There is a need for investigation of associations between intrasynovial medication and fracture risk in racehorses. To document the proportion and type of fracture occurring following intrasynovial medication. Retrospective longitudinal study. Review of clinical records from all Thoroughbred flat racehorses receiving intrasynovial medication under the care of a first-opinion veterinary practice between 2006 and 2011. Fractures pre- and post-medication were categorised by location, type, and severity; analysis of fractures sustained within 56-days of medication was undertaken. Survival analyses using Cox proportional hazards models were conducted for associations between medication used (corticosteroid vs. non-corticosteroid), age at first medication, sex and total number of medications and fracture. A total of 1488 horses received intrasynovial medication during the period of study; 8692 synovial spaces were medicated in 3925 sessions. There were 96 fractures that occurred within 56-days of medication, of which 44 were classified as 'serious' (44/1488; 3.0% total population, 95% CI 2.1, 3.8%) and 11 (11/1488, 0.7%, 95% CI 0.3, 1.2%) were euthanased due to severity of injury. Fifty-four of 96 cases (56.4%, 95% CI 46.3, 66.2%) returned to racing. Targeted imaging of the subsequent injury site was undertaken prior to injury in only 7 (7/96, 7.3%) horses that injured post-medication. Horses that had received ≥3 previous intrasynovial medication sessions had an increased hazard of sustaining a fracture within 56 days (HR 2.31, 95% CI 1.51, 3.54, P<0.001) compared with those receiving ≤3 medication sessions, adjusted for each increasing year of age (HR 0.67, 95% CI 0.53, 0.86, P = 0.001). Absence of an unexposed cohort of nonmedicated racehorses. Serious musculoskeletal injury following intrasynovial medication occurred in 3% (44/1488) horses. Reduction in injury rates may be possible through greater use of premedication diagnostic imaging.