AbstractObjectives(1) To estimate the prevalence of delayed union, non‐union and mal‐union in canine fractures; (2) to describe fracture, demographic, and treatment characteristics for these outcomes; (3) to identify risk factors for delayed or non‐union.Study designRetrospective study.Sample populationFour hundred and forty two dogs (461 fractures).MethodsA review was conducted of clinical records and radiographs from 2 teaching hospitals. “Union,” “delayed union,” “non‐union” and “mal‐union” were defined, and fracture, demographic, treatment, and outcome variables described. Differences in proportions or medians between “union,” “delayed union” and “non‐union” were tested using χ2 and Mann‐Whitney U‐tests for categorical and continuous variables respectively. Potential explanatory variables for “delayed or non‐union” were tested using logistic regression to identify risk factors.ResultsMedian radiographic follow up was 53 days (14‐282). Delayed union occurred in 13.9% of fractures (64/461), non‐union in 4.6% (21/461), and mal‐union in 0.7% (3/461). Risk factors for delayed or non‐union were age (OR 1.21, 95% CI 1.12‐1.31); comminuted fracture (OR 4.24, 95% CI 2.4‐7.5); treatment with bone graft (all types) (OR 3.32, 95% CI 1.3‐8.5); surgical site infection (OR 3.24, 95% CI 1.17‐8.97), and major implant failure (OR 12.94, 95% CI 5.06‐33.1).ConclusionOlder dogs, dogs with comminuted fractures, surgical site infection, or major implant failure were at increased odds of delayed or non‐union. Radius and ulna fractures in toy breed dogs were not at increased odds of delayed or non‐union.Clinical significanceThe identified risk factors should inform fracture planning and prognosticating. The prognosis for radial fractures in toy breeds appears better than historically believed.