Ottawa Ankle Rules Applied to Children with Mixed Results

Abstract & Commentary

Source: Clark KD, et al. Evaluation of the Ottawa ankle rules in children. Pediatr Emerg Care 2003;19:73-78.

The Ottawa Ankle Rules (OAR) first were introduced by Ian Stiell, et al in 1993.1 Stiell’s goal was a rule with 100% sensitivity for clinically significant fractures (> 3 mm fracture fragment), while not missing any clinically important fractures. The OAR met this goal with an associated specificity of 40% (i.e., the clinician will still obtain many negative radiographs, but no clinically significant fractures will be missed). The OAR, as originally tested, excluded patients younger than 18 years of age. Because there are significant developmental and physiologic differences between children and adults, including age-related variations in fracture epidemiology, the potential for physeal growth plate injury, and the inability of young children to localize pain, it has remained a question whether these rules can beapplied to a pediatric patient population. Salter-Harris injuries went unaddressed in Stiell’s study, as all patients were older than 18 and thus not at risk for such injuries. Finally, while fractures smaller than 3 mm were considered clinically insignificant in adults, it is unclear whether this same criterion can be applied to children.

The purpose of this prospective study was to evaluate the OAR in children younger than 18 years presenting to a pediatric ED. All fractures were considered clinically significant (specifically including Salter-Harris I [SHI] injuries and avulsion fractures < 3 mm). Data were collected on 195 patients enrolled in the study from April 1995-June 1997. Mean patient age was 12.6 years. Forty fractures (21%) were ultimately identified, with SHI fractures of the distal fibula being the most common fracture (15/40 [38%] ). Thirty-five patients were excluded from sensitivity and specificity calculations (the majority due to incomplete data collection forms). Overall, the OAR yielded a sensitivity of 83% (95% CI, 65-94%) and a specificity of 50% (positive predictive value = 28%, negative predictive value = 93%). When patients with avulsion fractures were excluded, the sensitivity and specificity remained essentially unchanged. Utilizing the OAR in this group would have decreased the number of radiographs by 44%, but five fractures (17%) would have been missed. Using a logistic regression method, the authors found three characteristics in this age group that could have increased sensitivity to 93% (with a specificity of 25%). These alternative three characteristics were: inability to walk immediately after the event, tender deltoid ligament, and swelling at the distal tibia. Using these revised OAR, the authors conclude that they could reduce unnecessary radiographs by 22%, but admit that they would have missed two fractures in doing so.

Commentary By Andrew D. Perron, MD, FACEP

There is no question that the OAR have become widely disseminated and adopted throughout the field of emergency medicine. Besides reducing the number of unnecessary radiographs, they can speed a patient’s course through the ED, ultimately resulting in improved ED throughput. The applicability of these rules to pediatric patients, as well as what constitutes a clinically significant fracture in this population, remains subject to debate. Two prior prospective studies have been published regarding the applicability of the OAR in children.2,3 Both studies found 100% sensitivity in predicting clinically significant fractures. One small study2 included all fractures as clinically significant, while another, large series3 excluded specifically the fractures included in this study (SHI and avulsion < 3 mm). The authors of this current study feel that these reported sensitivities are overly optimistic, given the question of whether all fractures are clinically significant, and whether the OAR successfully can be applied in children.

The orthopedic literature supports the belief that most SHI injuries resolve without any growth disturbance. Further, the radiographic identification of these injuries frequently is difficult to make acutely, and often becomes apparent only on subsequent radiographs. In clinical practice, avulsion fractures smaller than 3 mm are not intervened upon, even in the pediatric patient population. Thus, the two injuries that potentially can be missed remain clinically insignificant in the pediatric patient population. In regard to the revised OAR that Clark has suggested based on logistic regression, they have intuitive merit. (When was the last time you could get a 6-year-old to decide if he or she had posterior tibial tenderness limited to the distal 6 cm of the bone?) A prospective evaluation of these three criteria would be a welcome addition to the clinician’s armamentarium in evaluating ankle injuries in children. 

Dr. Perron, Associate Residency Director, Department of Emergency Medicine, Maine Medical Center,Portland, ME, is on the Editorial Board of Emergency Medicine Alert.


1. Stiell IG, et al. Decision rules for the use of radiography in acute ankle injuries. JAMA 1993;269:1127-1132.

2. Chande VT. Decision rules for roentgenography of children with acute ankle injuries. Arch Pediatr Adolesc Med 1995;149:255-258.

3. Plint AL, et al. Validation of the Ottawa Ankle Rules in children with ankle injuries. Acad Emerg Med 1999; 6:1005-1009.