A case-control study of factors associated with pelvic and tibial stress fractures in Thoroughbred racehorses in training in the UK

Few epidemiological studies have investigated risk factors for musculoskeletal injury occurring in Thoroughbred racehorses during training, although it is the major cause of wastage in the racing industry. We recently conducted a large-scale epidemiological study to estimate the incidence of fracture in racehorses in training in the UK and to identify associated risk factors. Thirteen racehorse trainers provided data on horses in their care, with daily recording of training information and provision of details on any fractures incurred. Data were collected for 2 years, including two consecutive flat racing seasons (1999 and 2000). This paper describes findings from a nested case-control study investigating factors associated with the occurrence of pelvic and tibial stress fractures in our study population. Cases were identified from the main study and defined as horses with a pelvic or tibial stress fracture, confirmed through routine diagnostic imaging. Randomly selected controls were matched on date of fracture in the case. Age and gender of the horse, its exercise history and training surfaces were examined as explanatory variables. Exercise was quantified as cumulative distances cantered and worked at high speed in 30- and 60-day periods prior to date of fracture in the case. Conditional logistic regression was used to construct multivariable models for the 30- and 60-day periods, respectively. We hypothesised that larger cumulative exercise distances would be associated with an increased risk of pelvic or tibial stress fracture and that different training surfaces would be associated with differences in fracture risk. In the 30-day period, when adjusting for trainer, the risk of pelvic or tibial stress fracture increased with increasing distance cantered, reaching a peak at around 50 km, after which the risk reduced. This trend was not obvious in the 60-day period, with no significant association between exercise distances and risk of stress fracture. Predominant use of one particular sand-based all-weather surface was related to an increased risk of pelvic or tibial stress injury, although this finding should be interpreted with caution. Surface maintenance and construction may play a role; variables that were not considered in the current analyses, which were based on a relatively small number of cases. Trainer was associated with differences in stress fracture risk after adjusting for exercise distances and surface but age and gender were not.