Exercise-induced pulmonary haemorrhage (EIPH) is associated with impaired short-term race performance but consequences to long-term race performance of Thoroughbred racehorses are unknown. Knowledge of effects of EIPH on performance over a prolonged time would inform the need for management and treatment of this disorder.
To determine the relationship between EIPH detected on a single occasion and long-term athletic performance in Thoroughbred racehorses.
Prospective, longitudinal, observational epidemiological study conducted from 2003 to 2012.
Seven hundred and forty-four Thoroughbred racehorses underwent a single tracheobronchoscopic examination to determine presence and severity of EIPH in 2003. Following retirement of all the horses from racing, career after examination and lifetime performance data (including career duration, earnings, starts, wins and placings) were abstracted from a commercial database. The association between presence and severity of EIPH and individual performance variables were analysed using survival analysis, linear regression and negative binomial regression.
No association was detected between the presence of any EIPH (grade ≥1) and all measures of long-term performance examined except that horses with EIPH (grade ≥2) had fewer earnings after endoscopic examination than did horses with mild EIPH or no EIPH (grades ≤1). Multiple associations were detected between the most severe form of EIPH (grade 4) and measures of long-term performance suggesting that the severity of EIPH commonly associated with epistaxis adversely affects either the horse's opportunity to race or physiologic capacity to race.
We conclude that there is no association between EIPH grades 0, 1, 2 and 3 and long-term racing performance of Thoroughbred racehorses. These results are based on a single endoscopic evaluation of EIPH grade made during a horse's career. Further studies are required to determine the reasons for an association of severe (grade 4) EIPH with impaired long-term race performance.
Exercise-induced pulmonary haemorrhage (EIPH) occurs commonly in horses undertaking high-intensity exercise. Although estimates of the prevalence of EIPH vary depending on the population of horses examined, the criteria for diagnosis and the frequency of examination, blood is detected by endoscopic examination in 43–75% of Thoroughbred horses after a single race [1-4]. The prevalence of EIPH increases to >95% with examination after a greater number of races . The consequences of EIPH in terms of missed training days, impaired performance and early retirement are perceived as important and substantial economic resources worldwide are expended on treating and managing the condition in Thoroughbred and Standardbred racehorses. In the USA alone, the annual estimated cost of medicating horses with EIPH exceeds $250 million . Despite the assumed adverse impact of EIPH, little scientific evidence exists to determine the importance of the condition to the long-term performance of equine athletes.
One method of determining the importance of EIPH is to assess the impact of the condition on race performance. Previous studies have demonstrated a significant negative association between EIPH and performance during a single race [1, 5, 6]. Horses with EIPH grade ≥2 are less likely to win or finish in the first 3 race positions and less likely to be in the 90th percentile or higher for earnings . Additionally, inferior finishing position was associated with both recurrence and severity of EIPH, including when horses were premedicated with furosemide . Two studies have identified a negative association between epistaxis, arguably the most severe form of EIPH, and number of starts during a career  and career duration . However, published information about whether the presence and severity of EIPH detected on a single endoscopic examination is associated with subsequent career performance or whether a single endoscopic determination of EIPH grade can be used as an indicator of lifetime racing potential of a horse is limited.
In order to determine the importance of EIPH to long-term athletic performance, indices of performance including career duration, earnings and number of starts, wins and placings were used as markers of career achievement. The purpose of this study was to determine the relationship between EIPH detected on a single occasion and long-term athletic performance in Thoroughbred racehorses. The aim of the study was to evaluate the potential association between the presence and severity of EIPH and long-term performance.
Materials and methods
This study was designed as a prospective, longitudinal, observational study utilising a convenience sample of Thoroughbred horses recruited while competing in flat races on metropolitan racetracks in Melbourne, Australia. Details of enrolment of horses and collection of data on EIPH have been reported . Briefly, horses underwent an endoscopic examination to determine the presence and severity EIPH on a single occasion. After retirement of the horses from racing, data characterising performance were used to determine if an association existed between EIPH detected on a single occasion and long-term athletic performance.
Horses enrolled in the study were Thoroughbred racehorses competing in flat races at one of 4 racetracks (Flemington, Caulfield, Sandown and Moonee Valley) around Melbourne, Australia.
Study subjects were prospectively enrolled by owners and trainers between 20 March and 16 June 2003 as part of a previously published study examining the association between EIPH and short-term racing performance . At the time of enrolment, horses competed in flat races over 1000–3200 m on turf; however, during the course of their careers many horses competed at alternative locations both nationally and internationally and performed over various distances on different track surfaces. Enrolment in the study was not restricted by age or sex. In accordance with the rules of racing in Victoria, Australia, medication with furosemide was not permitted on race day. The race career of each horse was examined after retirement of the last horse of the cohort to race, on the 31 May 2012.
Detection and quantification of EIPH
All horses underwent tracheobronchoscopic examination for evidence of EIPH, as detailed previously . Briefly, an endoscope was placed through one of the nares, and the nasopharynx, larynx and trachea to the level of the carina. Horses were not sedated for the procedure and all examinations were recorded on videotape for subsequent analysis. Three individuals, blinded to the identity of the horses and their race performance, independently reviewed the videotapes and recorded their categorical assessment without discussing their observations with each other. Severity of EIPH was graded on a scale of 0–4 as previously validated . Each score was recorded as a discrete number and the mode of the 3 scores was used in the statistical analysis of data . Trainers, handlers or both were present during the tracheobronchoscopic examination, and trainers received a faxed veterinary report the following morning detailing the results of the examination.
Race records for the horses included in the study were retrieved from a commercial racing database (SiRIUS). Variables abstracted from the database included career duration, earnings, starts, wins and places for each horse, as well as birth date, date of first race and date of last race. For the purpose of this study, long-term performance was considered in 2 ways: career after examination (CAE), where race data were used from date of tracheobronchoscopic examination to last race date and lifetime racing career (LRC), where race data were calculated from the date of the horse's first race to the date of its last race. A horse was considered to be retired when competitive racing had ceased for a period of one year. Raw data were entered into an Excel spreadsheet.
Data were structured for analysis in Excel and imported to Stata 12.1a for statistical analysis. Career duration, earnings, starts, wins and placings were determined for each horse for CAE and LRC. Presence of EIPH was defined as dichotomous (yes vs. no) in 2 ways: severity of grade ≥1 (yes) vs. severity of grade 0 (no) and ≥2 (yes) vs. severity grade ≤1 (no). Horses with EIPH severity grades ≥1 were referred to as having ‘any EIPH’ and horses with EIPH severity grades ≥2 as ‘moderate to severe’ EIPH. Exercise-induced pulmonary haemorrhage severity grades 0–4 were also considered as discrete levels.
Examination of summary and descriptive statistics and of graphs of the data was used to determine whether continuous data were normally distributed. Data that were not normally distributed were treated in 2 ways: via natural log transformation (x + 1) to yield an approximately normal distribution and a ratio of geometric means (GM) or through the use of negative binomial regression and rate ratios (RR) . All statistical analyses were performed for CAE and LRC data. Age (in years) and sex (male or female) were included as covariates in all models.
Survival analysis and career performance
Survival analysis using a Cox proportional hazards model was used to determine if the presence and severity of EIPH were associated with career duration. Because an effect of the most severe grade of EIPH (4) could have been obscured in the overall statistical analysis by the low proportion of horses with EIPH grade 4, the survival time of horses with EIPH grades 1, 2, 3 and 4 were each compared separately to horses without EIPH with Sidak's adjustment of the type 1 error rate for multiple comparisons (type 1 error rate of 0.013). Hazard ratios (HR) and confidence intervals (CI) were calculated from the Cox proportional hazards model.
Multiple linear regression was used to determine the association between the presence and severity of EIPH and the natural logarithm transformed continuous outcomes of (earnings + $1) and (starts + 1). Wins and placings per start were analysed via negative binomial regression using an offset, which was (natural logarithm [starts +1]) for CAE data and (natural logarithm [starts]) for LRC data. The log(starts +1) was analysed and the results reported as the ratio of geometric means to satisfy statistical assumptions about the distribution of data.
The analysis also used starts without any transformation and the results are reported as the difference in starts compared with EIPH grade 0. The differences in starts compared with grade 0 were adjusted for age and sex. Because the assumption of normality was not satisfied, the 95% CI and P values were not appropriate indicators of dispersion of the data and type 1 error rate, and therefore robust standard errors were used to produce robust 95% CI and P values.
Sensitivity analysis was performed to identify outliers or covariate patterns that could have had undue influence on the results and affected the statistical conclusions of the analysis. Where an outlier was determined to be influential on conclusions, it was reported in results. A P value <0.05 was considered significant in all analyses unless stated otherwise.
A total of 744 Thoroughbred racehorses, representing 52% of horses eligible for inclusion in the study, underwent tracheobronchoscopic examination after completion of 202 races at 26 race meets . The study population consisted of 375 geldings, 305 mares and 64 stallions with a mean age at enrolment of 4.1 years (range 2–10 years). The prevalence of EIPH in this population of horses was 55% (Fig 1). Descriptive statistics including earnings, starts, wins and placings were calculated for each EIPH grade (0, 1, 2, 3 and 4) using CAE and lifetime racing data (Figs 2, 3). Six horses in this study had epistaxis, as defined as presence of blood at one or both nares at the time of examination. Of these 6 horses, 5 horses had EIPH grade 4 and one horse had EIPH grade 2. The numbers of horses that did not start after endoscopic examination were: 13/332 (4%) horses with EIPH grade 0, 11/273 (4%) horses with EIPH grade 1, 6/101 (6%) horses with EIPH grade 2, 1/25 (4%) horse with EIPH grade 3 and 6/13 (46%) horses with EIPH grade 4.
Career after examination
There was no association detected between the presence of any EIPH (grade ≥1 vs. 0) or moderate to severe EIPH (grade ≥2 vs. ≤1) and career duration (Table 1). There was no association between EIPH grade (0, 1, 2, 3 or 4) and career duration (4 degrees of freedom; P = 0.24). However, examination of a plot of the data (Fig 4) revealed an apparent shorter career duration for horses with EIPH grade 4, which was supported in statistical analyses when comparing horses with EIPH grade 4 to those without EIPH (P = 0.011, Sidak cutoff P = 0.013; Tables 1 and 2, Fig 4). No association was detected between EIPH grades 1, 2, or 3 vs. 0 (P = 0.53, 0.74 and 0.67, respectively). Increasing age at enrolment was associated with shorter subsequent career duration (HR 1.3 [95% CI 1.20–1.40]; P<0.001) and males had significantly longer average careers when compared with females (HR 0.31 [95% CI 0.26–0.37]; P<0.001).
Table 1. Association between exercise-induced pulmonary haemorrhage (EIPH) and performance after endoscopic examination in 744 Thoroughbred racehorses competing in Australia between 2003 and 2012, adjusted for age and sex
Table 2. Median career duration of Thoroughbred horses (n = 744) after a single endoscopic examination to grade presence and severity of exercise-induced pulmonary haemorrhage (grade 0–4)
Number of horses
Median career duration (months) after endoscopic examination
Percentiles 25th 75th
Horses with EIPH grade ≤1 earned 3.3 times more prize money during their racing careers than horses with EIPH ≥2 (Table 1). Horses with EIPH grade 0 earned 135 times more prize money than horses with EIPH grade 4 (Table 1).
Starts and wins and placings per start
Horses with any EIPH (grade ≥1) had fewer starts than horses without EIPH, as did horses with moderate to severe (grade ≥2) EIPH when compared with horses with EIPH grade ≤1 (Table 1). The associations between EIPH grades ≥1 and ≥2 and starts were highly influenced by the low numbers of race starts of horses in the EIPH grade 4 category. Removal of grade 4 horses resulted in a P value that was not significant for any EIPH (P = 0.15) or moderate to severe EIPH (P = 0.30), respectively. Horses with EIPH grades 2, 3 and 4 raced 2.8 (95% CI 0.3–5.3; P = 0.028) fewer times over their career after diagnosis than did horses with EIPH grade ≤1. Horses with EIPH grade 4 had 12.1 (95% CI 5.9–18.4; P<0.001) fewer starts after diagnosis than did horses with EIPH grade 0.
Lifetime racing career
No association was detected between the presence of any EIPH (grade ≥1 vs. 0) or moderate to severe EIPH (grade ≥2 vs. ≤1) and career duration for lifetime racing career (Table 3). The Kaplan–Meier curve of horses with EIPH grades 0, 1, 2, 3 and 4 are presented in Figure 5 . Using a Sidak cutoff P value of 0.013, horses with EIPH grade 4 did not have a shorter duration (P = 0.014) of racing career than did horses without EIPH (Table 3).
Table 3. Associations between exercise-induced pulmonary haemorrhage (EIPH) and lifetime racing performance in 744 Thoroughbred racehorses competing in Australia between 2003 and 2012, adjusted for age and sex
Overall (4 degrees of freedom: 4df) EIPH P = 0.11.
Earnings and starts
Horses with EIPH grade ≥2 had fewer lifetime starts than horses with EIPH grade ≤1 (Table 3). The association between EIPH grade ≥2 and lifetime starts was highly influenced by horses in the EIPH grade 4 category, and after removal of these horses from the analysis the association was not significant (P = 0.33). The lifetime number of starts was also significantly different between horses with EIPH grade 4 vs. 0. Horses with EIPH grade 4 raced 15.2 (95% CI 7.4–23.0; P<0.001) fewer times than horses with EIPH severity grade 0.
Wins and placings per start
Increased wins and placings per start were detected in horses with EIPH grade 4, when compared with horses with EIPH grade 0 (Table 3). This finding was highly influenced by a single outlier grade 4 EIPH horse with exceptional racing performance. Removal of this outlier resulted in rates of wins (1.34 [95% CI 0.93–1.93]; P = 0.12) and placings (1.18 [95% CI 0.92–1.50]; P = 0.19) that were not significantly different to horses with EIPH grade 0.
This study does not provide evidence of an important effect of the most common forms of EIPH on long-term athletic performance of Thoroughbred horses racing without prerace administration of furosemide. We did identify multiple associations between the most severe form of EIPH (grade 4) and measures of long-term performance suggesting that the severity of EIPH commonly associated with epistaxis adversely affects either the horse's opportunity to race or capacity to race. These results could be the result of one or more biological factors that adversely influence the horse's capacity for competitive performance or management factors, such as decisions not to race the horse, that reduce the horse's opportunity to perform. Importantly, interpretation of the results pertaining to the horses with grade 4 EIPH is limited by the few horses in the study with EIPH in this category.
An important finding of this study was the lack of association between the presence of any EIPH (grade ≥1) and all measures of long-term performance examined with the exception that horses with moderate to severe EIPH (grade ≥2) had fewer earnings after endoscopic examination than did horses with mild EIPH or no EIPH (grades ≤1).The lack of significant associations between the most common forms of EIPH and long-term performance can be interpreted as indicating an absence of important clinical, physiological or management factors that limit performance in these horses. These results are unique in that we know of no other studies that have determined the long-term (almost one decade) effects of a single diagnosis of EIPH. Similarly, we are not aware of any studies that monitor the long-term performance of racehorses with EIPH using repeated examinations (see below for discussion of limitations of the current study).
Thoroughbred racehorses with severe (grade 4) EIPH had a shorter duration of racing career and fewer race starts during their career after endoscopic examination than did horses without EIPH (grade 0). This could plausibly be attributed to racing regulations in Victoria, Australia that reduce the horse's opportunity to race because of the mandatory 3 month ban from racing for the first episode of epistaxis and a permanent disqualification for a repeat episode of bleeding. However, only 6 horses in this study had epistaxis and examination of racing records of the 13 horses with severe (grade 4) EIPH determined that only 3 horses received a temporary ban during the course of their racing career and no horses were permanently disqualified due to epistaxis. This suggests that the negative association between severe EIPH and athletic performance is not consequent to a reduced opportunity to succeed because of racing regulations per se. However, trainers were aware of endoscopy results, and horses with severe EIPH, with or without epistaxis, might be trained, raced and managed differently to those that have not had epistaxis or severe EIPH. As such, trainers with knowledge of the horse's EIPH status could have influenced the subsequent racing career of those horses with severe EIPH. Although results of this study suggest that severe (grade 4) EIPH negatively impacts long-term performance, it is unclear whether this is due to management decisions based on the results of the endoscopic examination or a biological effect of EIPH.
Severe EIPH could have impaired the horse's capacity for intense exercise through persisting effects on pulmonary mechanics, gas exchange in the lungs, or the systemic effects of inflammation inthe lungs [11-13]. The relationship between severe EIPH and long-term performance, if related to biological factors, indicates that diagnosis of severe EIPH is either indicative of on-going or repeated haemorrhage in these horses, or of a persistent adverse effect of a single episode of severe EIPH. Elucidation of the reasons, be they management or biological, for the effect of severe EIPH on career performance awaits further study.
There was an association between moderate to severe EIPH (grade ≥2) and CAE earnings. This finding was robust and a significant association remained after removal of horses with EIPH grade 4. Horses with EIPH grade ≥1 did not earn significantly less than horses with EIPH severity grade 0 (P = 0.065). Examination of descriptive data for CAE earnings indicated a gradual decrease in career earnings as EIPH grade increased, which was not due to reduced opportunity to earn as horses with EIPH grade ≥2 did not have fewer wins or placings per start, nor were their careers of shorter duration. However, it is possible that horses with EIPH severity grade ≥2 competed in lower-quality races with less prize money to earn and therefore had reduced opportunity to earn. This suggestion implies an effect of EIPH on selection of races for affected horses and, although there is no evidence to support or refute this idea, we consider this suggestion implausible. It is more likely that there is a statistical anomaly related to the highly skewed distribution of earnings. Earnings are not a robust measure of the quality of racing career but are included here because they are commonly used as a measure of performance in epidemiological studies of race performance [14-17], including in recent studies investigating the association between epistaxis and long-term performance [7, 8] and presence and severity of EIPH and short-term performance .
Increased lifetime wins and placings per start were detected in horses with EIPH grade 4, when compared with horses without EIPH. This finding was, at least in part, due to the influence of a single outlier horse with grade 4 EIPH that had exceptional lifetime and CAE race performance. After removal of this outlier, wins and placings per start of horses with grade 4 EIPH was not significantly different when compared with horses with grade 0 EIPH.
Investigating the effects of EIPH on long-term performance is challenging because of the need to define the disease classification of all horses and to follow the horses for a period of time, preferably their entire career, which permits meaningful measures of long-term performance. The current study is strengthened by the large number of horses examined at the start of the study, the long duration of the follow-up period, and the availability to the authors of full data on long-term performance of all horses enrolled in the study. Determination of the EIPH status of the horses on one occasion is useful because it addresses the clinically relevant question of the relationship between EIPH diagnosed on a single occasion and the long-term athletic prospects of the horse. This differs from a study of the impact of development and progression of EIPH during a career on long-term performance. The latter requires repeated examinations of horses under standardised conditions (racing, race duration and length, quality of racing company, environmental conditions, prerace and other medications, and many others) with comprehensive follow-up. Such a study would provide information on the repeatability of endoscopic examinations of racehorses (data that are not currently available), evidence of progression or lack thereof of EIPH, and any association with diminished performance associated with progression of the condition. The current study was designed to investigate the first question and to avoid the shortcomings and logistical difficulties of the second.
Of the 744 Thoroughbred racehorses examined, only 13 were identified in this study with grade 4 EIPH. Within the grade 4 EIPH category, approximately half of the horses did not race again after endoscopic examination. The exact reason for retirement of grade 4 horses is not known and could have been due to several factors including persistent poor performance, perceived likelihood of future poor performance, financial considerations (costs vs. benefits), fear of future racing bans, or ethical concerns of trainers and owners to continue racing horses known to suffer from severe EIPH. Previous studies have determined that virtually all Thoroughbred racehorses experience EIPH at some time in their racing career . It is not known whether horses with mild EIPH progress to having more severe EIPH with time. A relationship was not detected between EIPH grade and age in a previous study investigating risk factors for EIPH in this population of horses  and the lack of detectable effect of milder grades of EIPH on performance in this study suggests the condition is not necessarily progressive.
The use of CAE and LRC in this study addressed 2 different epidemiological questions: 1) what happens to the racing career of Thoroughbred horses after they are diagnosed with EIPH; and 2) can a single endoscopic examination be used as a marker of disease presence and severity to determine if EIPH affects lifetime performance. In this population of horses, some associations detected between EIPH and CAE were not evident on analysis of LRC. This could be because a single endoscopic diagnosis of EIPH is not an accurate marker of disease presence or progression over time. Alternatively, career performance before EIPH might not reflect performance after development of the condition. Lastly, a single episode of EIPH might have a minimal effect on performance over a lifetime.
The current study investigated the career performance of Thoroughbred racehorses that were enrolled in the study while competing on metropolitan racetracks around Melbourne, Australia. Although a small proportion of horses subsequently raced overseas, the majority of horses included in this study completed their racing careers in Australia. Geographic differences in airway health; racing and training conditions, including track surface; prevalence of other respiratory disease; training techniques and medication use could have the potential to affect the predisposition to EIPH . Although associations between most of these variables and EIPH have not been demonstrated and we are not aware of any studies documenting geographic differences in the prevalence of EIPH, the possibility of regional difference should be considered when extrapolating results of the present study to conditions in other racing jurisdictions, especially those that permit race day use of furosemide .
On the basis of results of the current study, we conclude that there is no association between EIPH grades 0, 1, 2 and 3 and long-term racing performance of Thoroughbred racehorses. These results are based on a single endoscopic evaluation of EIPH grade made during a horse's career. Although results of this study suggest severe (grade 4) EIPH has a negative impact on long-term performance, it is unclear whether this is due to the biological effects of the condition or to management decisions based on knowledge of results of the endoscopic examination. Further studies are required to determine the importance of severe EIPH to long-term performance.
Authors’ declaration of interests
No competing interests have been declared.
Ethical animal research
The initial 2003 study was approved by the Animal Ethics Committee of the Faculty of Veterinary Science, University of Melbourne. As the current study involved only use of racing records, animal use approval was not required. Study subjects were prospectively enrolled by owners and trainers.
Source of funding
The Rural Industries and Research Development Corporation provided funding for the initial study in 2003. [Correction added on 30 September 2014 after first online publication: The Grayson-Jockey Club Research Foundation did not provide funding for this study and have therefore been removed from the source of funding statement.]
The authors gratefully acknowledge the regulatory racing authority in Australia, Racing Victoria, for the provision of data.
S.L. Sullivan was involved in study design, study execution, data analysis and interpretation, and preparation of the manuscript. G.A. Anderson contributed to data analysis and interpretation, and preparation of the manuscript. K.W. Hinchcliff contributed to study design, study execution, data analysis and interpretation, and preparation of the manuscript. P.S. Morley contributed to study design, study execution, data analysis and interpretation, and preparation of the manuscript. All authors gave their final approval of the manuscript.