- Email: [email protected]
Musculoskeletal injury rates in Thoroughbred racehorses following local corticosteroid injection
The Veterinary Journal 200 (2014) 71–76
Contents lists available at ScienceDirect
The Veterinary Journal journal homepage: www.elsevier.com/locate/tvjl
Musculoskeletal injury rates in Thoroughbred racehorses following local corticosteroid injection R.C. Whitton a,⇑, M.A. Jackson a, A.J.D. Campbell a, G.A. Anderson a, T.D.H. Parkin b, J.M. Morton c, L.A. Boden b a
Faculty of Veterinary Science, University of Melbourne, 250 Princes Hwy, Werribee 3030, Australia School of Veterinary Medicine, College of Medical, Veterinary and Life Sciences, University of Glasgow, 464 Bearsden Rd, Glasgow G61 1QH, UK c Jemora Pty Ltd., P.O. Box 2277, Geelong 3220, Australia b
a r t i c l e
i n f o
Article history: Accepted 5 September 2013
Keywords: Horse Corticosteroid Musculoskeletal injury Intra-articular Racehorse
a b s t r a c t A retrospective cohort study was performed to compare the rates of musculoskeletal injury (MSI) in horses receiving local corticosteroid injection (LCI) with those that were untreated and those prior to treatment. Of the 1911 study horses, 392 had been treated. A LCI was defined as any injection of corticosteroid into or adjacent to a synovial structure, muscle, or tendon/ligament. A MSI was defined as any limb injury identified by a veterinarian, following which the horse did not race for at least 6 months, or was retired. Hazard ratios (HR) comparing hazard of injury following injection to that in non-injected horses and prior to injection were calculated using Cox proportional hazards models. At least one LCI was administered to 392 horses (20.5%; median 2, range 1–16). Most LCIs were performed bilaterally (70.9%) and intra-articularly into the carpal (49.7%) or fore fetlock (29.3%) joints. There were 219 MSIs of which carpal injuries (47%), fore fetlock (22%) and forelimb tendon injuries (16%) were the most common. The incidence rate of MSI in untreated horses and those prior to injection was 1.22 (95% CI 1.04–1.44) injuries/100 horse-months, and following LCI the hazard of MSI was greater (HR 4.83, 3.54–6.61, P < 0.001). The hazard ratio returned to levels indistinguishable from before treatment after 49 days. The hazard of MSI in horses following second and subsequent LCIs in the data collection period was greater than in horses following their first LCI (HR 2.10, 1.31–3.36, P = 0.002). There was a positive association between LCI and subsequent musculoskeletal injury rates which was most likely due to progression of the musculoskeletal condition which prompted treatment. Assuming horses that received LCI were at increased risk of MSI subsequently, any beneficial effects of the LCI were insufficient to counter this increased risk for at least 49 days after the injection. Ó 2013 Elsevier Ltd. All rights reserved.
Introduction Musculoskeletal injuries (MSI) are common in racehorses and may disrupt training, often necessitating prolonged periods of rest or, in severe cases, retirement or euthanasia (Perkins et al., 2005a). MSIs are often treated with local injections of corticosteroids (LCIs). These may alleviate lameness allowing the affected horse to continue to train and race. Many jurisdictions do not allow horses with detectable serum levels of exogenous corticosteroids to race, but these may not be measurable when bioactive concentrations are still present at the site of injection, and clinical effects are observed for prolonged periods post injection (Derendorf et al., 1986; Chen et al., 1992; Foland et al., 1994; Frisbie et al., 1997). Horses with pre-existing musculoskeletal pathology are at ⇑ Corresponding author. Tel.: +61 3 97312268. E-mail address: [email protected] (R.C. Whitton). http://dx.doi.org/10.1016/j.tvjl.2013.09.003 1090-0233/Ó 2013 Elsevier Ltd. All rights reserved.
increased risk of subsequent MSIs developing (Cohen et al., 1999), and there is the potential for such horses to race and train while under the effect of LCIs. In horses, intra-articular corticosteroids increase hyaluronate concentrations in synovial fluid, reduce lameness, synovial inflammation, and maintain cartilage morphology in joint injury models (Ronéus et al., 1993; Foland et al., 1994; Frisbie et al., 1997). There are also no detectable adverse effects of intra-articular corticosteroids on equine subchondral bone (Kawcak et al., 1998; Murray et al., 2002). Many MSIs are unique to equine athletes that habitualy train at high speed. Horses moving at speed generate large loads in their joints, tendons and suspensory apparatus (Harrison et al., 2010). Stress fractures of long bones, articular fractures, subchondral bone injuries, tendon and suspensory ligament strains are all athletic injuries. There is rarely a history of direct impact, and pre-existing pathology is often identified (Birch et al., 1998; Norrdin and Stover,
72
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76
2006; Parkin et al., 2006). Fatigue damage is not routinely recognised by veterinarians due to limitations of current diagnostic techniques (Trope et al., 2011). In horses with pain due to the accumulation of fatigue damage, it is possible that allowing continued training and racing by alleviating pain through medication potentially exposes horses to the risk of MSI, although evidence for this is lacking (McIlwraith, 2010). Alternatively pain alleviation may avoid overloading of unaffected limbs thereby reducing the risk of subsequent MSI (McIlwraith, 2010). A placebo controlled trial examining the effect of LCI on MSI rates is impractical (and possibly unethical) in racing horses. However, it is possible to quantify the rate of MSI following the administration of LCIs in horses that continued to train and race. We compared MSI rates in horses receiving LCI with horses that were untreated and those prior to treatment in order to allow owners, trainers and veterinarians to make informed management decisions. Because horses generally receive LCIs to treat a musculoskeletal problem and pre-existing pathology increases the risk of subsequent MSI, we hypothesised that LCIs would not reduce this risk to that of horses not requiring LCI. Materials and methods Study design A retrospective cohort study using veterinary and racing data was performed. Horses were convenience sampled, being selected if they were trained by trainers that used one of three Australian veterinary practices selected because they were judged to have good quality records. Records were available for the following periods: practice 1, December 1994 to March 2010; practice 2, September 2003 to January 2010; practice 3, May 2001 to October 2009. Sample size The number of MSIs required to detect a signficant difference at the 0.05 level if the hazard ratio is 1.5 with a power of 80% was estimated to be 200 using the method of Schoenfeld (1983), as implemented in PASS (NCSS) software,1 and this required the inclusion of 2000 horses assuming an incidence of MSI of 10% (Bailey et al., 1999; Perkins et al., 2005b). Study population Horses were entered into the study when they had their first official trial or race start with a trainer that solely utilised one of the three selected veterinary practices. Horses exited the study at the first of the following events: changed to a trainer not using the selected veterinary practices, raced inter-state, exported internationally, trialled or raced in a jumps race, ceased racing (not raced for >12 months, or reported as retired), suffered a MSI, or the data collection period ended. Data collection and definitions Horse and racing data All horse and racing data were obtained from the Racing Victoria Sirius database.2 These data (or dates derived from these data) included: horse gender, foaling date, trainer, date of first official trial or race start, date of exit from the study, date of last start prior to a rest period, and date of first start back from a rest period. A rest period was considered to have occurred if a horse had longer than 28 days between race starts. The date of entry into the study was 28 days prior to the horse’s first official trial or race start. Duration of rest period was calculated by subtracting the date of last start prior to a rest period from the date 28 days prior to first start back from a rest period. Rest periods were treated as ‘gaps’ in statistical analyses and so did not contribute to time at risk of MSI. The time at risk of MSI was calculated as the time in the study (exit/censor date minus entry date) less duration of rest period(s). Horses’ ages at time of each event were calculated by subtracting the foaling date (1st August) from the event date and using the integer component of the age in years. Veterinary data Data pertaining to LCI and MSI were collated from each horse’s veterinary history. A horse was excluded from the study if there were unexplained gaps in its veterinary record. 1 2
See: http://www.ncss.com/software/pass/. See: http://Sirius.racingvictoria.net.au/roarrs/frmDefault.asp.
A LCI was defined as any injection containing corticosteroid into or adjacent to a synovial structure, muscle, tendon or ligament. Data collated for LCIs included: date of injection, corticosteroid type and dose injected, site or sites injected, the reason for injection, any recorded side effects or complications of injection, and addition of hyaluronate (HA). Towards the end of the data collection period, autologous conditioned serum (ACS) became available and in some instances replaced the use of corticosteroids. Data were recorded for these treatments as for corticosteroids. A MSI was considered to be any limb injury identified by a veterinarian where the attending veterinarian recommended that the horse should not continue training and required a rest period, retirement or euthanasia, and the horse did not trial or race for at least 6 months. Following the first occurrence of a MSI during the study period, the horse was excluded from contributing further time at risk. MSI was further classified as non-catastrophic or potentially catastrophic where the treatment options were internal fixation or euthanasia. Statistical analyses Stata v 10.1 software (StataCorp) was used for all analyses. Incidence rates were compared between groups by calculating incidence rate ratios and corresponding exact confidence intervals (Rothman, 1986). Associations between LCI variables and each of MSI and potentially catastrophic MSI were examined using Cox proportional hazards models with the horse as the unit of analysis and LCI, ACS, HA and age (continuous variable in 1 year units) fitted as time-varying covariates (Cox, 1972). Gender, age and veterinary practice were forced into all models as fixed effects. Trainer was included as a random effect via the shared option of the -stcox- command. The efron method for ties was used. Schoenfeld residuals were used to test the proportional hazards assumption via the -stphtest- command. A two-tailed P-value of <0.05 was considered significant. As risk of subsequent MSI after LCI was likely to be temporary, separate models examining the relationship between LCI and MSI were used to analyse data within serial time periods of 7 days, commencing on the date of the LCI, to investigate how the hazard changed with time since injection.
Results Horses There were 1911 horses trained by 36 trainers enrolled in the study; 922 (48.2%) fillies/mares, 915 (47.9%) geldings and 74 (3.9%) entire males. Horses known not to have experienced a MSI before exiting the study were right-censored (i.e. their records were ended before they experienced a MSI) for the following reasons: 689 (36.1%) for a trainer change, 485 (25.4%) finished racing, 294 (15.4%) raced inter-state, 202 (10.6%) study ended, 15 (0.8%) exported, and 7 (0.4%) jumps trialled or raced. The mean days at risk were 206.2 (median 159, range 28–1242). The mean age at entry into the study was 2.3 years (2; 2–6) and at exit was 3.6 years (3; 2–9). Veterinary data A total of 392 horses (20.5%) received at least one LCI and local corticosteroids were injected on 858 occasions (2.2 occasions/ horse, median 2, 1–16). The majority of treatments were performed bilaterally simultaneously (608/858; 70.9%; Table 1). Of these, triamcinalone acetonide (563 occasions; 65.6%) and betamethasone acetate (265 occasions; 30.9%) were the most commonly used whereas methylprednisolone acetate (17 occasions; 2.0%) and dexamethasone (13; 1.5%) were used rarely. On 524 (61.1%) occasions, sodium hyaluronate was combined with the corticosteroid. The total amounts administered at each individual injection of triamcinolone acetonide ranged from 5 to 40 mg (mean 12.4, median 10.0), for betamethasone acetate 3 to 17 mg (mean 6.1, median 5.7), for methylprednisolone acetate 20 to 200 mg (mean 112.9, median 100) and dexamethasone 13 to 150 mg (mean 46.9, median 25). ACS was injected in 16 horses on 63 occasions. Of horses receiving LCI, 197 (50.3%) were injected on multiple occasions (mean 3.4 median 3; 2–16) (Table 1). The specific joint for carpal and tarsal intra-articular injections was not recorded in
73
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76 Table 1 Sites of local corticosteroid injection on 858 occasions in 392 horses. Structure
Total number injected (%)
Left (%)
Right (%)
Bilateral (%)
Unknown (%)
Intra-articular carpus Intra-articular fore fetlock Intra-articular fore distal interphalangeal joint Intra-articular fore proximal interphalangeal joint Intra-articular hind fetlock Intra-articular stifle Intra-articular tarsus Intra-thecal tarsal sheath Other Intra-articular site not recorded
426 (49.7) 251 (29.3) 48 (5.6) 4 (0.5) 8 (0.9) 14 (1.6) 79 (9.2) 1 (0.1) 23 (2.7) 4 (0.5)
47 (11) 36 (14.3) 2 (4.2) 0 1 (12.5) 3 (21.4) 19 (24.1) 0 5 (21.7) 0
47 (11) 43 (17.1) 1 (2.1) 2 (50.0) 7 (87.5) 6 (42.9) 9 (11.4) 1 (100.0) 6 (26.1) 0
331 (77.7) 170 (67.7) 45 (93.8) 2 (50.0) 0 5 (35.7) 51 (64.6) 0 4 (17.4) 0
1 (0.2) 2 (0.8) 0 0 0 0 0 0 8 (34.8) 4 (100.0)
858
113
122
608
15
Total
Table 2 Sites and types of musculoskeletal injury in 219 horses affected of 1911 monitored horses. Injury site Carpus
Fore fetlock
Metacarpus Pelvis Fore foot Hind fetlock
Hind pastern Radius Tendon Suspensory ligament No diagnosis
Type of injury
Number not injected
Number1 LCI
Number>1 LCI
Total (%)
Osteoarthritis Chip fracture Slab fracture ACB fracture Osteoarthritis Chip fracture Sesamoid fracture MC3 complete fracture MC3 stress fracture Pelvic fracture Pedal bone fracture Chip fracture MT3 stress fracture Sesamoid fracture P1 fracture Radial fracture Tendon strain Desmitis
9 45 4 2 2 23 3 3 1 1 3 1 1 1 1 2 29 14 1
2 15 1 0 1 4 1 2 0 0 1 0 0 0 0 0 1 2 0
3 22 0 0 0 7 0 2 0 0 0 0 0 0 0 0 5 3 1
14 (6.4) 82 (37.4) 5 (2.3) 2 (0.9) 3 (1.4) 34 (15.5) 4 (1.8) 7 (3.2) 1 (0.5) 1 (0.5) 4 (1.8) 1 (0.5) 1 (0.5) 1 (0.9) 1 (0.5) 2 (0.9) 35 (16.0) 19 (8.7) 2 (0.9)
146
30
43
219
Total
MC3, third metacarpal bone; MT3, third metatarsal bone; P1, proximal phalanx; ACB, accessory carpal bone; LCI, local corticosteroid injection.
all cases. Other injection sites included splint bones (n = 4), metacarpi (n = 9) and flexor tendons (n = 2). Of the 392 horses receiving LCI, 248 were radiographed prior to treatment. There were 219 MSIs recorded during the study. Carpal injuries (47%) were the most common followed by fore fetlock (22%) and forelimb tendon injuries (16%) (Table 2). There were 19 complete fractures, including seven third metacarpal bone, five carpal slab fractures, two accessory carpal bone, two radius fractures, one proximal phalanx, one sesamoid and one pelvic, and two stress fractures (one of the 3rd metacarpal bone and one of the 3rd metatarsal bone). In two horses the site of lameness was not definitively localised. Eleven MSIs were considered potentially catastrophic; seven were fractures of the third metacarpal bone, including four with a confirmed condylar component and one where the sesamoids and proximal phalanx were also fractured, two radius fractures, one proximal sesamoid mid-body fracture and one comminuted hind proximal phalanx fracture. Association between local corticosteroid injection and musculoskeletal injury In 143 cases of MSI, there was no history of LCI whereas 76 cases had previously received at least one LCI. Median time after the last LCI to MSI was 17.5 days (range 2–398 days) with 64 of these injuries (84.2%) occurring within 8 weeks of their most re-
cent injection. In 56 horses (73.7%), the injury occurred at a site that had been injected. In 16 horses (21.1%), the injury was in a limb injected but at a different site, and in four horses (5.3%), the injury was in a different limb to that previously injected. In four horses, the musculoskeletal condition that prompted the use of LCI was ultimately the reason for a MSI. This included three horses with carpal chip fractures and one with osteoarthritis of the midcarpal joint. In 51/76 horses that had previously received a LCI there was evidence that the condition had developed post LCI; either a fracture that was confirmed not to be present prior to treatment or the MSI developed at a site distant to the injection site. The incidence rate of MSI in horses not receiving corticosteroids was 1.22 (95% 1.04–1.44) injuries/100 horse-months (Table 3). In horses that had received LCIs, the subsequent incidence of MSI was 5.5 injuries/100 horse-months, representing an incidence rate ratio between horses receiving and not receiving LCI of 4.6 (95% CI 3.4–6.1, P < 0.001; Table 3). When the analysis was restricted to only those horses where there was evidence in the veterinary record that the MSI developed post injection; i.e. radiographs were performed prior to injection or the subsequent MSI was at a site remote from any injection site, the incidence rate ratio relative to horses not receiving LCI was 3.3 (95% CI 2.4–4.6, P < 0.001). The directly-calculated univariable incidence rate ratios and the hazard ratios estimated from the multivariable Cox model were similar (Tables 3 and 4). Using the likelihood ratio test there was an overall
74
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76
Table 3 Total numbers of musculoskeletal injury (MSI) and univariable incidence rate ratios for intra-articular treatments in 1911 horses. Variable Non-injected Any LCI 1 LCI P2 LCIs LCI with HA LCI no HA ACS injection LCI and ACS
Total horses 1511 392 195 197 277e 175e 8 8
Number of MSIs
Days at risk a
141 74b 31c 43d 41b 33b 2b 2b
352,067 40,582b 24,184c 16,680d 25,230b 15,369b 731b 982b
Incidence rate ratio (95% CI)
P-value (compared to reference group)
Reference group 4.55 (3.39–6.07) 3.20 (2.10–4.75) 6.44 (4.46–9.11) 4.06 (2.79–5.78) 5.36 (3.55–7.88) 6.83 (0.82–25.14) 5.09 (0.61–18.71)
<0.001 <0.001 <0.001 <0.001 <0.001 0.040 0.068
LCI, local corticosteroid injection; ACS, autologous conditioned serum; HA, hyaluronate. a The total days that non-injected horses were in the study plus days between entry and first LCI for horses that were injected. b From first LCI or ACS to end of study. c From first LCI to either second LCI or end of study, whichever came first. d From the second LCI to the end of the study. e HA was not consistently combined with corticosteroid for each horse.
Table 4 Hazard ratios derived from multivariable Cox proportional hazards models for musculoskeletal injury (n = 1911 horses). Variable
Hazard ratio (95% CI)
Wald P-value (compared to reference group)
Any LCI Non-injected Any LCI Female Entire male Gelding Age Practice 1 Practice 2 Practice 3
Reference group 4.83 (3.54–6.61) Reference group 1.31 (0.63–2.72) 0.70 (0.53–0.93) 1.33 (1.10–1.66)a Reference group 1.68 (0.83–3.39) 1.71 (0.93–3.12)
Multiple LCIb Non-injected 1 LCI P2 LCIs
Reference group 3.38 (2.25–5.08) 7.08 (4.84–10.36)
<0.001 <0.001
ACS injectionsb Non-injected ACS only ACS and LCI
Reference group 3.74 (0.86–16.24) 2.91 (0.68–12.53)
0.078 0.15
LCI with/without HAb Non-injected LCI no HA LCI with HA
Reference group 5.12 (3.38–7.77) 4.54 (3.10–6.64)
<0.001 <0.001
<0.001 0.47 0.013 0.012 0.15 0.084
LCI, local corticosteroid injection; ACS, autologous conditioned serum; HA, hyaluronate. a Hazard ratio for each extra year of age. b horse sex and age, veterinary practice were included in each model but results only shown for any LCI.
effect of sex (P = 0.024) on MSI, with geldings having a hazard of 0.70 (P = 0.013) to that of females. Older horses experienced more injuries than younger horses, with the hazard of MSI increasing 1.3 times per additional year of age (P = 0.012). There was no difference between veterinary practices in rates of MSI (P = 0.23). There was marginal clustering of injury rates at the trainer level (shared frailty effect of trainer represented by variance theta = 0.057, P = 0.055). When the analysis was restricted to only those horses with evidence that the MSI developed post injection the hazard ratios for any LCI (HR 3.36, 95% CI 2.37–4.76, P < 0.001) and for single (HR 2.36, 95% CI 1.47–3.77, P < 0.001) and multiple LCI (HR 4.82, 95% CI 3.14–7.42, P < 0.001) remained significant. On restricting the time analysed to serial 7 day intervals post injection, the hazard of injury was approximately 6–12.5 times that of the untreated horses and those prior to treatment up to 49 days post LCI, but returned to levels indistinguishable from untreated horses thereafter (Fig. 1). The assumption of proportional
hazards was satisfied within all except the 1–7 and 8–14 day time periods. The hazard of MSI in horses after they received two or more LCIs in the data collection period was approximately twice the hazard in horses after receiving one injection (HR 2.1, 95% CI 1.3–3.4, P = 0.002). When hyaluronate was combined with corticosteroid, there was no change in the hazard of subsequent MSI compared with corticosteroid alone (P = 0.63). Musculoskeletal injury rates were highest following intra-articular injection of the carpus (HR 5.59, 95% CI 3.85–8.12, P < 0.001, reference = non-injected time at risk) and fetlock (HR 5.68, 95% CI 3.55–9.10, P < 0.001) and lower for all other sites injected (HR 2.40, 95% CI 1.20–4.82, P = 0.014). The rate of injury was higher following betamethasone than following triamcinolone injection (HR 1.91, 95% CI 1.16–3.13, P = 0.011). Of the 11 potentially catastrophic injuries, seven were in horses that did not receive local corticosteroids and four were in treated horses. None of these horses were radiographed prior to the injury. One horse sustained a compound fracture of the lower limb 20 days following its third bilateral carpal LCI, another a condylar fracture 20 days after its fifth bilateral fetlock LCI, one a fractured proximal phalanx, proximal sesamoid and distal metacarpal fracture 78 days after a single bilateral LCI of the fetlocks and a fourth a condylar fracture 21 days after a single bilateral fetlock LCI. Hazard of potentially catastrophic injuries following second and subsequent LCIs was higher than in horses that were not injected and those prior to injection (HR 6.9, 95% CI 1.2–39.8, P = 0.031) but this was not the case following one LCI (HR 2.5, 95% CI 0.44–13.5, P = 0.30). Horses receiving intra-articular ACS subsequently developed MSI at an increased hazard but the hazard ratio was not significant (P = 0.078); this estimate was imprecise due to the small number of these cases (Table 4). Discussion As expected, this study demonstrated that Thoroughbred racehorses that receive LCIs for the treatment of orthopaedic conditions and continue training have a greater subsequent incidence of MSI than horses that do not receive such treatment. Multiple LCIs and those involving the carpus and fetlock joints resulted in the highest injury rates. Progression of the musculoskeletal problem that prompted treatment appears to be an important contributor to this result and any beneficial effects of LCI are insufficient to counter the increased rate of MSI up to at least 49 days post injection. The aim of this study was not to determine whether the use of local corticosteroids was detrimental, as this was not possible from the study design, rather to quantify the association between the
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76
Fig. 1. Hazard ratios (solid line) and point-wise 95% confidence intervals (dashed lines) for musculoskeletal injury in horses after receiving local corticosteroid injections, compared to non-injected horses, estimated separately for sequential 7 day periods post local corticosteroid injection. After 49 days the hazard ratio was not significantly different from one.
use of LCI and subsequent musculoskeletal injury rates to allow veterinarians, owners and trainers to be better informed when deciding whether to treat a problem and continue to train and race a horse. It was not practical to compare LCIs with no treatment in horses with an identified limb problem, so the independent effect of LCI was not assessed. The greater hazard of injury identified following LCI is more likely due to veterinary examination identifying horses at risk of MSI rather than a direct adverse effect of the corticosteroid. Experimental evidence indicates that the two most commonly used corticosteroids in this study have beneficial effects when used locally to treat induced intra-articular pathology in horses that continue in training, and no detrimental effect of triamcinolone on equine subchondral bone has been demonstrated (Foland et al., 1994; Frisbie et al., 1997; Kawcak et al., 1998; Murray et al., 2002). In the current study, a minority of horses received LCI and there was generally a clear indication for treatment, which suggested that the prophylactic use of local corticosteroids was uncommon and that treated horses were likely to have a pre-existing orthopaedic condition. An increase in the subsequent risk of developing a MSI after injection in these horses was therefore not surprising. The increased hazard of MSI associated with LCI was time dependant. It is likely that horses with underlying problems either develop a MSI within 7 weeks of injection or require repeat treatments at 4–8 weekly intervals to remain in training and therefore return to the ‘less than 49 days since local injection’ groups. Those horses with minor or short term orthopaedic problems can remain in training without the need for repeat treatments beyond 49 days. The majority of MSIs were due to conditions where fatigue of tissues is an important part of the pathogenesis (Stover et al., 1993; Birch et al., 2008; Whitton et al., 2010). Consistent with this, exercise duration and intensity have been shown to be risk factors for MSIs in Thoroughbred racehorses (Estberg et al., 1998; Perkins et al., 2005b; Verheyen et al., 2006). Corticosteroids are effective for the treatment of inflammation and pain but have no ability to prevent fatigue of musculoskeletal tissues if loading continues. If the LCI contributes to allowing a horse that may be developing a MSI to continue to train and race either by its anti-inflammatory effects or by influencing the decision making of the trainer, this will keep the horse at risk of MSI.
75
Assessment of effects of LCI on musculoskeletal injury rates would require a placebo controlled trial. A number of injuries, including catastrophic fractures, occurred in horses that had minimal diagnostic workup prior to LCI. Detection of fatigue prior to fracture is challenging even with sophisticated diagnostic imaging but every effort should be made to identify incomplete fractures (Trope et al., 2011). Many horses receiving local corticosteroids do not develop MSI, and improved diagnostic methods may allow at-risk horses to be identified (Powell, 2012). Although this study was not designed to assess effects based on the expected frequency of catastrophic injury, an effect of multiple LCI (two or more sequential treatments) on the rate of potentially catastrophic injury was observed although the wide confidence interval makes it difficult to determine the magnitude of this effect. The pathogenesis of catastrophic injuries is similar to that of less severe injuries; they are associated with bone, tendon or ligament failure under the repeated loading induced by high speed exercise (Stover et al., 1993; Stepnik et al., 2004). In subchondral bone, the resultant damage is usually localised, but in a minority of horses localised damage may propagate into a complete fracture (Riggs, 1999; Norrdin and Stover, 2006). We found no evidence that adding HA decreased MSI rates following local medication. HA is commonly combined with a corticosteroid to alleviate some of its negative effects, although this is based on little evidence (McIlwraith, 2010). The combination of these medications is recommended in human patients because the longer duration but slower onset of action of HA combines well with the shorter term effects of the corticosteroid (Bannuru et al., 2009). Our failure to demonstrate any effect on MSI rates is consistent with our proposal that it is the underlying musculoskeletal condition that is the major contributor to the increased injury rate following LCI, and does not mean that HA is of no benefit to the synovial structure. Not all MSIs following LCI occurred at the site of injection. Possible reasons for a MSI at a distant site from injection within the same limb include a subsequent injury unrelated to the local injections, incorrect localisation of the original source of pain, or a remote effect of LCI (Frisbie et al., 1997). Possible reasons for a MSI in a different limb to that treated, include overload of the contralateral limb due to inadequate pain relief, unidentified bilateral pain, an unrelated injury or a remote effect of LCI. Because we could not be sure of the accuracy of the pre-injection diagnoses, and it is clear that the effect of continuing to train and race a horse is not localised to the site of injection, we included all injection sites and all injuries in the study. Limitations of this study include its retrospective design and dependence on the accuracy of records kept by veterinarians. We used practices that had good computer-based records and, where there were obvious deficiencies, we excluded the relevant horse. Not all joints were radiographed prior to treatment so we cannot be certain that an intra-articular fracture was not already present. However, in such cases there was clear deterioration of clinical signs consistent with progression of the injury and when these horses were excluded all results remained statistically significant. Also, it is possible that some horses developed MSIs and were spelled without veterinary examination, but to minimise this occurrence we selected trainers who worked closely with veterinarians and sought frequent veterinary advice. The length of rest periods used to define MSI was greater than the routine rest period horses receive without injury so it is likely that our case definition was specific and trainers were likely to obtain a diagnosis and prognosis where prolonged rest periods were required. Our study population had a relatively low proportion of local corticosteroid treatment and care should be taken in extrapolating these results to populations where the local corticosteroid treatment proportion is much higher. The failure to meet the proportional hazards
76
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76
assumption with some models was expected with the use of time varying covariates (Allison, 2010). If the presence of the original condition that prompts the use of LCI is a major contributor to the subsequent rate of MSI, new medications suitable for local treatment are unlikely to reduce this outcome. This is because any medication that allows a horse with an underlying problem to continue to train and race has the potential to indirectly increase injury rates unless that medication completely removes risk due to the pre-existing condition. Consistent with this, ACS, corticosteroids and HA/corticosteroid combination appeared to have similar associations with MSI. Conclusions Treatment of horses with musculoskeletal conditions by LCI does not return injury rates to levels observed in horses not under the influence of treatment. An otherwise identical group that did not receive LCI was not available so effects of LCI were not assessed. Further investigation should be directed at better identifying horses at risk of MSI following LCI. This information should be included in the decision making process when managing horses with musculoskeletal problems. Conflict of interest statement None of the authors has any financial or personal relationships that could inappropriately influence or bias the content of the paper. Acknowledgements We thank the veterinary practices that provided access to their veterinary records. This study was presented in part at the 50th British Equine Veterinary Association Congress, Liverpool, UK, 2011. References Allison, P.D., 2010. Survival Analysis using SAS. A Practical Guide, Second Ed. SAS Institute, Cary, NC, USA. Bailey, C.J., Reid, S.W., Hodgson, D.R., Rose, R.J., 1999. Impact of injuries and disease on a cohort of two- and three-year-old Thoroughbreds in training. Veterinary Record 145, 487–493. Bannuru, R.R., Natov, N.S., Obadan, I.E., Price, L.L., Schmid, C.H., McAlindon, T.E., 2009. Therapeutic trajectory of hyaluronic acid versus corticosteroids in the treatment of knee osteoarthritis: A systematic review and meta-analysis. Arthritis and Rheumatism 61, 1704–1711. Birch, H.L., Bailey, A.J., Goodship, A.E., 1998. Macroscopic ‘degeneration’ of equine superficial digital flexor tendon is accompanied by a change in extracellular matrix composition. Equine Veterinary Journal 30, 534–539. Birch, H.L., Wilson, A.M., Goodship, A.E., 2008. Physical activity: Does long-term, high-intensity exercise in horses result in tendon degeneration? Journal of Applied Physiology 105, 1927–1933. Chen, C.L., Sailor, J.A., Collier, J., Wiegand, J., 1992. Synovial and serum levels of triamcinolone following intra-articular administration of triamcinolone acetonide in the horse. Journal of Veterinary Pharmacology and Therapeutics 15, 240–246. Cohen, N.D., Mundy, G.D., Peloso, J.G., Carey, V.J., Amend, N.K., 1999. Results of physical inspection before races and race-related characteristics and their association with musculoskeletal injuries in Thoroughbreds during races. Journal of the American Veterinary Medical Association 215, 654–661. Cox, D., 1972. Regression models and life-tables (with discussion). Journal of the Royal Statistical Society B 34, 187–220.
Derendorf, H., Mollmann, H., Gruner, A., Haack, D., Gyselby, G., 1986. Pharmacokinetics and pharmacodynamics of glucocorticoid suspensions after intra-articular administration. Clinical Pharmacology and Therapeutics 39, 313–317. Estberg, L., Gardner, I.A., Stover, S.M., Johnson, B.J., 1998. A case-crossover study of intensive racing and training schedules and risk of catastrophic musculoskeletal injury and lay-up in California Thoroughbred racehorses. Preventive Veterinary Medicine 33, 159–170. Foland, J.W., McIlwraith, C.W., Trotter, G.W., Powers, B.E., Lamar, C.H., 1994. Effect of betamathasone and exercise on equine carpal joints with osteochondral fragments. Veterinary Surgery 23, 369–376. Frisbie, D.D., Kawcak, C.E., Trotter, G.W., Powers, B.E., Walton, R.M., McIlwraith, C.W., 1997. Effects of triamcinolone acetonide on an in vivo equine osteochondral fragment exercise model. Equine Veterinary Journal 29, 349– 359. Harrison, S.M., Whitton, R.C., Kawcak, C.E., Stover, S.M., Pandy, M.G., 2010. Relationship between muscle forces, joint loading and utilization of elastic strain energy in equine locomotion. Journal of Experimental Biology 213, 3998– 4009. Kawcak, C.E., Norrdin, R.W., Frisbie, D.D., Trotter, G.W., McIlwraith, C.W., 1998. Effects of osteochondral fragmentation and intra-articular triamcinolone acetonide treatment on subchondral bone in the equine carpus. Equine Veterinary Journal 30, 66–71. McIlwraith, C.W., 2010. The use of intra-articular corticosteroids in the horse: What is known on a scientific basis? Equine Veterinary Journal 42, 563– 571. Murray, R.C., Znaor, N., Tanner, K.E., DeBowes, R.M., Gaughan, E.M., Goodship, A.E., 2002. The effect of intra-articular methylprednisolone acetate and exercise on equine carpal subchondral and cancellous bone microhardness. Equine Veterinary Journal 34, 306–310. Norrdin, R.W., Stover, S.M., 2006. Subchondral bone failure in overload arthrosis: A scanning electron microscopic study in horses. Journal of Musculoskeletal and Neuronal Interactions 6, 251–257. Parkin, T.D., Clegg, P.D., French, N.P., Proudman, C.J., Riggs, C.M., Singer, E.R., Webbon, P.M., Morgan, K.L., 2006. Catastrophic fracture of the lateral condyle of the third metacarpus/metatarsus in UK racehorses – Fracture descriptions and pre-existing pathology. The Veterinary Journal 171, 157–165. Perkins, N.R., Reid, S.W., Morris, R.S., 2005a. Profiling the New Zealand Thoroughbred racing industry. 2. Conditions interfering with training and racing. New Zealand Veterinary Journal 53, 69–76. Perkins, N.R., Reid, S.W., Morris, R.S., 2005b. Risk factors for musculoskeletal injuries of the lower limbs in Thoroughbred racehorses in New Zealand. New Zealand Veterinary Journal 53, 171–183. Powell, S.E., 2012. Low-field standing magnetic resonance imaging findings of the metacarpo/metatarsophalangeal joint of racing Thoroughbreds with lameness localised to the region: A retrospective study of 131 horses. Equine Veterinary Journal 44, 169–177. Riggs, C.M., 1999. Aetiopathogenesis of parasagittal fractures of the distal condyles of the third metacarpal and third metatarsal bones: A review of the literature. Equine Veterinary Journal 31, 116–120. Ronéus, B., Lindblad, A., Lindholm, A., Jones, B., 1993. Effects of intraarticular corticosteroid and sodium hyaluronate injections on synovial fluid production and synovial fluid content of sodium hyaluronate and proteoglycans in normal equine joints. Journal of Veterinary Medicine 40, 10–16. Rothman, K.J., 1986. Modern Epidemiology. Little, Brown and Co., Boston. Schoenfeld, D., 1983. Sample-size formula for the proportional-hazards regression model. Biometrics 39, 499–503. Stepnik, M.W., Radtke, C.L., Scollay, M.C., Oshel, P.E., Albrecht, R.M., Santschi, E.M., Markel, M.D., Muir, P., 2004. Scanning electron microscopic examination of third metacarpal/third metatarsal bone failure surfaces in Thoroughbred racehorses with condylar fracture. Veterinary Surgery 33, 2–10. Stover, S.M., Ardans, A.A., Read, D.H., Johnson, B.J., Barr, B.C., Daft, B.M., Kindu, H., Anderson, M.L., Woods, L.W., Moore, J., et al., 1993. Patterns of stress fractures associated with complete bone fractures in racehorses. Proceedings of the American Association of Equine Practitioners 39, 131–132. Trope, G.D., Anderson, G.A., Whitton, R.C., 2011. Patterns of scintigraphic uptake in the fetlock joint of Thoroughbred racehorses and the effect of increased radiopharmaceutical uptake in the distal metacarpal/tarsal condyle on performance. Equine Veterinary Journal 43, 509–515. Verheyen, K., Price, J., Lanyon, L., Wood, J., 2006. Exercise distance and speed affect the risk of fracture in racehorses. Bone 39, 1322–1330. Whitton, R.C., Trope, G.D., Ghasem-Zadeh, A., Anderson, G.A., Parkin, T.D., Mackie, E.J., Seeman, E., 2010. Third metacarpal condylar fatigue fractures in equine athletes occur within previously modelled subchondral bone. Bone 47, 826– 831.
Contents lists available at ScienceDirect
The Veterinary Journal journal homepage: www.elsevier.com/locate/tvjl
Musculoskeletal injury rates in Thoroughbred racehorses following local corticosteroid injection R.C. Whitton a,⇑, M.A. Jackson a, A.J.D. Campbell a, G.A. Anderson a, T.D.H. Parkin b, J.M. Morton c, L.A. Boden b a
Faculty of Veterinary Science, University of Melbourne, 250 Princes Hwy, Werribee 3030, Australia School of Veterinary Medicine, College of Medical, Veterinary and Life Sciences, University of Glasgow, 464 Bearsden Rd, Glasgow G61 1QH, UK c Jemora Pty Ltd., P.O. Box 2277, Geelong 3220, Australia b
a r t i c l e
i n f o
Article history: Accepted 5 September 2013
Keywords: Horse Corticosteroid Musculoskeletal injury Intra-articular Racehorse
a b s t r a c t A retrospective cohort study was performed to compare the rates of musculoskeletal injury (MSI) in horses receiving local corticosteroid injection (LCI) with those that were untreated and those prior to treatment. Of the 1911 study horses, 392 had been treated. A LCI was defined as any injection of corticosteroid into or adjacent to a synovial structure, muscle, or tendon/ligament. A MSI was defined as any limb injury identified by a veterinarian, following which the horse did not race for at least 6 months, or was retired. Hazard ratios (HR) comparing hazard of injury following injection to that in non-injected horses and prior to injection were calculated using Cox proportional hazards models. At least one LCI was administered to 392 horses (20.5%; median 2, range 1–16). Most LCIs were performed bilaterally (70.9%) and intra-articularly into the carpal (49.7%) or fore fetlock (29.3%) joints. There were 219 MSIs of which carpal injuries (47%), fore fetlock (22%) and forelimb tendon injuries (16%) were the most common. The incidence rate of MSI in untreated horses and those prior to injection was 1.22 (95% CI 1.04–1.44) injuries/100 horse-months, and following LCI the hazard of MSI was greater (HR 4.83, 3.54–6.61, P < 0.001). The hazard ratio returned to levels indistinguishable from before treatment after 49 days. The hazard of MSI in horses following second and subsequent LCIs in the data collection period was greater than in horses following their first LCI (HR 2.10, 1.31–3.36, P = 0.002). There was a positive association between LCI and subsequent musculoskeletal injury rates which was most likely due to progression of the musculoskeletal condition which prompted treatment. Assuming horses that received LCI were at increased risk of MSI subsequently, any beneficial effects of the LCI were insufficient to counter this increased risk for at least 49 days after the injection. Ó 2013 Elsevier Ltd. All rights reserved.
Introduction Musculoskeletal injuries (MSI) are common in racehorses and may disrupt training, often necessitating prolonged periods of rest or, in severe cases, retirement or euthanasia (Perkins et al., 2005a). MSIs are often treated with local injections of corticosteroids (LCIs). These may alleviate lameness allowing the affected horse to continue to train and race. Many jurisdictions do not allow horses with detectable serum levels of exogenous corticosteroids to race, but these may not be measurable when bioactive concentrations are still present at the site of injection, and clinical effects are observed for prolonged periods post injection (Derendorf et al., 1986; Chen et al., 1992; Foland et al., 1994; Frisbie et al., 1997). Horses with pre-existing musculoskeletal pathology are at ⇑ Corresponding author. Tel.: +61 3 97312268. E-mail address: [email protected] (R.C. Whitton). http://dx.doi.org/10.1016/j.tvjl.2013.09.003 1090-0233/Ó 2013 Elsevier Ltd. All rights reserved.
increased risk of subsequent MSIs developing (Cohen et al., 1999), and there is the potential for such horses to race and train while under the effect of LCIs. In horses, intra-articular corticosteroids increase hyaluronate concentrations in synovial fluid, reduce lameness, synovial inflammation, and maintain cartilage morphology in joint injury models (Ronéus et al., 1993; Foland et al., 1994; Frisbie et al., 1997). There are also no detectable adverse effects of intra-articular corticosteroids on equine subchondral bone (Kawcak et al., 1998; Murray et al., 2002). Many MSIs are unique to equine athletes that habitualy train at high speed. Horses moving at speed generate large loads in their joints, tendons and suspensory apparatus (Harrison et al., 2010). Stress fractures of long bones, articular fractures, subchondral bone injuries, tendon and suspensory ligament strains are all athletic injuries. There is rarely a history of direct impact, and pre-existing pathology is often identified (Birch et al., 1998; Norrdin and Stover,
72
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76
2006; Parkin et al., 2006). Fatigue damage is not routinely recognised by veterinarians due to limitations of current diagnostic techniques (Trope et al., 2011). In horses with pain due to the accumulation of fatigue damage, it is possible that allowing continued training and racing by alleviating pain through medication potentially exposes horses to the risk of MSI, although evidence for this is lacking (McIlwraith, 2010). Alternatively pain alleviation may avoid overloading of unaffected limbs thereby reducing the risk of subsequent MSI (McIlwraith, 2010). A placebo controlled trial examining the effect of LCI on MSI rates is impractical (and possibly unethical) in racing horses. However, it is possible to quantify the rate of MSI following the administration of LCIs in horses that continued to train and race. We compared MSI rates in horses receiving LCI with horses that were untreated and those prior to treatment in order to allow owners, trainers and veterinarians to make informed management decisions. Because horses generally receive LCIs to treat a musculoskeletal problem and pre-existing pathology increases the risk of subsequent MSI, we hypothesised that LCIs would not reduce this risk to that of horses not requiring LCI. Materials and methods Study design A retrospective cohort study using veterinary and racing data was performed. Horses were convenience sampled, being selected if they were trained by trainers that used one of three Australian veterinary practices selected because they were judged to have good quality records. Records were available for the following periods: practice 1, December 1994 to March 2010; practice 2, September 2003 to January 2010; practice 3, May 2001 to October 2009. Sample size The number of MSIs required to detect a signficant difference at the 0.05 level if the hazard ratio is 1.5 with a power of 80% was estimated to be 200 using the method of Schoenfeld (1983), as implemented in PASS (NCSS) software,1 and this required the inclusion of 2000 horses assuming an incidence of MSI of 10% (Bailey et al., 1999; Perkins et al., 2005b). Study population Horses were entered into the study when they had their first official trial or race start with a trainer that solely utilised one of the three selected veterinary practices. Horses exited the study at the first of the following events: changed to a trainer not using the selected veterinary practices, raced inter-state, exported internationally, trialled or raced in a jumps race, ceased racing (not raced for >12 months, or reported as retired), suffered a MSI, or the data collection period ended. Data collection and definitions Horse and racing data All horse and racing data were obtained from the Racing Victoria Sirius database.2 These data (or dates derived from these data) included: horse gender, foaling date, trainer, date of first official trial or race start, date of exit from the study, date of last start prior to a rest period, and date of first start back from a rest period. A rest period was considered to have occurred if a horse had longer than 28 days between race starts. The date of entry into the study was 28 days prior to the horse’s first official trial or race start. Duration of rest period was calculated by subtracting the date of last start prior to a rest period from the date 28 days prior to first start back from a rest period. Rest periods were treated as ‘gaps’ in statistical analyses and so did not contribute to time at risk of MSI. The time at risk of MSI was calculated as the time in the study (exit/censor date minus entry date) less duration of rest period(s). Horses’ ages at time of each event were calculated by subtracting the foaling date (1st August) from the event date and using the integer component of the age in years. Veterinary data Data pertaining to LCI and MSI were collated from each horse’s veterinary history. A horse was excluded from the study if there were unexplained gaps in its veterinary record. 1 2
See: http://www.ncss.com/software/pass/. See: http://Sirius.racingvictoria.net.au/roarrs/frmDefault.asp.
A LCI was defined as any injection containing corticosteroid into or adjacent to a synovial structure, muscle, tendon or ligament. Data collated for LCIs included: date of injection, corticosteroid type and dose injected, site or sites injected, the reason for injection, any recorded side effects or complications of injection, and addition of hyaluronate (HA). Towards the end of the data collection period, autologous conditioned serum (ACS) became available and in some instances replaced the use of corticosteroids. Data were recorded for these treatments as for corticosteroids. A MSI was considered to be any limb injury identified by a veterinarian where the attending veterinarian recommended that the horse should not continue training and required a rest period, retirement or euthanasia, and the horse did not trial or race for at least 6 months. Following the first occurrence of a MSI during the study period, the horse was excluded from contributing further time at risk. MSI was further classified as non-catastrophic or potentially catastrophic where the treatment options were internal fixation or euthanasia. Statistical analyses Stata v 10.1 software (StataCorp) was used for all analyses. Incidence rates were compared between groups by calculating incidence rate ratios and corresponding exact confidence intervals (Rothman, 1986). Associations between LCI variables and each of MSI and potentially catastrophic MSI were examined using Cox proportional hazards models with the horse as the unit of analysis and LCI, ACS, HA and age (continuous variable in 1 year units) fitted as time-varying covariates (Cox, 1972). Gender, age and veterinary practice were forced into all models as fixed effects. Trainer was included as a random effect via the shared option of the -stcox- command. The efron method for ties was used. Schoenfeld residuals were used to test the proportional hazards assumption via the -stphtest- command. A two-tailed P-value of <0.05 was considered significant. As risk of subsequent MSI after LCI was likely to be temporary, separate models examining the relationship between LCI and MSI were used to analyse data within serial time periods of 7 days, commencing on the date of the LCI, to investigate how the hazard changed with time since injection.
Results Horses There were 1911 horses trained by 36 trainers enrolled in the study; 922 (48.2%) fillies/mares, 915 (47.9%) geldings and 74 (3.9%) entire males. Horses known not to have experienced a MSI before exiting the study were right-censored (i.e. their records were ended before they experienced a MSI) for the following reasons: 689 (36.1%) for a trainer change, 485 (25.4%) finished racing, 294 (15.4%) raced inter-state, 202 (10.6%) study ended, 15 (0.8%) exported, and 7 (0.4%) jumps trialled or raced. The mean days at risk were 206.2 (median 159, range 28–1242). The mean age at entry into the study was 2.3 years (2; 2–6) and at exit was 3.6 years (3; 2–9). Veterinary data A total of 392 horses (20.5%) received at least one LCI and local corticosteroids were injected on 858 occasions (2.2 occasions/ horse, median 2, 1–16). The majority of treatments were performed bilaterally simultaneously (608/858; 70.9%; Table 1). Of these, triamcinalone acetonide (563 occasions; 65.6%) and betamethasone acetate (265 occasions; 30.9%) were the most commonly used whereas methylprednisolone acetate (17 occasions; 2.0%) and dexamethasone (13; 1.5%) were used rarely. On 524 (61.1%) occasions, sodium hyaluronate was combined with the corticosteroid. The total amounts administered at each individual injection of triamcinolone acetonide ranged from 5 to 40 mg (mean 12.4, median 10.0), for betamethasone acetate 3 to 17 mg (mean 6.1, median 5.7), for methylprednisolone acetate 20 to 200 mg (mean 112.9, median 100) and dexamethasone 13 to 150 mg (mean 46.9, median 25). ACS was injected in 16 horses on 63 occasions. Of horses receiving LCI, 197 (50.3%) were injected on multiple occasions (mean 3.4 median 3; 2–16) (Table 1). The specific joint for carpal and tarsal intra-articular injections was not recorded in
73
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76 Table 1 Sites of local corticosteroid injection on 858 occasions in 392 horses. Structure
Total number injected (%)
Left (%)
Right (%)
Bilateral (%)
Unknown (%)
Intra-articular carpus Intra-articular fore fetlock Intra-articular fore distal interphalangeal joint Intra-articular fore proximal interphalangeal joint Intra-articular hind fetlock Intra-articular stifle Intra-articular tarsus Intra-thecal tarsal sheath Other Intra-articular site not recorded
426 (49.7) 251 (29.3) 48 (5.6) 4 (0.5) 8 (0.9) 14 (1.6) 79 (9.2) 1 (0.1) 23 (2.7) 4 (0.5)
47 (11) 36 (14.3) 2 (4.2) 0 1 (12.5) 3 (21.4) 19 (24.1) 0 5 (21.7) 0
47 (11) 43 (17.1) 1 (2.1) 2 (50.0) 7 (87.5) 6 (42.9) 9 (11.4) 1 (100.0) 6 (26.1) 0
331 (77.7) 170 (67.7) 45 (93.8) 2 (50.0) 0 5 (35.7) 51 (64.6) 0 4 (17.4) 0
1 (0.2) 2 (0.8) 0 0 0 0 0 0 8 (34.8) 4 (100.0)
858
113
122
608
15
Total
Table 2 Sites and types of musculoskeletal injury in 219 horses affected of 1911 monitored horses. Injury site Carpus
Fore fetlock
Metacarpus Pelvis Fore foot Hind fetlock
Hind pastern Radius Tendon Suspensory ligament No diagnosis
Type of injury
Number not injected
Number1 LCI
Number>1 LCI
Total (%)
Osteoarthritis Chip fracture Slab fracture ACB fracture Osteoarthritis Chip fracture Sesamoid fracture MC3 complete fracture MC3 stress fracture Pelvic fracture Pedal bone fracture Chip fracture MT3 stress fracture Sesamoid fracture P1 fracture Radial fracture Tendon strain Desmitis
9 45 4 2 2 23 3 3 1 1 3 1 1 1 1 2 29 14 1
2 15 1 0 1 4 1 2 0 0 1 0 0 0 0 0 1 2 0
3 22 0 0 0 7 0 2 0 0 0 0 0 0 0 0 5 3 1
14 (6.4) 82 (37.4) 5 (2.3) 2 (0.9) 3 (1.4) 34 (15.5) 4 (1.8) 7 (3.2) 1 (0.5) 1 (0.5) 4 (1.8) 1 (0.5) 1 (0.5) 1 (0.9) 1 (0.5) 2 (0.9) 35 (16.0) 19 (8.7) 2 (0.9)
146
30
43
219
Total
MC3, third metacarpal bone; MT3, third metatarsal bone; P1, proximal phalanx; ACB, accessory carpal bone; LCI, local corticosteroid injection.
all cases. Other injection sites included splint bones (n = 4), metacarpi (n = 9) and flexor tendons (n = 2). Of the 392 horses receiving LCI, 248 were radiographed prior to treatment. There were 219 MSIs recorded during the study. Carpal injuries (47%) were the most common followed by fore fetlock (22%) and forelimb tendon injuries (16%) (Table 2). There were 19 complete fractures, including seven third metacarpal bone, five carpal slab fractures, two accessory carpal bone, two radius fractures, one proximal phalanx, one sesamoid and one pelvic, and two stress fractures (one of the 3rd metacarpal bone and one of the 3rd metatarsal bone). In two horses the site of lameness was not definitively localised. Eleven MSIs were considered potentially catastrophic; seven were fractures of the third metacarpal bone, including four with a confirmed condylar component and one where the sesamoids and proximal phalanx were also fractured, two radius fractures, one proximal sesamoid mid-body fracture and one comminuted hind proximal phalanx fracture. Association between local corticosteroid injection and musculoskeletal injury In 143 cases of MSI, there was no history of LCI whereas 76 cases had previously received at least one LCI. Median time after the last LCI to MSI was 17.5 days (range 2–398 days) with 64 of these injuries (84.2%) occurring within 8 weeks of their most re-
cent injection. In 56 horses (73.7%), the injury occurred at a site that had been injected. In 16 horses (21.1%), the injury was in a limb injected but at a different site, and in four horses (5.3%), the injury was in a different limb to that previously injected. In four horses, the musculoskeletal condition that prompted the use of LCI was ultimately the reason for a MSI. This included three horses with carpal chip fractures and one with osteoarthritis of the midcarpal joint. In 51/76 horses that had previously received a LCI there was evidence that the condition had developed post LCI; either a fracture that was confirmed not to be present prior to treatment or the MSI developed at a site distant to the injection site. The incidence rate of MSI in horses not receiving corticosteroids was 1.22 (95% 1.04–1.44) injuries/100 horse-months (Table 3). In horses that had received LCIs, the subsequent incidence of MSI was 5.5 injuries/100 horse-months, representing an incidence rate ratio between horses receiving and not receiving LCI of 4.6 (95% CI 3.4–6.1, P < 0.001; Table 3). When the analysis was restricted to only those horses where there was evidence in the veterinary record that the MSI developed post injection; i.e. radiographs were performed prior to injection or the subsequent MSI was at a site remote from any injection site, the incidence rate ratio relative to horses not receiving LCI was 3.3 (95% CI 2.4–4.6, P < 0.001). The directly-calculated univariable incidence rate ratios and the hazard ratios estimated from the multivariable Cox model were similar (Tables 3 and 4). Using the likelihood ratio test there was an overall
74
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76
Table 3 Total numbers of musculoskeletal injury (MSI) and univariable incidence rate ratios for intra-articular treatments in 1911 horses. Variable Non-injected Any LCI 1 LCI P2 LCIs LCI with HA LCI no HA ACS injection LCI and ACS
Total horses 1511 392 195 197 277e 175e 8 8
Number of MSIs
Days at risk a
141 74b 31c 43d 41b 33b 2b 2b
352,067 40,582b 24,184c 16,680d 25,230b 15,369b 731b 982b
Incidence rate ratio (95% CI)
P-value (compared to reference group)
Reference group 4.55 (3.39–6.07) 3.20 (2.10–4.75) 6.44 (4.46–9.11) 4.06 (2.79–5.78) 5.36 (3.55–7.88) 6.83 (0.82–25.14) 5.09 (0.61–18.71)
<0.001 <0.001 <0.001 <0.001 <0.001 0.040 0.068
LCI, local corticosteroid injection; ACS, autologous conditioned serum; HA, hyaluronate. a The total days that non-injected horses were in the study plus days between entry and first LCI for horses that were injected. b From first LCI or ACS to end of study. c From first LCI to either second LCI or end of study, whichever came first. d From the second LCI to the end of the study. e HA was not consistently combined with corticosteroid for each horse.
Table 4 Hazard ratios derived from multivariable Cox proportional hazards models for musculoskeletal injury (n = 1911 horses). Variable
Hazard ratio (95% CI)
Wald P-value (compared to reference group)
Any LCI Non-injected Any LCI Female Entire male Gelding Age Practice 1 Practice 2 Practice 3
Reference group 4.83 (3.54–6.61) Reference group 1.31 (0.63–2.72) 0.70 (0.53–0.93) 1.33 (1.10–1.66)a Reference group 1.68 (0.83–3.39) 1.71 (0.93–3.12)
Multiple LCIb Non-injected 1 LCI P2 LCIs
Reference group 3.38 (2.25–5.08) 7.08 (4.84–10.36)
<0.001 <0.001
ACS injectionsb Non-injected ACS only ACS and LCI
Reference group 3.74 (0.86–16.24) 2.91 (0.68–12.53)
0.078 0.15
LCI with/without HAb Non-injected LCI no HA LCI with HA
Reference group 5.12 (3.38–7.77) 4.54 (3.10–6.64)
<0.001 <0.001
<0.001 0.47 0.013 0.012 0.15 0.084
LCI, local corticosteroid injection; ACS, autologous conditioned serum; HA, hyaluronate. a Hazard ratio for each extra year of age. b horse sex and age, veterinary practice were included in each model but results only shown for any LCI.
effect of sex (P = 0.024) on MSI, with geldings having a hazard of 0.70 (P = 0.013) to that of females. Older horses experienced more injuries than younger horses, with the hazard of MSI increasing 1.3 times per additional year of age (P = 0.012). There was no difference between veterinary practices in rates of MSI (P = 0.23). There was marginal clustering of injury rates at the trainer level (shared frailty effect of trainer represented by variance theta = 0.057, P = 0.055). When the analysis was restricted to only those horses with evidence that the MSI developed post injection the hazard ratios for any LCI (HR 3.36, 95% CI 2.37–4.76, P < 0.001) and for single (HR 2.36, 95% CI 1.47–3.77, P < 0.001) and multiple LCI (HR 4.82, 95% CI 3.14–7.42, P < 0.001) remained significant. On restricting the time analysed to serial 7 day intervals post injection, the hazard of injury was approximately 6–12.5 times that of the untreated horses and those prior to treatment up to 49 days post LCI, but returned to levels indistinguishable from untreated horses thereafter (Fig. 1). The assumption of proportional
hazards was satisfied within all except the 1–7 and 8–14 day time periods. The hazard of MSI in horses after they received two or more LCIs in the data collection period was approximately twice the hazard in horses after receiving one injection (HR 2.1, 95% CI 1.3–3.4, P = 0.002). When hyaluronate was combined with corticosteroid, there was no change in the hazard of subsequent MSI compared with corticosteroid alone (P = 0.63). Musculoskeletal injury rates were highest following intra-articular injection of the carpus (HR 5.59, 95% CI 3.85–8.12, P < 0.001, reference = non-injected time at risk) and fetlock (HR 5.68, 95% CI 3.55–9.10, P < 0.001) and lower for all other sites injected (HR 2.40, 95% CI 1.20–4.82, P = 0.014). The rate of injury was higher following betamethasone than following triamcinolone injection (HR 1.91, 95% CI 1.16–3.13, P = 0.011). Of the 11 potentially catastrophic injuries, seven were in horses that did not receive local corticosteroids and four were in treated horses. None of these horses were radiographed prior to the injury. One horse sustained a compound fracture of the lower limb 20 days following its third bilateral carpal LCI, another a condylar fracture 20 days after its fifth bilateral fetlock LCI, one a fractured proximal phalanx, proximal sesamoid and distal metacarpal fracture 78 days after a single bilateral LCI of the fetlocks and a fourth a condylar fracture 21 days after a single bilateral fetlock LCI. Hazard of potentially catastrophic injuries following second and subsequent LCIs was higher than in horses that were not injected and those prior to injection (HR 6.9, 95% CI 1.2–39.8, P = 0.031) but this was not the case following one LCI (HR 2.5, 95% CI 0.44–13.5, P = 0.30). Horses receiving intra-articular ACS subsequently developed MSI at an increased hazard but the hazard ratio was not significant (P = 0.078); this estimate was imprecise due to the small number of these cases (Table 4). Discussion As expected, this study demonstrated that Thoroughbred racehorses that receive LCIs for the treatment of orthopaedic conditions and continue training have a greater subsequent incidence of MSI than horses that do not receive such treatment. Multiple LCIs and those involving the carpus and fetlock joints resulted in the highest injury rates. Progression of the musculoskeletal problem that prompted treatment appears to be an important contributor to this result and any beneficial effects of LCI are insufficient to counter the increased rate of MSI up to at least 49 days post injection. The aim of this study was not to determine whether the use of local corticosteroids was detrimental, as this was not possible from the study design, rather to quantify the association between the
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76
Fig. 1. Hazard ratios (solid line) and point-wise 95% confidence intervals (dashed lines) for musculoskeletal injury in horses after receiving local corticosteroid injections, compared to non-injected horses, estimated separately for sequential 7 day periods post local corticosteroid injection. After 49 days the hazard ratio was not significantly different from one.
use of LCI and subsequent musculoskeletal injury rates to allow veterinarians, owners and trainers to be better informed when deciding whether to treat a problem and continue to train and race a horse. It was not practical to compare LCIs with no treatment in horses with an identified limb problem, so the independent effect of LCI was not assessed. The greater hazard of injury identified following LCI is more likely due to veterinary examination identifying horses at risk of MSI rather than a direct adverse effect of the corticosteroid. Experimental evidence indicates that the two most commonly used corticosteroids in this study have beneficial effects when used locally to treat induced intra-articular pathology in horses that continue in training, and no detrimental effect of triamcinolone on equine subchondral bone has been demonstrated (Foland et al., 1994; Frisbie et al., 1997; Kawcak et al., 1998; Murray et al., 2002). In the current study, a minority of horses received LCI and there was generally a clear indication for treatment, which suggested that the prophylactic use of local corticosteroids was uncommon and that treated horses were likely to have a pre-existing orthopaedic condition. An increase in the subsequent risk of developing a MSI after injection in these horses was therefore not surprising. The increased hazard of MSI associated with LCI was time dependant. It is likely that horses with underlying problems either develop a MSI within 7 weeks of injection or require repeat treatments at 4–8 weekly intervals to remain in training and therefore return to the ‘less than 49 days since local injection’ groups. Those horses with minor or short term orthopaedic problems can remain in training without the need for repeat treatments beyond 49 days. The majority of MSIs were due to conditions where fatigue of tissues is an important part of the pathogenesis (Stover et al., 1993; Birch et al., 2008; Whitton et al., 2010). Consistent with this, exercise duration and intensity have been shown to be risk factors for MSIs in Thoroughbred racehorses (Estberg et al., 1998; Perkins et al., 2005b; Verheyen et al., 2006). Corticosteroids are effective for the treatment of inflammation and pain but have no ability to prevent fatigue of musculoskeletal tissues if loading continues. If the LCI contributes to allowing a horse that may be developing a MSI to continue to train and race either by its anti-inflammatory effects or by influencing the decision making of the trainer, this will keep the horse at risk of MSI.
75
Assessment of effects of LCI on musculoskeletal injury rates would require a placebo controlled trial. A number of injuries, including catastrophic fractures, occurred in horses that had minimal diagnostic workup prior to LCI. Detection of fatigue prior to fracture is challenging even with sophisticated diagnostic imaging but every effort should be made to identify incomplete fractures (Trope et al., 2011). Many horses receiving local corticosteroids do not develop MSI, and improved diagnostic methods may allow at-risk horses to be identified (Powell, 2012). Although this study was not designed to assess effects based on the expected frequency of catastrophic injury, an effect of multiple LCI (two or more sequential treatments) on the rate of potentially catastrophic injury was observed although the wide confidence interval makes it difficult to determine the magnitude of this effect. The pathogenesis of catastrophic injuries is similar to that of less severe injuries; they are associated with bone, tendon or ligament failure under the repeated loading induced by high speed exercise (Stover et al., 1993; Stepnik et al., 2004). In subchondral bone, the resultant damage is usually localised, but in a minority of horses localised damage may propagate into a complete fracture (Riggs, 1999; Norrdin and Stover, 2006). We found no evidence that adding HA decreased MSI rates following local medication. HA is commonly combined with a corticosteroid to alleviate some of its negative effects, although this is based on little evidence (McIlwraith, 2010). The combination of these medications is recommended in human patients because the longer duration but slower onset of action of HA combines well with the shorter term effects of the corticosteroid (Bannuru et al., 2009). Our failure to demonstrate any effect on MSI rates is consistent with our proposal that it is the underlying musculoskeletal condition that is the major contributor to the increased injury rate following LCI, and does not mean that HA is of no benefit to the synovial structure. Not all MSIs following LCI occurred at the site of injection. Possible reasons for a MSI at a distant site from injection within the same limb include a subsequent injury unrelated to the local injections, incorrect localisation of the original source of pain, or a remote effect of LCI (Frisbie et al., 1997). Possible reasons for a MSI in a different limb to that treated, include overload of the contralateral limb due to inadequate pain relief, unidentified bilateral pain, an unrelated injury or a remote effect of LCI. Because we could not be sure of the accuracy of the pre-injection diagnoses, and it is clear that the effect of continuing to train and race a horse is not localised to the site of injection, we included all injection sites and all injuries in the study. Limitations of this study include its retrospective design and dependence on the accuracy of records kept by veterinarians. We used practices that had good computer-based records and, where there were obvious deficiencies, we excluded the relevant horse. Not all joints were radiographed prior to treatment so we cannot be certain that an intra-articular fracture was not already present. However, in such cases there was clear deterioration of clinical signs consistent with progression of the injury and when these horses were excluded all results remained statistically significant. Also, it is possible that some horses developed MSIs and were spelled without veterinary examination, but to minimise this occurrence we selected trainers who worked closely with veterinarians and sought frequent veterinary advice. The length of rest periods used to define MSI was greater than the routine rest period horses receive without injury so it is likely that our case definition was specific and trainers were likely to obtain a diagnosis and prognosis where prolonged rest periods were required. Our study population had a relatively low proportion of local corticosteroid treatment and care should be taken in extrapolating these results to populations where the local corticosteroid treatment proportion is much higher. The failure to meet the proportional hazards
76
R.C. Whitton et al. / The Veterinary Journal 200 (2014) 71–76
assumption with some models was expected with the use of time varying covariates (Allison, 2010). If the presence of the original condition that prompts the use of LCI is a major contributor to the subsequent rate of MSI, new medications suitable for local treatment are unlikely to reduce this outcome. This is because any medication that allows a horse with an underlying problem to continue to train and race has the potential to indirectly increase injury rates unless that medication completely removes risk due to the pre-existing condition. Consistent with this, ACS, corticosteroids and HA/corticosteroid combination appeared to have similar associations with MSI. Conclusions Treatment of horses with musculoskeletal conditions by LCI does not return injury rates to levels observed in horses not under the influence of treatment. An otherwise identical group that did not receive LCI was not available so effects of LCI were not assessed. Further investigation should be directed at better identifying horses at risk of MSI following LCI. This information should be included in the decision making process when managing horses with musculoskeletal problems. Conflict of interest statement None of the authors has any financial or personal relationships that could inappropriately influence or bias the content of the paper. Acknowledgements We thank the veterinary practices that provided access to their veterinary records. This study was presented in part at the 50th British Equine Veterinary Association Congress, Liverpool, UK, 2011. References Allison, P.D., 2010. Survival Analysis using SAS. A Practical Guide, Second Ed. SAS Institute, Cary, NC, USA. Bailey, C.J., Reid, S.W., Hodgson, D.R., Rose, R.J., 1999. Impact of injuries and disease on a cohort of two- and three-year-old Thoroughbreds in training. Veterinary Record 145, 487–493. Bannuru, R.R., Natov, N.S., Obadan, I.E., Price, L.L., Schmid, C.H., McAlindon, T.E., 2009. Therapeutic trajectory of hyaluronic acid versus corticosteroids in the treatment of knee osteoarthritis: A systematic review and meta-analysis. Arthritis and Rheumatism 61, 1704–1711. Birch, H.L., Bailey, A.J., Goodship, A.E., 1998. Macroscopic ‘degeneration’ of equine superficial digital flexor tendon is accompanied by a change in extracellular matrix composition. Equine Veterinary Journal 30, 534–539. Birch, H.L., Wilson, A.M., Goodship, A.E., 2008. Physical activity: Does long-term, high-intensity exercise in horses result in tendon degeneration? Journal of Applied Physiology 105, 1927–1933. Chen, C.L., Sailor, J.A., Collier, J., Wiegand, J., 1992. Synovial and serum levels of triamcinolone following intra-articular administration of triamcinolone acetonide in the horse. Journal of Veterinary Pharmacology and Therapeutics 15, 240–246. Cohen, N.D., Mundy, G.D., Peloso, J.G., Carey, V.J., Amend, N.K., 1999. Results of physical inspection before races and race-related characteristics and their association with musculoskeletal injuries in Thoroughbreds during races. Journal of the American Veterinary Medical Association 215, 654–661. Cox, D., 1972. Regression models and life-tables (with discussion). Journal of the Royal Statistical Society B 34, 187–220.
Derendorf, H., Mollmann, H., Gruner, A., Haack, D., Gyselby, G., 1986. Pharmacokinetics and pharmacodynamics of glucocorticoid suspensions after intra-articular administration. Clinical Pharmacology and Therapeutics 39, 313–317. Estberg, L., Gardner, I.A., Stover, S.M., Johnson, B.J., 1998. A case-crossover study of intensive racing and training schedules and risk of catastrophic musculoskeletal injury and lay-up in California Thoroughbred racehorses. Preventive Veterinary Medicine 33, 159–170. Foland, J.W., McIlwraith, C.W., Trotter, G.W., Powers, B.E., Lamar, C.H., 1994. Effect of betamathasone and exercise on equine carpal joints with osteochondral fragments. Veterinary Surgery 23, 369–376. Frisbie, D.D., Kawcak, C.E., Trotter, G.W., Powers, B.E., Walton, R.M., McIlwraith, C.W., 1997. Effects of triamcinolone acetonide on an in vivo equine osteochondral fragment exercise model. Equine Veterinary Journal 29, 349– 359. Harrison, S.M., Whitton, R.C., Kawcak, C.E., Stover, S.M., Pandy, M.G., 2010. Relationship between muscle forces, joint loading and utilization of elastic strain energy in equine locomotion. Journal of Experimental Biology 213, 3998– 4009. Kawcak, C.E., Norrdin, R.W., Frisbie, D.D., Trotter, G.W., McIlwraith, C.W., 1998. Effects of osteochondral fragmentation and intra-articular triamcinolone acetonide treatment on subchondral bone in the equine carpus. Equine Veterinary Journal 30, 66–71. McIlwraith, C.W., 2010. The use of intra-articular corticosteroids in the horse: What is known on a scientific basis? Equine Veterinary Journal 42, 563– 571. Murray, R.C., Znaor, N., Tanner, K.E., DeBowes, R.M., Gaughan, E.M., Goodship, A.E., 2002. The effect of intra-articular methylprednisolone acetate and exercise on equine carpal subchondral and cancellous bone microhardness. Equine Veterinary Journal 34, 306–310. Norrdin, R.W., Stover, S.M., 2006. Subchondral bone failure in overload arthrosis: A scanning electron microscopic study in horses. Journal of Musculoskeletal and Neuronal Interactions 6, 251–257. Parkin, T.D., Clegg, P.D., French, N.P., Proudman, C.J., Riggs, C.M., Singer, E.R., Webbon, P.M., Morgan, K.L., 2006. Catastrophic fracture of the lateral condyle of the third metacarpus/metatarsus in UK racehorses – Fracture descriptions and pre-existing pathology. The Veterinary Journal 171, 157–165. Perkins, N.R., Reid, S.W., Morris, R.S., 2005a. Profiling the New Zealand Thoroughbred racing industry. 2. Conditions interfering with training and racing. New Zealand Veterinary Journal 53, 69–76. Perkins, N.R., Reid, S.W., Morris, R.S., 2005b. Risk factors for musculoskeletal injuries of the lower limbs in Thoroughbred racehorses in New Zealand. New Zealand Veterinary Journal 53, 171–183. Powell, S.E., 2012. Low-field standing magnetic resonance imaging findings of the metacarpo/metatarsophalangeal joint of racing Thoroughbreds with lameness localised to the region: A retrospective study of 131 horses. Equine Veterinary Journal 44, 169–177. Riggs, C.M., 1999. Aetiopathogenesis of parasagittal fractures of the distal condyles of the third metacarpal and third metatarsal bones: A review of the literature. Equine Veterinary Journal 31, 116–120. Ronéus, B., Lindblad, A., Lindholm, A., Jones, B., 1993. Effects of intraarticular corticosteroid and sodium hyaluronate injections on synovial fluid production and synovial fluid content of sodium hyaluronate and proteoglycans in normal equine joints. Journal of Veterinary Medicine 40, 10–16. Rothman, K.J., 1986. Modern Epidemiology. Little, Brown and Co., Boston. Schoenfeld, D., 1983. Sample-size formula for the proportional-hazards regression model. Biometrics 39, 499–503. Stepnik, M.W., Radtke, C.L., Scollay, M.C., Oshel, P.E., Albrecht, R.M., Santschi, E.M., Markel, M.D., Muir, P., 2004. Scanning electron microscopic examination of third metacarpal/third metatarsal bone failure surfaces in Thoroughbred racehorses with condylar fracture. Veterinary Surgery 33, 2–10. Stover, S.M., Ardans, A.A., Read, D.H., Johnson, B.J., Barr, B.C., Daft, B.M., Kindu, H., Anderson, M.L., Woods, L.W., Moore, J., et al., 1993. Patterns of stress fractures associated with complete bone fractures in racehorses. Proceedings of the American Association of Equine Practitioners 39, 131–132. Trope, G.D., Anderson, G.A., Whitton, R.C., 2011. Patterns of scintigraphic uptake in the fetlock joint of Thoroughbred racehorses and the effect of increased radiopharmaceutical uptake in the distal metacarpal/tarsal condyle on performance. Equine Veterinary Journal 43, 509–515. Verheyen, K., Price, J., Lanyon, L., Wood, J., 2006. Exercise distance and speed affect the risk of fracture in racehorses. Bone 39, 1322–1330. Whitton, R.C., Trope, G.D., Ghasem-Zadeh, A., Anderson, G.A., Parkin, T.D., Mackie, E.J., Seeman, E., 2010. Third metacarpal condylar fatigue fractures in equine athletes occur within previously modelled subchondral bone. Bone 47, 826– 831.