• horse;
  • exercise;
  • equine exercise physiology;
  • lung;
  • haemorrhage;
  • EIPH;
  • exercise;
  • risk factor


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

Reasons for performing study: Risk factors for occult exercise-induced pulmonary haemorrhage (EIPH) are poorly defined or quantified.

Objectives: To investigate the importance of putative risk factors for EIPH amongst Thoroughbred racehorses in Australia.

Methods: Tracheobronchoscopy was used to determine EIPH status of 744 Thoroughbred racehorses after flat racing in Melbourne, Australia. Horses were identified for study before racing, and over 50% of horses racing during the study period were examined. Statistical analysis included use of bivariate and multivariate logistic regression analysis to account for simultaneous effects of a large number of variables.

Results: The only risk factor identified as associated with both EIPH ≥1 or ≥2 was ambient temperature, with horses racing at temperatures <20°C being at ∼ 2 times risk of occult EIPH. There was no association of EIPH with age, sex, weight carried, track hardness, speed of racing, or air quality.

Conclusions: There do not appear to be individual risk factors, amongst those examined in this study, that are strongly associated with EIPH.

Potential relevance: The risk of developing EIPH cannot be readily determined from a combination of age, race speed, race distance, track hardness or air quality. This study does not provide support for the hypotheses that racing on hard surfaces or in polluted air contributes to the development of EIPH.


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

Exercise-induced pulmonary haemorrhage (EIPH) occurs commonly in Thoroughbred and Standardbred racehorses throughout the world. Although estimates of the incidence of EIPH vary depending on the population of horses examined, the diagnostic method and the frequency of examination, blood can be detected by means of tracheobronchoscopic examination of the airways in >50% of Thoroughbred horses after a race (Pascoe et al. 1981; Hinchcliff et al. 2005a). Despite the high prevalence of the condition, variables that describe characteristics of the horse, track or environment associated with EIPH have not been well defined.

Studies reporting the assessment of risk factors for EIPH have often used epistaxis as a marker for the disorder (Takahashi et al. 2001; Weideman et al. 2003). However, epistaxis is commonly a marker for the most severe form of the disorder and, as such, results of studies that identify risk factors for epistaxis do not provide an assessment of occult EIPH (EIPH that can only be detected by tracheobronchoscopy or other invasive diagnostic techniques) (Hinchcliff et al. 2005a). There are few studies of the risk factors for EIPH detected by endoscopic examination (Pascoe et al. 1981; Speirs et al. 1982; Raphel and Soma 1982; Lapointe et al. 1994; Costa and Thomassian 2006) and those that exist have flaws in experimental design including low number of animals, nonrandom selection of study animals, or incorrect or incomplete statistical analysis of the data.

Determination of risk factors for EIPH might identify factors that can be altered in such a way as to minimise the severity or frequency of EIPH. This information could be used in the absence of a detailed understanding of the mechanisms underlying development of the disorder, which is currently the situation. Therefore, our hypothesis was that there are factors associated with horse demographics and performance, track characteristics, air quality and weather that are associated with the risk of horses having EIPH. The objectives of this study were to determine the prevalence and severity of EIPH in Thoroughbred racehorses and to investigate the association between these measures of EIPH and putative risk factors.

Materials and methods

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

Experimental protocol

The study was designed as a cross-sectional study of a convenience sample of horses racing at racetracks in Melbourne, Australia between March and June 2003, whose owners and trainers agreed to participate. Thoroughbred racehorses were enrolled 24–48 h prior to racing and were evaluated by tracheobronchoscopic examination after racing for evidence of EIPH. Data about putative risk factors were then abstracted from the appropriate databasesand associations with the presence of EIPH examined. Details of the study design are published elsewhere (Hinchcliff et al. 2005a,b).


Horses enrolled in the study were Thoroughbred racehorses of either sex competing in flat races at one of 4 racecourses in metropolitan Melbourne, Australia, between 1st March and 18th June 2003. All races were on turf and took place between 11.30 and 23.30 h. Administration of medications, including frusemide, was not permitted on the day of the race and this rule was stringently enforced by application of state-of-the-art drug testing procedures for detection of therapeutic and prohibited substances in blood and urine samples. Use of nasal dilator strips was not permitted. It is therefore unlikely that horses in this study were administered agents that could have affected performance or development or severity of EIPH.

To control for any potential enrolment bias, horses included in the study were identified before racing. Prior to enrolment of any horses in the study, information regarding the study was distributed to trainers and owners of Thoroughbred racehorses by means of facsimile transmission to all registered trainers in the state, publication of articles in trade newsletters and newspapers, broadcast interviews on radio and television, live presentations at the race tracks, and personal contacts with influential trainers. Horses to be studied on a particular race day were identified 24–48 h before the race. Lists of horses accepted to race were obtained from the body governing racing in this jurisdiction (Racing Victoria Ltd), and trainers of eligible horses were contacted by telephone to request permission to examine horses. The risks and benefits of the study were explained and verbal permission to examine horses was obtained. On the day of the race, the study was again discussed with the trainers, horses were visually identified by one of the investigators and written informed consent was obtained before the horse raced. After the race, horses were brought by their handlers to a central location at the racetrack and tracheobronchoscopic examination performed within 2 h after racing.

Detection and quantification of EIPH

All horses underwent tracheobronchoscopic examination for evidence of EIPH. Briefly, an endoscope (Model CF-100TL [1.7 m in length and 1.1 cm in diameter])1 was passed through one of the nares, and the nasopharynx, larynx and trachea to the level of the carina were examined. Horses were not sedated for this procedure and all examinations were recorded on videotape for subsequent analysis. Three individuals, blinded to the identity of horses and their race performance, independently reviewed the videotapes and recorded their categorical assessments without discussing their observations with each other. Severity of EIPH was graded on a scale from 0–4 as described (Hinchcliff et al. 2005a). The presence of tracheal or pharyngeal mucus or dirt was recorded and graded as: 0) no dirt or mucus detected; 1) 1–5 flecks of mucus or dirt; 2) >5 flecks of mucus or particles of dirt, but which are not confluent; 3) >5 flecks or strands of mucus or particles of dirt, which frequently abut other dirt particles or flecks of mucus; and 4) multiple, long strands of mucus with pooling of mucus at the thoracic inlet, abundant dirt in clumps and distributed around the circumference of the airway.

Race records

Race records for horses included in the study were retrieved from a commercial database (i-RIS)2. Variables recorded on the day of study, abstracted from the database or obtained from the Bureau of Meterology3 or Environment Protection Authority4 included: horse name, age, sex, trainer, jockey; race date and time; race track, distance, purse; weight carried; whether the horse finished the race; finishing position; finishing time of the winner; margin or distance finished behind the winner; speed rating for the race; number of horses in the race; days since last race; earnings for this race; lifetime earnings prior to this race; lifetime starts prior to this race; lifetime wins prior to this race; lifetime third-place finishes prior to this race. Information obtained about the weather during the 24 h preceding the race and at the time of race included ambient temperature, humidity, rainfall, wind speed and wind direction. Indicators of air quality at the time of racing that were recorded were particulate count (µg/m3), and concentrations of sulphur dioxide, nitrogen dioxide and carbon monoxide (all in parts per billion). The air quality data were determined at the measuring station nearest to the racetrack, which was <5 km in all instances.

Data analysis

Risk factors for EIPH were determined by multivariable logistic regression using the dependent variable of either EIPH of Grade 0 vs. ≥1 or EIPH ≤1 vs. EIPH ≥2. For the initial analysis, each variable was considered individually while controlling for track as a fixed effect. Variables were initially examined in a bivariable analysis and those variables with a P≤0.250 were examined using backwards logistic regression for repeated measures to develop a final multivariable model using SAS Proc Genmod5. Repeated measures controlled in the analysis were track and race date. Variables were categorised before analysis because of the non-Gaussian distribution of the data. Track, age and sex were included in the final modelling as we believed that there were biological reasons, or evidence from previous studies, that these variables could be associated with the risk of EIPH. Variables were included in the final model if P<0.05. Odds ratios (ORs) and 95% confidence intervals (CIs) derived from likelihood ratio statistics were calculated from the logistic regression models.


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

Tracheobronchoscopic examinations were performed on 744 horses competing in 202 races at 26 race meetings at 4 racetracks. Horses were from the stables of 214 trainers, with no trainer contributing more than 41 horses (median, 2 horses, range, 1–41 horses) or 5.5% of the total number of horses examined. During the period of study, there were 2396 race starts by 1428 horses in flat races at the race meets during which horses were examined. Mean ± s.d. number of horses in each race was 11.9 ± 2.5 horses. Overall, 52.1% of horses eligible for participation in the study were examined. Horses examined ranged in age from 2–10 years (median, 4 years) and included 306 females, 375 geldings and 63 sexually intact males.

Forty variables describing characteristics of the horse, track and environment were examined of which 16 were tested in the multivariable model for presence of any EIPH (EIPH ≥1) and 8 were tested in the multivariable model for presence of EIPH ≥2 (Tables 1–3). There were 412 horses with EIPH ≥1, and 139 had EIPH ≥2.

Table 1. Summary statistics for variables examined in 744 Thoroughbred racehorses for association with EIPH
VariableMeanMedianLower 5%Upper 5%
Race-examination time difference (min)31301946
Weight carried (kg)54.25451.557.5
Number of starters11.912816
Winner's speed (m/s)16.5116.5317.3215.51
Sectional time (s per final 600 m)35.3635.2933.2737.67
Race distance (m)1,4721,40010002,400
Track conditionN/AN/AN/AN/A
Wind speed (knots)10.711418
Dry bulb temperature (°C)17.91812.922.9
Relative humidity (%)63.464.938.1283.4
Pressure (hPa)1020.11019.91010.71034.4
No2 (ppb)10.077428
So2 (ppb)0.22001
Co (ppm)0.350.312500.96
Pm10 [particles <10 µm in size] (µg/m3)16.513.57.533
Age (years)4.2427
Sex (M, G, F)N/AN/AN/A 
Lifetime starts18.714249
Lifetime firsts3.2308
Lifetime earnings86,36544,550715283,746
Days since last start49.5177195
Days since second to last start90.34220259
Age at first start (years)
No. starts as 2-year-old1.6106
Pharyngeal dirt0.77003
Mucus score1.5203
Tracheal dirt0.8003
Table 2. Variables identified by bivariable analysis for inclusion in final multivariable modeling for EIPH ≥1
VariableCategoriesOdds ratioLower 2.5%Upper 2.5%P valueP value
Race-examination time difference (min)0.0004
Weight carried (kg) 0.960.921.01 0.19
No. of starters0.055
Winner's speed (m/s)0.041
Sectional time (s)0.13
Race distance (m)0.15
Track condition0.17
Penetrometer (mm)0.12
Dry bulb temperature (°C)0.03
PM10 [particles <10 µm in size] (µg/m3)0.036
 Data missing0.770.471.240.28 
Age (years)0.35
Lifetime starts0.055
No. starts as 2-year-old0.157 
Pharyngeal dirt0.049
Mucus score0.059
Tracheal dirt0.06
Table 3. Variables identified by bivariable analysis for inclusion in final multivariable modeling for EIPH ≥2
VariableCategoriesOdds ratioLower 2.5%Upper 2.5%P valueP value
Race-examination time difference (min)0.006
Number of starters0.06
Race distance (m)0.07
Dry bulb temperature (°C)0.008
Age (years)0.2
Race prize money - total purse (Aus$)0.037
Tracheal dirt0.03

After accounting for confounding factors and adjusting for factors that could influence the risk of EIPH, horses with >50 lifetime starts were 1.8 times as likely to have EIPH ≥1 than were horses with <40 lifetime starts (P = 0.015). Horses racing in ambient temperatures <20°C were 1.9 times more likely to have EIPH ≥1 than horses racing at warmer temperatures (P = 0.014). Horses with greater amounts of dirt in the trachea were half as likely to be detected as having EIPH ≥1 (P = 0.004). The longer the period of time after racing, up to 60 min, when horses were examined increased the likelihood that they would have EIPH ≥1 (P = 0.0003)) (Table 4).

Table 4. Variables retained in final model for occurrence of EIPH ≥1
VariableOdds ratio (95% CI)P value
Life time starts 0.015
 ≥501.78 (1.11–2.84)0.0164
 40–491.35 (0.71–2.60)0.3625
Ambient temperature (°C) 0.014
 <201.85 (1.35–2.54)0.0001
Tracheal dirt (category) 0.004
 0–10.45 (0.29–0.70)0.0004
Racing–examination period (min) 0.0003
 30–592.13 (1.68–2.70)<0.0001
 60–1201.72 (0.69–4.26)0.2419

Horses racing 1400–2400 m were less likely to have EIPH ≥2 than those racing <1400 m (P = 0.03). Horses with no lifetime earnings were 20% as likely to have EIPH ≥2 than were horses with lifetime earnings of $1–149,000 (P = 0.008). Presence of dirt in the trachea was associated with a decreased likelihood of EIPH ≥2 (P = 0.025), and racing at ambient temperature <20°C was associated with a 2-fold increase in the risk of EIPH ≥2 (P = 0.002). The period between finishing racing and examination (P = 0.012) was associated with EIPH ≥2 (Table 5).

Table 5. Variables retained in final model for occurrence of EIPH ≥2
VariableOdds ratio (95% CI)P value
Race distance (m) 0.03
 ≥24001.10 (0.52–2.33)0.803
 1400–24000.72 (0.54–0.96)0.0261
Lifetime earnings ($) 0.008
 00.19 (0.05–0.77)0.0202
 ≥$150,0000.72 (0.41–1.24)0.2358
Tracheal dirt (category) 0.025
 0–11.86 (1.10–3.15)0.0214
Ambient temperature (°C) 0.002
 <201.98 (1.46–2.68)<0.0001
Racing–examination period (min) 0.012
 60–1201.52 (0.42–5.52)0.5234
 30–592.04 (1.35–3.10)0.0008


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

The only variables consistently associated with the presence of EIPH either ≥1 or ≥2 in this study were ambient temperature, presence of dirt in the trachea, and the time between finishing the race and tracheobronchoscopic examination. The number of lifetime starts was only associated with risk of EIPH ≥1, whereas distance raced and the lifetime earnings were associated only with EIPH ≥2. The age of the horse and its sex were not associated with the risk of EIPH when the effect of other factors was taken into account during the statistical analysis. Similarly, the speed of racing, track surface, and air quality were not associated with risk of EIPH. Risk factors identified by this study are not consistently similar to those identified in other studies of occult EIPH or of epistaxis. Results of this study suggest that there are no readily identified factors that are strongly associated with risk of EIPH, and that it is therefore problematic to use epidemiological information to develop hypotheses for the pathogenesis of EIPH.

Of the variables associated with both EIPH ≥1 and EIPH ≥2, only ambient temperature is likely to have biological relevance. The presence of tracheal dirt was significantly associated with EIPH such that horses with a greater amount of dirt in the trachea were less likely to be detected as having EIPH. We believe that the reason for this observation was that dirt in the trachea obscured the small amounts of blood that resulted from EIPH of grades 1 or 2. Failure to detect small amounts of blood because they were obscured by abundant dirt would have reduced the frequency with which horses with tracheal dirt were diagnosed as having EIPH, and would therefore have resulted in an artificially reduced risk of EIPH in these horses. The frequency of EIPH varied somewhat with the time between the horse finishing the race and the tracheobronchoscopic examination. This is not likely to be a biological effect on the severity of EIPH but rather a reflection of the time taken for rostral movement of blood in the airways after racing.

We are unaware of other reports of a relationship between ambient air temperature and risk of EIPH in Thoroughbred racehorses. However, there are 2 studies that investigated an association between ambient temperature and EIPH in Standardbred racehorses. A study in Australia did not identify an association between EIPH and ambient temperature (9–22° C) in Standardbred horses (Speirs et al. 1982) whereas there was a significant correlation between the proportion of Standardbred horses with EIPH on a given day in Canada and the ambient temperature (-15–25°C) (Lapointe et al. 1994). Horses were at increased risk of EIPH when horses raced at colder temperatures (r2= 0.39, P = 0.0007). Thoroughbred horses in the current study racing in an ambient temperature of <20°C were 1.8–2.0 times as likely to have EIPH as were horses racing at temperatures >20°C. There is no clear biological explanation for this observation, although inhalation of cold air (-5°C) during strenuous exercise induces inflammation of airways in horses (Davis et al. 2007). The relevance of this observation to development of EIPH is unclear. It is worth noting that the range of temperatures to which horses were exposed during this study was relatively small (mean 17.9°C, 95% range 12.9–22.9°C).

Increasing age has been suggested to be associated with epistaxis in horses, with horses aged ≥5 years being up to 6.4 times more likely to have epistaxis compared to 2-year-old horses (Takahashi et al. 2001; Weideman et al. 2003) although neither study controlled for the confounding effect of increased numbers of lifetime starts, which would be expected to be highly correlated with age in most horses. Age has been associated with presence of occult EIPH (Pascoe et al. 1981; Raphel and Soma 1982) but again the confounding effect of lifetime starts was not accounted for during statistical analysis of data. Results of the current study indicate that age, per se, is not a risk factor for EIPH, but measures of racing intensity (lifetime starts, lifetime earnings) are associated with risk of EIPH, albeit not consistently using either criteria of EIPH, suggesting that there could be a cumulative effect of racing on risk of developing EIPH (Pascoe et al. 1981).

Epistaxis is more common after races <1600 m than in longer races (Takahashi et al. 2001) whereas others have found that the risk of occult EIPH increases with increasing race distance (Raphel and Soma 1982;Costa and Thomassian 2006). The risk of EIPH ≥2 was lower (odds ratio = 0.72) for horses racing 1400–2400 m compared with horses racing 1000–1200 m in the current study. As there was no association between race speed and risk of EIPH, interpretation of these results in light of a pathophysiological explanation for EIPH is unclear.

Female Thoroughbreds are 1.4 times more likely than stallions to have epistaxis (Takahashi et al. 2001) although this is not a consistent finding, with others finding that geldings are at greater risk than mares or intact males (Weideman et al. 2003). There was no detectable association of sex with risk of occult EIPH in the current study, which is consistent with other reports (Pascoe et al. 1981; Raphel and Soma 1982; Costa and Thomassian 2006). It does not appear that sex of the horse is associated with risk of occult EIPH.

There are anecdotal reports that the risk of EIPH is associated with air quality, and concern that racing horses in air that is polluted will increase the risk of EIPH. There is no association between air pollutants, including particulates, CO, SO2, NO2, NO, oxidants and lead, and incidence of EIPH in Standardbred racehorses in Canada (Lapointe et al. 1994). We examined a number of indicators of air quality and were not able to detect any association with EIPH. However, it might be that the effect of poor air quality is cumulative and that a single exposure is not sufficient to detect an effect of air pollution on risk of EIPH. The results of this study should not be interpreted to suggest that air quality is unimportant as a risk factor for EIPH when horses are chronically exposed to polluted air, but rather that a single exposure to poor quality air is unlikely to be associated with increased risk of EIPH.

Foot strike and transmission of impact shock to the thorax has been proposed as a mechanism for EIPH (Schroter et al. 1998). Support for this hypothesis is provided by the observation that horses racing on hard ground are more likely to develop epistaxis (Newton et al. 2005). This is in contrast to the current study in which neither subjective assessments of track hardness, nor objective quantification by penetrometer measurements were associated with risk of occult EIPH. Results of the current study, and the anecdotal observation that EIPH occurs in swimming horses, do not support the hypothesis that racing on hard surfaces increases the risk of EIPH.

Studies of EIPH are always dependent on the methodology used to detect blood in the airways. Detection of EIPH not associated with epistaxis is usually achieved by endoscopic examination of the upper airways, or airway lavage (tracheal lavage or bronchoalveolar lavage). The relative merits of each of these techniques have been debated in various fora and will not be rehearsed here. Suffice to say that the sensitivity of detecting EIPH varies between diagnostic methods with no one method being demonstrably more clinically relevant. The current study used a rigorous and blinded system to evaluate the presence and severity of EIPH in a large number of horses. The reproducibility of the methodology has been well documented in this population of horses, supporting use of this technique. It must be borne in mind that the risk factors identified in this study are those for EIPH detected by endoscopic examination. The results might not apply for EIPH detected using other methodologies (or for epistaxis).

Several well designed studies have not detected consistently similar risk factors for epistaxis as were detected for occult EIPH in the current study. It is possible and indeed plausible that risk factors for epistaxis differ from those for occult EIPH. There is substantial evidence in the scientific literature, based on incongruous findings between studies, to support this contention. Therefore, results of this study should not be inferred to indicate risk factors for epistaxis.

In summary, this prospective, blinded, observational study conducted using rigorous methodology and statistical analysis identified only ambient temperature, presence of dirt in the trachea, and the time between finishing the race and tracheobronchoscopic examination as being consistently associated with presence of EIPH either ≥1 or ≥2. Further studies are needed to refine these findings, and to determine whether they are reflective of risk factors in other areas of the world or codes of racing.

Manufacturers' addresses

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

1 Shirakawa Olympus Co Ltd, Tokyo, Japan.

2 Racing Victoria Ltd, Flemington, Victoria, Australia.

3 Bureau of Meteorology and Environmental Protection Authority, Melbourne, Victoria, Australia.

4 Environment Protection Authority, Melbourne, Victoria, Australia.

5 SAS version 9.1, SAS Institute, Cary North Carolina, USA.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conflicts of interest
  8. Manufacturers' addresses
  9. References
  • Costa, M.F. and Thomassian, A. (2006) Evaluation of race distance, track surface and season of the year on exercise-induced pulmonary haemorrhage in flat racing Thoroughbreds in Brazil. Equine vet. J., Suppl. 36, 87-89.
  • Davis, M.S., Williams, C.C., Meinkoth, J.H., Malayer, J.R., Royer, C.M., Williamson, K.K. and McKenzie, E.C. (2007) Influx of neutrophils and persistence of cytokine expression in airways of horses after performing exercise while breathing cold air. Am. J. vet. Res. 68, 185-189.
  • Hinchcliff, K.W., Jackson, M.A., Brown, J.A., Morley, P.M., McCaffrey, J.P., O'Callaghan, P.A., Dredge, A.F., Slocombe, R.F. and Clarke, A.F. (2005a) Tracheobronchoscopic assessment of exercise-induced pulmonary hemorrhage in Thoroughbred racehorses. Am. J. vet. Res. 66, 596-598.
  • Hinchcliff, K.W., Jackson, M.A., Brown, J.A., Morley, P.M., McCaffrey, J.P., O'Callaghan, P.A., Dredge, A.F., Slocombe, R.F. and Clarke, A.F. (2005b) Association between exercise-induced pulmonary hemorrhage and performance in Thoroughbred racehorses. J. Am. vet. med. Ass. 227, 768-774.
  • Lapointe, J.M., Vrins, A. and McCarvill, E. (1994) A survey of exercise-induced pulmonary haemorrhage in Quebec standardbred racehorses. Equine vet. J. 26, 482-485.
  • Newton, J.R., Rogers, K., Marlin, D.J., Wood, J.L. and Williams, R.B. (2005) Risk factors for epistaxis on British racecourses: Evidence for locomotory impact-induced trauma contributing to the aetiology of exercise-induced pulmonary haemorrhage. Equine vet. J. 37, 402-411.
  • Pascoe, J.R., Ferraro, G.L. and Cannon, J.H. (1981) Exercise-induced pulmonary hemorrhage in racing Thoroughbreds: A preliminary study. Am. J. vet. Res. 42, 703-707.
  • Raphel, C.F. and Soma, L.R. (1982) Exercise-induced pulmonary hemorrhage in Thoroughbreds after racing and breezing. Am. J. vet. Res. 43, 1123-1127.
  • Schroter, R.C., Marlin, D.J. and Denny, E. (1998) Exercise-induced pulmonary haemorrhage (EIPH) in horses results from locomotory impact induced trauma - a novel, unifying concept. Equine vet. J. 30, 186-192.
  • Speirs, V.C., Van Veenendaal, J.V., Harrison, I.W., Smith, G.B., Anderson, G.A., Wilson, D.V. and Gilbo, B. (1982) Pulmonary haemorrhage in standardbred horses after racing. Aust. vet. J. 59, 38-40.
  • Takahashi, T., Hiraga, A. and Ohmura, H. (2001) Frequency of and risk factors for epistaxis associated with exercise-induced pulmonary hemorrhage in horses: 251,609 race starts (1992–1997). J. Am. vet. med. Ass. 218, 1462-1464.
  • Weideman, H., Schoeman, S.J., Jordaan, G.F. and Kidd, M. (2003) Epistaxis related to exercise-induced pulmonary haemorrhage in South African Thoroughbreds. S. Afr. vet. Ass. 74, 127-131.