Prevalence and severity of exercise‐induced pulmonary hemorrhage in 2‐year‐old Thoroughbred racehorses and its relationship to performance

Abstract

Background

Little has been reported regarding the prevalence and severity of exercise‐induced pulmonary hemorrhage (EIPH) in 2‐year‐old Thoroughbred racehorses.

Objectives

Evaluate EIPH prevalence and severity and its association with performance, speed index, furosemide administration, race distance, and track surface.

Animals

A total of 830 2‐year‐old Thoroughbreds.

Methods

Prospective blinded observational study. Videoendoscopy was performed 30 to 60 minutes postrace at 15 American racetracks. Three blinded observers independently assigned an EIPH grade (0‐4) to each video, and prevalence and severity of EIPH were determined. Relationships of EIPH grade to performance, speed index, race distance, track surface, and prerace administration of furosemide were evaluated using Pearson's chi‐squared test for categorical variables and analysis of variance (ANOVA) for numerical variables. Multivariable logistic regression assessed relationships between EIPH prevalence and severity, respectively, and the aforementioned independent variables. A P < .05 was considered significant.

Results

A total of 1071 tracheoendoscopies were recorded. The EIPH prevalence was 74% and for EIPH grade ≥3 was 8%. Speed index (P = .02) and finishing place (P = .004) were lower with EIPH ≥3. The EIPH prevalence and severity were lower at 2 tracks where postrace tracheoendoscopy was mandatory rather than voluntary (P < .001). Probability of observing EIPH was negatively associated with speed index (P = .01) at tracks where postrace tracheoendoscopy was mandatory. Prerace furosemide administration decreased the probability of EIPH occurrence (P = .007) and severity (P = .01) where study participation was voluntary.

Conclusions and Clinical Importance

Prevalence and severity of EIPH in 2‐year‐old racehorses were consistent with that of older racehorses. An EIPH grade ≥3 was associated with decreased performance. Prerace furosemide administration was associated with a decreased likelihood, but not severity, of EIPH at most tracks.

Abbreviations

ANOVA

analysis of variance

CI

confidence interval

EIPH

exercise‐induced pulmonary hemorrhage

IQR

interquartile range

OR

odds ratio

1. INTRODUCTION

Exercise‐induced pulmonary hemorrhage (EIPH) is defined as the presence of blood detected on tracheoendoscopic examination, bronchoalveolar lavage, or both, after exercise and occurs in most horses undergoing strenuous exercise. ¹ , ² The majority of Thoroughbreds experience EIPH during a racing career, ³ and EIPH's relationship to performance is an important topic in the racing industry today. Some studies reported that an increase in severity of EIPH is correlated with inferior performance as reflected by an increased likelihood of finishing worse than third in a race ⁴ , ⁵ and less race earnings. ⁵ Although the severity and prevalence of EIPH are associated with a horse's total number of race starts, ⁶ , ⁷ it is unclear as to when Thoroughbred racehorses first experience EIPH.

Furosemide has been used for over 40 years to attenuate the severity of EIPH in performance horses. ⁸ It effectively decreases the severity or occurrence of endoscopic evidence of EIPH in most horses when administered at the dose of 0.5 to 1.0 mg/kg intravenously (IV) 4 hours before strenuous exercise, ¹ with 68% of horses experiencing a decrease in severity of EIPH after the prerace administration of the drug. ⁸ Furosemide's prerace use in North America has become increasingly controversial in recent years and has been prohibited in 2‐year‐old races and graded stakes races for older horses at many racetracks across the United States since 2020 and 2021, respectively. ⁹

It is unclear whether or not 2‐year‐olds experience EIPH to the same degree as older horses in terms of prevalence and severity, because no reports to date have specifically focused on horses of this age. Similarly, little has been reported regarding the use and efficacy of furosemide in the 2‐year‐old age group. Therefore, our purpose was to determine the prevalence and severity of EIPH in American 2‐year‐old Thoroughbred racehorses, to assess the relationship between EIPH and performance, and to evaluate the relationship of prerace administration of furosemide to EIPH severity.

2. MATERIALS AND METHODS

The study was approved by the Washington State University Institutional Animal Care and Use Committee.

2.1. Study design

Ours was a prospective blinded observational study involving 2‐year‐old Thoroughbred racehorses that raced at 15 American racetracks (identified by the letters A‐O) between August 2020 and August 2021. These tracks were located in 10 states and distributed from the east to west coasts and from near the Canadian border to southern California and south Florida. The rules of racing related to the pharmaceutical agents that horses could and could not receive before racing were determined by the individual states' racing commissions and differed among some states. The assistance of American racetrack veterinarians was solicited in Maryland where postrace tracheoendoscopy of 2‐year‐olds was legislatively mandated in 2020, and at other tracks where veterinarians routinely performed postrace tracheoendoscopy for trainer‐clients. Also, trainers were given the opportunity to voluntarily enroll horses in the study at 4 tracks. Tracheoendoscopy was performed by a participating veterinarian 30 to 60 minutes postrace and a videorecording was made of the examination. Races took place on turf, dirt, and all‐weather surfaces. Some horses at tracks identified by the letters A, G, I, J, L, and M received 250 mg furosemide IV 4 hours before racing.

2.2. Detection and grading severity of EIPH

Videorecordings of the tracheoendoscopic examinations were uploaded to a secure web site by the veterinarians. All information identifying each horse, trainer, and the track on which the horse raced was anonymized. Each videorecording then was sent independently to 3 veterinarians with experience evaluating EIPH for the assignment of an EIPH severity grade using a previously described scale of 0 to 4. ⁴ If there was not unanimity regarding the EIPH grade for a video, the grade assigned by 2 of the 3 veterinarians was the grade recorded.

2.3. Recorded variables

Race day records for each horse from which a videorecording was received were obtained from online racing databases. Thoroughmanager (http://www.tmracingdata.com) was used to gather the speed index and the elapsed time for the race for each horse. The higher the speed index, the better a horse performed. Another online racing database Equibase (http://www.equibase.com) was used to collect the date, track, weather, track condition and surface, race distance and total prize money (purse), presence or absence of furosemide administration, and the horse's finishing position and earnings in the race in question. The number of days between races was calculated for horses from which sequential postrace videorecordings were obtained and grouped as follows: short (7‐14 days between races), moderate (14‐30 days), and prolonged (≥30 days). Race distances were categorized as sprint (900‐1200 m), intermediate (1300‐1500 m), and long (1600‐1800 m). The average speed in meters per second (m/s) was calculated for each horse in every race in which it competed for the study.

2.4. Data analysis

The prevalence of each grade of EIPH severity and the overall prevalence of EIPH were tabulated for each racetrack and the entire study population. Simple descriptive results were reported for speed index and race distance for each EIPH grade. Residual plots were used to check for variance equality and QQplot, and the Shapiro‐Wilk test of normality was used to identify any departures from statistical assumptions. If data was not normally distributed, the Kruskal‐Wallis nonparametric test was used to compare data from different groups of numerical variables (speed index, finish time, race distance, race purses, and horse earnings). For normally distributed data, the analysis of variance (ANOVA) F‐test was used for numerical variables, and Tukey's simultaneous confidence interval (CI) was used post hoc for pairwise comparisons when the F statistic was significant. Univariate analysis was used to evaluate the association between EIPH severity and each independent variable. For the purposes of analyzing the effects of EIPH on performance, EIPH grades 3 and 4 were combined because of the relatively small number of videos with grade 4 (n = 15). Pearson's chi‐square test of independence was applied to the analysis of categorical variables (surface, track location, track condition, race distance category, and furosemide) associated with each race. The Pearson's chi‐square test was used to compare EIPH grades of severity recorded from horses racing in Maryland to those racing without prerace administration of furosemide at the other tracks from which videorecordings were generated.

Two separate multivariable logistic regression models were implemented to assess the relationship between multiple variables and the outcomes of EIPH prevalence and severity (grade <3 vs grade ≥3), respectively. Initially, all of the variables that were significantly associated with EIPH grade from the univariate analysis (track surface, speed index, and race distance) were included together with track condition, the prerace administration of furosemide, and the interaction between speed index and furosemide. Because there was no all‐weather race surface at the Maryland racetracks where tracheoendoscopy was mandated, and the administration of furosemide was prohibited, the analysis was split to account for tracks where tracheoendoscopy was voluntary vs mandatory. A random effect for horses was added to account for possible repeated measures within multiple races run by the same horse. The stepwise model selection approach based on Akaike information criterion and maximum likelihood was used to further identify significant variables in the fixed effects. The Monte Carlo simulation method was used for the power analysis. Outliers were identified using predictive performance checking. Model fitness was assessed by checking the deviance residuals. The analyses were performed using R (R Studio version 3.4.2, R Core Team, 2020; https://cran.r-project.org/). In all instances, significance was deemed to exist when P < .05.

3. RESULTS

Results are presented as the mean ± SD, unless stated otherwise. The number and purses of races after which horses were examined, the number of horses and their speed index (range, mean, and median), and these horses' earnings in those races are summarized for each track in Table 1. The EIPH grade was determined for 1071 videorecordings from 830 2‐year‐old horses (430 females and 400 males) after 376 races. More than 1 video was available from 174 horses with these horses accounting for 415 of the videorecordings; 83% of them were generated at 2 tracks in Maryland and the other 17% came from 9 veterinary practices where postrace tracheoendoscopy was routinely performed. There were no repeated examinations of horses that had been voluntarily enrolled by the trainer.

TABLE 1.

Descriptive information regarding 15 American racetracks at which 2‐year‐old Thoroughbred racehorses were tracheoendoscopically examined 30 to 60 minutes postrace with reference to the whether or not they raced after the IV administration of 250 mg furosemide (Fur), the number of horses evaluated and the number of races after which examinations took place, the purse for that race, the money earned by the horses that were examined after the race and the speed index of the horses that were examined.

Track	Fur (Y/N)	Horses (n)	Races (n)	Purse ($)		Speed index		Earnings/horse ($)
				Mean	Median	Mean	Range	Mean	Median
A	Y	44	20	33 403	35 000	40	2‐78	4558	1200
B	N	79	32	76 299	80 000	53	4‐89	9683	3200
C	N	49	31	258 036	69 943	58	5‐104	16 654	3500
D	N	6	6	77 600	80 000	55	38‐64	6084	4000
E	N	302	103	45 352	44 352	49	1‐93	4532	1680
F	N	72	53	69 406	83 508	48	1‐85	6096	1116
G	Y	2	1	25 250	25 250	17	6‐27	875	875
H	N	32	18	65 101	60 500	57	29‐96	9128	1400
I	Y	8	2	39 360	39 360	45	5‐73	8834	6200
J	Y	18	11	27 357	20 770	42	3‐77	5275	2992
K	N	52	10	61 911	50 120	50	5‐92	9560	2400
L	Y	29	18	44 633	48 500	57	8‐80	4925	2760
M	Y	24	8	37 256	36 040	40	4‐74	3374	750
N	N	115	56	47 678	40 000	45	2‐87	5802	4000
O	N	45	7	54 217	62 000	49	5‐83	6694	1860

Note: 830 different horses were examined. Multiple horses competed and were examined at more than one racetrack.

Evidence of EIPH was observed in 66% of the videorecordings. Six‐hundred and eleven of the horses examined were positive for EIPH at least once, resulting in a prevalence of 74%. The number of tracheoendoscopic examinations per horse ranged from 1 to 5 over the course of the study. Repeated postrace examinations of individual horses significantly increased the likelihood of identifying EIPH in that horse (P < .001; Table 2), with >90% of horses that were examined 3 to 5 times exhibiting EIPH at least once. Horses had from 1 to 9 race starts (median, 2) at the time of their tracheoendoscopic examinations. The number of starts was not associated with EIPH grade (P = .84; Table 3).

TABLE 2.

The prevalence of EIPH increased with the number of postrace tracheoendoscopic (TE) examinations performed per horse on 830 individual horses (P < .001).

Number of TE exams	n	Prevalence of EIPH (%)
1	656	71
2	123	79
3	38	95
4	10	90
5	3	100

TABLE 3.

The distribution of EIPH grades in relation to the number of race starts at the time of tracheoendoscopic examination and total number and prevalence of each EIPH grade based on 1071 postrace examinations of 830 2‐year‐old Thoroughbred racehorses.

Number of starts	EIPH grade
	n	0	1	2	3	4
1	340	113	127	75	21	4
2	313	113	98	71	28	3
3	206	79	71	43	10	3
4	107	33	39	24	9	2
5	65	15	27	19	3	1
6	23	7	10	4	0	2
7	12	3	3	6	0	0
8	4	3	0	0	1	0
9	1	0	0	1	0	0
Total	1071	366	375	243	72	15
Prevalence (%)		34	35	23	7	1

There were 197 instances in which horses were tracheoendoscopically examined after consecutive races. The interval between these races ranged from 6 to 78 days (median, 24; interquartile range [IQR], 18‐30) with 9% of intervals being <14 days and 27% being longer than 30 days. The categorical interval between successive races had no effect on EIPH grade (P = .06; power = 0.84). There was an association between race distance and EIPH grade with horses racing over intermediate distances (1300‐1500 m) having higher EIPH grades than those racing over shorter or longer distances (P = .03).

Five‐hundred and nineteen videorecordings from 321 horses that competed in 113 races for 2‐year‐olds were received from Maryland racetrack veterinarians. The severity of EIPH was lower at the 2 Maryland tracks (tracks E and K) when compared with that in horses that raced elsewhere without prerace administration of furosemide and for which tracheoendoscopy was not mandatory (P < .001). No or mild (grade 1) EIPH was observed in 80% of the videorecordings from the Maryland tracks compared with 55% from the other tracks, whereas severe EIPH (grades 3 and 4) was seen in 4% of the videos from Maryland and 13% of the videos from the other tracks (Figure 1).

FIGURE 1.

The difference in grades of EIPH severity assigned to videorecordings of postrace tracheoendoscopic examinations of 2‐year‐old Thoroughbreds in Maryland (n = 519) where such examinations were compulsory, compared with grades assigned at other tracks where tracheoendoscopy was optional (n = 425). Overall, the severity of EIPH in Maryland was less than that at the other tracks (P < .001). Prerace medication with furosemide was prohibited in Maryland and each of the other tracks represented here.

3.1. Performance

Speed indices of the horses ranged from 1 to 104 and were distributed as follows: 0 to 30, n = 197 (18%); 31 to 50, n = 309 (29%); 51 to 70, n = 422 (40%); and 71 to 104, n = 141 (13%). Two horses received no speed index. Speed index was negatively associated with EIPH grade (P = .02; Table 4).

TABLE 4.

The speed index and race distance for each EIPH grade based on EIPH grades from 1071 postrace tracheoendoscopic examinations.

EIPH grade	n	Speed index ± SD	Distance (m) ± SD
0	366	51.3 ± 20.4	1315 ± 239
1	375	47.9 ± 21.1	1331 ± 233
2	243	49.2 ± 19.6	1367 ± 233
≥3	87	44.6 ± 20.9	1387 ± 216

Note: Higher EIPH grades were associated with slower speed indexes (P = .02) and longer race distances (P = .01).

Horses with EPIH grade ≥3 were less likely to place in 1 of the first 3 finishing positions than were those with EIPH grade 0 (P = .004). Of the races after which the winner underwent tracheoendoscopy (n = 125), 43% were won by horses with grade 0 EIPH although only 34% of all of the tracheoendoscopic findings were assigned this grade. This finding contrasted with that for EIPH grades ≥3, which were associated with only 4% of the race winners despite representing 8% of all tracheoendoscopic grades. Thirty‐four percent of the winners had grade 1 EIPH and 19% had grade 2 EIPH. These numbers were consistent with the total number of examinations to which each of these 2 EIPH grades was assigned (grade 1: 35%, grade 2: 23%).

3.2. Furosemide

Prerace IV administration of furosemide at 6 tracks preceded 127 of the 1071 postrace tracheoendoscopic examinations. The distribution of the EIPH grades assigned to the associated videorecordings is shown in Table 5. Univariate statistical analysis failed to identify an effect of furosemide on EIPH prevalence and severity (P = .2) but lacked adequate statistical power (=0.35). This result was attributed to the marked disparity in numbers of furosemide‐treated and untreated horses. Analysis of the effect of furosemide indicated an association with speed index (P = .004) with that of horses receiving furosemide being lower (furosemide: 44.2 ± 21.4; no furosemide: 49.8 ± 20.4). The Kruskal‐Wallis nonparametric test was used to compare the average earnings received by horses for the race associated with the tracheoendoscopic examination, according to whether they raced with or without IV furosemide administration, because the normality assumption for a 1‐way ANOVA was violated. The analysis indicated that the earnings of horses treated with furosemide were lower than those of horses that raced without furosemide (P < .001; Table 1). Additionally, the purses for these races also were less (P < .001) than those for horses competing without prerace furosemide administration. In a small number of races (n = 7) at 4 tracks, some horses received prerace furosemide (n = 17) and others in the same races did not (n = 12). No effect of furosemide on EIPH grade was detected in these races (median, [IQR]: 1, [0‐2] with no furosemide; 1, [0‐1] when furosemide was given; P = .16), although this analysis also was underpowered (=0.45).

TABLE 5.

The respective speed indexes (mean ± SD) and prevalence of EIPH grades in horses that did and did not receive IV furosemide on race‐day.

EIPH grade	n	No Furosemide		n	Furosemide
		%	Speed index		%	Speed index
0	313	33	52.9 ± 19.3	53	41	41.7 ± 23.8
1	332	35	48.4 ± 21.1	43	34	44.5 ± 19.9
2	219	23	49.2 ± 19.8	24	20	49.7 ± 18.2
≥3	80	8	44.8 ± 20.6	7	5	42.3 ± 20.3

Note: Overall, speed index was higher in horses competing without the prerace administration of furosemide (P = .004).

3.3. Racetrack factors

Three tracks each contributed <10 videorecordings to the study (range, 2‐8) and were excluded from the analyses aimed at evaluating the relationship between EIPH severity and track location, surface, or condition. The geographic location of the other 12 racetracks was associated with severity of EIPH, with the EIPH grades at some tracks being higher than at others (P < .001; Table 6). More races were run on dirt tracks (n = 840) than either turf (n = 207) or all‐weather surfaces (n = 24, track M only). The EIPH grades varied with track surface, with those associated with races on all‐weather tracks being lower than those of the other 2 surfaces (P < .001; Table 7).

TABLE 6.

The distribution of EIPH grades at the 12 tracks at which >10 postrace tracheoendoscopic examinations were performed.

Track	n	Grade 0	Grade 1	Grade 2	Grade 3	Grade 4
A	50	23 (46%)	14 (28%)	8 (16%)	5 (10%)	0 (0%)
B a	84	26 (31%)	20 (24%)	28 (33%)	7 (8%)	3 (4%)
C	50	26 (52%)	15 (30%)	6 (12%)	3 (6%)	0 (0%)
E	467	178 (38%)	194 (42%)	74 (16%)	18 (4%)	3 (1%)
F a	76	8 (11%)	34 (45%)	23 (30%)	10 (13%)	1 (1%)
H a	34	13 (38%)	9 (26%)	10 (29%)	2 (6%)	0 (0%)
J	24	17 (71%)	4 (17%)	2 (8%)	1 (4%)	0 (0%)
K	52	18 (35%)	23 (44%)	9 (17%)	2 (4%)	0 (0%)
L a	30	3 (10%)	14 (47%)	10 (33%)	1 (3%)	2 (7%)
M	22	5 (23%)	12 (54%)	5 (23%)	0 (0%)	0 (0%)
N	120	37 (31%)	24 (20%)	44 (37%)	12 (10%)	3 (2%)
O a	46	7 (15%)	10 (22%)	19 (41%)	8 (18%)	2 (4%)
Total	1055	361 (34%)	373 (35%)	238 (23%)	69 (7%)	14 (1%)

^a Tracks at which EIPH grade was higher (P < .001).

TABLE 7.

The numerical and percentile distribution of EIPH grades according to track surface.

	GRADE 0	GRADE 1	GRADE 2	GRADE 3	GRADE 4	n
All weather	17 (71%)	4 (17%)	2 (8%)	1 (4%)	0 (0%)	24
Dirt	273 (32%)	316 (38%)	185 (22%)	58 (7%)	8 (1%)	840
Turf	76 (37%)	55 (27%)	56 (27%)	13 (6%)	7 (3%)	207

Note: P < .001 when all 3 surfaces were compared, and P = .004 when only comparing results for dirt and turf tracks, with EIPH grades being worse after races on turf.

3.4. Multivariable logistic regression analysis

3.4.1. EIPH prevalence

The multivariable logistic regression model to evaluate the outcome of EIPH prevalence used the following variables: speed index, furosemide/no furosemide, race distance category, track condition, track surface, and whether or not postrace tracheoendoscopy was mandatory. Track location was not included in the model because of the large number of tracks involved in the study, the large range in sample sizes from the tracks, and because predictors such as track surface and track condition that were included in the model also likely were related to track location. When checking model fitness, we found that a single track (track L) had a prediction accuracy that was 23% lower than the overall accuracy of the model and had shown an unusual pattern in the scatterplot for speed index vs EIPH grade. This track had 28/30 horses that raced after receiving furosemide, and from which 90% of horses had evidence of EIPH. Fifty percent of the horses evaluated at this track had speed indices >60 compared with 21% of the horses that raced after furosemide administration at the other tracks at which it was permitted, and 32% of all the horses that raced without receiving furosemide. The faster horses from track L also were found to have a higher probability of developing EIPH than slower horses. This result contrasted with the finding for all of the other racetracks at which the faster horses were less likely to experience EIPH regardless of whether they received furosemide before racing or not (Figure 2). We therefore decided to remove the data from track L and hereafter report the results based on data from the remaining tracks.

FIGURE 2.

The relationships between the probability of any grade of EIPH occurring and speed index for horses that were or were not treated with furosemide 4 hours before racing, except for horses racing at track L. The result of the logistic regression model for the subset of horses from track L is shown separately because of its difference from the other two lines.

The analysis indicated that the prevalence of EIPH was significantly associated with speed index, furosemide/no furosemide, track surface and whether or not postrace tracheoendoscopy was mandatory (Table 8). The prevalence of EIPH was not associated with race distance category or track condition. Specifically, EIPH prevalence differed significantly according to whether postrace tracheoendoscopy was mandatory or not (P < .001), with the prevalence being lower on the tracks where the procedure was required. Because there was no all‐weather race surface at tracks E or K in Maryland, and the administration of furosemide was prohibited at the tracks at which tracheoendoscopy was mandated, we divided the results according to whether tracheoendoscopy was voluntary or mandatory at the respective tracks.

TABLE 8.

Odds ratio and 95% confidence interval (CI) for variables contributing to the final multivariable mixed‐effect logistic regression model for which EIPH prevalence was the outcome, after splitting the analysis into voluntary or mandatory postrace tracheoendoscopy.

	P value	Odds ratio	95% CI
Voluntary tracheoendoscopy
Speed index	.17	0.99	0.97‐1.00
Furosemide administration vs no administration	.01	0.22	0.07‐0.67
Track surface: Dirt vs all weather	.02	10.15	1.48‐69.62
Track surface: Turf vs all weather	.06	3.34	0.91‐37.83
Track surface: Turf vs dirt	.13	0.60	0.28‐1.18
Mandatory tracheoendoscopy
Speed index	.01	0.98	0.97‐0.99
Track surface: Turf vs dirt	.01	0.40	0.20‐0.81

For tracks where involvement in the study was voluntary, the prevalence of EIPH was significantly associated with administration of furosemide (P = .01) and track surface (P = .01), but not speed index (P = .17; Table 8). Specifically, the probability of observing EIPH of any grade was lower in horses that received the prerace administration of furosemide when compared with horses that raced without furosemide (odds ratio [OR], 0.215; 95% CI, 0.069‐0.665).

For the tracks where postrace tracheoendoscopy was mandatory, there was no all‐weather surface and the probability of developing EIPH was lower on turf than on dirt (P = .01; OR, 0.398; 95% CI, 0.196‐0.807). The probability of observing EIPH was negatively associated with speed index (P = .01), indicating that faster horses had lower probability of displaying any tracheoendoscopic evidence of EIPH (Table 8). Because administration of furosemide was prohibited at the tracks at which tracheoendoscopy was mandated, it was not possible to determine whether furosemide administration had any effect on the prevalence of EIPH at those tracks.

3.4.2. EIPH severity

We also applied the multivariable logistic regression model to the outcome of EIPH severity (whether EIPH was severe [grade ≥3] vs not severe [grade <3]). The analysis indicated that the severity of EIPH differed significantly according to whether or not furosemide had been administered before the race (P = .01), and if tracheoendoscopy was mandatory (P < .001). We therefore also divided the results for tracks where tracheoendoscopy was voluntary vs mandatory. For tracks where involvement in the study was voluntary, the probability of severe EIPH was lower in horses receiving prerace administration of furosemide when compared with horses that raced without furosemide (P = .01; OR, 0.11; 95% CI, 0.019‐0.69). For the tracks where postrace tracheoendoscopy was mandatory, no significant association was found between EIPH severity and any of the model predictors.

4. DISCUSSION

Our results indicated that the prevalence of EIPH in this age group was similar to that reported for Thoroughbred racehorses of all ages and that severe EIPH (grades ≥3) also occurs in 2‐year‐olds at a similar prevalence (approximately 8%) to that of an all‐age population. ⁴ , ⁵ Another previous study also reported that the odds of 2‐year‐olds (actual number of horses not reported) having EIPH grade ≥1 were no different than for those of Thoroughbred racehorses of other ages. ⁷ The prevalence of EIPH grades = 0 in our study was approximately 10% less and that of mild to moderate EIPH (grades 1 and 2) 10% higher than that reported in a similar but larger study of Thoroughbreds of all ages. ⁵

All the horses in a previous study ⁵ ran on turf tracks whereas 78% of the tracheoendoscopic examinations in our study were performed after races on dirt tracks. The results of our study regarding the association between track surface and the prevalence of EIPH were equivocal. When track surface was treated as a single independent variable, no difference was found in the prevalence or severity of EIPH associated with races run on dirt or turf, but EIPH was less common at the track that only held races on an all‐weather surface. This finding was confirmed by multiple logistic regression analysis for the 14 tracks where enrollment in the study was voluntary and prerace administration of furosemide was permitted at 6 of them. However, the prevalence of EIPH was significantly higher after races run on dirt when compared with turf in Maryland where postrace tracheoendoscopy of 2‐year‐olds was legislatively mandated and the prerace administration of furosemide was banned, and reasons for the different results in this situation could not be ascertained. The finding regarding the lower prevalence and severity of EIPH at the track that raced on an all‐weather surface should not be overinterpreted. The horses racing at this track (track M) only accounted for approximately 2% of all the postrace examinations and tended to have slower speed indices and lower median earnings in the races after which they were examined than many of those racing on dirt and turf at other tracks. Most of them also received furosemide before they raced. Consequently, there were many factors besides the all‐weather track surface that might have contributed to the lower EIPH prevalence and severity at that track.

The finding that EIPH was common in young racehorses is compatible with reported evidence that 19 to 24 month‐old horses that were in training but had not raced showed histologic evidence of EIPH at necropsies conducted after their deaths because of nonpulmonary‐related reasons. ¹⁰ Sixty‐seven percent of the 340 horses in our study that were examined after their first race start had tracheoendoscopic evidence of EIPH. This number was only slightly less than the overall prevalence (74%) of EIPH, further emphasizing the fact that EIPH develops very early in Thoroughbreds' athletic careers. Furthermore, although not every horse had tracheoendoscopic evidence of EIPH on its first examination, the likelihood that a 2‐year‐old horse eventually would be EIPH positive increased with the number of times it was examined after racing. This observation was also in keeping with the findings of previous studies of Thoroughbreds and Standardbreds ³ , ¹¹ and is further evidence that EIPH is as common an event in 2‐year‐olds as it is in older Thoroughbred racehorses.

Before 2020, most 2‐year‐old Thoroughbreds racing in America competed after prerace IV administration of 250 mg furosemide and there was concern that, after the decisions by a large number of state racing commissions to prohibit this usage in 2020, 2‐year‐olds would experience bouts of EIPH that were more severe than had previously been the case in this age group. Our findings allay those concerns. Although no comment can be made about the prevalence and severity of EIPH in 2‐year‐olds before we conducted of our study, these results indicate that 2‐year‐olds racing without administration of furosemide did not experience EIPH more frequently or severely than has been observed in racehorses overall. ⁵ Prerace administration of furosemide to a large population of Thoroughbred racehorses previously has been shown to attenuate the severity of EIPH in the majority of them. ¹² Unfortunately, only approximately 12% of the horses that had postrace tracheoendoscopies in our study competed after receiving a prerace injection of furosemide. The lack of balance between the respective sizes of the 2 groups resulted in the analysis of the data being underpowered and it was not possible to determine whether prerace furosemide did or did not affect the severity of EIPH. Moreover, because of the prohibition placed on the administration of furosemide at most of the racetracks from which videorecordings were obtained, the data presented here lacks comparison between horses racing with and without furosemide administration at the same racetrack. However, results of the multivariable analysis indicated that, except for track L, the probability that horses racing after receiving furosemide would experience EIPH of any grade was lower than that of horses that raced without being treated with the drug. The results of the multivariable analysis that pertained to the use of furosemide were also difficult to interpret because the populations of horses running after receiving the drug were different than those running where its use was banned, because instances in which horses that were treated with furosemide ran against horses that did not receive it were negligible. Based on the lower speed indices, and the lower race purses and earnings, horses racing after being treated with furosemide were less capable athletes than those that competed without receiving the drug. Consequently, the apparent finding that the probability of developing EIPH was lower for any given speed index in horses racing after treatment with furosemide when compared with those that did not receive it must be regarded as equivocal and in need of additional study, especially because of the conflicting results related to horses racing at track L.

There was no apparent explanation for why the probability of EIPH increased for horses at track L as their speed index increased. The rules of racing in the state where track L was located specified that several drugs could be administered legally to horses before they raced, with the aim of mitigating or preventing EIPH. Whether any horses included in our study received any of these permitted drugs as well as furosemide is not known and, if that was the case, whether or not such an occurrence could have contributed to the finding can only be speculated upon. The probability that a horse would experience EIPH while racing had no direct relationship to the severity of the EIPH if EIPH occurred. Consequently, although our study lacked enough power to determine whether prerace administration of furosemide affected the severity of EIPH, it was apparent that, with the exception of track L, prerace treatment with furosemide decreased the overall probability that a horse would experience EIPH of grades ≥1.

The multivariable analysis confirmed that horses that performed better as reflected by their speed index, were more likely to be assigned an EIPH score = 0. Horses with lower speed indices had a higher prevalence of more severe EIPH. The finding that 2‐year‐old Thoroughbreds with EIPH grade ≥3 were less likely to finish in the first 3 places in a race was consistent with previous reports involving older Thoroughbred racehorses. ⁴ , ⁵ When the results regarding the association between EIPH of grades ≥3 and speed index and the likelihood of placing in the first 3 places in a race are taken together, there is reason to believe that EIPH of this severity can decrease a horse's performance.

Two‐year‐olds running distances that were categorized as intermediate (1300‐1500 m) had more marked EIPH than horses that ran at faster speeds over short distances and those that ran farther but slower. The intermediate distance category bridged 2 race distance groupings from another study that found that horses aged 2 to 10 years (median, 4 years) racing <1400 m were more likely to have EIPH grades ≥2 than those racing 1400 to 2400 m. ⁷ Another study of horses aged 2 to 7 years conducted in Rio de Janeiro, Brazil found that horses competing in long races (1600‐3500 m) had more severe EIPH than those racing shorter distances. ¹³ No specific information was provided regarding the number of races at each distance or the ages of horses competing in them. Consequently, the relevance of that study to any association between race distance and EIPH grade in 2‐year‐olds cannot be gauged.

Ninety percent of the 2‐year‐olds that underwent postrace tracheoendoscopy ≥3 times were observed to have EIPH on at least 1 occasion. This finding was consistent with those of other studies ³ , ¹¹ , ¹⁴ and reinforced the observation that the EIPH grade associated with 1 race is not predictive of the EIPH grade associated with the next race. It is possible that the findings from the repeated tracheoendoscopies of horses may have reflected an inherent bias because trainers had the option to enroll some animals, such as those known or suspected to have EIPH, and not others. This potential limitation was largely avoided because of the legislated mandate that all 2‐year‐olds racing in Maryland undergo postrace tracheoendoscopy, which accounted for the majority of horses examined multiple times (83%), and the assistance of veterinarians who routinely evaluated horses postrace for tracheal evidence of EIPH for trainer‐clients. Overall, these results supported the impression that EIPH is ubiquitous in racehorses. ¹⁵

Forty‐eight percent of the videorecordings and 38% of the horses in our study were from races in Maryland, and the severity of EIPH observed in those recordings and horses were lower than was seen in recordings obtained from the other tracks at which furosemide was prohibited. Postrace tracheoendoscopy was legislatively mandated for all 2‐year‐olds that raced in Maryland, but not at the other tracks in 2020. Previous reports related to the prevalence and severity of EIPH all have relied on voluntary enrollment of horses in those studies. One of the limitations of such studies has been the possibility that trainers who enroll their horses do so because they suspect that they have experienced EIPH and wish to have their suspicion confirmed, thus potentially biasing the findings of the study. Existing information regarding the prevalence and severity of EIPH is based on the results of these studies. The conditions under which 2‐year‐old racing was conducted in Maryland in 2020 provided an opportunity to examine every horse in the state's 2‐year‐old Thoroughbred racing population, regardless of any bias of a trainer. Although compliance with the legislative mandate might not have been 100%, >500 videorecordings from horses that ran in 113 races were submitted to our study. The results indicated that the prevalence and severity of EIPH in these horses were lower than they were in horses from the other tracks. One possible interpretation of this finding is that the prevalence and severity of EIPH in Thoroughbred racehorses are less than previously reported because those reports were based on results of investigations that had been unconsciously biased by the way in which the studies' subject populations had been selected. All of the horses in our study were 2‐year‐olds and in the earliest stages of their athletic careers, and there are other possible reasons for the differences between the results from Maryland and the tracks in other states. Ideally, a study of a large population of Thoroughbreds of all ages that is not based on voluntary enrollment should be conducted to provide a clearer understanding of the overall prevalence and severity of EIPH in Thoroughbred racehorses.

The finding that EIPH grade varied with the location of the racetrack was unexpected because there have been no prior reports suggesting such a relationship. However, previous reports that alluded to or specifically investigated possible non‐aged‐related risk factors for EIPH were undertaken in specific cities or localized regions ⁶ , ⁷ , ¹³ , ¹⁶ whereas, our study involved locations that covered the breadth of the United States and there was considerable environmental variability associated with this geographical diversity. The possible association of these potential risk factors with EIPH grade was beyond the scope of our study and warrants further investigation.

5. CONCLUSION

Two‐year‐old American Thoroughbred racehorses experience EIPH at a similar prevalence and level of severity as older horses. Prerace administration of furosemide lowered the probability that a horse would experience any grade of EIPH, but it remains unclear as to whether this treatment mitigated the severity of EIPH in horses of this age, and additional study is needed to further investigate this possibility. Exercise‐induced pulmonary hemorrhage of grades ≥3 was associated with worse racing performance in 2‐year‐old Thoroughbreds and was more likely to occur in races of 1300 to 1500 m. The location of the racetrack was associated with the prevalence and severity of EIPH for reasons that could not be determined and requires additional study.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF‐LABEL ANTIMICROBIAL DECLARATION

Authors declare no off‐label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Approved by the Washington State University IACUC, ASAF#6843.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

Supporting information

Data S1. Supporting Information.

Shoemaker S, Wang Y, Sellon D, et al. Prevalence and severity of exercise‐induced pulmonary hemorrhage in 2‐year‐old Thoroughbred racehorses and its relationship to performance. J Vet Intern Med. 2024;38(2):1167‐1176. doi: 10.1111/jvim.17003