biochemistry;
  • metabolic predictors

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Conflicts of interest
  10. Manufacturers’ addresses
  11. References

Reasons for performing study: Endurance races are the competition with the biggest metabolic demands for the sport horse. During races, some horses show homeostasis alterations, having repercussions in diverse biochemical parameters and negative consequences on performance and health.

Objectives: To evaluate the utility of biochemical analysis in the early diagnosis of metabolic stress and to determine cut-off values of biochemical parameters to assist in the prevention of metabolic alterations in endurance horses.

Methods: This study involved 36 CEI races and 283 horses (41 eliminated because of metabolic disturbances). Blood samples were taken before competition, after the vet-gates and after finishing the race or veterinary disqualification. Packed cell volume (PCV), activities of CK, AST and LDH, and concentrations of total plasma proteins (PP), urea, creatinine (Cr), uric acid (UA) and plasma lactate were determined. Successful horses were compared with horses eliminated due to metabolic conditions in the values obtained in the phase prior to being removed from the competition. Factors associated with metabolic elimination were further analysed using multiple logistic regression analysis. Dichotomisation for each variable was made using the receiver-operating characteristic curve to enter into the model.

Results: PCV>52%, PP>82 g/l, standardised Cr>30 mg/l 100 km, UA>72 mg/l, standardised CK>12.6 ui/l km and standardised AST>6.2 ui/l km were associated with the development of metabolic alterations. Of the horses with an imbalance between PCV and PP, 30% had metabolic elimination in the following phases. Muscle enzymes and Cr were directly related to the distance covered.

Conclusions: Selected biochemical markers are evident in some endurance horses before their elimination. However, most horses developed metabolic disturbances without any important alterations in the variables determined in this study.

Potential relevance: Analysis of selected plasma biochemical parameters could be useful in the prevention and early diagnosis of metabolic stress in endurance horses.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Conflicts of interest
  10. Manufacturers’ addresses
  11. References

The sport of endurance riding is a competitive ride taking place over 80–160 km, divided into various phases, where the winner is the horse and rider team who successfully complete the ride in the shortest time. Competition regulations require that veterinary judges protect the health and welfare of the horses, based on the measurement of heart rate and the assessment of some clinical signs related to hydration status. However, so far, the accuracy of the score system used to evaluate hydration in endurance horses in competition has not been proven.

When detected, the condition must be monitored and when deterioration develops, the horse's welfare must be protected by removing it from the competition. Heart rate measurement appeared to be a reliable indicator of the metabolic status of endurance horses, suggesting that veterinary examinations, according to the official Fédération Équestre Internationale (FEI) rules, are adequate to protect the health and welfare of horses competing in endurance races (Sloet van Oldruitenborgh-Oosterbaan et al. 1991). However, the thin line between a horse eliminated for metabolic reasons and a fit horse is not always clear and disputes occasionally arise, as well as unnecessary medical treatments in some cases and the development of life-threatening diseases that could have been detected earlier in others.

Of all equine competitions, endurance races have the greatest metabolic demands for the sport horse, requiring substantial energy production for many hours (Treiber et al. 2006). Such sustained energy demands cause the cardiorespiratory, endocrine and neuromuscular systems to operate at an elevated level for an unnatural length of time (Flaminio and Rush 1998; Schott et al. 2006). During races, some horses show homeostasis changes, such as energy depletion and alterations in fluids, electrolytes and acid-base balance, with negative consequences on both the performance and the health of the horse, which might show significant changes in diverse biochemical parameters (Flaminio and Rush 1998; Castejon et al. 2006; Fielding et al. 2009).

The results of various blood analyses on horses competing in different distances have been reported (Grosskopf and van Rensburg 1983; Rose et al. 1983; Sloet van Oldruitenborgh-Oosterbaan et al. 1991; Lindinger and Ecker 1995; Schott et al. 1997, 2006; Barton et al. 2003; Castejon et al. 2006), and a reasonably clear picture can be obtained from the changesthat occur in horses performing in endurance exercises. However, the information concerning horses eliminated due to metabolic conditions is not so extensive (Fielding et al. 2009).

The objectives of this study were to evaluate the utility of biochemical analysis in the early diagnosis of metabolic stress in endurance horses and to determine the cut-off values of biochemical parameters as predictors of elimination for metabolic reasons in horses subjected to a prolonged effort.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Conflicts of interest
  10. Manufacturers’ addresses
  11. References

Animals

This retrospective study involved 283 horses competing in 36 endurance races (15*, 9** and 12***CEI level). Races followed FEI rules and were predominantly held in Spain, but also in France, Portugal and Germany. Characteristics of the races used in this study are summarised in Table 1.

Table 1. Characteristics of the races used in this study
 CEI*CEI**CEI***
Races (n)15912
Average speed (km/h range)13.5–19.213.8–18.611–17.1
Length (km - range)72–86110–124151–169
All horses at competitions   
 Total horses (n)531255364
 Eliminated due to metabolic condition (n)422836
Sampled horses   
 Total of sampled horses (n)10374106
 Eliminated due to metabolic condition (n)131018

The animals were selected after consent was obtained from the owners or riders of the horses for enrolling in the research. Of the 283 horses, 242 concluded the race adequately and did not show any disease associated with the exercise during the following days (successful, S), while 41 horses were eliminated because of metabolic disturbances during the race and needed medical treatment (metabolic, M).

The diagnoses of the veterinary treatments were: lack of recovery of the heart rate (n = 11 horses), exhausted horse syndrome with clinical signs of exhaustion without pathognomonic signs of disease (n = 7), myopathies (n = 7), synchronous diaphragmatic flutter (n = 6), cardiac arrhythmia (n = 3), colic (n = 3), hypovolaemic shock (n = 2), diarrhoea (one horse) and central nervous symptoms (one horse). All the horses needed medical assistance, but samples in all cases were taken before any treatment. Two M horses died in the following days.

Horses that were eliminated because of metabolic reasons without medical treatment, eliminated because of lameness and horses eliminated because of metabolic reasons in the first phase, were all excluded from this study.

Sample collection

Jugular blood samples (10 ml) were taken before competition, immediately after the mandatory veterinary examination (vet-gate) of each phase and immediately after finishing the race in the successful horses (S horses) or after veterinary disqualification for metabolic reasons (M horses).

Packed cell volume (PCV) was measured with the microhaematocrit method. Blood samples were then placed on heparinised tubes, plasma was obtained by centrifugation within the first 10 min after collection and was immediately refrigerated (4–8°C) during the competitions and transport for later analysis within the following 48 h.

Determination of activities of creatine kinase (CK), aspartate aminotransferase (AST) and lactate dehydrogenase (LDH) and concentrations of uric acid (UA), urea, creatinine (Cr), plasma lactate (L) and total plasma proteins (PP) were performed by enzymatic methods with spectrometry1 using reagents designed for each determination2,3.

Statistical analysis

Data are summarised as means ± s.d. Normality was tested using the Shapiro-Wilks test. Changes amongst the groups were evaluated with ANOVA. A post hoc Fisher's protected LSD test was performed to test for differences between means.

The S horses were compared according to the length of the races (80, 120 and 160 km) in order to check differences due to the distance, and to identify distance related variables. When observed, Pearson correlation coefficient was used to assess the relationship and appropriate standardisations (St variable = variable/covered distance for enzymes and St Cr/100 km = Cr x 100/covered distance) were made for further analysis.

The S horses (combined data obtained at the end of the competitions) were compared with the M horses after elimination (M Post) and with the values obtained in the previous phase before being removed from the competition (M Pre). Appropriate transformations were assessed for continuous variables in an attempt to improve model fit. Dichotomisation was made using the receiver-operating characteristic curve for each variable, being the cut-off value with the highest sensitivity + specificity value. Specificity, sensitivity, positive (PPV) and negative (NPV) predictive values, and likelihood ratios were calculated to estimate the prediction ability conferred by each variable. Factors associated with metabolic elimination were further analysed using multiple logistic regression analysis. Variables with a P value <0.20 were then entered into the logistic regression model. Excluded variables were re-investigated for any confounding effect by single introductions into the preliminary start model before final exclusion (Hosmer and Lemeshow 2000).

Relative speed (RS) of the last phase was included as an independent variable in these steps to control this important risk factor throughout the model-building process. As phase speed is related to many factors (terrain, length, weather, etc.), a conversion factor to determine RS was used as follows: RS = (individual speed of the phase - individual speed of the previous phase) - (average speed of the phase - average speed of the previous phase). Positive values were considered as high relative speed (HRS dichotomic variable).

The SPSS4 statistical software package version 12.0 was used for all of the statistical analyses. The level of significance was set at P<0.05.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Conflicts of interest
  10. Manufacturers’ addresses
  11. References

The prevalence of horses eliminated due to metabolic conditions (including all horses competing as well as horses that were not treated medically) was 8, 11 and 10% for 80, 120 and 160 km, respectively (Table 1). Eight horses eliminated due to metabolic conditions did not require medical treatment, and 7 horses were eliminated in the first phase and were therefore also eliminated from the study. Two of the M horses died as consequence of their metabolic condition in the days following the race.

Results from PCV, PP, Urea, Cr, L, UA, CK, AST and LDH in S horses during different length races are shown in Table 2. CK, AST, LDH and Cr had significant differences in all distance groups, and are directly related with the distance covered. These variables were then standardised (St) to each distance for further analysis.

Table 2. Biochemical variables after 80, 120 and 160 km in successful endurance horses
 80 km (n = 90)120 km (n = 64)160 km (n = 88)r
  1. r, Pearson's correlation coefficient between increment of each variable and race length. Different letters mean differences among groups.

PCV (%)44.3 ± 7a45.7 ± 446.4 ± 5b 
PP (g/l)75 ± 4a77 ± 4b77 ± 5b 
Urea (mg/l)360 ± 40a390 ± 50b400 ± 50b 
Cr (mg/l)17 ± 2a20 ± 3b22 ± 3c0.42 (P = 0.015)
L (mmol/l)1.8 ± 0.32.1 ± 0.62.0 ± 0.7 
UA (mg/l)42 ± 5.0a45 ± 8.0b47 ± 6.0b 
CK (ui/l)683 ± 165a911 ± 204b1160 ± 296c0.64 (P<0.001)
AST (ui/l)315 ± 63a475 ± 101b682 ± 146c0.81 (P<0.001)
LDH (ui/l)512 ± 95a630 ± 166b818 ± 237c0.73 (P<0.001)

Post exercise plasma biochemistry from S horses (combined data), M Post horses and M Pre horses are compared in Table 3. The following parameters showed significant higher means in M Post horses in comparison to S horses: PCV, PP, UA, standardised CK, standardised AST and standardised LDH. In comparison with S horses, M Pre horses had significantly higher PP. When comparing S horses and M Pre horses, PCV, PP, UA, St CK and St AST showed a P value <0.20, and were then selected for entry into the logistic regression model.

Table 3. Biochemical variables in successful endurance horses at the end of the competitions (combined data of the different types of competitions), M Post (metabolic horses after elimination and M Pre (metabolic horses before elimination)
 Successful horses (n = 242)M PostM PreP
  1. P, significance value between successful and M Pre. Different letters mean differences among groups.

PCV %45.2 ± 7a47.1 ± 7b46.4 ± 50.18
PP g/l76 ± 5a79 ± 6b78 ± 7b0.04
Urea mg/l370 ± 60360 ± 50370 ± 50>0.20
St Cr mg/l 100km17 ± 2a21 ± 6b18 ± 50.15
L mmol/l1.9 ± 0.62.0 ± 0.62.0 ± 0.7>0.20
UA mg/dl45 ± a58 ± 8b51 ± 60.06
St CK ui/l km7.5 ± 1.5a9.1 ± 3.9b8.3 ± 2.80.06
St AST ui/l km4.1 ± 0.8a4.8 ± 1.8b4.5 ± 1.80.17
St LDH ui/l km5.5 ± 1.8a6.1 ± 2.6b5.6 ± 1.9>0.20

Likelihood ratios, sensitivity and specificity for each biochemical variable (dichotomised) to predict the development of metabolic alterations (univariate analysis) are shown in Table 4.

Table 4. Likelihood ratios, sensitivity and specificity of individual dichotomised variables
 SensitivitySpecificityPositive likelihood ratioPositive likelihood ratio
  1. Numbers within parentheses indicate 95% CI.

PCV >52%0.34 (0.21–0.49)0.87 (0.82–0.91)2.81 (1.6–4.73)0.75 (0.6–0.94)
PP >82 g/l0.39 (0.24–0.55)0.90 (0.85–0.93)3.94 (2.31–6.8)0.67 (0.52–0.86)
St Cr >30 mg/l km0.21 (0.11–0.38)0.92 (0.88–0.95)3.12 (1.49–6.53)0.83 (0.71–0.99)
UA >72 mg/l0.36 (0.22–0.53)0.91 (0.86–0.94)4.21 (2.37–7.48)0.69 (0.54–0.87)
St CK >12.6 ui/l km0.24 (0.13–0.39)0.78 (0.73–0.83)1.15 (0.63–2.08)0.95 (0.79–1.15)
St AST >6.2 ui/l km0.14 (0.06–0.29)0.81 (0.75–0.85)0.78 (0.35–1.72)1.04 (0.91–1.2)

Multicollinearity was assessed by examining the associations among the biochemical markers to be included in the multiple logistic regression analysis. There were high correlation coefficients between PCV and PP, CK and AST, and CK and LDH (0.52, 0.41 and 0.45, respectively). Therefore, PCV, AST and LDH were not included in the multiple logistic regression analysis. Moderate to low correlations were found between UA and CK and UA and PP (0.17 and 0.23, respectively).

The results of the logistic regression analysis are shown in Table 5. The parameters CK, UA, HRS and PP made significant independent contributions to the development of metabolic alterations, whereas St Cr did not contribute significantly to the model.

Table 5. Risk factor categories for development of metabolic elimination
 PPVNPVSensitivitySpecificity
  1. PPV, positive predictive value; NPV, negative predictive value. Numbers in parentheses indicate 95% CI.

CK + UA0.63 (0.31–0.87)0.87 (0.82–0.91)0.17 (0.07–0.32)0.98 (0.95–0.99)
CK + UA + HRS0.85 (0.42–0.99)0.87 (0.82–0.90)0.14 (0.06–0.29)0.99 (0.97–0.99)
PP + HRS0.64 (0.38–0.84)0.88 (0.84–0.92)0.26 (0.14–0.43)0.97 (0.94–0.98)

In some cases (n = 13), we found a marked imbalance between PCV and PP. These horses showed PCV >52% and PP <73 g/l. Four of the 13 developed metabolic conditions during the competition (laminitis, cardiac arrhythmias and 2 colic cases) and were eliminated, 3 were withdrawn from the competition, while the remaining 6 completed the race.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Conflicts of interest
  10. Manufacturers’ addresses
  11. References

To the authors’ knowledge, this is the largest study to examine the biochemical findings in horses eliminated due to metabolic conditions and in comparison with those that successfully completed the competitions.

We excluded horses eliminated in the first phase, because in these cases the data prior to elimination were the resting values. We excluded horses eliminated due to metabolic conditions but untreated, because these metabolic diagnoses were considered doubtful.

A wide range of biochemical perturbations occur during endurance riding. These results confirm previous individual studies on biochemical and haematological variables after endurance rides. Some of these changes are exercise induced and should not be confused with indicators of disease (Rose et al. 1983; Foreman 1998). Differences related to the distance were also described by other authors (Lindinger and Ecker 1995; Hoffman et al. 2002; Barton et al. 2003). Some variables (PCV, PP, urea and UA) showed significant differences between 80 and 120 km, but not between 120 and 160 km, suggesting that changes in biochemical parameters related to distance continue up to 120 km races but reach a plateau thereafter. This is coherent if it is taken into account that the distance between 80 and 120 is proportionally larger than between 120 and 160 km.

Differences in several biochemistry variables between horses eliminated due to metabolic conditions and successful horses have been described previously (Grosskopf and van Rensburg 1983). In a recent study, Fielding et al. (2009) found minor differences in laboratory parameters (Na+, K+, Cl-, HCO3-, lactate, PCV, total proteins) when endurance horses that were eliminated were compared with those that successfully completed the competitions. In this paper, sampling time is not controlled. If the blood samples were taken when the horses arrived at a veterinary hospital, the laboratory values found could have been significantly different from those obtained immediately after elimination in the place of the endurance event (Grosskopf and van Rensburg 1983). In addition, the most important cause of elimination described by Fielding et al. (2009) was colic, whereas we found that the lack of cardiovascular recovery was the primary elimination cause related to metabolic conditions. It seems plausible to think that those horses that required additional diagnostic and treatment techniques, not applicable in field conditions, such as colic cases that did not respond to proper and effective fluid therapy, were probably referred to an intensive facility. In fact, according to Fielding et al. (2009), horses with colic showed mild biochemical alterations when compared with horses with poor recovery.

Grosskopf and van Rensburg (1983) suggested that values of PCV >55%, PP >90 g/l, and L >3 mmol/l are risk indicators and therefore reasons to eliminate endurance horses due to metabolic conditions. In our study, in agreement with other authors (Castejon et al. 2006; Fielding et al. 2009), L did not differ in M horses. However, in a recent study, L concentrations were higher in M than in S horses covering different distances (Muñoz et al. 2010). Grosskopf and van Rensburg (1983) sampled endurance horses in a ‘stop-and-go ride’ where the rider with a sound horse could depart from the checkpoints without the compulsory rest periods. As L values during endurance races are very close to maximal lactate steady state, resting periods could facilitate its clearance (Jones et al. 2010), preventing a possible accumulation in horses at risk of developing metabolic alterations. In the current study, L was measured in heparinised samples instead of using sodium fluoride tubes in order to inhibit glycolytic pathway and L production by the red blood cells. However, blood samples were centrifuged as soon as possible and red blood cells were separated. The same procedure was followed in all the samples and therefore it does not seem plausible that this would have influenced the comparison between S and M horses, although caution should be taken when comparing these values of L with those of other authors using different anticoagulants.

The physiological significance of the imbalance between PCV and PP is not easy to explain. Some of these horses developed severely painful metabolic conditions (colic, laminitis), so we can suspect that moderate pain was present in the previous phase. However, it is hard to believe that these horses were able to recover their heart rate and pass the vet-gate with moderate pain (Friend 2000; Foreman et al. 2008).

There was a correlation between PCV and PP. Both are indicators of dehydration. However PCV is influenced also by stress (Friend 2000). In this study, both values had high prediction ability. However PP had higher specificity and sensitivity and therefore, higher PP concentrations seem to be better predictor of metabolic alterations in endurance horses in competition.

The observed correlations between CK, AST and LDH were as expected. Moderate correlations between UA and CK and UA and PP were mentioned previously (Castejon et al. 2006).

Some laboratorial alterations were associated with the development of metabolic alterations: PCV >52%, PP >82 g/l, St Cr >30 mg/l km, UA >72 mg/l, St CK >12.6 ui/l km, St AST >6.2 ui/l km. There were strong correlations between PCV and PP, CK and AST, and CK and LDH and low correlations between UA and CK and UA and PP. The rest of variables were independently associated with the development of metabolic alterations. These results are consistent with our current understanding of the pathophysiology of exhausted horse syndrome. It could be suggested that the biochemical markers reflect different pathogenic mechanisms with independent contributions to the development of metabolic alterations (Rose et al. 1983; Foreman 1998; Barton et al. 2003).

The results of this analysis indicate that increased PP, St Cr, UA and St CK, alone or combined, helped to identify which horses have an increased risk of developing metabolic alterations. The specificity of each of these biochemical markers taken alone was limited (<0.90). Combined, however, the specificity value increased considerably, with an acceptable NPV and PPV. HRS significantly improves specificity of the model. However this variable could not be known before the last phase and thus has no practical use. There is, however, a clear demonstration of the involvement of the rider and the development of metabolic alterations in their horse.

The blood parameters analysed in the present study, unfortunately, were not measured in field conditions. Samples were transported and analysed in the laboratory. However, there are portable blood analysis systems able to quickly and accurately perform blood analysis in field situations (Tschudi 1998). If a careful interpretation of the data reported in our study is made, our findings will be important to assist veterinarians in their judgement during endurance competitions. Most horses developed significant metabolic disturbances without any important alterations in the variables determined in this study. However, we have to keep in mind that, although all these horses were considered fit to continue in the vet-gate when the samples were taken, they were eliminated in the next vet-gate or during the next loop and all of them required medical treatment. Although the sensitivity of these biochemical markers was low, they were able to detect 1–4 of every 10 horses with very high specificity (>0.94), whereas the physical examination failed to detect any of these horses at this time.

Another limitation in this study and its possible application is the fact that sensitivity, specificity, likelihood ratios, PPV and NPV of a test are dependent on disease prevalence (in this case % of horses eliminated due to metabolic conditions). Several studies have revealed substantial variation of these measures in different populations for the same test (Levy et al. 1990; van der Schouw et al. 1995).

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Conflicts of interest
  10. Manufacturers’ addresses
  11. References

The results of this study suggested that biochemical findings, alone or combined, could help identify horses with an increased risk of developing metabolic alterations before their elimination during competitive endurance races of different distances.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Conflicts of interest
  10. Manufacturers’ addresses
  11. References

The authors thank all the riders who allowed their horses be sampled during the competition, and Elizabeth Plowright for reviewing the language of the manuscript.

Manufacturers’ addresses

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Conflicts of interest
  10. Manufacturers’ addresses
  11. References

1 Quick Lab, Hameln, Germany.

2 BioSystems S.A., Barcelona, Spain.

3 Roche Diagnostics Corp, Indianapolis, Indiana, USA.

4 SPSS, Chicago, Illinois, USA.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Conflicts of interest
  10. Manufacturers’ addresses
  11. References
  • Barton, M.H., Williamson, L., Jack, S. and Norton, N. (2003) Body weight, hematologic findings, and serum and plasma biochemical findings of horses competing in a 48-, 83-, or 159-km endurance ride under similar terrain and weather conditions. Am. J. vet. Res. 64, 746-753.
  • Castejon, F., Trigo, P., Muñoz, A. and Riber, C. (2006) Uric acid responses to endurance racing and relationships with performance, plasma biochemistry and metabolic alterations. Equine vet. J. 36, 70-73.
  • Fielding, C.L., Magdesian, K.G., Rhodes, D.M., Meier, C.A. and Higgins, J.C. (2009) Clinical and biochemical abnormalities in endurance horses eliminated from competition for medical complications and requiring emergency medical treatment: 30 cases (2005–2006). J. vet. emerg. crit. Care 19, 473-478.
  • Flaminio, M.J. and Rush, B.R. (1998) Fluid and electrolyte balance in endurance horses. Vet Clin. N. Am.: Equine Pract. 14, 147-158.
  • Foreman, J.H. (1998) The exhausted horse syndrome. Vet. Clin. N. Am.: Equine Pract. 14, 205-219.
  • Foreman, J.H., Barange, A., Lawrence, L.M. and Hungerford, L.L. (2008) Effects of single-dose intravenous phenylbutazone on experimentally induced, reversible lameness in the horse. J. vet. Pharmacol. Therap. 31, 39-44.
  • Friend, T.H. (2000) Dehydration, stress, and water consumption of horses during long-distance commercial transport. J. anim. Sci. 78, 2568-2580.
  • Grosskopf, J.F.W. and Van Rensburg, J.J. (1983) Some observations on the haematology and blood biochemistry of horses competing in 80 km endurance rides. In: Equine Exercise Physiology, Eds: D.H.Snow, S.G.B.Persson and R.J.Rose, Granta Editions, Cambridge. pp 425-431.
  • Hoffman, R.M., Hess, T.M., Williams, C.A., Kronfeld, D.S., Griewe-Crandell, K.M., Waldron, J.E., Graham-Thiers, P.M., Gay, L.S., Splan, R.K., Saker, K.E. and Harris, P.A. (2002) Speed associated with plasma pH, oxygen content, total protein and urea in an 80 km race. Equine vet. J. 34, 39-43.
  • Hosmer, D.W. and Lemeshow, S. (2000) Applied Logistic Regression, 2nd edn., Wiley, New York.
  • Jones, A.E., Shapiro, N.I., Trzeciak, S., Arnold, R.C., Claremont, H.A., Kline, J.A. and Emergency Medicine Shock Research Network (EMShockNet) Investigators (2010) Lactate clearance vs. central venous oxygen saturation as goals of early sepsis therapy: A randomized clinical trial. J. Am. med. Ass. 303, 739-746.
  • Levy, D., Labib, S.B., Anderson, K.M., Christiansen, J.C., Kannel, W.B. and Castelli, W.P. (1990) Determinants of sensitivity and specificity of electrocardiographic criteria for left ventricular hypertrophy. Circulation 81, 815-820.
  • Lindinger, M.I. and Ecker, G.I. (1995) Ion and water loss from body fluids during a 163 km endurance ride. Equine vet. J. 18, 314-322.
  • Muñoz, A., Riber, C., Trigo, P., Castejon-Riber, C. and Castejon, F.M. (2010) Dehydration, electrolyte imbalances and renin-angiotensin-aldosterone-vasopressin axis in successful and unsuccessful endurance horses. Equine vet. J., Suppl. 38, -.
  • Rose, R.J., Hodgson, D.R., Sampson, D. and Chan, W. (1983) Changes in plasma biochemistry in horses competing in a 160 km endurance ride. Aust. vet. J. 60, 101-105.
  • Schott, H.C. 2nd, McGlade, K.S., Molander, H.A., Leroux, A.J. and Hines, M.T. (1997) Body weight, fluid, electrolyte, and hormonal changes in horses competing in 50- and 100-mile endurance rides. Am. J. vet. Res. 58, 303-309.
  • Schott, H.C., 2nd, Marlin, D.J., Geor, R.J., Holbrook, T.C., Deaton, C.M., Vincent, T., Dacre, K., Schroter, R.C., Jose-Cunilleras, E. and Cornelisse, C.J. (2006) Changes in selected physiological and laboratory measurements in elite horses competing in a 160 km endurance ride. Equine vet. J., Suppl. 36, 37-42.
  • Sloet van Oldruitenborgh-Oosterbaan, M.M., Wensing, T., Barneveld, A. and Breukink, H.J. (1991) Heart rate, blood biochemistry and performance of horses competing in a 100 km endurance ride. Vet. Rec. 128, 175-179.
  • Treiber, K.H., Hess, T.M., Kronfeld, D.S., Boston, R.C., Geor, R.J., Friere, M., Silva, A.M. and Harris, P.A. (2006) Glucose dynamics during exercise: dietary energy sources affect minimal model parameters in trained Arabian geldings during endurance exercise. Equine vet. J., Suppl. 36, 631-636.
  • Tschudi, P.R. (1998) Evaluation of the portable blood analyser i-STAT. Schweiz. Arch. Tierheilkd. 140, 507-512.
  • Van der Schouw, Y.T., Van Dijk, R. and Verbeek, A.L.M. (1995) Problems in selecting the adequate patient population from existing data files for assessment studies of new diagnostic tests. J. clin. Epidemiol. 48, 417-422.