<%BANNER%>

Epidemiology of Lameness and Athletic Performance in Thoroughbred Pinhooked Horses

Permanent Link: http://ufdc.ufl.edu/UFE0043535/00001

Material Information

Title: Epidemiology of Lameness and Athletic Performance in Thoroughbred Pinhooked Horses
Physical Description: 1 online resource (109 p.)
Language: english
Creator: Preston, Stephanie D
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: epidemiology -- horse -- lameness -- pinhook -- racetrack -- thoroughbred
Veterinary Medicine -- Dissertations, Academic -- UF
Genre: Veterinary Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This PhD research encompassed two main objectives. The first was to investigate associations between lameness and athletic performance in Thoroughbred pinhooked horses. Study results identified a high incidence of lameness during training, most injuries were diagnosed in joints, lameness was more frequent in hind than forelimbs, and speed was associated with sales price. Most study horses with a high purchase price had a reduced number of furlongs galloped and were classified as fast at the sales. Lameness alone was not associated with sales prices, however, lameness was associated with reduced exercise, which was more evident among lame horses with a high purchase price compared to lame horses with a low purchase price. These results support the conclusion that yearling pinhooked horses suffer a high incidence of lameness and that financial return was significantly different among horses classified as fast, average or slow at the sales and that yearling purchase price affects the trainers management decision on the amount of training allocation prescribed to lame horses. The second overall objective was to determine the prevalence of abnormal radiographic findings and their association to future racing performance. Study results determined prevalence rates for radiographic abnormalities in the fore and hind limbs, established the prevalence rate of presale arthroscopic surgery, and reported sales price was higher in horses with a history of presale surgery than in horses without. Statistical modeling was conducted to adjust for gender, history of arthroscopic surgery and purchase price and the odds of starting a race were lower in horses with forelimb proximal sesamoid osteophytes/enthesophytes and in horses with hind limb P1 OC fragments compared to horses that did not have these changes. Selected radiographic findings were not associated with outcome variables for starts placed, money earned, or money earned per start. These results support the conclusions establishing the prevalence of abnormal radiographic finds and presale arthroscopic surgery and determined that most radiographic abnormalities do not affect future racing performance.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Stephanie D Preston.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Hernandez, Jorge A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043535:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043535/00001

Material Information

Title: Epidemiology of Lameness and Athletic Performance in Thoroughbred Pinhooked Horses
Physical Description: 1 online resource (109 p.)
Language: english
Creator: Preston, Stephanie D
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: epidemiology -- horse -- lameness -- pinhook -- racetrack -- thoroughbred
Veterinary Medicine -- Dissertations, Academic -- UF
Genre: Veterinary Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This PhD research encompassed two main objectives. The first was to investigate associations between lameness and athletic performance in Thoroughbred pinhooked horses. Study results identified a high incidence of lameness during training, most injuries were diagnosed in joints, lameness was more frequent in hind than forelimbs, and speed was associated with sales price. Most study horses with a high purchase price had a reduced number of furlongs galloped and were classified as fast at the sales. Lameness alone was not associated with sales prices, however, lameness was associated with reduced exercise, which was more evident among lame horses with a high purchase price compared to lame horses with a low purchase price. These results support the conclusion that yearling pinhooked horses suffer a high incidence of lameness and that financial return was significantly different among horses classified as fast, average or slow at the sales and that yearling purchase price affects the trainers management decision on the amount of training allocation prescribed to lame horses. The second overall objective was to determine the prevalence of abnormal radiographic findings and their association to future racing performance. Study results determined prevalence rates for radiographic abnormalities in the fore and hind limbs, established the prevalence rate of presale arthroscopic surgery, and reported sales price was higher in horses with a history of presale surgery than in horses without. Statistical modeling was conducted to adjust for gender, history of arthroscopic surgery and purchase price and the odds of starting a race were lower in horses with forelimb proximal sesamoid osteophytes/enthesophytes and in horses with hind limb P1 OC fragments compared to horses that did not have these changes. Selected radiographic findings were not associated with outcome variables for starts placed, money earned, or money earned per start. These results support the conclusions establishing the prevalence of abnormal radiographic finds and presale arthroscopic surgery and determined that most radiographic abnormalities do not affect future racing performance.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Stephanie D Preston.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Hernandez, Jorge A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043535:00001


This item has the following downloads:


Full Text

PAGE 1

1 EPIDEMIOLOGY OF LAMENESS AND ATHLETIC PERFORMANCE IN THOROUGHBRED PINHOOKED HORSES By STEPHANIE DYER PRESTON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Stephanie Dyer Preston

PAGE 3

3 To my son Cole, thank you for making me the person I am and believing in all that I do. I strive to be a better person beca use of you. Remember, no one can ever take yo ur education away and that makes it priceless.

PAGE 4

4 ACKNOWLEDGEMENTS I would like t o thank Dr. Jorge Hernandez for serving as my advisor and providing guidance and perspective throughout the good and bad persona l and professional times I experienced while pursuing this degree I would like to thank the rest of my committee Drs. Murray Brown, Terese Chmielews ki, Troy Trumble, and Dana Zimmel, for providing additional knowledge, and support. I also appreciate the A ckerly Brothers, Oxley Foundation, the Siegel f amily the Casner family and all of the other contributors to the Equine Soundness Program (ESP) for providing financial support for these studies. I am grateful for Dr s. Ed Murray and Eleanor Green for introd ucing the idea of ESP and my entire program; the College of Veterinary Medicine and all VMS students are lucky to have you. Many individuals have inspired me throughout my education and I would likely not be where I am today without their influence and encouragement. My early education in the Santa Barbara Gifted and Talented Education prog ram inspired me to think outside of the box. My parents encouraged me to leave high school early to follow my dream of becoming a Doctor. My beloved, late mother always encouraged me to be fiercely ruly love your partner, cook passionately, and best of all, have more children. How you surrendered I will never understand but I wish you peace and happiness wherever you are. My Master of Science graduate committee members Dr s William Moyer and Noah Co hen initiated my interest in understanding racing injuries and the issues pertaining to the welfare of race horses. I thank Dr. J ames Benjamin who taught m e how to be successful in

PAGE 5

5 academia while still having fun On both a personal and professional level I owe sincere gratitude to Dr. David Feinberg. During the most difficult time in my life, y ou made me appreciate the difference between having a PhD and not. I wish I had never compromised on your recommendations but I did not know then what I know now I cannot express enough gratitude for my family and friends, who have given me so much love and support over the years. My parents have encouraged all of my pursuits, and I am grateful for the sacrifices they have made to help me reach my potential. I look f orward to giving back now that this milestone has been achieved. I thank Jimmie and Gina Price. Your love has been an inspiration that a soul mate exists and helped me find mine. Your incredible loyalty and professionalism protected us from the evil empire and proved sometimes the guys in white hats win Cole and I will always be grateful for your support and loyalty. Most of all I thank the love of my life, Cole Dyer Preston my only son You provide me with the perspective of what is important in life and drive me to be the best that I can be. Your smile and happiness are worth more than anything else in the world to me. I thank Ashlin and Linden Meuser for their acceptance during a difficult transition in their lives. F inally I thank my dear husband Mike Meuser, the second love of my life and only true partner You gave me inspiration and the belief that tomorrow will be better than today. You always built me up instead of put ting me down. You made me a bett er, stronger, professional, spouse and human. I cannot wait to have our baby.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGEMENTS ................................ ................................ ............................... 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 14 2 REVIEW OF THE LITERATURE ................................ ................................ ............ 18 Pinhooked Thoroughbred Yearling Horses ................................ ............................. 18 Pathogenesis of Lameness and Osteoarthritis ................................ ....................... 22 Clinical Diagnostics Used to Identify Lameness ................................ ..................... 25 Epide miologic Studies of Lameness and Athletic Performance in TB Yearlings ..... 28 Epidemiologic Studies of Yearling Sale Purchase Price, Exercise History, Lameness and Speed as Predisposing Factors for Sales Price in 2 year old in Training Horses ................................ ................................ ............................... 29 Epidemiologic Studies of Presale Abnormal Radiographic Findings and History of Arthroscopic Surgery in TB Yearlings ................................ .............................. 30 Epidemiologic Studies of Presale Abnormal Radiographic Findings in TB Yearlings and Future Racing Performance ................................ .......................... 31 3 LAMENESS, ATHLETIC PERFORMANCE, AND F INANCIAL RETURNS IN THOROUGHBRED PINHOOKED YEARLING HORSES ................................ ....... 33 Materials and Methods ................................ ................................ ............................ 34 Results ................................ ................................ ................................ .................... 38 Discussion ................................ ................................ ................................ .............. 39 4 EPIDEMIOLOGIC ANALYSIS OF YEARLING SALE PURCHASE PRICE, EXERCISE HISTORY, LAMENESS, AND SPEED AS PREDISPOSING FACTORS FOR SALES PRICE IN 2 yea r old IN TRAINING THOROUGHBRED HORSES ................................ ................................ ................................ ................. 48 Materials and Methods ................................ ................................ ............................ 49 Results ................................ ................................ ................................ .................... 53 Discussion ................................ ................................ ................................ .............. 54

PAGE 7

7 5 ASSOCIATIONS OF PRESALE RADIOGRAPHIC FINDINGS AND PRESALE ARTHROSCOPY WITH SALES PRICE IN THOROUGHBRED YEARLINGS SOLD IN KENTUCKY ................................ ................................ ............................. 60 Materials and Methods ................................ ................................ ............................ 61 Results ................................ ................................ ................................ .................... 67 Discussion ................................ ................................ ................................ .............. 69 6 EPIDEMIOLOGIC ANALYSIS OF SELECTED YEARLING PRESALE RADIOGRAPHIC FINDINGS AND THEIR ASSOCIATION WITH RACING PERFORMANCE IN 2 year old THOROUGHBRED HORSES .............................. 77 7 SUMMARY AND RESEARCH CONCLUSIONS ................................ ..................... 97 Summary ................................ ................................ ................................ ................ 97 Research Conclusions ................................ ................................ ............................ 97 APPENDIX A SOURCES OF MEDICATION: ................................ ................................ .............. 102 B RADIOGRAPHIC VIEWS ACQUIRED FOR EACH HORSE (32 TOTAL VIEWS): 103 REFERENCES ................................ ................................ ................................ ............ 105 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 109

PAGE 8

8 LIST OF TABLES Table page 3 1 Anatomic location of lameness that develop ed during training in 37 of 40 yearling Thoroughbreds that were bought for the purpose of resale for profit and were classified as fast, average, or slow at the 2 year old in training sales. ................................ ................................ ................................ .................. 44 3 2 Comparison of sex, age, purchase price, financial returns, and exercise variables in 40 yearling Thoroughbreds that were bought for the purpose of resale for profit and were classified as fast, average, or slow at the 2 year old in training sales. ................................ ................................ ................................ .. 45 4 1 Univariable analysis comparing median sales price in study horses sold at 2 year old in training sales according to gender and other investigated exposure factors. ................................ ................................ ................................ 57 4 2 Multivariable analysis (ANOVA). Analyses of median sales price in horses sold at 2 year old in training sales. ................................ ................................ ..... 58 4 3 Median number of furlongs galloped and breezed 1 to 60 days before the sale among horses with high or low yearling sale purchase price. ..................... 59 4 4 Median number of furlongs galloped and breezed 1 to 60 days before the sale among horses affected or not affected with lameness. ............................... 59 4 5 Observed combined effect between yearling sale purchase price and lameness during training on number of furlongs galloped and numbe r of furlongs breezed 1 to 60 days before the sale in study horses. .......................... 59 5 1 Frequency distribution of surgery site for horses offered for sale at the Keeneland September 2006 yearling sale tha t had a history of presale arthroscopy. ................................ ................................ ................................ ........ 74 5 2 Sales price of Thoroughbreds offered for sale at the Keeneland September 2006 yearling sale. ................................ ................................ ............................. 75 5 3 Comparison of sales prices for Thoroughbred yearlings sold at the September 2006 Keeneland yearling sale with and without various presale radiographic findings or a history of presale arthroscopy. ................................ .. 76 6 1 Distribution of gender, history of arthroscopic surgery, sale day, purchase price and radiographic findings in Thoroughbred yearlings that did or did not race as 2 year olds ................................ ................................ ............................. 92

PAGE 9

9 6 2 Adjusted odds ratios (OR) and 95% confidence intervals (CI) for logistic regression analysis of the association between pre sale radiographic findings in Thoroughbred yearlings and failure to start a race as 2 year olds .................. 93 6 3 Distribution of gender, history of arthroscopic surgery, sale day, purchase price and radiographic findings in Thoroughbred yearlings that finished 1 st 2 nd or 3 rd place as 2 year olds ................................ ................................ ........... 94 6 4 Comparison of money earned among horses that raced as 2 year olds by gender, history of arthroscopic surgery, sale day, purchase price and radiographic findings ................................ ................................ .......................... 95 6 5 Comparison of money earned per start among horses that raced as 2 year olds by gender, history of arthroscopic surgery, sale day, purchase price and radiographic findings ................................ ................................ .......................... 96

PAGE 10

10 LIST OF FIGURES Figure page 3 1 Time to first event of lameness among 40 yearling Thoroughbreds that were bought for the purpose of resale for profit after training ................................ ...... 46 3 2 Distribution of new cases of lameness within 1 to 150 days prior to sale* in 37 of 40 yearling Thoroughbreds that were bought for the purpose of resale for profit after training. ................................ ................................ ........................ 47

PAGE 11

11 LIST OF ABBREVIATION S CI Confidence interval DMD Dorsal metacarpal disease MC3 Third metacarpal bone MCP Metacarpophalangeal MT3 Third metatarsal bone MTP Metatarsophalangeal OC Osteochondral OCD Osteochondritis dissecans OR Odds r atio P1 Proximal phalanx RR Risk ratio TB Thoroughbred

PAGE 12

12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EPIDE MIOLOGY OF LAMENESS AND ATHLETIC PERFORMANCE IN THOROUGHBRED PINHOOKED HORSES By S tephanie D yer P reston December 2011 Chair: Jorge A. Hernandez Major: Veterinary Medical Sciences This research ha d two main objectives. The first was to investigate associations between lameness and athletic performance in Thoroughbred pinhooked horses. Study results identified a high incidence of lameness during training, most injuries were diagnosed in joints, lameness was more frequent in hind than forelimbs, and speed was associated with sale price. Most study horses with a high purchase price had a reduced number of furlongs galloped and were classified as fast at the sales. Lameness alone was not associated with sale prices but lameness was associated with redu ced exercise, which was more evident among lame horses with a high purchase price compared to lame horses with a low purchase price. These results support the conclusion that yearling pinhooked horses suffer a high incidence of lameness and that financial ret urn was significantly different among horses classified as fast, average or slow at t he sales and that yearling purchase price affects the management decision on the amount of training allocation prescribed to lame horses The second overall objective was to deter mine the prevalence of abnormal ra diographic findings and their association with future racing performance. Study results determined prevalence of radiographic abnormalities in the fore and hind limbs,

PAGE 13

13 establi shed the prevalence of pr esale arthroscopic surgery, and reported sale price s w ere higher in horses with a history of presale surgery than in horses without. Statistical modeling was conducted to adjust for gender, history of arthroscopic su rgery and purchase price The odds of s tarting a race were lower in horses with forelimb proximal sesamoid osteophytes/enthesophytes and in horses with hind limb P1 OC fragments compared to horses that did not have these changes. Selected radiographic findings were not associated with outcome v ariables for starts placed, money earned, or money earned per start.

PAGE 14

14 CHAPTER 1 INTRODUCTION In Thoroughbred racing circles, practice of buying a horse specifically for resale and profit. Pinhooked yearling TB s are equine athletes 16 20 months old that are purchased at public auctions during summer months for the purpose of selling them the following spring at 2 year old in training sales. The price of yearling TBs can vary significantly based on their pedigree and conformation. At the Keeneland September yearling sales, elite horses wit h prestigious pedigree s are offered for sale during the first four days of the sale; the value of these horses is higher, compared to horses offered for sale during the next 10 d ays of the sale. After the yearling sales, trainers have approximately 5 months to break, condition, and prepare their horses to work at high speed during training in preparation for the 2 year old in training sales. At these sales, trained TB yearlings ra ce at high speed (i.e., breeze) for 1 or 2 furlongs (1/8 or 1/4 mile; ~ 200 or 400 meters) to assess their athletic performance (speed). After the 2 year old in training sales, the TB yearlings start racing as 2 year olds. Associations between yearling sal e purchase price in the US and such exposu re factors during training as lameness, training patterns (eg, exercise distance accumulated during training), and speed recorded at the 2 year old in training sales have not been well investigated. Based on indust ry empi rical knowledge and beliefs, a TB with a stronger pedigree, good conformatio n, and a high purchase price is expe cted to be sold for a high sale price at the 2 year old in training sales, compared to horses without these characteristics However, the criteria used to formulate purchase decisions are different at the yearling sale and 2 year old in training sale. At the 2 year

PAGE 15

15 old sale buyers have the added benefit of assessing the breeze performance of all horses offered for sale in addition to confo rmation, pedigree, and veterinary pre purchase exam results. Although it seems intuitive that high priced yearlings would be high priced 2 year olds there is no objective data in the literature to identify and quantify factors associated with sale price at both sales and what intrinsic (i.e., pedigree, conformation, etc.) or extrinsic factors (i.e., lameness, exercise history etc.) may be similar or different for individual horses sold in both sales. For example, a yearling sold with a desirable pedigree m ay later be sold as a 2 year old by a sire that is no longer fashionable due to lack of racing performance of offspring in the interim. In addition, a middle quality horse based on yearling sale price, pedigree, and conformation may prove exceptional durin g the breeze at the 2 year old sale. Lameness may not be evident in yearlings due to medication use or because of the relative lack of cyclic trauma that accrues during training. These horses may be susceptible to lameness as they enter training for 2 year old sales due to conformation defects or pre existing injury that was not observable at the yearling sale. It is important to understand what variables influence price and athletic performance for buyers, consignors, and veterinarians to better assess inf ormation available to them at the sales and to include objective knowledge in the scientific literature Historically, the economic value of TB yearlings offered for sale as future racehorses has been influenced by their pedi gree and conformation. P rospec tive buyers have paid increased attention to the health status of yearlings, with emphasis on the results of presale radiographic examinations, a history of presale arthroscopic surgery, and the results of endoscopic evaluation of the upper airway. In the TB industry, there is

PAGE 16

16 a perception that results of presale radiograph ic examinations affect the sale price of yearlings. However, knowledge on the association between the presence of specific radiographic findings and yearling sale price is limited to one study of 348 TB horses offered for sale in Texas in 2002 and 2003. 1 Results from that study were inconclusive, as study horses were not randomly selected, and only horses with lesions of the proximodorsal aspect of the sagittal ridge of MC3 or MT3 on pre sa le radiographs had lower sale price s compared to h orses without such lesions. Another previous stud y 2 investigated association s between presale radiographic findings in TB yearlings and future racing performance in the US but study conclusions were differ ent from the Texas study 1 In a study 2 of 1,162 TB yearlings offered for sale at two different sales in Kentucky between 1993 and 1996, yearlings with moderate or extreme palmar supracondylar lysis of MC3, dorsal medial middle carpal joint disease, entheso phyte formation on the fore proximal sesamoid bones, or hindlimb P1 OC fragments were less likely to st art a race as 2 or 3 year olds compared to yearlings without these radiographic findings. However, in another study 1 of 348 TB yearlings offered for sal e in Texas between 2002 and 2003, no radiographic findings were associated with future racing performance as 2 and 3 year olds In each study, the sample of horses used was not randomly selected from each corresponding sale. These studies were based on rad iographic examinations of yearlings obtained from private veterinary practices, and the study populations were different. Because previous s tudies have produced inconclusive and contradictory results, the epidemiologic associations between presale radiogra phic findings in TB yearlings and future racing performance in the US have not been well established.

PAGE 17

17 The objectives of the PhD research reported here were: 1) to characterize lameness during training and compare exercise variables and financial returns am ong yearling TBs that were bought for the purpose of resale for prof it; 2 a) to examine the relationships between yearling sale purchase price, exercise history, lameness, and speed a s predisposing factors for sale price in 2 year old in training TBs; 2 b) t o compare exercise distance accumulated during the last 60 days before the sales between horses with high and low purchase price and between horses affected and not affected with lameness during training; 3) to determine prevalence of various presale radio graphic findings and of presale arthroscopy in horses offered for sale at the 2006 Keeneland September ye arling sale and to compare sale price between yearlings with and without various presale radiographic findings or a history of arthroscopy; 4) to exami ne the associations between selected yearling presale radiographic findings and racing performance parameters in 2 year old TBs.

PAGE 18

18 CHAPTER 2 REVIEW OF THE LITERA TURE Pinhooked Thoroughbred Yearling Horses In racing circles, pinhooking is a term used to des cribe the practice of buying a horse specifically for resale and profit. Pinhooked yearling TBs are purchased at public auctions from July through October for the purpose of selling them the following year (February through April) at what are known as 2 ye ar old in training sales. Trainers have approximately 5 months to train the horse to accept a rider and condition, and prepare their horses to work at high speed during training in preparation for the sales. At the sales, the trained 2 year old horses bree ze (race individually) 1 or 2 furlongs (1/8 or 1/4 of a mile). Among the 2 year old in training sales that occur in a given year, the earlier sales (eg, sales in February) are considered the most important because the mean sale price is higher, compared wi th that of later sales. After purchase, the horses start racing as 2 year old s. Knowledge on lameness during training in pinhooked horses is limited. An epidemiologic study 3 was conducted at a training center in Florida that purchased 40 yearlings solely f or the purpose of resale at 2 year old s in training sales. The study was designed to compare financial returns between 20 yearling horses that respectively had mild training failure (i.e., 1 to 11 days lost) versus 20 yearlings with severe training failur e (i.e., 13 to 108 days lost); and between 8 and 32 yearlings that had planned and non planne d training failure 3 Horses with planned training failure were those identified with radiographic abnormalities during routine pre purchase examinations at yearlin g sales and were treated surgically after purchase but prior to training; horses with non planned training failure were horses that were considered to have normal radiographs at

PAGE 19

19 the yearling sales. Median financial returns were lower for horses that had s evere training failure ($1,000), compared to horses with mild training failure ($24,000) (P < 0.05). In addition, median returns were lower in horses with radiographic abnormalities ( $2,000), compared to normal horses ($10,000). 3 Lameness, planned traini ng failure, respiratory disease, and ringworm were the most commonly reported causes of training failure defined by training days lost 3 A d iagnosis of training failure due to lameness, respiratory disease, or ringworm was presumably det ermined by the trai ner but this was not reported so it is unknown whether a veterinarian confirmed the diagnosis, which ability to assess the potential for misclassification bias. Another limitation of this study 3 was that breezing speed during the 2 yea r old in training sales, which is perceived to b e an important variable in sale price throughout the TB industry, was not recorded. In addition, lameness was not characterized and risk factors for lameness were not explored. Knowledge on exercise, adaptat ion to exercise, and the impact of specific training regimens on lameness in TB horses is limited. In the US, forced exercise and other sales prepping techniques are tho ught to be wide spread but only one previous study 4 documents the prevalence of exercis e an d impact of lameness in horses sold at public auctions. The previous study was limited to horses in Texas; included mostly Quarter horses (64%); and most were weanlings (66%) rather than yearlings. 4 Previous studies have shown that structural damage re sults when the rate of tissue repair is exceeded by micro damage caused by the strain of ongoing exercise specific adaptation to a particular function strain caused by exercise can exceed the rate of tissue repair (i.e., bone remodeling rate) resulting in structural damage. 5 9 Early exercise

PAGE 20

20 has been reported to increase mass, cortical thickness, and structural bone strength in horses. 10 Recent studies report providing young, growing horses with an environment that enhances functional adaptation of the skel etal structure to exercise, without adversely affecting growth or inducing lameness, may result in greater structural integrity and durability of musculoskeletal tissues later as racehorses. 10 12 I n a previous study, 13 box stall confinement combined with e xercise was reported as a risk factor for overstimulation of bone mineral density in weanling horses but this association has not been reported in yearlings. Another study, 1 0 reported that weanling horses exercised at a medium trot for 20 minutes 5 days per week s uffered no lameness problems and improved the stress bearing characteristics (3 rd metacarpal density and circumference) without affecting growth. O ther studies 14 1 6 reported sub maximal loading as optimal for healthy musculoskeletal development and heavie r more strenuous exercise adverse to long term viability of tissues. Previous studies have reported differing results due to different exercise loads, frequencies, duration, and age of onset. 10 15 Overall, there seems to be a consensus in previous reports about the positive impact of early exercise on later musculoskeletal tissue health and adaptive strength but the amount, frequency, duration, intensity, and age of onset for the most optimal exercise training regimens requires further study. Previous findi ngs demonstrate the opportunity and importance of developing management strategies for exercise to optimize musculoskeletal health and reduce lameness. However, before modifications to training programs can be made, current industry practices must be know n and quantified in a standardized method (Chapters 3, 4).

PAGE 21

21 As reported in Chapter 3 of our study, lameness, exercise, and other sales preparation techniques prior to the yearling sale were not disclosed in the repository and therefore, not assessed in our studies We expected that lameness in pinhooked yearling to 2 year old in training horses may be characteriz ed differently from lameness in yearlings or later as racehorses due to the accelerated training regimens typically used in preparation for the 2 y ear old in training sales. Identifying specific anatomic sites and frequencies of lameness may also help to determine whether training and clinical treatment and management practices can be improved to identify sites of commonly diagnosed musculoskeletal i njuries before lameness occurs, which would likely result in improved well being of the animal and economic benefits. The 2 year old in training sales process is an ideal model to study lameness issues pertaining to racing horses where the economic impact, humane considerations, horse and jockey safety, as well as public perception can help formulate lameness prevention strategies. It is critical to study lameness in TB yearling horses to more completely understand the exposure factors associated with lamen ess health and safety concerns for both horse and rider, and to formulate effective lameness prevention strategies. This PhD research contains four epidemiological studies performed to better understand what horse related (intrinsic) and exposure factors increase risk for lameness in yearlings during training for 2 year old sales, and what population (extrinsic) factors are associated with increased lamene ss. The literature consists of one study performe d on TB pinhooked yearlings. This first study 3 provid ed insight into the association between financial returns and lameness but left many important questions unanswered. Numerous studies exist on the risks and frequency of catastrophic injury in race

PAGE 22

22 horses 1 7 2 5 yet, we have reportedly made little progress i n reducing lameness as measured by training days lost. 2 6 Knowledge o f catastrophic injuries is important to further our understanding of risk factors for sustaining fatal injuries, but there is a lack of data describing the pathogenesis of musculoskeletal injury prior to the point of failure. If risk factors for the incidence and progression of lameness are understood early in the disease process then prevention policies can be implemented before fatalities occur. The research report ed in Chapters 3, 4, 5, and 6 includes associations of lameness, exerc ise, athletic performance, sale price, and presence of radiographic abnormalities in TB yearlings, 2 year old in training horses, and later as 2 year old racehorses in the US. This knowledge would contribute t o the formulation of prevention and control strategies for lameness reductions to protect the health and welfare of 2 year old in training horses. Pathogenesis o f Lamenes s a nd Osteoarthritis In normal joints, the pathogenesis of athletic injuries is report edly due to chronic fatigue damage. 27 30 The biomechanical properties of articular cartilage and subchondral bone play crucial roles in the efficient performance of a healthy joint. Disruption of normal joint homeostasis involving any joint component may r esult in a cascade of inflammatory and degradation events, which may result in pain, lamenes s, and eventually progress ion to osteoarthritis (OA) of the joint. Cyclic trauma incurred during exercise causes alterations in the homeostasis of the normal health y joint from repetitive cyclic stress of abnormal intensity, duration, or frequency. 2 7 30 When exercise induced trauma exceeds the rate of repair, injury results. Results from previous studies in racehorses support the hypothesis that many injuries result from accumulation of

PAGE 23

23 long term insidious skelet al and soft tissue damage, 3 1 a result of repeated loads during high speed exercise. 27 3 1 The response of articular cartilage to strenuous exercise appears to be influenced by the state of the cartilage before exercise, intensity of the exercise program, and by species variability. 2 7 30 3 2 Healthy chondrocytes respond to mechanical loading and mai ntain the homeostatic metabolism of the extra cellular matrix. 27 30 3 2 When the joint is overloaded, maintaining matrix homeostasis becomes difficult. Changes in articular cartilage, synovial membrane, subchondral bone, and SF initiate OA when they singula rly or collectively are no longer able to absorb and dissipate loads applied across the joint. 2 7 30 3 2 All the joint components are interrelated, so that any alteration to one structure potentially changes properties of the other structures in the joint. 2 7 Initially, proteoglycan and protein destruction expose cartilage collagen fib e r s and depress the hyalurona n ( HA ) content. 27 30 3 2 Eventually, OA can develop which is a common musculoskeletal problem in racehorses. 2 7 30 3 2 The development of OA is complex involving the subchondral bone, synovium and articular cartilage 2 7 30 3 2 lea d ing to erosion of cartilage, sclerosis of subchondral bone, and inflammation of the synovial membrane leaving the joint unabl e to distribute loads across that joint. It is a painful condition that progresses to narrowing of the joint spaces and difficulty with movement. Currently, the diagnosis of OA in horses is limited to clinical diagnostics, including physical examination and diagnostic blocks, and imaging techniques such as radiograph s, ultrasound, nuclear scintigraphy, and magnetic resonance imaging (MRI) Therapeutic interventions primarily attempt

PAGE 24

24 palliative means to control clinical signs of inflammation and pain but do little to reduce or reverse the joint destruction. 2 7 30 3 2 The two most commonly injured joints in racehorses are the carp us (kn ee) an d fetlock (ankle) 1 7 ,2 6 ,35 Osteochondral fragmentation occurs in athletic horses after repeated trauma or chronic fatigue exceeds the remodeling capacity of the subchondral bone Evidence of a relationship between intra articular osteochondral fractu res and pre existing subchondral bone injury has rece ntly been investigated 3 2 The development of lytic lesions in subchondral bone was presumed to be associated with microdamage due to cyclic trauma. Subchondral bone disease commonly manifests in osteocho ndral chip fractures. 3 2 Resulting inflammation in the joint lead s to a cascade of degradation events. Accumulated joint damage causes chondrocyte injury, enzymatic degradation of proteoglycans and collagen, and decreased synthesis of extracellular matrix, all of which are important steps in articular cartilage degeneration, and the pathogenesis of OA. 27 30 3 2 Understanding the normal and pathologic responses of the skeletal system is as important to preventing OA as it is to treating the disease. 3 2 When a horse suffers an osteochondral chip fracture, the injury eventually progresses to OA. Surgical removal of the bone fragments may help alleviate inflammation, degradation, and clinical lameness. D isease modifying OA drugs (DMOADs) are frequently used to con trol the clinical manifestations of inflammation and pain in the joint s of racehorses In racehorses, unlike humans, the typical patient affected with OA is a young skeletally immature animal that experiences performance limitations which may eventually lead to athletic retirement. Ideally, intra articular therapy resol ves clinical

PAGE 25

25 signs rapidly, resulting in chondroprotecti on and disease modification. Treatment is complicated by the dynamic nature of OA Structural damage may precede clinical signs and l ate stage diagnosis precludes satisfactory resolution. Variable p rogress ion rates and individual responses to similar lesions and treatments make accurate prognostication difficult Clinical Diagnostics Used t o Identify Lameness The diagnosi s of lameness has been described as an art (i.e., skilled workmanship, craft, or studied action) since the late 1800s. 29 A thorough physical examination includes, but is not limited to, basic vital statistics, conformation, symmetry and posture assessment, palpation, fl exion tests, and observation of movement. Initially, a passive exam ination is performed that evaluates the entire horse for signs of heat, swelling effusion, and/or pain. 27 30,32 T he horse is observed at rest to assess conformation and symmetry and then in motion to assess symmetry of movement Often the horse is led in a straight line over a hard surface at the walk and trot. Additionally, the veterinarian may wish to see the horse circle or exercise under tack to identify indicators of lameness such as a head nod, hip hike, shortness of stride, or decreased flexion. L ameness is then graded by the attending veterinarian by use of a 0 to 5 grade lameness scoring system adopted by the American Association of Equine Practitioners. 33 In this scoring system, g rade 0 = no lameness; 1= lameness is difficult to observe and not consistently apparent, regardless of circumstances (eg, weight carrying, circling, or traveling on inclines and hard surfaces); 2 = lameness is difficult to observe at a walk or when trottin g in a straight line but is consistently apparent under certain circumstances (eg, weight carrying, circling, or traveling on inclines and hard surfaces); 3 = lameness is consistently observable at a trot under all circumstances; 4 =

PAGE 26

26 lameness is obvious at a walk; and 5= lameness results in minimal weight bearing in motion or at rest or complete inability to m ove. Both the passive and active examinations provide clues used in isolating the area of lameness, which can then be further confirmed with ancillary diagnostics. 27 30,32 Perineural or intra articular analgesia may be used to further localize the anatomic location of lameness in horses without obvious evidence of joint effusion or pain elicited during flexion of a joint. 27 30,32 Once the location has been identified appropriate imaging techniques are then utilized to further characterize the problem Currently, imaging techniques include: r adiographs, ultrasound, nuclear scintigraphy and magnetic resonance imaging (MRI). Although MRI is the best te chnique used for musculoskeletal pain in humans, its use in horses is limited. Current systems are available that al low images to be acquired in a standing, awa ke horse but the image qua lity is limited and the technique is sensitive to any movement. As a result, general anesthesia is required to obtain diagnostic MRI proximal to MCP/ MTP joints, which greatly increases the risk of complications and increases diagnostic costs. As a consequence, veterinarians often use other diagnostics such as radiography or ultrasonography to confirm the injury. 27 30 Although radiographs provide limited information for subtle changes or the progression of OA they are the gold standard for the equine industry for assessing the musculoskeletal system after the lameness ha s been localized 32 Radiographs provide identification of pathologic changes that have already occurred to the musculoskeletal system (primarily bone). However, these changes may or may not be related to the actual cause of lameness. Therefore, radiograph s must be combined with a thorough

PAGE 27

27 lameness exam to be able to interpret the significance of any changes identified. Radiographs are often used to screen horses for the presence of pathologic changes and are common at many TB horse auctions (yearling and 2 year old in training (post breeze) sales as required by the sales companies). It is important to note that even though radiographs are taken, a thorough lameness exam as described above is not traditionally performed. However, the potential for compariso ns between films of the yearling and 2 year old in training sales may help identify any changes in bone and the presence of skeletal adaptation to the training regimens This technique has not been validated by objective research methods. During training of young TB racehorses it is common practice for horses to be examined daily for signs of lameness by the rider or trainer on the racetrack during or immediately after an exercise event (gallop ing or breezing). If there are any signs of unsoundness identif ied, the attending veterinarian is often asked to examine the horse. However, because there are minimal diagnostic techniques that can be performed at a barn in a standing horse many injuries may remain undiagnosed and are often treated empirically Thus the horse may continue to train or race with mild clinical lameness and as a result, create further injury to the affected structures. C atastrophic injuries in TB racehorses have been previously reported to result from injuries secondary to undiagnosed o r untreated primary injuries. 31 If primary injuries can be identified and treated catastrophic failure s could potentially become less common 31 Therefore, many o wners and trainers often prefer to rest a valuable horse or reduce levels of training if a hor se demonstrates lameness, even if a definitive diagnosis c annot be made. 27 30,32 The resulting loss of training days may negatively affect resale price or future

PAGE 28

28 performance, emphasizing the need for timely and accurate diagnoses for optimal management of training schedules Epidemiologic Studies o f Lameness a nd Athletic Performance i n TB Yearlings Knowledge on lameness in yearling horses is limited to one previous study. 3 After yearlings are purchased for pinhooki ng trainers have approximately five mo nths to break, condition, and prepare their horses to work at high speed during training in preparation for the 2 year old in training sales. This intense accelerated schedule places unique demands on pinhooked yearling horses compared with yearlings that are more conventionally developed into racehorses. 3 An epidemiologic study 3 was conducted at a training center in Florida that purchased 40 yearlings solely for the purpose of resale at 2 year old s in training sales. The study was designed to compare fina ncial returns between 20 yearling horses that had mild training failure (i.e., 1 to 11 days lost) versus 20 yearlings with severe training failure (i.e., 13 to 108 days lost); and between 8 and 32 yearlings that had planned and non planned training failure respectively. Horses with planned training failure were those identified with radiographic abnormalities during routine pre purchase examinations and were treated surgically; horses with non planned training failure were horses that were considered to ha ve normal radiographs at the yearling sales. Median financial returns were lower in horses that had severe training failure ($1,000), compared to horses with mild training failure ($24,000) (P < 0.05). In addition, median returns were lower in horses wit h radiographic abnormalities ( $2,000), compared to normal horses ($10,000). Lameness, planned training failure, respiratory disease, and ringworm were the most common and important causes of training failure. 3 In the US, exercise and other sales prepping techniques are perceived to be wide spread in the TB industry, however, only one previous study 4

PAGE 29

29 documents the prevalence, frequency, and duration of exercise and incidence of lameness in yearling horses sold at public auctions in Texas. This study was lim ited to horses sold in Texas, biased with non random sampling methods, included more Quarter horses than TBs, and more weanlings than yearlings. 4 Therefore, a gap in the literature exists regarding what happens to yearling TB horses during training in prep aration for the 2 year old in training sales. Epidemiologic Studies o f Yearling Sale Purchase Pri ce, Exercise History, Lameness a nd Speed a s Predisposing Factors for Sales Price in 2 year old i n Training Horses Knowledge of yearling sale purchase price an d exposure factors such as lameness during training and speed recorded at the 2 year old in trai ning sales associated with sale price in 2 year old in training TB horses in the US is very limited. The only study 3 conducted on yearling pinhooked horses reve aled that incidence of lameness was highest during the high speed exercise preparation before the sales among yearling TBs entered into 2 year old in training sales. In general, TBs with a high pedigree, good conformation and a high purchase price are expe cted to be sold for a high sales price at the 2 year old in training sales, compared to horses without these characteristics. Speed is an additional variable perceived to influence the sales price in TBs sold at the 2 year old in training sales. The relat ionship between training patterns (such as exercise distance accumulated during training) and sales price in TBs sold at the 2 year old in training sales have not been well investigated. It is known that lameness can affect the training schedule of yearlin g TBs by causing days lost during training 3 or reduced exercise. However, it is not known if purchase price has an effect on training schedules prescribed to yearling TBs affected with lameness. Therefore, a gap in the literature

PAGE 30

30 exists regarding predispos ing factors for 2 year old in training sales prices. The objectives of the study in Chapter 5 were (i) to examine the relationships between yearling sale purchase price, exercise history, lameness, and speed as predisposing factors for sales price in 2 yea r old in training TB horses and (ii) to compare exercise distance during the last 60 days before the sales between horses with high and low purchase price and between horses affected and not affected by lameness during training. Epidemiologic Studies o f P resale Abnormal Radiographic Findings and History of Arthroscopic Surgery i n TB Yearlings In the TB industry, there is a perception that results of presale radiographic examinations affect the sales price of yearlings. Objective data on the prevalence of p resale radiographic findings and presale arthroscopy, and the association of these factors with the sales price of yearlings may help guide sales decisions and the subsequent management of yearlings in training. Current knowledge of the prevalence of vario us radiographic findings in TB yearlings offered for sale is limited. In a study 34 of 1,162 horses offered for sale in Kentucky between 1993 and 1996, vascular channels in the proximal sesamoid bones of the forelimbs (98% of horses) and hind limbs (93% of horses) were the most common radiographic findings. However, study horses were not randomly selected. Radiographic abnormalities were validated by using more than 1 observer but misclassification bias remained possible. In another study 1 that evaluated 348 TB yearlings offered for sale in Texas in 2002 and 2003, the most common radiographic finding (8% of horses) was an abnormal sagittal ridge of MC3 or MT3. The relationship between the presence of specific radiographic findings and yearling sales price has been investigated. In the previous study of TB yearlings in Texas, 1 horses with

PAGE 31

31 lesions of the proximodorsal aspect of the sagittal ridge of MC3 or MT3 on presale radiographs had lower median sales prices, compared with horses that did not have such lesio ns. To our knowledge, relationships between the presence of various radiographic findings, presale arthroscopy, and sales price at the Keeneland September yearling sale have not been objectively investigated. Further estimates on true prevalence rates of v arious radiographic findings and presale arthroscopy in horses offered for sale at the 2006 Keeneland September yearling sale, and comparisons of sales price between yearlings with and without various presale radiographic findings or a history of arthrosco py are needed to aid owners, trainers, and veterinarians in pre purchase examinations. Epidemiologic Studies o f Presale Abnormal Radiographic Findings i n TB Yearlings a nd Future Racing Performance Yearling TB horses are purchased at public auctions from Ju ly through September for the purpose of racing them the following year as 2 year old racehorses. The price of yearlings can vary based on their pedigree, conformation, and disclosed health information such as history of arthroscopic surgery and presale rad iographic findings. Because the presence of presale abnormal radiographic findings may affect the ability of a yearling to race as a 2 year old or limit racing performance, veterinarians are hired by horse buyers to evaluate presale radiographs for abnorma lities. Two previous studies that investigated the association between presale radiographic findings in TB yearlings and future racing performance in the US came to opposite conclusions. In a study 2 of 1,162 TB yearlings offered for sale at two different s ales in Kentucky between 1993 and 1996, yearlings with moderate or extreme palmar supracondylar lysis of MC3, dorsal medial middle carpal joint disease, enthesophyte

PAGE 32

3 2 formation on the fore proximal sesamoid bones, or hindlimb P1 OC fragments were less likel y to start a race as 2 or 3 year olds, compared to yearlings without these radiographic findings. 2 However, in another study 1 of 348 TB yearlings offered for sale in Texas between 2002 and 2003, no radiographic findings were associated with future racing performance. In each study, the sample of horses used was not randomly selected from each corresponding sale. These studies were based on radiographic examinations of yearlings obtained from private veterinary practices, and the study populations were diff erent. For example, the median sales price of TB yearlings was $40,000 in the study in Kentucky 2 and $9,000 in the Texas study. 1 In addition; the number of horses in cohorts with hind P1 OC fragments in these 2 studies was 25 and 10, respectively (a limiti ng factor that could have produced false negative results, particularly in the study in Texas). It is important to study lameness in TB yearling horses to more completely understand the exposure facto rs associated with lameness, health and safety concerns for both horse and rider, and to formulate effective lameness prevention strategi es. The literature consists of one study pe rformed on TB Pinhooked horses. 3 It provided insight into the association between financial returns and lameness but left many important questions unanswered. Numerous studies exist on the frequency of lameness in race horses and yet we have made little prog ress in reducing lameness. 2 6 This knowledge would contribute to the formulation of prevention and control strategies to protect the health and welfare of 2 year old in training horses. The research contained in Chapters 3, 4, 5, and 6 assess these associati ons in TB yearlings, 2 year old in training horses, and later as 2 year old racehorses.

PAGE 33

33 CHAPTER 3 LAMENESS, ATHLETIC P ERFORMANCE, AND FINA NCIAL RETURNS IN THOROUGHBRED PINHOOK ED YEARLING HORSES In racing circles, pinhooking is a term used to describe the practice of buying a horse specifically for resale and profit. Pinhooked yearling Thoroughbreds are purchased at public auctions from July through September for the purpose of selling them the following year (February through April) at what are known as 2 year old in training sales. Trainers have approximately 5 months to break, condition, and prepare their horses to work at high speed during training in preparation for the sales. At the sales, the trained 2 year old horses race 1 or 2 furlongs (1/8 or 1/4 of a mile). Among the 2 year old in training sales that occur in a given year, the earlier sales (eg, sales in February) are considered the most important because the mean sale price is higher, compared with that achieved at later sales. After purchase, th e horses start racing as 2 year old s. Knowledge regarding the relationship between lameness and athletic performance in yearling horses in the US is limited. A previous study 3 conducted at a training center in Florida revealed that lameness was the most im portant cause of training failure in yearling horses, as measured by the number of days on which training could not be performed. In horses affected with lameness, training failure during high speed exercise in preparation for 2 year old in training sales was considered severe (eg, 13 to 108 days of training lost). Financial returns for horses with severe training failure were significantly lower than returns for horses with mild training failure (eg, 1 to 11 days of training lost). In that study, 3 however, the anatomic location, diagnosis, and treatment of lameness were not characterized. The purpose of the study reported here was to characterize lameness during training and compare exercise variables and financial

PAGE 34

34 returns among yearling Thoroughbreds that were bought for the purpose of resale for profit (ie, pinhooked horses). On the basis of the racing completed at 2 year old in training sales, the performance speed of the horses in the study population was classified as fast, average, or slow, and variabl es were compared among these 3 groups. Materials and Methods Horses : Forty Thoroughbreds (age range, 14 to 21 months) purchased from July through September 2004 for resale at 2 year old in training sales (February through April 2005) were included in the s tudy. All 40 horses were from 1 training center in Ocala, Florida. This center was selected because of the number of horses trained and willingness of the trainer to participate. On arrival, horses were vaccinated by the attending veterinarian against equi ne herpesvirus types 1 and 4, influenza virus types A 1 and A 2 West Nile virus, Eastern and Western equine encephalitis viruses and Clostridium tetani In addition, horses were vaccinated for Streptococcus equi on 3 separate occasions (October 2004, Februa ry 2005, and April 2005). Training procedures : All horses were broken to ride and underwent the same training program starting October 2004. Briefly, all horses were trotted to the racetrack, fast galloped (22 seconds/furlong) for a distance of 1.5 miles, and walked back to their stalls 6 days each week; 1 day each week was a rest day. In December, horses were introduced to high speed exercise (breezing; ie, running at racing speed during training; 1 day each week (Saturday), followed by 1 day of rest (Sunday), and galloping (at speeds of approx 18 to 22 seconds/furlong) on 5 days each week.

PAGE 35

35 Diagnosis of lameness : During training, horses were examined daily for signs of lameness by the rider or train er on the racetrack during or immediately after an exercise event (galloping or breezing). Diagnosis of lameness was confirmed by the attending veterinarian by use of a 0 to 5 grade lameness scoring system adopted by the American Association of Equine Prac titioners. 3 3 In this scoring system, grade 0 = no lameness; 1 = lameness is difficult to observe and not consistently apparent, regardless of circumstances (eg, weight carrying, circling, or traveling on inclines and hard surfaces); 2 = lameness is difficu lt to observe at a walk or when trotting in a straight line but consistently apparent under certain circumstances (eg, weight carrying, circling, or traveling on inclines and hard surfaces); 3 = lameness is consistently observable at a trot under all circu mst ances; 4 = lameness is obvious at a walk; 5 = lameness results in minimal weight bearing in motion or at rest or complete inability to move. For purposes Anatomic location of lameness : Am ong horses confirmed as lame by the attending veterinarian, joint inflammation injury was localized by the presence of 1 clinical sign: joint heat, joint effusion, or signs of pain elicited during flexion of the joint. Perineural or intraarticular analgesi a was used to localize the source of lameness in horses without evidence of joint heat, effusion, or signs of pain elicited during flexion of a joint. Radiographic or ultras onographic examinations were used in 10 horses (with a lameness) to confirm the source of lameness. Treatment of lameness : Injuries to the distal interphalangeal, metacarpophala ngeal, middle carpal or antebrachiocarpal, distal intertarsal,

PAGE 36

36 tarsometatarsal, femorotibial, and sacroiliac joints were treated in the affected limbs with intra articular injection of sodium hyaluronate a or methyl prednisolone acetate b and amikacin. c Estr adiol cypionate d was administered IM to horses in which patellar fixation was diagnosed. Suspensory ligament injuries were diagnosed by heat in the proximal palmar region of the distal limb, signs of pain elicited on palpation of the suspensory origin, lam ultrasonographic examination. Injuries to the proximal suspensory ligaments of the fore and hind limbs were treated with extracorporeal shock wave therapy every 7 days fo r 3 weeks in conjunction with intraligamentous injection of triamcinolone, e amikacin, c and a distillation of pitcher plants ( Sarraceniaceae family). f Other soft tissue injuries such as synovitis of the distal digital tendon sheath were treated with injecti on to the affected area of triamcinolone, amikacin, and pitcher plant distillate. Athletic performance at the 2 year old in training sales : The official time of individual breezes in each horse was monitored by the electric timer operated by the racetrack clocker. At each sale, study horses were assigned into 1 of 3 groups (fast, average, slow) on the basis of their recorded official speed, compared to the calculated average speed of the same sale Fast horses were those for which the race speed was faster than the sale average speed. Average horses were those for which the race speed was equal to or slower than the sale average speed by 0.4 seconds. Slow horses were those for which the race speed was slower than the sale average speed by > 0.4 seconds.

PAGE 37

37 Fin ancial returns : For each horse, this variable was calculated as the sale price as a 2 year old minus the purchase price as a yearling minus the cost of training ($10,000) minus the cost of treatment of lameness (if any). One horse was not entered into a 2 year old in training sale because of a carpal chip fracture that occurred during training; the sale value of this horse was recorded as $0. Data collection : For each horse the following additional data were recorded: yearling identification number, foaling date, age (months), sex (colt or filly), purchase price, arrival date, breaking date, interval (days) from arrival to being broken, interval (days) from being broken to first day on the racetrack for training, interval (days) first day on the racetrack fo r training to first breeze, furlongs galloped, furlongs breezed, lameness, days lost during training, athletic performance at the sales (fast, average, or slow performance speed), an d sale price. A day lost in training was defined as a day when a yearling did not undertake training because of an injury, disease, rain, or absence of the rider. Statistical analysis : Survival analysis was used to estimate median time to the first event o f lameness during training. Simple linear regression was used to assess the relationship between number of new cases of lameness and days prior to the sale. Horses were assigned into 1 of 3 groups (fast, average, and slow) on the basis of the speed recorde d at the 2 year old in training sales in relation to the sale average performance speed. New cases of lameness 30 or 60 days before the sale were compared among groups by use of a X 2 test. Financial returns were compared among groups by use of a Kruskal Wa llis test. Other continuous variables (age, purchase price,

PAGE 38

38 interval from arrival to breaking, interval from being broken to first day at the racetrack for training, interval from first day at the racetrack to first breeze, furlongs galloped, days of tra ining lost, and days on which the horse was lame) were compared among groups by use of a Kruskal Wallis test. Values of P < 0.05 were considered significant. Results Thirty seven of 40 horses were classified as lame during training. Joint injury was the mo st common cause of lameness (27/37 horses). In contrast, only 4 horses had lameness associated with DMD (Table 3 1) Eighteen of 37 horses had an injury to hind limbs only; 5 other lame horses had an injury to both fore and hind limbs. Median time to firs t event of lameness was 137 days (Figure 3 1) The frequency of new cases of lameness increased as time (days) approached the date of 2 year old in training sales (Figure 3 2) Four of 11 horses classified as lame within 30 days prior to the sales were sol d in February; those horses had lameness associated with the sacroiliac joint (n = 2), proximal suspensory ligament of the left forelimb (1), or DMD (1). The remaining 7 lame horses were sold in March. Three of the 7 horses had lameness in the hind limbs. Affected regions included distal tarsal, distal interphalangeal and sacroiliac joint in 1 horse; femorotibial joint in 1 horse; and 1 metatarsophalangeal and distal tarsal joints in 1 horse. Four of the 7 horses had lameness in the forelimbs. Affected regi ons included the metacarpophalangeal and antebrachiocarpal joints in 1 horse; DMD in 1 horse; antebrachiocarpal and intertubercular bursa in 1 horse; and antebrachiocarpal joint in 1 horse. At the sales, 4 horses were classified as fast, 21 as average, and 15 as slow (Table 3 2) Median financial return was slightly but significantly ( P < 0.05) different between horses classified as fast ($14,000) or average ($0) and slow ( $8,000). The median number of days of training lost in horses with an injury in the forelimb (14

PAGE 39

39 horses; 13 days lost), in the hind limb (18 horses; 16 days lost), or in both limbs (5 horses; 15 days lost) was not significantly different ( P = 0.83). Discussion The scope of the present study was descriptive. It was limited to 40 horses in 1 training center, the training procedures of which may not be representative of the population of horses trained in Ocala, Florida. The sample size in our study was low, and it may have contributed to the failure to detect significant differences in exerc ise variables (eg, number of furlongs galloped 30 and 60 days before the sales) among horses classified as fast, average, or slow. Another limitation was that diagnosis of lameness was conducted by 1 veterinarian, whose accuracy for diagnosis of lameness b y use of the 5 grade lameness scoring system was not assessed. Overall, the study results revealed that joint injury was the most common cause of lameness among the horses entered into 2 year old in training sales, the incidence of lameness in hind limbs w as high, the frequency of new cases of lameness increased as time (days) approached the date of the 2 year old in training sales, and median financial return was slightly but significantly different between horses classified as fast ($14,000) or average ($ 0) and slow ( $8,000) at the sales. In the present study, incidence of lameness caused by DMD was low and incidence of lameness caused by joint injury was high. The observed low incidence of lameness caused by DMD may be explained by the training program, which included fast galloping (at a rate of 18 to 22 seconds/furlong). In a previous study, 8 incorporation of breezing (at a rate of 13 to 14 seconds/furlong) during early training of Thoroughbred racehorses was associated with a decreased incidence of DMD Slow galloping (at a rate of 30 seconds/furlong) may lead to tension on the dorsal surface of the third

PAGE 40

40 metacarpal bone, whereas breezing will result in compression on that surface. 7,8 Horses that train at a slow speed gallop will acquire bone that is ad apted to that training modality and is susceptible to DMD. 7,8 If the training pattern resembles racing, then the acquired bone will become adapted to racing without the horse developing DMD. 7,8 Although daily breezing was not part of the training protocol of horses in the present study, the fast galloping (at a rate of 18 to 22 seconds/furlong) used at the study training center is considered faster than that used at other training centers in Ocala (typically, a rate of 30 seconds/furlong). In contrast, the high incidence of lameness caused by joint injury in our study can be explained by repetitive loading during exercise. Traumatic synovitis and capsulitis are common joint injuries of athletic horses, presumably resulting from repeated trauma during exercis e. 27 3 0 Acute synovitis affects synovial membrane diffusion and synoviocyte metabolism. 27 3 0 Mechanically damaged synoviocytes release degradative enzymes, prostaglandin E 2 and cytokines, which cause inflammation and disruption of joint homeostasis. 27 3 0 I nflammation results in distention of the synovial membrane and effusion into the synovial cavity; because of increased intra articular pressure in the affected joint, indirect stimulation of the capsular tissue receptors and subchondral bone receptors caus es signs of pain. 2 7 3 0 Among the horses of the present study, the incidence of lameness in the hind limbs during high speed exercise training was considered high, 35 compared with the incidence in adult racehorses, 17 which most often injure or develop lamen ess in forelimbs. 17 We offer 3 possible explanations for the observed high incidence of hind limb lameness in our study horses. First, the gait known as the racing gallop is

PAGE 41

41 asymmetric the muscles of hind limbs initiate propulsion and the diagonal forelimb s complete deceleration to support the center of body mass. 3 6 ,37 A brief period of suspension occurs as the body is propelled forward during each stride. 3 6 ,37 Ground reaction forces in the forelimbs account for 60% of the total vertical impulse to support the center of body mass. 36,37 However, the impulse (force X time) of horizontal propulsion for the body mass is greatest in the hind limbs. 3 6 ,37 In the horses of our study, an asymmetric racing gallop combined with immature muscle tone may explain a higher incidence of lameness in hind limbs, compared with the incidence in adult racehorses. Second, in yearling horses, it is known that the frequencies of radiographic changes and arthroscopic surgery are higher in hind limbs than in forelimbs. 35 However, the associations between radiographic changes, history of arthroscopic surgery, and lameness in pinhooked horses have not been established. Finally, the high frequency of injury in hind limbs may also relate to the racetrack surfaces used in private training c enters in Ocala. It is possible that tracks with a deep sand cushion and with limited daily pre and post training maintenance are 2 potential factors that can lead to poor track surface quality. Poor track surface quality could demand greater effort by th e hind limbs for horizontal propulsion; however, these associations have not been investigated by use of objective research methods to our knowledge. Median time to the first event of lameness during training was 137 days, which coincides with high speed e xercise training 30 days prior to the first sales. Visual assessment of a plot of the incidence of lameness in pinhooked horses revealed a linear increase of new cases during training, with the highest incidence occurring within 30 days prior to the sales. Cyclic loading to an immature joint during exercise 27 29 and

PAGE 42

42 cumulative racing speed exercise (breezing and racing) during the last 30 to 60 days prior to injury have been identified as risk factors for musculoskeletal injury in racehorses. 18 22 Median fi nancial return was slightly (but significantly) different between horses with performance speeds classified as fast ($14,000) or average ($0) and slow ( $8,000). This association was expected, because speed is an important determinant in the selling price of 2 year old horses in training. Median purchase price was similar among 2 year old horses classified as fast ($41,000), average ($30,000), and slow ($40,000), which is an indication that the association between speed and financial returns was not confoun ded by purchase price. In a previous study, 3 financial returns in 2 year old s were associated with training failure, whereas financial returns were significantly lower in 2 year old s that had a median of 38 days of training lost ($2,000), compared with 2 y ear old s that had a median of 10 days of training lost ($10,000). In that study, 3 lameness was the main cause of training failure. In the present study, total number of training days lost during the training period or total days of training lost 30 days pr ior to the sale among 2 year old s classified as fast, average, or slow was not different. Differences in management of injured or diseased (eg, ringworm infested) horses between the 2 study populations may explain this discrepancy. For example, type and fr equency of medication used in pinhooked yearling horses during training was not reported in the previous study. 3 In our study, horses that had mild lameness (grades 1 and 2) were medicated in an attempt to avoid loss of training days. Similarly, in our stu dy, horses affected with ringworm were allowed to train, which resulted in no loss of

PAGE 43

43 training because of the parasitic infestation. In contrast, horses affected with ringworm in the previous study 3 were not allowed to train to reduce risk of disease trans mission.

PAGE 44

44 Table 3 1. Anatomic location of lameness that developed during training in 37 of 40 yearling Thoroughbreds that were bought for the purpose of resale for profit and were classified as fast, average, or slow at the 2 year old in training sales. A natomic location Fast (n = 4) Average (n = 19) Slow (n = 14) Forelimb Distal interphalangeal joints Dorsal third metacarpal bone Second metacarpal bone Antebrachiocarpal joints Right antebrachiocarpal joint Distal interphalangeal joint s and left proximal suspensory ligament Right digital tendon sheath Left metacarpophalangeal and right antebrachiocarpal joints Right intertubercular bursa and antebrachiocarpal joint Left proximal suspensory ligament Metacarpophalangeal j oints Hind limb R digital tendon sheath Medial femorotibial joints Plantar tarsai desmitis Medial femorotibial joints Medial and lateral femorotibial joints Medial and lateral femorotibial joints, tarsi Medial femorotibial joints, sacr oiliac joint Tarsi Medial femorotibial joints Medial femorotibial joints, tarsi Tarsi, sacroiliac joints Sacroiliac joints Fore and hind limb Tarsi and metacarpophalangeal joints Tarsi and right dorsal third metacarpal bone Tarsi and middle carpal joints Tarsi and right forelimb proximal suspensory ligament Tarsi, distal interphalangeal joints, and sacroiliac joints 1 1 1 1 1 3 2 2 5 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1

PAGE 45

45 Table 3 2 Comparison of sex, age, purchase price, financial returns, and exercise variables in 40 yearling Thoroughbreds that were bought for the purpose of resale for profit and were classified as fast, average, or slow at the 2 year old in training sales. Variable Fast (n = 4) Median Average (n = 21) Range Slow (n = 15) Median Range Median Range P value Age at start of study (mo ) 16 18 18 18 14 20 18 16 21 0.32 Purchase price (X $10 3 ) 41 30 85 30 4 120 40 3 80 0.60 Financia l return (X $10 3 ) 14a 11 108 0a,b 30 29 8c 52 39 0.03 Sex* (colt) 3 (75) 18 (86) 12 (80) 0.84 Before training on track Interval from arrival to breaking (d) 4 2 51 20 2 41 20 2 51 0.63 Interval from being broken to first day on racetrack 13 13 33 30 2 46 28 12 97 0.45 Interval from first day on racetrack to 1st episode of breezing 57 44 64 57 42 89 57 33 64 0.97 Galloping before 1st breeze Furlongs galloped 225 130 320 300 120 380 300 200 380 0.22 No. of days of training lost 10 0 35 8 1 30 13 2 53 0.36 No. of days lame 4 0 29 2 0 29 5 0 51 0.58 Lameness* (yes) 1 (25) 1 (5) 1 (13) 0.88 30 days before the sale Furlongs galloped 115 90 160 130 50 200 140 100 170 0.49 Furlongs breezed 6 1 7 3 0 11 6 2 12 0.21 No. of days of training lost 2 0 14 2 0 13 1 0 4 0.45 No. of days lame 2 0 14 2 0 13 0 0 9 0.27 Lameness* (yes) 4 (100) 18 (86) 11 (73) 0.39 60 days before the sale Furlongs galloped 265 260 330 280 90 350 300 120 360 0.88 Furlongs breezed 14 8 18 15 1 22 18 2 22 0.09 No. of days of training lost 5 0 15 3 1 24 3 0 26 0.40 No. of days lame 4 0 14 2 0 24 1 0 26 0.32 Lameness* (yes) 4 (100) 18 (86) 14 (93) 0.73 Lameness events 2 1 3 2 0 6 2 0 2 0.30 Cost of lameness ($) 605 373 1,183 667 0 3,01 0 655 0 2,619 0.92 2 year old sale price (X $10 3 ) 65a 30 205 37b 7 140 20c 0 90 0.01 a,b,c Within row, median values with different superscripts are significantly ( P < 0.05) different. *Data are reported as No. of horses (%).

PAGE 46

46 Figure 3 1 Time to fi rst event of lameness among 40 yearling Thoroughbreds that were bought for the purpose of resale for profit after training Days in training

PAGE 47

47 Figure 3 2 Distribution of new cases of lameness within 1 to 150 days prior to sale* in 37 of 40 y earling Thoroughbreds that were bought for the purpose of resale for profit after training. The graph includes linear regression line (R = 0.37; R2 = 0.14; P = 0.16).

PAGE 48

48 CHAPTER 4 EPIDEMIOLOGIC ANALYSIS OF YEARLING SALE PURCHASE PRICE, EXERCISE HISTORY, LAMENESS, AND SPEED AS PREDISPOSING FACTORS FOR SALES PRICE IN 2 YEAR OLD IN TRAINING THOROUGHBRED HORSES Knowledge of yearling sale purchase price and exposure factors such as lameness during training and speed recorded at the 2 year old in training sale s associated with sales price in 2 year old in training Thoroughbred horses in the US is very limited. In a study 3 conducted in 1 training center in Florida, lameness was the most important cause of days lost during training in yearling Thoroughbreds in pr eparation for the 2 year old in training sales. Another study 3 5 in Florida revealed that incidence of lameness was highest during the last 60 days before the sales, and joint injury was the most common cause of lameness among yearling Thoroughbreds entered into 2 year old in training sales. In general, Thoroughbreds with a high pedigree, good conformation and a high purchase price are expected to be sold for a high sales price at the 2 year old in training sales, compared to horses without these characteris tics. Speed is an additional variable that can influence the sales price in Thoroughbreds sold at the 2 year old in training sales. 3 5 The relationship between training patterns (such as exercise distance accumulated during training) and sales price in Tho roughbreds sold at the 2 year old in training sales have not been well investigated. In one study, 3 5 the number of furlongs galloped 60 days before the sales was lower in horses classified as fast and sold for more money ($65,000), compared to horses class ified as slow and sold for less money ($20,000) at the sales. The underlying reasons for this observed relationship were not identified. It is known that lameness can affect the training schedule of yearling Thoroughbreds by causing days lost during traini ng 3 or reduced exercise. However, it is

PAGE 49

49 not known if purchase price has an effect on training schedules prescribed to yearling Thoroughbreds affected with lameness. The objectives of the study reported here were (i) to examine the relationships between yea rling sale purchase price, exercise history, lameness, and speed as predisposing factors for sales price in 2 year old in training Thoroughbred horses and (ii) to compare exercise distance during the last 60 days before the sales between horses with high a nd low purchase price and between horses affected and not affected by lameness during training. Materials and Methods Study horse s: Fifty one yearling Thoroughbreds purchased at the 2006 Keeneland September yearling sale and trained in Florida for resale a t 2 year old in training sales (February through April 2007) were included in this study. The median purchase price of study horses sold at the yearling sales was $55,000 (range = $7,000 to $725,000). All horses were shipped from Lexington, Kentucky to 1 t raining center in Ocala, Florida in late September 2006. Days after arrival, horses were vaccinated against equine herpes virus types 1 and 4, influenza virus types A1 A2, West Nile virus, Eastern and Western equine encephalomyelitis, tetanus and strangles (October 2006, February 2007, April 2007) by the attending veterinarian. The training center was selected because of the number of horses available and willingness of the trainer to participate. Training procedures : All horses were trained to ride and su bjected to similar exercise training programs starting in October 2006. Briefly, all horses trotted to the racetrack, galloped 1 to 1 miles (~ 1.6 to 2.4 km) and walked back to their stalls 6 days per week; 1 day per week was a rest day. Starting Decembe r 2006, yearlings were introduced to galloping 5 days per week (~25 to 30 seconds per furlong) and high speed

PAGE 50

50 exercise (breezing) 1 day per week (Saturday; 1 or 2 furlongs in ~ 15 seconds per furlong), followed by 1 day of rest (Sunday). One furlong = 1/8 mile or ~ 200 meters. Outcomes of interest : For the first objective, the outcome of interest was sales price of study horses sold at the 2 year old in training sales. For the second objective, the outcomes of interest were number of furlongs galloped and n umber of furlongs breezed 1 to 60 days before the sales in study horses. Demographic data : For each horse, the following data were recorded: yearling identification number, gender (colt or filly), history of arthroscopic surgery (yes, no), yearling sale pu rchase price, date of arrival, first day (date) a horse was first trained to ride, and first day (date) a horse was first on the racetrack for training. Exercise data : During training, horses were monitored daily by the trainer to measure exercise distance (ie, number of furlongs galloped and breezed during training). In this study, the time period of interest was 1 to 60 days before the 2 year old in training sales because it was shown that the incidence of lameness in yearling Thoroughbreds is highest dur ing that time period. 35 Lameness : Horses were identified as lame by the rider or trainer on the racetrack during or immediately after an exercise event (galloping or breezing). Horses identified as lame were further examined by the attending veterinarian by using a 0 to 5 grade lameness scoring system adopted by the American Association of Equine Practitioners. 33 lame. Anatomic location of lameness and individual results of radiographic or ultrasonographic examinations in lame horses were not available for this study. The

PAGE 51

51 sensitivity and spec affected with lameness were not determined. Speed : At the 2 year old in training sales, the official time of individual high speed sessions (breezes) for horses was measured by an electronic timer operated by the racetrack clocker. The trainer assigned horses into 1 of 2 groups on the basis of their speed recorded at the sales; the two groups were designated as fast or non fast based on a previous study 3 5 with some modifications. Br iefly, fast horses were those for which the time trial was faster than the mean time for all horses at the sale. Non fast horses were those for which the race speed was equal to or slower than the sale mean speed. Statistical analyses : For the first object ive, the null hypotheses that median sales price is not different (i) in horses with a low or high yearling sale purchase price, (ii) in horses with a low or high number of furlongs galloped during the last 60 days before the sale, (iii) in horses with a l ow or high number of furlongs breezed during the last 60 days before the sale, (iv) in horses with or without lameness, and (v) in horses classified as fast or non fast at the sale were tested by using the Wilcoxon rank sum test and multivariable ANOVA. Th e continuous variables for yearling sale purchase price, total number of furlongs galloped and total number of furlongs breezed 1 to 60 days before the sale were categorized into 2 groups (high, low) on the basis of its frequency of distribution (median). In the univariable analysis, the association between the outcome of interest (sales price) and purchase price, exercise distance, lameness, speed and additional variables such as: gender (colt, filly) and history of arthroscopic surgery (yes, no) were exam ined by use of the Wilcoxon rank sum test. In the multivariable analysis, ANOVA was used to examine the relationship between sales price (rank transformed

PAGE 52

52 data) and investigated exposure variables (factors). Initially, exposure variables that were associat ed (ie, P an ANOVA model, and a forward stepwise approach was used to identify variables associated with the main outcome of interest (sales price), with a 2 sided P value of 0.05 used to i nclude variables in the model and a 2 sided P value of 0.10 used to remove variables. The variables for yearling sale purchase price and speed at the 2 year old s in training sale were forced in final models because they can affect the sales price of horses at the 2 year old in training sales. In addition, in the final models for sales price, because the exposure variable for distance of furlongs galloped (or breezed) during training can be affected by lameness 3 5 two different models using the variable for number of furlongs galloped (Model 1) or the variable for lameness (Model 2) were examined. For the second objective, the null hypotheses that number of furlongs galloped and number of furlongs breezed in the 60 days prior to the sale were not different be tween horses with low and high purchase price and between horses affected and not affected with lameness during training were tested by using the Wilcoxon rank sum test. Finally, in order to determine if purchase price and lameness had a combined effect on exercise distance accumulated during the last 60 days before the sales in study horses, exercise distance during that time period was compared between (i) horses with a high purchase price and affected with lameness (n = 13); (ii) horses with a high purch ase price but not affected with lameness (n = 13); (iii) horses with a low purchase price and affected with lameness (n = 7); and (iv) horses with a low purchase price but not

PAGE 53

53 affected with lameness (n = 18) by using the Kruskal Wallis test. For all statis tical analyses, values of P Results Overall, the median sale price of all 51 study horses at the 2 year old in training sales was $37,000 (minimum = $3,000; 25 th percentile = $20,000; 75 th percentile = $150,000; maximum = $1,000,000). Twenty of 51 (39%) horses were diagnosed with lameness during the 60 days before the sale; the median number of days lame = 1 (25 th percentile = 1 day; 75 th percentile = 10 days). In the univariable analysis, median sales price of study hors es was associated with yearling sale purchase price, number of furlongs galloped 1 to 60 days before the sales, and speed recorded at the 2 year old in training sales ( P < 0.05) ( Table 4 1 ). In the multivariable analysis ( Table 4 2 ), after controlling for the variables for yearling sale purchase price and speed at the 2 year old s in training sale, median sales price was higher in horses with a low number of furlongs galloped 60 days before the sale, compared to horses with a high number of furlongs galloped ( P < 0.05; Model 1). Variables included in the model explained 63% of the variation in sales price ( R 2 = 0.63). Lameness was not associated with sales price in study horses ( P = 0.90) (Model 2). Horses with a high yearling sale purchase price or horses af fected with lameness were significantly associated with a reduced number of furlongs galloped during the last 60 days before the sales ( Tables 4 3 and 4 4 ). Finally, we examined the combined effect of purchase price and lameness on exercise distance accumu lated during the last 60 days before the sales. Study results revealed that the number of furlongs galloped during this time period was significantly lower in horses with a high purchase price and

PAGE 54

54 affected with lameness (250 furlongs), compared to horses w ith a low purchase price and not affected with lameness (412 furlongs) ( Table 4 5 ). Discussion This study was limited to 51 yearling Thoroughbreds from 1 training center in Florida. Thus, study results cannot be extrapolated to other training centers in Fl orida or in the US. Furthermore, the median purchase price of study horses was $55,000 compared to $45,000 in a total of 3,355 yearlings sold at the 2006 Keeneland September yearling sale; an indication that the quality and purchase price of study horses c ould have been different from that in the population of yearling Thoroughbreds sold at the sale. All horses that the rider or trainer thought were lame were confirmed by to correctly identify horses as lame was not assessed. Thus, it is difficult to determine the magnitude of exposure (lameness) misclassification potentially present in this study. It is possible that horses with mild signs of lameness may not have detected as lame (false negative results). If this scenario were an important bias, the associations observed between lameness and sales price or number of furlongs galloped would apply only to horses with signs of lameness commonly recognized by the rider or trai ner and confirmed by the attending veterinarian at the training center used in this study. A scenario where horses with severe lameness were not detected by the rider or the trainer is less likely. For the first objective, because the variables for yearlin g sale purchase price or speed at the 2 year old in training sales can influence sales price of Thoroughbreds at the sales, the discussion of study results is focused on exercise history and lameness as predisposing factors for sales price.

PAGE 55

55 Overall, study results revealed that (after controlling for yearling sale purchase price and speed recorded at the 2 year old in training sales) sales price of study horses was associated with a low number of furlongs galloped during the last 60 days before the sale. In addition, number of furlongs galloped was associated with purchase price of yearling Thoroughbreds and lameness during training. This study is relevant because it decisi on on the amount of training allocation prescribed to individual horses, particularly among horses affected with lameness. In the analysis for sales price, median sales price was higher in horses with a low number of furlongs galloped 60 days before the s ale, compared to horses with a high number of furlongs galloped. This association can be better explained by the observed combined effect of yearling sale purchase price and lameness on the number of furlongs galloped during the last 60 days before the sal es in study horses. Among 13 lame horses with a high purchase price (median = $225,000), it is possible that the trainer made a decision to reduce exercise for galloping to prevent a musculoskeletal injury that would prevent affected horses from performing at the 2 year old in training sales, thus protecting a high initial investment; if this explanation is valid, then a reduced exercise for galloping did not apply to 7 lame horses with a low purchase price ($20,000) because the initial investment was relat ively low. A previous study 8 identified an increased risk of dorsal metacarpal disease associated with an increase in exercise allocation to galloping in 2 year old Thoroughbred racehorses. Other studies 18 2 0,24 in racehorses have identified increased risk of catastrophic musculoskeletal injury following high speed distance accumulation within 60 days.

PAGE 56

56 Although we identified an association between sales price and exercise distance accumulated during the last 60 days before the sales, we failed to identify lameness alone as a risk factor associated with reduced sales price in yearling Thoroughbreds. One possible explanation is that horses affected with lameness were identified early and responded well to treatment, as the duration of lameness in most affecte d horses was 1 to 10 days. In this study, the anatomic location and cause of lameness was not analyzed; however, the duration of lameness was similar to a group of yearling Thoroughbreds investigated in a previous study in Florida, 3 5 where joint injury was the main cause of lameness, and treatment consisted of intra articular injections of sodium hyaluronate or methyl prednisolone acetate and amikacin. Finally, it is also known that horses affected with early acute dorsal metacarpal disease can continue to train after 5 to 10 days of rest and anti inflammatory analgesics. 27 30 In summary, after controlling for yearling sale purchase price and speed recorded at the 2 year old s in training sale, sales price in study horses was associated with less exercise dis tance accumulated during the last 60 days before the sales. Lameness alone was not associated with sales price. However, lameness was associated with a subsequent, reduced exercise allocation for galloping which was more evident among lame horses with a hi gh purchase price, compared to lame horses with a low purchase price; this finding suggests that the level of risk tolerance for a musculoskeletal injury associated with lameness was different among study horses with a low or high purchase price.

PAGE 57

57 Table 4 1 Univariable analysis comparing median sales price in study horses sold at 2 year old in training sales according to gender and other investigated exposure factors. Variable No. of horse s Median sales price ($) Mean SE sales price rank* P Gender Co lt Filly 25 26 40,000 35,000 26 3 25 3 0.92 History of arthroscopic surgery Yes No 4 47 155,000 32,000 33 8 25 2 0.36 Purchase price at the yearlings sale High ($45,000 to $725,000) 26 25 145,000 22,000 36 2 16 2 < 0 .01 Lameness Yes No 20 31 56,000 30,000 29 4 24 3 0.19 Days lame during training High (1 to 26) Low (0) 20 31 56,000 30,000 29 4 24 3 0.19 Total furlongs galloped High (366 to 480) Low (50 to 365) 26 25 25,000 85,000 19 3 33 3 < 0. 01 Total furlongs breezed High (7 to 22) Low (0 to 6)** 26 25 30,000 62,000 23 3 30 4 0.11 Speed at 2 year old in training sales Fast Non fast 9 42 350,000 30,000 45 2 22 2 < 0.01 *In each variable, a higher mean sales price rank correspo nds to a higher sales price. ** Three horses recorded 0 furlongs breezed 1 30 or 1 60 days prior to sale.

PAGE 58

58 Table 4 2. Multivariable analysis (ANOVA). Analyses of median sales price in horses sold at 2 year old in training sales. Variable No. of horses Mea n SE sales price rank P Model 1 Total number of furlongs galloped High (191 to 250) Low (16 to 190) 26 25 26.9 2.2 34.0 2.3 0.01 Purchase price at the yearlings sale High ($45,000 to $725,000) 26 25 36.8 1.9 24.2 2.7 < 0.01 Speed at 2 year old in training sales Fast Non fast 9 42 37.6 3.4 23.4 1.4 < 0.01 Model 2 Lameness Yes No 20 31 31.3 2.6 29.7 2.1 0.90 Purchase price at the yearlings sale High ($4 5,000 to $725,000) 26 25 38.1 2.0 22.9 2.9 < 0.01 Speed at 2 year old in training sales Fast Non fast 9 42 37.6 3.6 23.5 1.5 < 0.01 R 2 (Model 1) = 0.63; R 2 (Model 2) = 0.58

PAGE 59

59 Table 4 3 Median number of furlongs galloped and b reezed 1 to 60 days before the sale among horses with high or low yearling sale purchase price. Yearling sale purchase price High $45,000 to $725,000 N = 26 Low N = 25 P Number of furlongs galloped 268 410 0.03 Number of furlongs breezed 6 7 0.65 Table 4 4 Median number of furlongs galloped and breezed 1 to 60 days before the sale among horses affected or not affected with lameness. Lameness Yes N = 20 No N = 31 P Number of furlongs galloped 256 409 0.04 Number of furlongs breeze d 6 7 0.62 Table 4 5 Observed combined effect between yearling sale purchase price and lameness during training on number of furlongs galloped and number of furlongs breezed 1 to 60 days before the sale in study horses. Purchase price Lame No. of ho rses No. of furlongs galloped No. of furlongs breezed High ($225,000)* Yes 13 250 a 6 a High ($120,000) No 13 372 a,b 7 a Low ($20,000) Yes 7 412 b 4 a Low ($25,000) No 18 408 a,b 8 a *Purchase price data in parenthesis are reported as the median. a,b Within each column, groups with different superscripts are significantly ( P < 0.05) different.

PAGE 60

60 CHAPTER 5 ASSOCIATIONS OF PRESALE RADIOGRAPHIC FINDINGS AND PRESALE ARTHROSCOPY WITH SALES PRICE IN THOROUGHBRED YEARLINGS SOLD IN KENTUCKY The Keeneland September yearling sale is typically the largest sale in the yearling Thoroughbred market. In the past 12 years, the number of horses sold has increased from 2,949 in 1996 to 3,605 in 2008, while median value increased from $22,000 to $37,000. 38 For many years, the economic value of Thoroughbred yearlings offered for sale as future racehorses has been influenced by their pedigree and conformation. Recently, prospective buyers have paid increased attention to the health status of yearlings, with emphasis on the resul ts of presale radiographic examinations, a history of presale arthroscopic surgery, and the results of endoscopic evaluation of the upper airway. In the Thoroughbred industry, there is a perception that results of presale radiographic examinations affect t he sales price of yearlings. Objective data on the prevalence of presale radiographic findings and presale arthroscopy and the association of these factors with the sales price of yearlings may help guide sales decisions and the subsequent management of ye arlings in training. Current knowledge of the prevalence of various radiographic findings in Thoroughbred yearlings offered for sale is limited. In a study 34 of 1,162 horses offered for sale in Kentucky between 1993 and 1996, vascular channels in the proxi mal sesamoid bones of the forelimbs (98% of horses) and hind limbs (93% of horses) were the most common radiographic findings. However, study horses were not randomly selected. In another study 1 that evaluated 348 Thoroughbred yearlings offered for sale in Texas in 2002 and 2003, 1 the most common radiographic finding (8% of horses) was an abnormal sagittal ridge of MC3 or MT3.

PAGE 61

61 The relationship between the presence of specific radiographic findings and yearling sales price has been investigated. In the previ ous study of Thoroughbred yearlings in Texas, 1 horses with lesions of the proximodorsal aspect of the sagittal ridge of MC3 or MT3 on presale radiographs had lower median sales prices, compared with horses that did not have such lesions. To our knowledge, relationships between the presence of various radiographic findings, presale arthroscopy, and sales price at the Keeneland September yearling sales have not been investigated. The objective of the study reported here was to estimate prevalences of various radiographic findings and presale arthroscopy in horses offered for sale at the 2006 Keeneland September yearling sale and compare sales price between yearlings with and without various presale radiographic findings or a history of arthroscopy. Materials and Methods Horse s: Yearling Thoroughbreds offered for sale at the 2006 Keeneland September yearling sale (n = 5,161) were considered for inclusion in the study. This sale was selected because of the large number of horses offered, compared with other sale s in the US. The 2006 sale was conducted from September 11 through 25, 2006. Design : The study was designed as a cross sectional study. A sample size calculation performed with standard software indicated that given a population size of 5,161 horses, an as sumed expected prevalence of P1 fragments in the MCP joint 3 4 of 10% 3%, and a desire to calculate 95% CIs with 80% power, 357 horses would be required in the study. To account for possible missing data in medical records at the time of sale, it was decid ed to include 40 (approximately 10%) additional horses. Thus,

PAGE 62

62 397 horses randomly selected by use of a computer software program from among all horses offered for sale were included in the study. Evaluation of radiographs : For each horse offered for sale during the 2006 Keeneland September yearling sale, a repository consisting of digital radiographs (taken within 30 days prior to the sale), veterinary certificates, and any other information that the seller wished to disclose was established. For the prese nt study, repository radiographs of horses selected for inclusion in the study were evaluated by an equine veterinarian with 26 years of professional experience in yearling pre purchase examinations (JCB). The veterinarian evaluating the radiographs was no t employed by the sellers or prospective buyers, and was engaged to work exclusively for the purposes of the present study. A list of selected hip numbers for identification of study horses was provided to this individual for review of radiographs, which w as performed prior to the sale of any horses. Therefore, the evaluator was not aware of the sales price of any horse at the time of evaluation of the radiographs. A radiographic finding was defined as any congenital, developmental, or acquired bony lesion or any abnormality but may have included normal variations in the radiographic appearance or development of bone structures. Anatomic locations of all radiographic finding were recorded on standardized data collection forms classifying lesions as affecting the left, right, or both forelimbs or the left, right, or both hind limbs. Specific abnormalities that were recorded were selected on the basis of previous reports 1 34 and clinical experience of the veterinarian evaluator. As for a previous study, 3 5 a rad iographic set for each horse was considered complete when it contained the 32 radiographic views recommended by the American Association of Equine Practitioners ( Appendix ).

PAGE 63

63 Data regarding presale surgery were listed by surgery site and date on the standard ized data collection form. As set forth in the conditions of sale, 39 surgical procedures designed to produce a permanent change in physical characteristics of the horse, including transphyseal bridging surgery and periosteal transection and elevation were disclosed by sellers voluntarily. However, more invasive surgeries, such as arthroscopy, other invasive joint surgeries, and abdominal surgery of any type, were required to be disclosed according to the conditions of sale. All disclosed surgeries were reco rded, but for the present study, only data on presale arthroscopy were analyzed. Data collection : For each horse, the following data were recorded: hip number, sales day, whether the horses was sold (yes vs no), sales price, sire name, dam name, age (mont hs), sex (colt vs filly), whether the stated reserve was attained (yes vs no), and whether the horse was withdrawn (yes vs no). Categorization of radiographic findings : A standardized form was developed for collection of radiographic findings in each study horse. The classification system was based on that reported in a previous study 2 of radiographic abnormalities in 1,162 Thoroughbred yearlings, with some modifications. For each study horse, each radiographic view of each limb was evaluated for specific a bnormalities and data were recorded. Thus, multiple abnormalities may have been recorded for individual study horses. For the MCP and MTP joints, radiographic findings that were recorded included proximodorsal and proximopalmar or proximoplantar fragments of P1. A distinction between articular and non articular fragments was not made because of the potential for misclassification. Cysts were identified and defined as any area of increased lucency

PAGE 64

64 that extended into subchondral bone. The presence of a well defined semicircular lucency on the proximodorsal aspect of the sagittal ridge of MC3 or MT3 was recorded as OCD. Other radiographic findings included bony fragments, lucencies, and supracondylar lysis involving the palmarodistal or plantarodistal aspect o f MC3 or MT3. The medial and lateral proximal sesamoid bones were categorized as elongated if they were > 2 mm longer than the other proximal sesamoid bone of the same limb. Abnormal shape of the sesamoid was defined as proximal, distal, or abaxial enlarg ement. Fracture of the proximal sesamoid bone was classified as apical, abaxial, basilar, or other. Periarticular bone formation at the proximal or distal articulation of the MCP and MTP joints was recorded as an osteophyte, and bone production within the suspensory or distal sesamoidean ligaments was recorded as enthesophytes. Other radiographic findings included the presence and number of defined circular lucencies in the proximal sesamoid bones, and the presence of linear lucencies (ie, vascular channels ) in the proximal sesamoid bones. Regular vascular channels were defined as Irregular vascular channels were defined as linear lucencies that had non parallel sides for any portion of their length or were > 2 mm in width. For the carpal joints, radiographic findings that were identified included abnormalities of the radial carpal bone, third carpal bone, or both (ie, thickened dorsal cortex, proliferative changes, en thesophytes, and bony fragments). The presence of circular lucencies involving the palmar region of the ulnar carpal bone, circular lucencies involving the accessory carpal bone, fragments, osteophytes, cysts, accessory carpal bone fractures, and wedging o r collapse of the carpal bones were also recorded.

PAGE 65

65 For the tarsal joints, radiographic findings that were identified included lucencies involving the medial malleolus, bone fragments, concavities of the distal intermediate ridge of the tibia, and abnormali ties of the lateral and medial ridges of the talus (ie, flattened areas, lucencies, and fragments). Variations in the radiographic appearance of the distomedial aspect of the trochlear ridge were recorded, along with fragments, osteophytes, and enthesophyt es involving the distal intertarsal or tarsometatarsal joint margins, and subchondral lucencies and wedging or collapse of the tarsal bones. For the stifle joints, radiographic findings that were identified included flattened areas involving the lateral an d medial trochlear ridges of the femur, subchondral defects with or without fragmentation, lucencies or fragmentation in the trochlear groove, and subchondral cysts involving the patella, femoral condyles, and proximal aspect of the tibia. Statistical ana lysis : Prevalence of horses with various radiographic findings and prevalence of horses with presale arthroscopy were calculated by dividing the number of affected horses by the total number of study horses; 95% CIs were calculated as described. 6 The corre lation between median daily sales price of all horses offered at the sale (n = 5,161) and median daily sales price of study horses (n = 397) was estimated by means of linear regression analysis. Initially, variables that were associated (ie, P 0.20) with the outcome of interest (ie, sales price) in univariable analyses were entered into the model, and a forward stepwise approach was used to identify variables associated with the outcome of interest, with a 2 sided P value of 0.05 used to inclu de variables in the model, and a 2 sided P value of 0.10 used to remove variables.

PAGE 66

66 The association between sales price and radiographic findings was examined by use of the Wilcoxon rank sum test. Multivariable ANOVA was applied to examine the relationship between sales price (rank data) and radiographic findings, controlling for sales day, presale arthroscopy, and other radiographic findings. Initially, variables that were associated (ie, P 0.20) with the outcome of interest (ie, sales price) in univariable analyses were entered into the model, and a forward stepwise approach was used to identify variables associated with the outcome of interest, with a 2 sided P value of 0.05 used to inclu de variables in the model, and a 2 sided P value of 0.10 used to remove variables. In analyses of the association between sales price and radiographic findings, dorso and palmaroproximal fragments of P1 in either forelimb were grouped together into 1 vari able (forelimb P1 fragments). Similarly, dorso and plantoproximal fragments of P1 in either hind limb were grouped into 1 variable (hind limb P1 fragments). Osteochondritis dissecans of the lateral or medial trochlear ridge of the femur and of the medial femoral condyle were grouped into 1 variable (OCD of the stifle joint). Sales price was compared between horses with and without a history of presale arthroscopy, matched by sales day, by use of the Wilcoxon signed rank test. The overall median sales pric e was compared between horses with and without a history of presale arthroscopy by use of the Wilcoxon rank sum test. The association between sales price (ranked data) and presale arthroscopy (controlling for sales day and radiographic findings) was examin ed by use of multivariable ANOVA. Analysis of the relationship between sales price and radiographic findings or a history of presale arthroscopy, matched by sales day and sire, was not possible because the number of horses with or without radiographic fin dings or a history of

PAGE 67

67 presale arthroscopy with the same sire was small. For all statistical analyses, values of P Results Of the 397 horses included in the study, 291 (73%) were sold, 95 (24%) were listed as reserve not attained, and 11 (3%) were withdrawn before entering the sales ring. In the forelimbs, the most common radiograph ic findings were vascular channels in the proximal sesamoid bones (23%; 95% CI, 19% to 27%), enthesophytes or osteophytes in the radiocarpal joints (22%; 95% CI, 18% to 26%), and OCD of the sagittal ridge of MC3 (20%; 95% CI, 16% to 24%). In addition, 12 o f the 397 horses (3%; 95% CI, 1% to 5%) had forelimb P1 fragments, including 10 horses with fragments of the proximodorsal aspect of P1 and 2 horses with fragments of the proximopalmar aspect. In the hind limbs, the most common radiographic findings were e nthesophytes or osteophytes involving the proximal sesamoids (39%; 95% CI, 34% to 44%), radiographic abnormalities of the distodorsal aspect of MT3 (36%; 95% CI, 31% to 40%), enthesophytes or osteophytes involving the distal intertarsal joint (27%; 95% CI, 22% to 31%), P1 fragments (9%; 95% CI, 6% to 11%), and OCD involving the stifle joints (8%; 95% CI, 6% to 11%). Prevalence estimates for specific radiographic findings in the forelimbs and hind limbs were not significantly different when horses sold durin g the first 4 days were compared with horses sold during the later days of the sale, except that vascular channels in the proximal sesamoid bones ( P = 0.03) and OCD of the distal sagittal ridge of MC3 ( P = 0.05) were significantly more common in horses sol d during the later days of the sale.

PAGE 68

68 Thirteen percent (95% CI, 10% to 16%) of horses had a history of presale arthroscopy ( Table 5 1 ). The most common anatomic site of presale arthroscopy was the MTP joint (31/52). Horses having arthroscopy in multiple an atomic sites were also common (14/52), and all of those horses had surgery in the MTP joint. For all horses offered for sale, median price the first day of the sale (ie, September 11) was $310,000 and median price the fourth day of the sale (ie, September 14) was $120,000. By comparison median price the fifth day of the sale (ie, September 16) was $80,000, and median price the last day of the sale (ie, September 25) was $4,600 (Table 5 2) Daily median prices of all horses offered for sale (n = 5,161) and of study horses (n = 397) were highly correlated throughout the 14 day sales period ( R = 0.94; P < 0.01). In univariable analyses, the following forelimb radiographic findings were associated ( P P = 0.08), vascular c hannels in the proximal sesamoid bones ( P = 0.12), and enthesophytes or osteophytes involving the radiocarpal joint ( P = 0.16). Similarly, in univariable analyses, the following hind limb radiographic findings were associated with sales price: P1 fragments ( P = 0.05), vascular channels in the proximal sesamoid bones ( P = 0.17), enthesophytes or osteophytes involving the distal intertarsal joint ( P = 0.11), and OCD involving the stifle joints ( P = 0.07). In the multivariable analysis, forelimb P1 fragments a nd hind limb P1 fragments were significantly ( P rank for horses with forelimb P1 fragments (mean SE sales price rank, 170 29) was significantly ( P = 0.01) lower than mean sales price rank for horses without forelimb P1 fragments (238 13), after controlling for sales day, history of arthroscopy, hind limb P1

PAGE 69

69 fragments, and OCD involving the stifle joints (Table 5 3) Means sales price rank for horses with hind limb P1 fragments (186 22) was signifi cantly ( P = 0.05) lower than mean sales price rank for horses without hind limb P1 fragments (222 18). Mean sales price rank was lower for horses with OCD involving the stifle joints (186 24) than for horses without this radiographic finding (222 15) but this difference was not significant ( P = 0.06). In the univariable analysis, median sales price was not significantly ( P = 0.46) different when horses with ($50,000) and without ($35,000) a history of presale arthroscopy were compared (Table 3). Howe ver, multivariable ANOVA indicated that after controlling for sales day and the presence of P1 fragments and OCD involving the stifle joints, mean sales price rank was significantly ( P < 0.01) lower in horses without a history of presale arthroscopy (180 18), compared with horses with a history of presale arthroscopy (229 20). Discussion In the present study, the most common radiographic findings in the forelimbs of Thoroughbred yearlings offered for sale at the 2006 Keeneland September yearling sale we re vascular channels in the proximal sesamoid bones, enthesophytes or osteophytes in the radiocarpal joint, and OCD of the sagittal ridge of MC3. In the hind limbs, the most common radiographic findings were proximal sesamoid bone enthesophytes or osteophy tes, abnormalities of the distodorsal aspect of MT3, and enthesophytes or osteophytes involving the distal intertarsal joint. There was a significant association between the presence of P1 fragments in forelimbs or hind limbs and sales price. Yearlings wit h P1 fragments in forelimbs had lower sales prices than did yearlings

PAGE 70

70 without such fragments, after controlling for sales day, history of presale arthroscopy, and the presence of P1 fragments in the hind limbs, OCD in the stifle joints, or both. Similarly, yearlings with P1 fragments in the hind limbs had lower sales prices than did yearlings without such fragments, after controlling for sales day, history of presale arthroscopy, P1 fragments in the forelimbs, and OCD in the stifle joints. A random sampling approach was used to estimate the prevalence of various radiographic findings in horses offered for sale. Because median daily price of all horses offered for sale and median daily price of the study horses was highly correlated during the 14 day sales pe riod, we suggest that this indicates that the random sample of horses in the present study was representative of the population of horses offered for sale. Current knowledge of the association between the presence of P1 fragments in Thoroughbred yearlings and future athletic performance is limited. In a study 1 of Thoroughbred yearlings sold in Texas, the median sales price ($6,500) was lower in yearlings with P1 fragments in the forelimbs or hind limbs, compared with price for horses without P1 fragments ( $9,000). However, the analysis did not account for the potential confounding effects of additional radiographic findings on sales price, and no distinction was made between forelimbs and hind limbs. In a study 2 conducted in Kentucky, the presence of P1 fra gments in the forelimbs did not affect the ability of Thoroughbred yearlings to start a race as 2 year old s, whereas the odds of starting a race were lower in Thoroughbred yearlings with P1 fragments in the hind limbs. However, in that study, 2 any associat ion between the presence of specific radiographic findings and future ability to start a race as a 2 year old was difficult to assess because the occurrence of injuries during training prior to racing could not be measured.

PAGE 71

71 In the present study, median sa les price was lower in horses with OCD involving the stifle joints ($20,000), compared with horses without this radiographic finding ($38,500), after controlling for sales day, presale arthroscopy, and P1 fragments in the forelimbs or hind limbs. However, the difference in price was not significant. Because OCD of the stifle joint often causes lameness, 40 it is possible that a small sample size was the reason that a significant difference was not observed in our study. If the observed difference were real, then the power of detecting this difference given the sample size used was < 50% at the 5% level of significance. A comparison of estimates of prevalence of radiographic findings in Thoroughbred yearlings from prior reports 1, 34 and the present study is dif ficult because the sampling methods differed between studies. In the previous study conducted in Kentucky, 34 study horses were not randomly selected, and the median sales price of study horses was significantly higher ($40,000) than that for all yearlings offered for sale ($22,000), suggesting that the study yearlings did not represent the sale population. In the study conducted in Texas, 1 study horses were selected on the basis of whether health records were available for review. Therefore, it is possible that prevalence estimates for radiographic findings may have been under or overestimated. For horses in the present study, the prevalence of presale arthroscopy was 13%. However, compliance with surgery disclosure policies was not known, and if surgery d isclosure was perceived to negatively affect sales price, then some surgeries may not have been disclosed. Therefore it is possible that the prevalence of presale arthroscopy may have been underestimated in this study. Most horses had arthroscopy of the hi nd limb MTP joint. In yearlings, osteochondral fragments most frequently occur in the hind

PAGE 72

72 limbs and may develop early in life (eg, < 3 months of age). 29 3 0 In addition, some yearlings undergoing sales preparation may kick objects (eg, stalls or exercisers ) with their hind limbs, causing traumatic osteochondral P1 fragments. It is also possible that hind limb MTP joint arthroscopy may be perceived to mitigate potential detrimental effects of osteochondral P1 fragments on yearling sales price or future athle tic performance. In a previous study, 2 yearlings with proximodorsal P1 fragments in the hind limb MTP joint at the time of sale were less likely to start a race as 2 or 3 year olds, compared with yearlings without this radiographic finding. Results from a study 3 conducted at a training center in Florida revealed that arthroscopy (after sale but before training) may have a detrimental effect on athletic performance in Thoroughbred yearlings during training. The number of days lost during training was 28 day s higher and financial returns were $12,000 lower for Thoroughbred yearlings that had arthroscopy, compared with horses that did not have surgery. In the present study, yearlings with a history of presale arthroscopy sold for a higher price than did yearli ngs without a history of surgery, after controlling for sales day, P1 fragments, and OCD of the stifle joints. Most often, surgery to remove fr agments is performed before pre p urchase radiographs are obtained. As a result, the radiographic findings that ne cessitated surgery are no longer evident on radiographs of these horses that are placed in the repository at the time of the yearling sale. Because yearlings with a history of presale arthroscopy sold for more money than did yearlings without a history of surgery in our study, we suggest that veterinarians advising buyers may assume that presale arthroscopy is unlikely to compromise future racing

PAGE 73

73 performance. However, a higher sales price in yearlings with a history of surgery could also be explained by ped igree related factors, which were not assessed. The present study had several limitations. All radiographs were evaluated by a single veterinarian, and the sensitivity and specificity of the method of radiographic evaluation was not assessed. Misclassific ation of mild radiographic findings was possible, which may have biased the prevalence estimates for specific radiographic findings (eg, vascular channels in proximal sesamoid bones and presence of enthesophytes and osteophytes in radiocarpal and intertars al joints). In addition, it was difficult to assess the effect of pedigree (eg, sire) on the relationship between sales price and radiographic findings and between sales price and presale arthroscopy, because the number of horses presented for sale during the first 4 days with or without radiographic findings or a history of presale arthroscopy with the same sire was small. Therefore, the analyses of sales price included the variable of sales day (ie, the first 4 days vs subsequent days) because sales price s were higher during the first 4 days of the sales, when elite horses were offered for sale. Despite these limitations, the information provided by the present study may assist consignors, buyers, and veterinarians to more objectively evaluate Thoroughbred yearlings offered for sale.

PAGE 74

74 Table 5 1 Frequency distribution of surgery site for horses offered for sale at the Keeneland September 2006 yearling sale that had a history of presale arthroscopy. Surgery site Left Right Bilateral No. of horses MCP join t 3 4 0 7 Carpal joint 1 1 0 2 MTP joint 7 8 2 17 Tarsal joint 1 4 2 7 Stifle joint 2 3 0 5 Multiple sites* NA NA NA 14 Total 14 20 4 52 NA= Not applicable. *Includes 1 horse each with arthroscopy of the right MCP and MTP joints, left MTP joint and left tarsal joint, left MTP joint and both stifle joints, left MTP joint and right tarsal joint, right MCP joint and right MTP joint, right MTP and right tarsal joint, both MTP joints and right tarsal joint, right MTP joint and left stifle joint, right MTP joint and right stifle joint, left MCP joint and right stifle joint, left MCP joint and left MTP joint, right tarsal joint and other joint, right MCP joint and left MTP joint, and right MCP joint and right MTP joint.

PAGE 75

75 Table 5 2 Sales price of Thoroughb reds offered for sale at the Keeneland September 2006 yearling sale. Day of the month Median sale price of all yearlings ($) Study yearlings Median sale price ($) IQR No. of horses 11 310,000 277,000 145,000 352,500 18 12 300,000 437,000 170,000 743,000 18 13 150,000 132,000 91,000 200,000 30 14 120,000 75,000 37,000 170,000 31 16 80,000 70,000 47,000 135,000 31 17 72,000 57,000 37,000 100,000 30 18 55,000 45,000 27,000 75,000 31 19 40,000 33,000 13,000 50,000 30 20 30,000 21,000 12,000 30 ,000 31 21 20,000 19,000 7,000 37,000 31 22 13,000 10,000 6,000 35,000 31 23 11,000 14,000 9,000 21,000 30 24 6,500 3,000 2,000 13,000 28 25 4,600 3,000 2,500 7,000 27 IQR = Interquartile range (25th to 75th percentile)

PAGE 76

76 Table 5 3 Comparison of sal es prices for Thoroughbred yearlings sold at the September 2006 Keeneland yearling sale with and without various presale radiographic findings or a history of presale arthroscopy. Radiographic finding No. of horses Median sales price ($) Mean SE sales pr ice rank P value Forelimb P1 fragments 0.01 No 280 38,000 238 13 Yes 11 15,000 170 29 Hind limb P1 fragments 0.05 No 263 40,000 222 18 Yes 28 16,000 186 22 OCD of stifle joint 0.06 No 266 38,500 222 15 Yes 25 20,000 1 86 24 Presale arthroscopy < 0.01 No 253 35,000 180 18 Yes 38 50,000 229 20 Sales day < 0.01 1 to 4 68 170,000 285 20 5 to 14 223 150,000 123 17 Mean sales price rank was adjusted for sales day and history of presale arthrosc opy.

PAGE 77

77 CHAPTER 6 EPIDEMIOLOGIC ANALYSIS OF SELECTED YEARLING PRESALE RADIOGRAPHIC FINDINGS AND THEIR ASSOCIATION WITH RACING PERFORMANCE IN 2 year old THOROUGHBRED HORSES Yearling TB horses are purchased at public auctions from July through September for t he purpose of racing them the following year as 2 year old racehorses. The price of yearlings can vary based on their pedigree, conformation, and disclosed health information such as history of arthroscopic surgery and pre sale radiographic findings. 1, 3 4,4 1 Because the presence of pre sale abnormal radiographic findings may affect the ability of a yearling to race as a 2 year old or limit racing performance, veterinarians are hired by horse buyers to evaluate pre sale radiographs for abnormalities. Two prev ious studies 1, 2 investigated the association between pre sale radiographic findings in TB yearlings and future racing performance in the United States; however, the study conclusions were different. In a study 2 of 1,162 TB yearlings offered for sale at two different sales in Kentucky between 1993 and 1996, yearlings with moderate or extreme palmar supracondylar lysis of MC3, dorsal medial middle carpal joint disease, enthesophyte formation on the fore proximal sesamoid bones, or hindlimb P1 OC fragments wer e less likely to start a race as 2 or 3 year olds, compared to yearlings without these radiographic findings. However, in another study 1 of 348 TB yearlings offered for sale in Texas between 2002 and 2003, no radiographic findings were associated with fut ure racing performance. In each study, the sample of horses used was not randomly selected from each corresponding sale. These studies were based on radiographic examinations of yearlings obtained from private veterinary practices, and the study population s were different. For example, the median sales price

PAGE 78

78 of TB yearlings was $40,000 in the study in Kentucky and $9,000 in the Texas study. In addition, the number of horses in cohorts with hindlimb P1 OC fragments in these 2 studies was 25 and 10, respectiv ely (a limiting factor that could have produced false negative results, particularly in the study in Texas). Because previous studies have produced discrepant results, the epidemiologic associations between pre sale radiographic findings in TB yearlings an d future racing performance in the United States have not been well established. The objective of this study was to examine the relationships between selected pre sale abnormal radiographic findings in TB yearlings and future racing performance using a ran dom sample of yearlings offered for sale at the 2006 Keeneland September yearling sale. The variables used to assess racing performance were: starting a race, starts placed (ie, when a horse finished a race in first, second, or third place), money earned, and money earned per start. Materials and Methods Horses : Yearlings offered for sale at the 2006 Keeneland September Yearling Sale were considered for inclusion in the study (n = 5,161). Sample size : A sample of 397 horses from a previous study 35 was used to examine the relationship between exposure to selected yearling pre sale radiographic findings and future racing performance. Given an available sample size of 33 horses diagnosed with hindlimb P1 OC fragments and 362 without this finding, this study had 95% confidence and 80% power to declare a difference of 50% (horses without fragments that failed to start a race) versus 75% (horses with fragments that failed to start a race) as significant. 42

PAGE 79

79 Design : This investigation was designed as an observatio nal cohort study. Horses were grouped based on the presence (yes, no) of selected pre sale radiographic findings on forelimbs or hindlimbs. The outcome variables for starting a race, starts placed, money earned, and money earned per start were used to comp are future racing performance between horses with or without presence of selected abnormal radiographic findings. Potential confounding factors that may affect racing performance such as gender, history of pre sale arthroscopic surgery, purchase price at t he yearling sale and sales day were examined in the analysis. The follow up period included the 2 year old racing year only. Demographic data : For each horse, the following data were recorded at the yearling sales: hip number, gender, history of pre sale a rthroscopic surgery, yearling sales price and sale day. Evaluation and categorization of radiographic findings : For each horse offered for sale during the 2006 Keeneland September yearling sale, repository of digital radiographs of horses were evaluated by an equine veterinarian with 26 years of experience in yearling pre purchase examinations. Evaluation and categorization of radiographic procedures were the same as those reported in a previous study 43 and included the standard 32 radiographic views recom mended by a panel of veterinarians selected by the Keeneland September yearling sale, based on guidelines provided by the American Association of Equine Practitioners. 33 In this study, specific radiographic findings of interest were, in forelimbs: P1 OC fr agments, proximal sesamoid osteophytes/enthesophytes, and middle carpal joint osteophytes/enthesophytes. In hindlimbs: P1 OC fragments and proximal sesamoid osteophytes/enthesophytes. In

PAGE 80

80 addition, osteochondritis dissecans of the lateral or medial trochlea r ridge of the femur and subchondral bone cysts of the medial femoral condyle were grouped into 1 variable (OCD in stifle joint). These 6 radiographic findings were selected because they have been previously reported to affect yearling sales price 1, 35 or fu ture racing performance. 2 R acing performance : Racing performance data were collected electronically from The Jockey Club Information Systems 4 4 and included the following: horse identification, started a race, number of starts, starts placed (when a horse f inished a race in first, second, or third place: yes, no), money earned (US dollars) during the 2 year old racing year, and money earned per start. Money earned per start was calculated as the total amount of money earned during the 2 year old racing year divided by the total number of races in that year. Statistical analysis : Median sales price at the yearling sale was compared between horses that started a race and those that did not by using the Wilcoxon rank sum test. The frequency of investigated exp osure factors including gender (colt, filly), history of arthroscopic surgery (yes, no), purchase price at the yearling sale (high, low), sale day (1 to 4 vs 6 to 14), and presence of selected radiographic findings on fore or hindlimbs (yes, no) was compar ed between horses that raced as two year olds (yes, no) and between horses that finished 1 st 2 nd or 3 rd place as two year olds (yes, no) by using a chi square test. The continuous variable for purchase price at the yearling sale was categorized into 2 gr oups (high, low) on the basis of its frequency of distribution (median).

PAGE 81

81 Among horses that raced as two year olds, median money earned and median money earned per start were compared by gender, history of arthroscopic surgery, purchase price at the yearli ng sale (high, low), sale day and presence of selected radiographic findings on fore or hindlimb by using the non parametric Wilcoxon rank sum test. The relationships between selected radiographic findings and the outcome variables for money earned and mon ey earned per start were not examined using multi ANOVA because these two variables did not meet the assumption of normality (including rank data, log 10, natural log, or square root transformations). Instead, these relationships were examined by using log istic regression. Unconditional logistic regression was used to model the odds of starting a race (yes, no), starts placed (yes, no), money earned (high, low), and money earned per start (high, low). The continuous variables for money earned and money earn ed per start were categorized into 2 groups (high, low) on the basis of its frequency of distribution (median). In the univariable analysis, variables with a P value < 0.20 were considered eligible for further multivariable logistic regression modeling to identify radiographic changes associated with the four outcomes of interest. To determine the best fitting model, the variable with the smallest P value in the univariable analysis was entered into the model first. Thereafter, each of the remaining variabl es was added to the model containing the first variable to test whether its addition significantly improved the fit of the model. The variable with the highest likelihood ratio statistic ( X 2 test with one degree of freedom) was selected for addition to the model and the process was then repeated; variables had to have a P history of arthroscopic surgery, purchase price at the yearling sale and sale day were

PAGE 82

82 examined for confounding by adding one of the variables and assessing the changes in the adjusted OR of the r emaining variables (ie, > 10%). Following the fitting of the main effects model, the interaction term between forelimb proximal sesamoid osteophytes/enthesophytes and hindlimb P1 OC fragments was tested for significance by using the likelihood ratio statis tic. The logistic regression model for failure to start a race was assessed for goodness of fit using the Hosmer Lemeshow test, while the ability to discriminate between horses that started a race versus those that did not was determined by using the recei ver operating characteristic (ROC) curve. 4 5 In the final model, adjusted OR and 95% confidence intervals were reported. In the analyses, proximal dorsal and palmar P1 OC fragments were grouped into 1 variable (forelimb P1 OC fragments). Similarly, in the h indlimb, proximal dorsal and plantar P1 OC fragments were grouped into 1 variable (hindlimb P1 OC fragments). Osteochondritis dissecans of the lateral or medial trochlear ridge of the femur and subchondral bone cysts of the medial femoral condyle were grou ped into 1 variable (OCD in stifle joint). In all analyses, values of P Results Overall, 192 of 397 (48%) study horses started in at least one race during the 2 year old season. Median sales price at the yearling sale o f all study horses was $35,000 (25 th percentile = $10,000; 75 th percentile = $85,000). Median sales price was not different between horses that started a race ($37,000) and those that did not ($30,000) ( P = 0.39). Number of horses that started a race : In t he univariable analysis, the number of horses that did not start a race was higher in horses with forelimb proximal sesamoid

PAGE 83

83 osteophytes/enthesophytes (63%) than horses without these findings (50%; P = 0.04) ( Table 6 1 ). In addition, 67% of horses with hin dlimb P1 OC fragments (dorsal or plantar combined) did not race as 2 year olds, compared to 50% of horses that did not have this finding ( P = 0.07). Finally, 46/78 or 59% of horses with diagnosed with two or more abnormal radiographic findings did not race as 2 year olds, compared to 159/319 or 50% of horses with one or no findings ( P = 0.14). Most horses (39/78) with multiple findings had fore and hindlimb proximal sesamoid osteophytes/enthesophytes, followed by hindlimb P1 OC fragments and forelimb proxi mal sesamoid osteophytes/enthesophytes (16/78). Using logistic regression, after adjusting for a diagnosis of hindlimb P1 OC fragments, the odds of failure to start a race as a 2 year old were 1.78 greater in horses with forelimb proximal sesamoid osteophy tes/enthesophytes, compared to horses without this finding ( P = 0.04; Table 6 2 ). In addition, after adjusting for a diagnosis of forelimb proximal sesamoid osteophytes/enthesophytes, the odds of failure to start a race as a 2 year old were 2.02 times grea ter in horses with hindlimb P1 OC fragments, compared to horses without this finding ( P = 0.06). While this finding did not reach statistical significance at the 5% level, nearly the entire area of the 95% CI (0.95, 4.31) was above the null value (1.0). In the analysis, adding of the variables for gender, history of arthroscopic surgery, purchase price at the yearling sale, or sale day did not result in a change in the ORs for forelimb proximal sesamoid osteophytes/enthesophytes or hindlimb P1 OC fragments by > 10%, an indication that the associations between diagnosis of these 2 abnormal radiographic findings and failure to start a race were not confounded by gender, history of arthroscopic surgery,

PAGE 84

84 purchase price at the yearling sale, or sale day. Addition to the model of the interaction term between forelimb proximal sesamoid osteophytes/enthesophytes and hindlimb P1 OC fragments did not produce a significant likelihood ratio statistic. The Hosmer Lemeshow goodness of fit test showed that the model fit the data ( X 2 = 0.0005; df = 1; P = 0.98). The model, however, did not have a good ability to discriminate between horses that started a race versus those that did not (area under the ROC curve ~ 0.60). Number of horses with starts placed : In the univariable a nalysis, the number of horses that finished a race in first, second, or third place was higher in horses with OCD of the stifle joint (81%) compared to horses without this finding (61%); but this difference was not significant ( P = 0.08) ( Table 6 3 ). Using logistic regression, gender, history of arthroscopic surgery, purchase price at the yearling sale, and sale day were examined as potential confounders and not found to have an influence on the association between diagnosis of OCD of the stifle and starts placed. Money earned : Using the Wilcoxon rank sum test ( Table 6 4 ), median money earned was lower in horses with forelimb P1 OC fragments ($114) than in horses without this finding ($4540); but this difference was not significant ( P = 0.19). Using logistic regression, gender, history of arthroscopic surgery, purchase price at the yearling sale, and sale day did not have an influence on the association between diagnosis of forelimb P1 OC fragments and money earned (high, low). Money earned per start : In the Wilcoxon rank sum test ( Table 6 5 ), median money earned per start was lower in horses with forelimb P1 OC fragments ($114) compared to horses without this finding ($1560); but this difference was not significant ( P = 0.13). Using logistic regression, gende r, history of arthroscopic surgery, purchase

PAGE 85

85 price at the yearling sale, and sale day did not have an influence on the association between diagnosis of forelimb P1 OC fragments and money earned per start (high, low). Discussion This study provides new info rmation on the associations between yearling pre sale radiographic findings and racing performance in 2 year old TB horses. In contrast to previous studies 2,3 that have examined racing performance in horses during the 2 and 3 year old seasons combined, th is investigation was focused on the 2 year old season only. Overall, study findings revealed that a diagnosis of forelimb proximal sesamoid osteophytes/enthesophytes or hindlimb P1 OC fragments was associated with failure to start a race during the 2 year old racing year of TB horses. In this study, 192 of 397 (48%) horses started at least 1 race during the 2 year old racing year. This frequency is higher when compared to racing starts for the total number of TB horses foaled in the United States in 2005 (1 1,197 / 35,045 or 33%) 4 4 which was the same year horses in this study were foaled. Yearlings offered for sale at Keeneland, KY have a higher median sales price compared to yearlings sold elsewhere. 1 3 Authors of a previous study 2 proposed that more expens ive horses likely have better opportunities to race, enter high quality training stables, and benefit from better opportunities in racing. They found that 946 (81%) of the horses in their study raced during the 2 or 3 year old racing years. Median sales pr ice of study horses was significantly higher than that in all horses offered for sale at two different sales in Kentucky between 1993 and 1996. The authors suggested that pedigree, conformation, and other unknown factors related to future racing performanc e may have contributed to the increased likelihood of those study horses starting a race. Comparison of our results

PAGE 86

86 with the previous study conducted in Kentucky is difficult because in that study the frequency of horses racing as 2 year olds only was not reported. The results of the present study do not support the belief that higher priced yearlings probably benefit from better opportunities in racing. 1,2 Median yearling sales price was not different ( P = 0.39) between horses that raced ($37,000) and hor ses that did not ($30,000). It is possible that this finding resulted from the random sample selection of horses used in this study, thus avoiding over representation of horses with a high yearling sale price. However, a recent study 4 1 conducted on 51 TB y earlings in Florida provided evidence that TB yearlings with a higher commercial value may indeed benefit from better opportunities in racing. In that study, 4 1 lameness in TB yearlings was associated with reduced galloping during training. Reduction in exe rcise distance was more evident among lame horses with a high yearling sales price (ie, median sales price = $225,000; galloping = 250 furlongs or 50 kilometers), compared to lame horses with a low yearling sales price ($20,000; 412 furlongs or 82 kilomete rs). This finding suggests that, in some training centers, more emphasis on prevention of musculoskeletal injury is given to horses with a high commercial value, compared to horses with low commercial value. As a result, more horses with a high commercial value may have more opportunities to start a race because management standards of lameness are more stringent, compared to horses with a low commercial value. In this study, the odds of failure to start a race as a 2 year old were 1.78 times greater in hor ses with forelimb proximal sesamoid osteophytes/enthesophytes, compared to horses without this finding. Albeit a smaller magnitude of association, our study confirms findings of a previous study in TB yearlings sold in Kentucky during

PAGE 87

87 1993 1996, in which o dds of racing were 3 times lower for yearlings with forelimb proximal sesamoid enthesophyes. 2 One difference, however, is that the frequency of horses with this radiographic finding was higher in our study (15%) compared to the previous study in Kentucky ( 1%) making the association observed in this study more relevant. In our study, osteophytes/enthesophytes were combined to eliminate possible misclassification and because of the low prevalence observed in TB yearlings. 1,34 43 While our study provides furth er support that this finding in yearlings is associated with reduced future racing performance, it is unknown what type of effect this radiographic change really has on soundness of the horse since this could be associated with MCP osteoarthritis or suspen sory ligament changes. In this study, the odds of failure to start a race as a 2 year old were 2.02 times greater in horses with hindlimb P1 OC fragments, compared to horses without this finding. Our study confirmed results from a previous study in Kentuck y 2 but not from another study in Texas. 1 The latter study had several limitations including a small number of horses (n = 10) with P1 OC fragments. We offer several explanations for the observed association between this condition on hindlimbs and failure to race as a 2 year old. First, the equine MTP (and MCP) joints have the largest number of traumatic and degenerative lesions of all joints of the appendicular skeleton. 46 Fragments present in the joint initiate an inflammatory cascade and joint destructio n through the action of metalloproteinases, aggrecanases, and cathepsins released from cartilage and synovial tissue. 46 These proteases cleave type II collagen disrupting the extracellular matrix, which results in fluid distention of the affected joint, cl inical signs of mild to moderate lameness, pain on palpation, and the progression towards osteoarthritis. 46 The MTP

PAGE 88

88 joints are high motion, which may, in part, explain why an OC fragment may perpetuate the inflammatory process especially in racehorses that are prone to hyperextension of the joint at racing speed. Second, it is known that the frequency of pre sale radiographic changes or arthroscopic surgery is higher in MTP joints than in MCP joints in TB yearlings. 34,43 In addition, there is evidence that lameness is more commonly diagnosed in hindlimbs than in forelimbs in TB yearlings and is attributable to joint injury. 35 Third, options for clinical management of yearlings with plantar P1 OC fragments are debatable since they are often embedded in fibrou s tissue making their clinical significance questionable and surgical removal more difficult. 27 Finally, P1 OC fragments in horses can be treated surgically after the yearling sales and before training. A previous study 3 conducted in a training center in Florida, however, revealed that median number of days lost during training was higher in 8 TB yearlings with P1 OC fragments treated surgically after the yearling sale but before training (38 days lost), compared to 32 yearlings without P1 fragments (19 da ys lost); presence of dorsal and plantar fragments were not reported separately in that study. A high number of days lost during training due to surgical recovery alone or in combination with other causes of training failure could detrimentally affect the ability of TB yearlings to race as 2 year olds. In this study, history of arthroscopic surgery performed before the yearling sale was not associated with number of horses that started a race. However, history of arthroscopic surgery performed in yearlings with P1 OC fragments after the yearling sale and before training was not measured. In this study, among the 192 (48%) horses that started a race, the outcomes for starts placed, money earned, and money earned per start during the 2 year old racing

PAGE 89

89 year wer e not different between horses with or without forelimb proximal sesamoid osteophytes/enthesophytes or hindlimb P1 OC fragments. This may be an indication that the consequences of these pre sale radiographic findings on racing performance are more relevant in TB yearlings in training than later as 2 year olds. It is possible that these radiographic findings in combination with risk factors for athletic performance not measured in this study can be contributing factors that prevent yearlings from starting a race as 2 year olds. For example, this and previous studies 1 2 had a limitation in that known risk factors (ie, lameness and exercise history) 3 35, 4 1 for athletic performance in TB yearlings during training, before the 2 year old racing year, and their a ssociation with future racing performance could not be measured. Finally, while yearling pre sale radiographic findings of forelimb proximal sesamoid osteophytes/enthesophytes or hindlimb P1 OC fragments are associated with failure to start a race as 2 yea r olds, their long term effects on future racing performance requires further evaluation during the 3 year old racing year. This study had several limitations. The sample size used (n = 397 horses) was initially designed to estimate the prevalence of radio graphic findings in TB yearlings in a previous study, 43 and not to examine the relationship between diagnosis of selected radiographic findings and future racing performance in TB horses in this cohort study. For example, the sample size of horses with for elimb P1 OC fragments during the yearling pre purchase examinations (12/397) was too small to adequately examine the relationship of this exposure and future racing performance in TB horses. In a previous study 2 fewer TB yearlings with forelimb P1 OC frag ments started a race as 2 or 3 year olds than yearlings without this finding. Another study limitation was that all radiographs

PAGE 90

90 were evaluated by a single veterinarian, and the accuracy (sensitivity and specificity) of the method of radiographic evaluati on were not assessed. 35 Sales day (early: days 1 to 4 vs late 6 to 14) was used as a proxy variable to examine the relationship between horse quality (pedigree and conformation) on future performance. While this approach to measure horse quality has some v alue, its accuracy needs further evaluation. Another limitation in this and other studies, 1 2 was that exercise history prior to racing as 2 year olds was not available for evaluation. While the logistic regression model for failure to start a race (which included 2 abnormal radiographic findings) fit the data well, groups which are desired for assessment of model fitness). 4 5 This result likely affected the ability of the m odel to discriminate between horses that started a race versus those that did not. A model with a larger sample size of horses diagnosed with selected abnormal radiographic findings and additional exposure factors known to be associated with athletic perfo rmance (ie, exercise history during training, before horses start racing as 2 year olds) should improve model performance. Study results apply to TB yearlings sold at the 2006 Keeneland September yearling sale, and caution must be observed when trying to e xtrapolate study finding to more recent generations of TB yearlings sold at the Keeneland sale. Overall, this is the third observational study that has examined associations between selected yearling pre sale radiographic findings and future racing perform ance in Thoroughbreds in the United States. Despite the study limitations identified above, results of the epidemiologic analysis presented here confirmed findings from a previous observational study 2 that a yearling pre sale radiographic diagnosis of prox imal sesamoid osteophytes/enthesophytes on forelimbs or P1 OC fragments on

PAGE 91

91 hindlimbs is associated with failure to start a race during the 2 year old racing year of Thoroughbred horses.

PAGE 92

92 Table 6 1 Distribution of gender, history of arthroscopic surgery, sale day, purchase price and radiographic findings in Thoroughbred yearlings that did or did not race as 2 year olds Variable Category Raced P* Yes No No. horses % No. horses % Gender Colt Filly 101 91 52 45 94 111 48 55 0.17 History of ar throscopic surgery No Yes 163 29 47 56 182 23 53 44 0.25 History of arthroscopic surgery: MTP or other location No Yes Other 163 18 11 47 60 55 184 12 9 53 40 45 0.33 Sale day 1 to 4 6 to 14 52 140 54 47 45 160 46 53 0.23 Purchase price at the yearling sale High Low 101 91 51 46 98 107 49 54 0.33 Multiple abnormal radiographic findings No Yes** 160 32 50 41 159 46 50 59 0.14 Forelimbs : P1 OC fragments No Yes 187 5 49 42 198 7 51 58 0.63 P1 OC dorsal No Yes 188 4 49 40 199 6 51 60 0.59 Pl OC palmar No Yes 191 1 48 50 204 1 52 50 0.96 Proximal sesamoid osteophytes/enthesophytes No Yes 170 22 50 37 167 38 50 63 0.04 Osteophytes/enthesophytes in middle carpal joint No Yes 175 17 48 55 191 14 52 45 0.45 Hindlimbs : P1 OC fragments No Yes 181 11 50 33 181 22 50 67 0.07 P1 OC dorsal No Yes 187 5 49 31 194 11*** 51 69 0.16 P1 OC plantar No Yes 186 6 49 33 193 12*** 51 67 0.19 Proximal sesamoid osteophytes/enthesophytes No Yes 120 72 50 46 122 83 50 54 0.54 OCD of stifle joint No Yes 171 21 47 58 190 15 5 3 42 0.20 P values were calculated using a chi square test **Most horses (39/78) had fore and hindlimb proximal sesamoid osteophytes/enthesophytes ***1 horse had both P1 OC dorsal and plantar abnormalities

PAGE 93

93 Table 6 2 Adjusted odds ratios (OR) and 95% confidence intervals (CI) for logistic regression analysis of the association between pre sale radiographic findings in Thoroughbred yearlings and failure to start a race as 2 year olds Variable Category Adjusted odds ratio (95% CI) P Forelimb proximal se samoid osteophytes/enthesophytes No Yes 1.00 1.78 (1.01, 3.16) NA 0.04 Hindlimb P1 OC fragments No Yes 1.00 2.02 (0.95, 4.31) NA 0.06 Hosmer Lemeshow goodness of fit statistic chi square = 0.0005; df = 1; P = 0.98).

PAGE 94

94 Table 6 3 Distribution of gender, history of arthroscopic surgery, sale day, purchase price and radiographic findings in Thoroughbred yearlings that finished 1 st 2 nd or 3 rd place as 2 year olds Variable Category Starts placed P* Yes No No. horses % No. horses % Gender Col t Filly 63 59 62 65 38 32 48 35 0.72 History of arthroscopic surgery No Yes 102 20 63 69 61 9 37 31 0.51 History of arthroscopic surgery: MTP or other location No Yes Other 102 14 6 63 78 55 61 4 5 37 22 45 0.36 Sale day 1 to 4 6 to 14 22 74 56 66 23 47 44 34 0.17 Purchase price at the yearling sale High Low 64 58 64 64 37 33 36 36 0.95 Multiple abnormal radiographic findings No Yes 100 22 63 69 60 10 37 31 0.50 Forelimbs : P1 OC fragments No Yes 120 2 64 40 67 3 36 60 0.26 P1 OC dorsal No Yes 120 2 6 4 50 68 2 36 50 0.32 Pl OC palmar No Yes 122 0 37 0 204 1 63 100 0.18 Proximal sesamoid osteophytes/enthesophytes No Yes 107 15 63 68 63 7 37 32 0.23 Osteophytes/enthesophytes in middle carpal joint No Yes 112 10 64 59 63 7 36 41 0.67 Hindlimbs : P1 O C fragments No Yes 113 9 62 82 68 2 38 18 0.19 P1 OC dorsal No Yes 118 4 63 80 69 1 47 20 0.43 P1 OC plantar No Yes 117 5 63 83 69 1 37 17 0.30 Proximal sesamoid osteophytes/enthesophytes No Yes 75 47 60 62 45 25 40 38 0.69 OCD of stifle joint No Yes 1 05 17 61 81 66 4 39 19 0.08 P values were calculated using a chi square test.

PAGE 95

95 Table 6 4 Comparison of money earned among horses that raced as 2 year olds by gender, history of arthroscopic surgery, sale day, purchase price and radiographic findings V ariable Category No. horses Money earned Median (25 th P 75 th P)* P** Gender Colt Filly 101 91 3,094 (416 13,613) 5,165 (1,194 14,673) 0.16 History of arthroscopic surgery No Yes 163 29 4,080 (520 13,700) 5,200 (1,377 15,419) 0.89 History of ar throscopic surgery: MTP or other location No Yes Other 163 18 11 4,080 (520 13,700)) 7,418 (1,669 18,325) 2,174 (110 7,620) 0.57 Sale day 1 to 4 6 to 14 52 140 2,011 (289 12,257) 4,690 (856 14,203) 0.22 Purchase price at the yearling sale High Low 101 91 5,565 (406 19,335) 3,280 (840 10,805) 0.34 Multiple Abnormal radiographic findings No Yes 160 32 4,464 (476 13,423) 4,863 (691 20,175) 0.48 Forelimbs : P1 OC fragments No Yes 187 5 4,540 (620 13,853) 114 (100 26,972) 0.19 P1 OC dor sal No Yes 188 4 4,532 (579 13,815) 1,687 (95 38,826) 0.45 Pl OC palmar No Yes 191 1 4,523 (565 13,853) 114 ND Proximal sesamoid osteophytes/enthesophytes No Yes 170 22 3,968 (543 13,389) 8,675 (595 26,262) 0.25 Osteophytes/enthesophytes in mi ddle carpal joint No Yes 175 17 4,600 (565 14,635) 3,940 (450 8,679) 0.39 Hind limbs : P1 OC fragments No Yes 181 11 4,405 (538 13,406) 10,270 (905 34,800) 0.30 P1 OC dorsal No Yes 187 5 4,523 (557 13,700) 3,940 (1,632 34,774) 0.73 Pl OC plan tar No Yes 186 6 4,242 (548 13,505) 12,472 (734 44,325) 0.29 Proximal sesamoid osteophytes/enthesophytes No Yes 120 72 4,226 (422 13,179) 5,190 (722 20,175) 0.31 OCD of stifle joint No Yes 171 21 3,995 (450 13,853) 8,365 (1,927 13,418) 0.32 *Data presented as median (25th percentile, 75th percentile). **P values were calculated using the Wilcoxon rank sum test.

PAGE 96

96 Table 6 5 Comparison of money earned per start among horses that raced as 2 year olds by gender, history of arthroscopic surgery, sale day, purchase price and radiographic findings Variable Category No. horses Money earned per start Median (25 th P 75 th P)* P Gender Colt Filly 101 91 1,215 (250 3,603) 1,850 (409 4,220) 0.08 History of arthroscopic surgery No Yes 163 29 1,508 ( 349 4,025) 1,680 (370 4,076) 0.92 History of arthroscopic surgery: MTP or other location No Yes Other 163 18 11 1,508 (349 4,025) 2,359 (391 4,253) 1,524 (110 2,174) 0.56 Sale day 1 to 4 6 to 14 52 140 1,627 (250 5,524) 1,510 (398 3,510) 0. 97 Purchase price at the yearling sale High Low 101 91 1,848 (312 5,485) 1,220 (398 2,831) 0.11 Multiple abnormal radiographic findings No Yes 160 32 1,565 (355 4,017) 1,391 (287 4,317) 0.95 Forelimbs : P1 OC fragments No Yes 187 5 1,560 (397 4,067) 114 (100 3,824) 0.29 P1 OC dorsal No Yes 188 4 1,560 (382 4,056) 259 (95 5,532) 0.30 Pl OC palmar No Yes 191 1 1,560 (377 4,067) 114 ND Proximal sesamoid osteophytes/enthesophytes No Yes 170 22 1,511 (348 4,035) 2,044 (442 4,895) 0.42 Osteophytes/enthesophytes in middle carpal joint No Yes 175 17 1,575 (349 4,133) 1,364 (325 1,890) 0.34 Hindlimbs : P1 OC fragments No Yes 181 11 1,524 (362 3,867) 2,567 (226 14,580) 0.47 P1 OC dorsal No Yes 187 5 1,560 (348 4,025) 788 (204 10,866) 0.94 Pl OC plantar No Yes 186 6 1,519 (370 4,001) 4,592 (197 15,285) 0.31 Proximal sesamoid osteophytes/enthesophytes No Yes 120 72 1,488 (342 3,473) 1,812 (397 4,318) 0.30 OCD of stifle joint No Yes 171 21 1,468 (300 4,085) 2,091 (95 7 3,880) 0.27 *Data presented as median (25 th percentile, 75 th percentile).

PAGE 97

97 CHAPTER 7 SUMMARY AND RESEARCH CONCLUSIONS Summary Th e overall objectives of this research were to characterize lameness and presale radiographic findings in TB yearlings duri ng training to assess their effects on athletic performance in TB yearlings during training and later as two year old race horses. The research approach used a series of observational epidemiologic studies comparing yearling sale price, two year old in tr aining sale price, and future racing performance as 2 year old Thoroughbreds between horses classified as lame (yes, no), and with or without the presence of abnormal radiographic findings. Analytical epidemiologic methods and techniques were used to contr ol for potential confounding effects of multiple exposure factors that could have an effect on selected outcomes of interest such as yearling sale price, 2 year old in training sale price, and future racing performance as 2 year old Thoroughbreds. Research C onclusions Objective 1: to characterize lameness during training and compare exercise variables and financial returns among yearling Thoroughbreds that were bought for the purpose of resal e for profit. Thirty seven of forty horses became lame during tra ining, most commonly because of joint injury. Eighteen of the lame horses ha d injuries in hind limbs only; five horses had injuries both in fore and hind limbs. The frequency of new cases of lameness increased as time (days) approached the date of the 2 ye ar old in training sales. At the sales, 4, 21, and 15 horses were classified as fast, average, or slow, respective ly; median financial returns were slightly (but significantly) different among

PAGE 98

98 horses classified as fast ($14,000) or average ($0) and slow ( $8,000) ( P = 0.03). Incidence of lameness during training in yearling horses purchased for the purpose of resale and profit was high. Lameness was more commonly diagnosed in hind limbs than in forelimbs. Lameness was attributable to joint injury in most a ffected horses. Financial returns were different between horses classified as fast or average and slow at the 2 year old in training sales. T he incidence of lameness was high in this group of 40 pinhook horses and most injuries were diagnosed in joints. La meness was not associated with financial returns and most responded favorably to anti inflammatory treatment. It is likely that yearlings in this study may have benefitted from rest alone as a treatment option if the sales schedule had not dictated resumpt ion of training as quickly as possible T his study reveal ed that injuries in pinhooked yearlings are characterized differently than those experienced by older race horses. Objective 2: (i) to examine the relationships between yearling sale purchase price, exercise history, lameness, and speed as predisposing factors for sales price in 2 year old in training Thoroughbreds; and (ii) to compare exercise distance accumulated during the last 60 days before the sales between horses with a high and low purchase pr ice and between horses affected and not affecte d with lameness during training Median sales price of study horses at the 2 year old in training sales was $37,000. Sal e price at the 2 year old in training sales was associated with yearling sale purchase p rice, number of furlongs galloped 1 to 60 days before the sales, and speed recorded at the 2 year old in training sales ( P < 0.05). Most study horses with a high sales price

PAGE 99

99 were horses with a high purchase price, had a reduced number of furlongs galloped, and were classified as fast at the sales. Lameness alone was not associa ted with sales price. L ameness however, was associated with a reduced exercise distance, which was more evident among lame horses with a high purchase price, compared to lame horses with a low purchase price; this finding suggests that yearling sale purchase price can affect the trainers management decision s on the amount of training allocation prescribed to horses affected with lameness. We observed that higher priced yearlings tend to have reduced exercise schedules when compared with less expensive horses. It is possible that this training decision is made arbitrarily on the part of the trainer to reduce the risk of musculoskeletal injury in more expensive horses. This finding raise s welfare and safety issues for lower priced TB pinhooked yea rlings, which may continue to train despite being lame thus plac ing them at greater risk for injury. Different training strategies need to be evaluated to determine the most effective and safe m ethods for preparing TB racehorses. Objective 3: to estimate prevalences of various presale radiographic findings and presale arthroscopy in horses offered for sale at the 2006 Keeneland September yearling sale and compare sale price between yearlings with and without various presale radiographic findings or a history of arthroscopy. Results from this study indicate the most common radiographic findings in the fore limbs were vascular channels in the proximal sesamoid bones (23%), enthesophytes or osteoph ytes in the radiocarpal joint (22%), and osteochondritis lesions involving the sagittal ridge of the third metacarpal bone (20%). In the hind limbs, the most common

PAGE 100

100 radiographic findings were enthesophytes or osteophytes involving the proximal sesamoids (3 9%), abnormalities of the distodorsal aspect of the third metatarsal bone (36%), enthesophytes or osteophytes involving the distal intertarsal joint (27%), and osteochondritis lesions involving the stifle joint (8%). Thirteen percent of horses had a histor y of arthroscopy prior to sale. Median sales price was significantly lower in horses with fragments of the proximal phalanx than in horses without. Median sales price was significantly higher in horses with a history of presale arthroscopy than in horses w ithout. Results revealed associations between a diagnosis of fragments of the proximal phalanx presale arthroscopy, and sale price in yearlings. The association between specific ARF, such as forelimb sesamoid osteophytes/ enthesophytes may be more relevan t due to higher prevalence rates observed in our study (15%) compared to the previous KY study (1%). 34 Our results indicate that arthroscopic surgery prior to the sale does not have a negative impact on yearling sale price which suggest s that veterinarian s subjectively evaluating yearlings do not foresee a negative association between surgery and future racing performance. However, there is no objective data in the literature to support this mindset This study makes prevalence estimates of ARF and surgery available to guide consignors, buyers, and veterinarians in purchase decisions at yearling sale s and future management during training Objective 4: To examine the associations between selected yearling presale radiographic findings and future racing perf ormance in 2 year old Thoroughbreds. Overall, 192 of 397 (48%) study horses started in at least one race during the 2 year old racing year. After adjusting for gender, history of presale arthroscopic surgery

PAGE 101

101 and purchase price at the yearling sale, horses with forelimb proximal sesamoid osteophytes/enthesophytes were 1.36 times less likely to start a race, compared to horses that did not have these changes (RR = 1.36; 95% CI = 0.97, 1.96; P = 0.07). Horses with hindlimb first phalanx osteochondral fragmen ts were 1.56 times less likely to start a race, compared to horses that did not have these changes (RR = 1.56; 95% CI = 0.95, 2.56; P = 0.08). Selected radiographic findings were not associated with outcome variables for starts placed, money earned, or mon ey earned per start. Presale radiographic diagnosis of proximal sesamoid osteophytes/enthesophytes on forelimbs or first phalanx osteochondral fragments on hindlimbs in yearlings was associated with an increased risk of failure to start a race during the 2 year old racing year of Thoroughbred horses. In this study, y earlings with fore limb sesamoid osteophytes/enthesophytes were less likely to start a race; 37% of these horses did not race as a 2 year old. Similarly, 33% of yearlings with hind limb P1 OC fr agments failed to start a race as a 2 year old. S urg ery is not done for proximal sesamoid osteophytes/enthesophytes, but is an option for P1 fragments. However, we did not determine the effect of surgery on the likelihood of starting a race. We suggest th at the association of these 2 ARFs with the reduced likelihood of starting a race should be considered in the decision to purchase TB yearlings.

PAGE 102

102 APPENDIX A SOURCES OF MEDICATIO N: a. Hylartin V, 20 to 40 mg (10 mg/mL), Pfizer Inc, New York, NY. b. Depo Medrol, 40 to 240 mg (20 mg/mL), Pharmacia & Upjohn Co, a division of Pfizer Inc, New York, NY. c. Amiglyde, 0.5 to 2mL (250 mg/mL). Fort Dodge Animal Health, a division of Wyeth, Overland Park, Kan. d. Depo Estradial, 5 mg (5 mg/mL), Pharmacia & Upjohn Co, a division of Pfizer Inc, New York, NY. e. Vetalog, 3 to 24 mg (6 mg/mL), Fort Dodge Animal Health, a division of Wyeth, Overland Park, Kan. f. Sarapin, 2 to 10 mL (0.17 mg/mL distillate), High Chemical Co, Levittown, Pa.

PAGE 103

103 APPENDIX B RADIOGRAPHIC VI EWS ACQUIRED FOR EACH HORSE (32 TOTAL VIEWS): Right and left metacarpophalangeal (fetlock) joints Dorsoproximal Dorsomedial Dorsolate ral Lateromedial (obtained with the joint flexed) Right and left metatarsophalangeal (fetlock) joints Dorsoproximal Dorsoproximomed ial Dorsoproximolateral and 30o lateral to the dorsoplantar line) Lateromedial (obtained with the horse standing) Right and left carpal joints Dorsolateral palmaromedial oblique (30 o lateral to the dorsopalmar line; D45L PaMO) Dorsomedial palmarolateral oblique (30 o medial to the dorsopalmar line; D35M PaLO) Lateromedial (obtained with joint flexed) Right and left tarsal (hock) joints Dorsomedial Dorsolateral

PAGE 104

104 Lateromedial Right and left stifle joints Lateromedial Caudolateral the craniocaudal line; must include medial femoral condyle in its entirety) *Alternatively, the plantarolateral dorsoplantar line) may be used.

PAGE 105

105 REFERENCES 1. Cohen ND, Carter GK, Watkins JP, et al. Associatio n of racing performance with specific abnormal radiographic findings in Thoroughbred yearlings sold in Texas. J Equine Vet Sci 2006;10:462 474. 2. Kane AJ, McIlwraith CW, Park RD, et al. Radiographic changes in Thoroughbred horses. Part 2: Association wit h racing performance. Equine Vet J 2003;35:366 374. 3. Hernandez J and Hawkins DL. Training failure among yearling horses. Am J Vet Res 2001;62:1418 1422. 4. Gibbs P and Cohen ND Early management of race bred weanlings and yearlings on farms. Journal o f Equine Veterinary Science. 2001 ;21(6):279 283. 5. Norwood GL. The bucked shin complex in Thoroughbreds. Proc Am Assoc Equine Practners 1978;24:319 335. 6. Jeffcott LB, Buckingham HW, McCarthy RN, et al. Non invasive measurement of bone: A review of clin ical and research applications in the horse. Equine Vet J 1988;S6:71 79. 7. Nunamaker DM, Butterweck DM, Provost MT. Fatigue fractures in thoroughbred racehorses: relationships with age, peak bone strain, and training. J Ortho Res Jul 1990;8(4):604 611. 8. Boston RC and Nunamaker DM. Gait and speed as exercise components or risk factors associated withonset of fatigue injury of the third metacarpal bone in 2 year old Thoroughbred racehorses. Am J Vet Res 2000;61:602 608 9. Reilly GC, Currey JD, Goodship AE Exercise of young Thoroughbred horses increases impact strength of the third metacarpal bone. Orth Res Soc 1977;15:862 868. 10. Raub RH, Jackson SG, Baker JP. The effect of exercise on bone growth and development in weanling horses. J An im Sci 1989;67:25 08 2514. 11. Smith RK, Goodship AE. The effect of early training and the adaption and conditioning of skeletal tissues. Vet Cli n Equine 2008;24:37 51. 12. Perkins NR, Rogers CW, Firth EC, et al. Musculoskeletal responses of 2 year old Thoroughbred horses t o early training. 3. In vivo ultrasonographic assessment of the cross sectional area and echogenicity of the superficial digital flexor tendon. New Zeal Vet J 2004;52(5):280 284.

PAGE 106

106 13. Cornelissen BP, van Weeren PR, Ederveen AG, et al. Influence of exercise on bone mineral density of immature cortical and trabecular bone of the equine metacarpus and proximal sesamoid bone. Osteochon Musculoskel Dev 1999;31S:79 85. 14. Cherdchutham W, Meershoek LS, van Weeren PR, et al. Effects of exercise on biomechanical pro perties of superficial digital flexor tendon in foals. Am J Vet Res 2001;62:1859 1864. 15. Firth EC, Rogers CW, Anderson BH. Musculoskeletal responses of 2 year old Thoroughbred horses to early training. 4. Morphometric, microscopic and biomechanical prop erties of the digital tendons of the forelimb. New Zeal Vet J 2004;52(5):285 292. 16 Barneveld A, van Weeren PR. Conclusions regarding the influence of exercise on the development of the equine musculoskeletal system with special reference to osteochondro sis. Equine Vet J 1999;31(S):112 119. 17. Peloso JG, Mundy GD, Cohen ND. Prevalence of, and factors associated with, musculoskeletal racing injuries of Thoroughbreds. J Am Vet Med Assoc 1994;204:620 626. 18 Estberg L, Gardner IA, Stover SM et al. A cas e crossover study of intensive racing and training schedules and risk of catastrophic musculoskeletal injury and lay up in California Thoroughbred racehorses. Prev Vet Med 1998;33:159 170. 19 Estberg L, Gardner IA, Stover SM, et al. Cumulative racing spe ed exercise distance cluster as a risk factor for fatal musculoskeletal injury in Thoroughbred racehorses in California. Prev Vet Med 1995;24:253 263. 20 Estberg L, Stover SM, Gardner IA, et al. High speed exercise history and catastrophic racing fractur e in Thoroughbreds. Am J Vet Res 1996;57:15479 1555. 21 Hernandez J, Hawkins DL, Scollay MC. Race start characteristics and risk of catastrophic musculoskeletal injury in Thoroughbred racehorses. J Am Vet Med Assoc 2001;218:83 86. 22 Cohen ND, Berry SM Peloso JG, et al Association of high speed exercise with racing injury in thoroughbreds. J Am Vet Med Assoc 2000;216(8):1273 1278. 23 Mohammed HO, Hill T, Lowe J. Risk factors associated with injuries in thoroughbred horses. Equine veterinary journal. N ov 1991;23(6):445 448.

PAGE 107

107 24 Hernandez JA, Scollay MC, Hawkins DL. Evaluation of horseshoe characteristics and high speed exercise history as possible risk factors for castostrophic musculoskeletal injury in Thoroughbred horses. Am J Vet Res 2005;66(8):1314 20. 25 Kane AJ, Stover SM, Gardner IA, et al. Hoof size, shape, and balance as possible risk factors for catastrophic musculoskeletal injury of Thoroughbred horses. Am J Vet Res 1998;5912)1545 52. 26 Dyson PK, Jackson BF, Pfeiffer DU, Price JS. Days l ost from training by two and three year old Thoroughbred horses: a survey of seven UK training yards. Equine Vet J 2008;40(7):650 657. 2 7 Pool R R. Pathologic m anifestations of j oint d isease in the Athletic Horse. In: Joint Disease in the Horse McIlwrai th CW and Trotter GW, ed. Philadelphia, PA: W. B. Saunders Company; 1996. 28 Caron JP. Understanding the pathogenesis of equine osteoarthritis. Brit Vet J 1992;148(5):369 371. 29 Ros s MW, Dyson, SJ. Noninfectious a rthritis. In: Diagnosis and Management of Lameness in the Horse. Ross MW and Dyson SJ, ed s St. Louis, MO: Saunders; 2003. 30 Bertone, AL. The metacarpus and metatarsus. In: Stashak, TS (Ed). Philadelphia: Lippincott Williams and Wilkins, 2002; 800 805. 31 Stover SM, Johnson BJ, Daft BM, et al. An association between complete and incomplete stress fractures of the humerus in racehorses. Equine Vet J 1992;24(4):260 263. 3 2 McIlwraith CW. Current concepts in equine degenerative joint disease. J Am Vet Med Assoc 19 92;180:239 250. 33 American Association of Equine Practit ioners (AAEP). Lameness scale. D efinition and classification of lameness. In: Guide for V ete rinary Service and J udging of E questrian E vents 4 th ed. Lexington, Ky: American Association of Equine Pr actitioners, 1991;19. 34 Kane AJ, Park RD, McIlwraith CW, et al. Radiographic changes in Thoroughbred yearlings. Part 1: prevalence at the time of the yearling sales. Equine Vet J 2003;35:354 65. 35 Preston SA, Trumble TN, Zimmel DN, et al. Lameness, at hletic performance, and final returns in yearling Thoroughbreds bought for the purpose of resale for profit. J Am Vet Med Assoc 2008;232:85 90.

PAGE 108

108 36 Payne RC, Veenman P, Wilson AM. The role of the extrinsic thoracic limb muscles in equine locomotion. J Ana t 2004;205:479 90. 37 Payne RC,Jutchinson JR, Robilliard JJ, et al. Functional specialization of pelvic limb anatomy in horses (Equus caballus) The role of the extrinsic thoracic limb muscles in equine locomotion. J Anat 2005;206:557 74. 3 8 Keeneland R acecourse, Lexington, KY. Available at: http://www.keeneland.com/data/sales /lists/copy/recap.aspx. Accessed November 1, 2007. 39. Keeneland Racecourse, Lexington, KY. Available at: http://www.keeneland.com/data/sales /lists/copy/conditions.aspx. Accessed November 1, 2007. 40. Dabareiner RM, Sullins KE, White NA. Progression of femoropatellar osteochondrosis dissecans in nine young horses. Vet Surg 1993; 22:515 523. 4 1 Preston SA, Brown MU, Trumble TN, et al. Epidemiologic analysis of yearling sale purchase price, exercise history, lameness, and speed as predisposing factors for sales price in 2 year old in training Thoroughbred horses. J Am Vet Med Assoc 2011;in press. 42. SamplePower 2.0, SPSS software products, Chicago, IL. 4 3. Preston SA, Brown MU, Trumble TN, et al. Prevalence of various presale radiographic findings and association of findings with sales price in Thoroughbred yearlings sold in Kentu cky J Am Vet Med Assoc 2010 ; 236:440 445 44. The Jockey Club Information Systems, Lexington, KY. Available at: www.jockeyclub/factbook.com Accessed Mar 20, 2011. 45. Hosmer DW, Lemeshow S. Assessing the fit of the model. In: Applied Logistic Regression 2nd ed. Hosmer DW, Lemeshow S, eds. Hoboken: John Wiley & Sons, 2000;143 202. 46. McIlwraith CW. General pathobiology of the joint and response to injury. In: Joint Disease in the Horse. McIlwraith CW and Trotter GW, eds. Philadelphia: W.B. Saunders Co., 1996;40 70.

PAGE 109

109 BIOGRAPHICAL SKETCH Stephanie Pres ton received a Bachelor of Science from Texas A&M University in Animal Science in 1990. After graduation she trained race horses and became a profes sional jockey. After suffering from a terrible racing accident and extensive injuries she returned to her studies at Texas A&M and recei ved a Master of Science in Veterinary Medicine and Surgery in 200 0. After finishing her m she again returned to t he horse racing industry as an owner and trainer of hundreds of Thoroughbreds over a 10 year period. Her two favorites, Flashy Bull and Ermine were everything she thought them to be and both won Grade 1 stakes races. In 2011 she expects to complete all th e requirements for her PhD from the University of Florida in Veterinary Medical Sciences. Stephanie currently resides in Lexington, KY with her son and new husband. She is actively involved in the Thoroughbred racing business and is pursuing research oppor tunities in the same area.