EXTREMITY injury is a common cause for visits

Size: px

Start display at page:

Download "EXTREMITY injury is a common cause for visits"

Debra Beasley
6 years ago
Views:

1 ACADEMIC EMERGENCY MEDICINE October 1999, Volume 6, Number Validation of the Ankle Rules in Children with Ankle Injuries AMY C. PLINT, MD, BLAKE BULLOCH, MD, MARTIN H. OSMOND, MD, IAN STIELL, MD, HAL DUNLAP, MD, MARTIN REED, MD, MILTON TENENBEIN, MD, TERRY P. KLASSEN, MD Abstract. Objectives: The Ankle Rules (OAR) have been found to be 100% sensitive in adult patients with ankle injuries, and application of the OAR has resulted in a 28% reduction in the number of x-rays ordered. The objectives of this study were to determine the sensitivity and specificity of the OAR in children and to determine the potential change in x-ray utilization. Methods: Children, aged 2 16 years, presenting to the EDs of two children s hospitals, with an ankle injury in the previous 48 hours, were enrolled. All patients were assessed by either staff physicians or fellows. X-rays were ordered according to standard clinical practice. Prior to reviewing x-rays, the physical examination was recorded on a standardized form. Positive outcomes (clinically significant) were defined as fractures with fragments 3 mm. Patients not x-rayed and asymptomatic at five to seven days postinjury were considered to have no significant fracture. Results: Six hundred seventy patients were enrolled. The OAR were 100% sensitive (95% CI = 95% to 100%) for significant ankle fractures, with a specificity of 24% (95% CI = 20% to 28%). The OAR were 100% sensitive (95% CI = 82% to 100%) for the midfoot, with a specificity of 36% (95% CI = 29% to 43%). If the OAR had been followed, there would have been a reduction of ankle x-rays by 16% and foot x-rays by 29% without missing any clinically significant fracture. However, analysis of the two hospitals showed that if the rules had been applied, one would have a reduction in x-rays, while the other center would have an increase. Conclusions: This study demonstrates the OAR to be sensitive for detecting clinically significant ( 3 mm) ankle and midfoot fractures in children. The application of these rules may reduce the number of x-rays ordered. A further study is required to determine the effect of using the OAR in clinical practice. Key words: child; ankle; injury; Ankle Rules; guidelines; x-rays; radiology. ACADEMIC EMERGENCY MEDICINE 1999; 6: EXTREMITY injury is a common cause for visits to EDs. 1 At the Children s Hospital of Eastern Ontario (CHEO), more than 1,000 children present each year with ankle injuries. Patients with extremity injuries frequently have radiographic studies. In one study, 97% of children presenting with extremity injuries to the ED had From the Pediatric Emergency Medicine Section, Department of Pediatrics, and the Pediatric Radiology Section, Children s Hospital of Eastern Ontario, University of,, Ontario, Canada (ACP, MHO, HD, TPK); Department of Medicine and Department of Epidemiology and Community Medicine, University of,, Ontario, Canada (IS); and the Pediatric Emergency Medicine Section, Department of Pediatrics, Children s Hospital, University of Manitoba,, Manitoba, Canada (BB, MR, MT). Received March 15, 1999; revision received May 27, 1999; accepted June 18, Presented at the Pediatric Academic Society annual meeting, Washington, DC, May 1997; and the SAEM annual meeting, Washington, DC, May 1997 (Best Fellow s presentation 1997). Supported by Children s Hospital of Eastern Ontario Research Institution, Grant (96/02R/S (E)). Address for correspondence and reprints: Dr. Amy C. Plint, MD, Division of Emergency Medicine, Children s Hospital of Eastern Ontario, 401 Smyth Avenue,, Ontario, Canada K1H 8L1. Fax: ; plinta@travel-net.com radiographic studies. 2 Clinical rules that are sensitive for fracture detection, but reduce the number of radiographs ordered, would be useful in reducing radiation exposure, health care costs, and waiting time in the ED. Studies in adult populations (aged years) have found that the Ankle Rules (OAR) approached 100% sensitivity for detecting clinically significant fractures in adult patients with ankle injuries, 3,4 and application of these rules resulted in a 28% reduction in the number of x-rays ordered. 5,6 The OAR state that ankle x-rays are necessary only if there is pain near the malleoli and one of the following: 1) inability to bear weight immediately after the injury and in the ED (four steps) or 2) bone tenderness at the posterior edge or tip of either malleolus. Foot x-rays are necessary only if there is pain in the midfoot and one of the following: 1) inability to bear weight as defined above and 2) bone tenderness over the navicular or base of the fifth metatarsal (Fig. 1). Clinically significant fractures are defined as those with bone fragments 3 mm in width. Ankle injuries in children have several unique factors. Children have growth plates and may sus-

2 1006 OTTAWA ANKLE RULES Plint et al. OTTAWA ANKLE RULES IN CHILDREN Figure 1. The Ankle Rules. Reproduced with permission from: Stiell IG, Greenberg GH, McKnight RD, et al. Decision rules for the use of radiography in acute ankle injuries: refinement and prospective validation. JAMA. 1993; 269: tain Salter-Harris fractures. Assessment of pain and weight bearing may also be difficult in children. Given these issues, prior to the utilization of the OAR in children, one must determine the sensitivity and specificity of the rules in this population. One prospective study has applied the OAR to children with ankle injuries; however, the sample size (n = 71) was small and, hence, the 95% confidence interval was large. This study found the OAR to be 100% sensitive (95% CI = 77% to 100%). 7 Application of the rule in children would have resulted in a reduction of x-rays ordered by 25% without missing any fractures. We proposed that the adult-derived OAR would be sensitive in the pediatric population with ankle injuries and developed a prospective study to test this hypothesis. METHODS Study Design. This was a prospective evaluation of the OAR when applied to a pediatric population. This study was approved by the ethics committee at both centers and since the study included only collection of data normally obtained during a visit, written informed consent was not required. Study Setting and Population. Patients aged 2 to 16 years presenting to the ED at the CHEO or the Children s Hospital,, Manitoba, with an ankle injury sustained in the preceding 48 hours were eligible. Both of these hospitals are tertiary care centers staffed primarily with pediatric emergency physicians (EPs). There are 50,000 ED visits per year at CHEO and 34,000 ED visits per year in. The centers are similar in patient acuity and admission rates. Ankle injury was defined as any acute soft-tissue or bony injury to the distal tibia, distal fibula, talus, malleoli, or midfoot zone. All ankle injury patients, whether or not x-rays were taken, were eligible. Children less than 2 years old were excluded to reduce difficulties in cooperation. All children with multiple injuries, obvious open fractures, evidence of neurovascular compromise, underlying disease predisposing to fractures (e.g., osteogenesis imperfecta),

3 ACADEMIC EMERGENCY MEDICINE October 1999, Volume 6, Number or underlying disease with sensory abnormalities (spina bifida), patients with isolated injuries of the skin, patients returning for reassessment of the same injury, patients referred to the ED with x-rays, and clinically intoxicated patients were excluded. Study Protocol. All patients were assessed by ED staff and fellows oriented and trained in the application of the OAR. The physical examination was recorded on a clinical data form prior to viewing the x-ray. This form included information regarding age of the patient, mechanism of injury, location of tenderness, ability to bear weight, EP interpretation of the x-ray, diagnosis, and treatment. The examining physicians were not asked to interpret whether the OAR were positive or negative on the data form. The principal investigator reviewed all data forms and determined whether the rules were positive or negative. X-rays were ordered by each center s current practice. In, physicians ordered all x-rays. This is contrasted to, where extremity x-rays are usually ordered by the triage nurse based on the Brand protocol. 8 The Brand protocol uses signs such as bone deformity, instability, crepitation, and point tenderness to determine the need for x-ray of ankle and foot. The x-ray was reviewed by the staff physician for immediate clinical management. The staff physicians were also asked to answer questions regarding mechanism of injury, and to record their diagnosis and management as well as their opinion regarding the need for an x-ray. In, physicians were asked to examine the patient prior to reviewing the x-ray; however, compliance with this suggestion was not monitored. Patients who did not have x-rays were telephoned at five to seven days after their visits to determine their outcomes. Measures. A positive outcome (significant outcome) was defined as any fracture in which the fragment was 3 mm or larger as diagnosed by a pediatric radiologist on call for the day who was blinded to the clinical data form. This definition of significant fracture reflects the clinical management at our institutions since we do not usually treat avulsion fractures of 3 mm or less with plaster immobilization. Salter-Harris type 1 fractures were not included as positive outcomes for the purpose of analysis, but these patients have plaster immobilization of their affected extremities at our institutions. Salter-Harris type 1 fractures were diagnosed either by changes on x-ray (such as widening of the growth plate) or clinically. To assess interobserver reliability, 20% of the radiographs were reviewed by a second pediatric radiologist blinded to the interpretation of the first radiologist TABLE 1. Characteristics of the Patient Population No. patients enrolled No. patients eligible, not enrolled Age median (25 75%) 12.5 ( ) yr 12.9 ( ) yr Gender male 288 (56%) 71 (48%) Method of outcome assessment X-ray studies 517 (>99%) 126 (84%)* Telephone followup 3 (<1%) 24 (16%) Mechanism of injury Twisting 382 (73.5%) 102 (68%) Free fall >4 ft 34 (6.5%) 12 (8%) Direct blow 43 (8.3%) 12 (8%) Motor vehicle collision 0 1 (0.7%) Other 61 (11.8%) 14 (15.3%) *Percentage difference of x-ray rate = 15% (CI = 10% to 21%). and the clinical data form. A negative outcome was defined as no fracture or a fracture fragment less than 3 mm or ability to return to normal physical activity with no pain when followed five to seven days after injury. Telephone follow-up of patients who did not have x-rays taken was by a study nurse or one of the investigators, and patients were asked standardized questions regarding pain, activity, and further investigations/ physician visits. Data Analysis. Sensitivity and specificity of the OAR were calculated in the usual manner, with the 95% confidence intervals being calculated by the method of Gardner and Altman. 9 All data were entered and analyzed using SPSS (Chicago, IL), version 6. Interrater agreement for radiographs was measured using the weighted kappa. RESULTS Patient Population. Five hundred twenty patients were enrolled in and 150 patients were enrolled in (Table 1). Patients were enrolled in from August 95 to October 96 and in from October 95 to May 96. There were no significant differences between the two centers, with the exception of the proportion of patients who had x-ray studies ordered. Ninety-six patients had both ankle and midfoot x-rays. Eighty-six patients had significant fractures and 119 patients were diagnosed as having Salter-Har-

4 1008 OTTAWA ANKLE RULES Plint et al. OTTAWA ANKLE RULES IN CHILDREN TABLE 2. Number of Fractures No. patients No. patients with nonsignificant fractures* 29 (4.6%) 3 (2.0%) No. patients with significant fractures 67 (12.9%) 19 (12.7%) No. patients with Salter-Harris type 1 fractures 81 (15.6%) 38 (25.3%) *Nonsignificant fractures are all fractures <3 mm (does not include diagnosed Salter-Harris type 1 fractures). Significant fractures are all fractures 3 mm (does not include Salter-Harris type 1 fractures). p < TABLE 3. Types of Significant Fractures Fibula Tibia 26 5 Fifth metatarsal 7 3 Talar 2 0 First metatarsal 2 0 Cuboid 1 0 Navicular 1 0 Calcaneus 0 1 TOTAL FRACTURES ris type 1 fractures (Table 2). Details of significant fractures are listed in Table 3. The radiology interobserver reliability was high, with a kappa of Three hundred five eligible patients were not enrolled during the study period. These patients met eligibility criteria but the examining physicians were not compliant with completing the clinical data form. The missed patients differed from the study group in that the incidence of fracture was lower at 4%, while 22% were diagnosed clinically as having Salter-Harris type 1 fractures and 74% were diagnosed as having soft-tissue injuries. Three eligible patients were subsequently excluded because telephone follow-up was not completed. The number of patients with ankle and midfoot injuries who did not meet the eligibility criteria was not tracked. Validity of the OAR. The OAR were 100% sensitive (95% CI = 95% to 100%) for 69 significant fractures of the ankle, with a specificity of 24% (95% CI = 20% to 28%). The OAR were 100% sensitive (95% CI = 82% to 100%) for 17 significant fractures of the midfoot, with a specificity of 36% (95% CI = 29% to 43%) (Table 4). As expected, OAR criteria were positive in all children diagnosed as having Salter-Harris type 1 fractures, since bony tenderness is a requirement for both. Including these patients in the data analysis of the OAR sensitivity would further reduce the confidence intervals. The sensitivity of the OAR for any fracture (including those <3 mm) was 91% for ankle fractures (95% CI = 84 to 95%) and 82% for midfoot fractures (95% CI = 65% to 93%). Change in X-ray Rate Using the OAR. The possible reduction in x-rays was calculated on a retrospective basis. If only patients with positive rules were x-rayed, the number of ankle x-rays done would be reduced by 16% (95% CI = 14% to 19%) and the number of foot x-rays would be reduced by 29% (95% CI = 22% to 35%). However, since the two centers had differed in the manner in which x-rays were ordered, it was decided to review each center independently. In, the baseline x-ray rate is 99%, and using the OAR, it could be reduced to 76% (Table 3). s baseline x-ray rate is 84%, and retrospectively, using the OAR would increase this to 94%. However, it is interesting to note that by using the OAR, would obtain a lower x-ray rate than s baseline rate. Separating x-rays by ankle and midfoot, would see a greater reduction in foot x-rays (37%) than ankle (23%). Physician Sensitivity and Specificity. Using clinical acumen alone, the physicians would have missed three clinically significant fractures. Thus, physician sensitivity was 95% (95% CI = 87% to 99%) and specificity was 37% (95% CI = 32% to 41%). The three fractures that would have been missed were an undisplaced fracture of the tibia and fibula, a nondisplaced fracture of the fifth metatarsal, and an avulsion of the tip of the fibula. TABLE 4. Classification Performance of the Ankle Rules for Identifying Significant Ankle and Foot Fractures ( 3 mm) among 670 Ankle Injury Patients Decision Rule Positive DISCUSSION To our knowledge, this is the largest study to date examining the sensitivity and specificity of the OAR in children. This study also included children who were deemed not to need x-rays in the ED and were followed by telephone. In children, as in adults, these rules have now been found to be Yes Ankle Fracture No Foot Fracture Yes Yes No Sensitivity (95% confidence interval) 100% (95, 100) 100% (82, 100) Specificity (95% confidence interval) 24% (20, 28) 36% (29, 43) No

5 ACADEMIC EMERGENCY MEDICINE October 1999, Volume 6, Number % sensitive for significant fractures (fragment 3 mm). These results also confirm the findings of a smaller study in children by Chande. 7 The eligibility criteria used in our study differ from those of the adult studies in patient age and that the injury was less than 48 hours old. Physicians using these rules need to be comfortable with the definition of significant fracture and the confidence intervals surrounding the sensitivity (95% CI = 95% to 100% for ankle injuries and 82% to 100% for midfoot injuries). Overall, using the OAR would have retrospectively reduced the number of x-rays by 16%. Such a reduction should result in a significant reduction in health care costs, patient radiation exposure, and waiting time. Interestingly, this reduction is less than the 25% found by Chande 7 in children and less than the prospective reduction found by Stiell et al. in adults. 3 However, in our study there was a variation between the two hospitals in reduction in x-ray rate would see a reduction in x-ray rate while would see an increase in x-ray rate. The possible reduction in may reflect that the triage nurses are responsible for ordering x-rays. In it appears that clinicians are more specific in detecting significant fractures than the OAR, but this may simply reflect the small number of patients enrolled. Interestingly, this study also gives a higher rate of nonsignificant fractures than previous adult studies. Small avulsion fractures are usually considered more common in adults. We hypothesize that this may reflect parents being cautious with their children and presenting to the ED with more minor injuries than in adult hospitals. LIMITATIONS AND FUTURE QUESTIONS There are several limitations to this study. All patients with ankle/midfoot injuries were not x-rayed since we believed it unethical to subject patients to radiation when the examining physician deemed such investigations unnecessary. These patients were contacted at five to seven days after presentation. We assumed that patients who returned to normal activity without pain had no significant fracture. Furthermore, no criterion standard for the definition of a significant fracture size was found in a review of the literature, and the adoption of fracture size 3 mm was based on our current clinical practice. Patients with these fractures are routinely placed in plaster immobilization at our hospitals. The use of a standardized clinical data form may have affected practice by influencing clinicians to not order x-rays. Furthermore, some physicians may have already been using the OAR to determine the need for x-rays. However, this should not have affected the validation of the rules but may have influenced the possible reduction in x-ray rate. A large number of eligible patients were missed. These patients had more minor injuries (only 4% had significant fractures) than the study population. We attempted to collect interobserver reliability data by planning to have 10% of all patients reviewed by a second physician, but poor physician compliance made this impossible. The majority of the patients in this study were between 9.7 and 15 years of age, and this may limit the applicability of the OAR to younger children. However, this age range is also reflective of the population with ankle and midfoot injuries seen in two pediatric EDs. Future studies could address several questions: 1) Does use of the OAR prospectively reduce x-ray utilization in pediatric patients with ankle and midfoot injuries? 2) Does use of the OAR in pediatric patients prospectively reduce health care costs and patient waiting time? 3) Is there good interobserver reliability of the OAR in pediatric patients? CONCLUSIONS This study has prospectively validated the Ankle Rules in children with ankle and midfoot injuries. The rules have been shown to be highly sensitive for identifying significant fractures and have the potential to reduce the use of radiography in children with ankle injuries. References 1. Gallagher SS, Finison K, Guyer B. The incidence of injury among 87,000 Massachusetts children and adolescents: results of the statewide childhood injury prevention program surveillance system. Am J Public Health. 1984; 74: Klassen TP, Ropp LJ, Sutcliffe T et al. A randomized controlled clinical trial of radiograph ordering for extremity trauma in a pediatric emergency department. Ann Emerg Med. 1993; 18: Stiell IG, Greenberg GH, McKnight RD, Nair RC, McDowell I, Worthington JR. A study to develop clinical decision rules for the use of radiography in acute ankle injuries. Ann Emerg Med. 1992; 21: Stiell IG, Greenberg GH, McKnight RD, et al. Decision rules for the use of radiography in acute ankle injuries: refinement and prospective validation. JAMA. 1993; 269: Stiell IG, McKnight RD, Greenberg GH, et al. Implementation of the Ankle Rules. JAMA. 1994; 271: Stiell IG, Wells G, Laupacis A, et al. Multicentre trial to introduce the Ankle Rules for use of radiography in acute ankle injuries. BMJ 1995; 311: Chande VT. Decision rules for roentgenography of children with acute ankle injuries. Arch Pediatr Adolesc Med. 1995; 149: Brand DA, Frazier WH, Kohlepp WC, et al. A protocol for selecting patients with injured extremities who need x-rays. N Engl J Med. 1982; 306: Gardner MJ, Altman DG (eds). Statistics with Confidence. London: BMJ, 1989.

Journal reading. Introduction. Introduction. Ottawa Ankle Rules. Method

Journal reading. Introduction. Introduction. Ottawa Ankle Rules. Method Journal reading Presenter: PGY 林聖傑 Supervisor: Dr. 林俊龍 102.12.23 The accuracy of ultrasound evaluation in foot and ankle trauma Salih Ekinci, MD American Journal of Emergency Medicine 31 (2013) 1551 1555