1074 Ottawa Ankle rules cannot be safely used to rule out ankle fractures in patients who present ≥10 days post-injury

Abstract

Aims/Objectives/Background The Ottawa ankle rules (OAR) have been validated as a highly sensitive tool to rule out ankle fractures and reduce need for radiography. However, datasets validating OAR to date have excluded patients presenting ≥10 days post-injury and there is a need to ascertain if OAR can be safely used to rule out ankle fractures in this population. Methods/Design Patients presenting with ankle injuries to an emergency department (ED) in England between June 2015 and November 2020 were identified retrospectively through a clinical-coding search. Patient records were used to confirm the number of days between injury and presentation; those who presented ≥10 days post-injury were included for further analysis. Data was collected from ED documentation including region of pain, bony tenderness and weight-bearing status. OAR were used to categorise patients as ‘Ottawa-positive’, ‘Ottawa-negative’ or insufficient documentation. It was recorded whether the patient underwent radiography and whether the formal radiograph report confirmed a clinically-significant fracture. Patients who didn’t undergo radiography and didn’t subsequently re-present were deemed not to have a fracture. Data collected for each patient was checked and agreed by two authors. Results/Conclusions 6782 patients presented with ankle injuries, of which 126 patients presented ≥10 days post-injury. Of these 126 patients, 9 were Ottawa-positive, 90 were Ottawa-negative and 27 patients had insufficient documentation. 85 patients underwent radiography and 19 were found to have clinically-significant fractures. Of these fracture patients, 4 were Ottawa-positive and 15 were Ottawa-negative. Within our dataset, OAR demonstrated a sensitivity of 21.05%, specificity 93.75%, PPV 44.40% and NPV 83.30%. Using Fishers exact test, p=0.0658. OAR demonstrate poor sensitivity and cannot be safely used to rule out ankle fractures in patients who present ≥10 days post-injury. However, due to the p-value and low power there may be a risk of type 2 error and a larger study may prove otherwise.