COMMENTARY

Can We Safely Apply the Ottawa Ankle Rules

to Children?

T
he Ottawa Ankle Rules (OAR) were derived, the pooled data suggest that application of the rules in
refined, validated, and implemented in the early children could potentially result in a 24.8% reduction in
1990s by Stiell et al.1–3 to guide the clinical assess- x-ray utilization. It is important to consider the lower
ment of patients after blunt ankle trauma. The original age range of those to whom the rules may be safely
rules were intended to identify all clinically significant applied. As the authors point out, weight bearing is a
ankle and midfoot fractures, while maintaining sufficient requirement of the rule; therefore, it can only be
specificity to allow a safe reduction in x-ray utilization applied to those children able to walk independently
with its attendant ionizing radiation exposure, increased prior to the injury. A preplanned subgroup analysis of
health care costs, and diminished efficiency in already patients less than 6 years of age was not possible due
overcrowded emergency departments (EDs). The vali- to insufficient data, and thus the authors appropriately
dated rules state that an ankle x-ray series is only indi- recommend caution when applying the OAR to this age
cated after an acute blunt injury when there is pain in the group.
malleolar zone and bony tenderness to palpation over So, what types of fractures do the OAR miss in chil-
the posterior edge or tip of either the medial or lateral dren? Further characterization was available for 4 of
malleolus or if the patient is unable to bear weight both the 10 missed fractures. Of these, one was a Salter-
immediately and in the ED. Likewise, a foot x-ray series Harris (SH)-I fracture, 1 was a SH-IV fracture, and 2
is indicated only when there is pain in the midfoot and were deemed ‘‘insignificant,’’ defined by the authors of
bony tenderness to palpation of either the navicular or the original study as either SH-I or avulsion fractures
the base of the fifth metatarsal or if the patient is unable less than 3 mm. Of note, the latter injuries were also
to bear weight both immediately and in the ED. The deemed clinically insignificant in the original work by
authors define weight bearing in the ED as the ability to Stiell et al., and the OAR were never intended to iden-
transfer weight twice (a total of four steps), regardless of tify them. The issue of the clinical significance of SH-I
limping. Subsequent multicenter implementation of the fractures remains a source of debate, and while the
OAR demonstrated decreases in radiography, waiting authors of the present work note this fact, their final
times, and costs, without an increased rate of missed analyses and calculations included all reported frac-
fractures.4 tures.
In international surveys, large numbers of emergency Some clinicians may be understandably reluctant to
physicians (EPs) report familiarity with and routine use rely on a clinical decision rule (CDR) that misses even a
of the OAR.5,6 While the original investigations specifi- small number of pediatric ankle fractures; however,
cally excluded patients under 18 years of age, the rules routine radiographs are likewise imperfect at diagnos-
have been subsequently studied in various adult and ing these injuries.9,10 In fact, the yield of initial x-ray
pediatric populations. A meta-analysis of 15,581 studies is likely lowest in patients with a low-risk clini-
patients (adults and children) from 27 studies demon- cal examination, such as those in whom the OAR are
strated an overall sensitivity and specificity for the OAR ‘‘negative.’’ This is in contrast to delayed radiographs,
of 97.6 and 31.5%, respectively.7 obtained 7–10 days after the injury, which may be help-
In this issue of Academic Emergency Medicine, Dow- ful in visualizing occult injuries by showing evidence of
ling et al.8 present a well-designed and executed meta- bone remodeling and fracture healing. While no CDR
analysis of the Ottawa Ankle Rules for the identification should be used indiscriminately to supplant clinical
of ankle and midfoot fractures in children. In a pooled judgment, the current work provides evidence that it is
analysis of 12 studies representing 3,130 children with reasonable to apply the OAR to help guide clinical deci-
671 fractures, the authors report a pooled sensitivity of sion-making in children 6 years or older with blunt
98.5% (95% confidence interval = 97.3% to 99.2%) and ankle trauma and that doing so may decrease the rate
a negative likelihood ratio of 0.11. When these diagnos- of unnecessary radiographs. Certainly, it is important
tic indices are applied to the average fracture preva- to involve the patient and family in the clinical decision-
lence among the included studies (21.4%), the authors making process, clearly explain that a small percentage
estimate a missed fracture rate of 1.2%. Furthermore, of fractures will be missed by both the OAR and initial
x-rays, and reinforce the need for follow-up evaluation
and consideration of delayed radiographs if symptoms
A related article appears on page 277.

ISSN 1069–6563 ª 2009 by the Society for Academic Emergency Medicine

352 PII ISSN 1069–6563583 doi: 10.1111/j.1553-2712.2009.00370.x
ACAD EMERG MED • April 2009, Vol. 16, No. 4 • www.aemj.org 353

persist despite conservative management. This is analo- comparable to that of initial x-ray studies (which also
gous to what many EPs currently do in cases of sus- miss a small percentage of bony injuries) and that con-
pected radiographically occult fractures in children. servative management and delayed radiographs for
As with all good systematic reviews, the main continued pain is a safe and effective alternative. It
strengths of the present work lie in the study methodol- remains to be seen whether the rules will enjoy the
ogy. The authors devised a protocol that asked clinically same widespread acceptance for use in pediatric
important questions and defined their intended sub- patients, with a resultant decrease in radiographic utili-
group analyses a priori. They used a comprehensive zation and favorable cost–benefit profile, as they have
search strategy with no language restriction. In addi- in the adult population.
tion to querying the large electronic databases of
Michael S. Runyon, MD
published studies, they examined meeting abstracts,
(michael.runyon@carolinashealthcare.org)
conference proceedings, and trial registries; reviewed
Department of Emergency Medicine
the references lists of identified studies; and contacted
Carolinas Medical Center
the first authors of select investigations. This search
Charlotte, NC
of the so-called ‘‘gray literature’’ yielded an additional
15 studies for review that were not identified by other
means. Two reviewers independently screened all References
works utilizing a standardized inclusion and exclusion
form, and study quality was independently assessed in 1. Stiell IG, Greenberg GH, McKnight RD, Nair RC,
a structured manner. The authors performed an appro- McDowell I, Worthington JR. A study to develop
priate statistical analysis, presented their data clearly, clinical decision rules for the use of radiography in
and were careful not to overstate their conclusions. acute ankle injuries. Ann Emerg Med. 1992; 21:384–
McGinn et al.11 have published guidelines on the use 90.
of CDRs. In their work they describe four strata that 2. Stiell IG, Greenberg GH, McKnight RD, et al. Deci-
illustrate the hierarchy of evidence for such rules. The sion rules for the use of radiography in acute ankle
current work is consistent with Level 2 evidence for the injuries: refinement and prospective validation.
use of the OAR in children in that it presents data dem- JAMA. 1993; 269:1127–32.
onstrating accuracy of the rules across several different 3. Stiell IG, McKnight RD, Greenberg GH, et al. Imple-
settings. By definition, this level of evidence heralds a mentation of the Ottawa ankle rules. JAMA. 1994;
rule that may be used with confidence in various set- 271:827–32.
tings. To achieve Level 1 evidence requires an impact 4. Stiell I, Wells G, Laupacis A, et al. Multicentre trial
analysis demonstrating that the CDR is effective in to introduce the Ottawa ankle rules for use of radi-
altering clinician behavior and improving patient care. ography in acute ankle injuries. BMJ. 1995; 311:594–
Toward this end, successful and widespread implemen- 7.
tation of the OAR in children hinges on whether clini- 5. Graham ID, Stiell IG, Laupacis A, et al. Awareness
cians will be comfortable with a CDR that misses a and use of the Ottawa Ankle and Knee Rules in 5
small number of injuries and whether the projected countries: can publication alone be enough to
benefits of decreased radiation exposure and improved change practice? Ann Emerg Med. 2001; 37:259–66.
resource utilization can be realized without compromis- 6. Brehaut JC, Stiell IG, Visentin L, Graham ID. Clinical
ing patient care. decision rules ‘‘in the real world’’: how a widely dis-
Another important consideration in evaluating the seminated rule is used in everyday practice. Acad
impact and performance of a CDR is whether clinicians Emerg Med. 2005; 12:948–56.
can accurately recall the elements of the rule and apply 7. Bachmann LM, Kolb E, Koller MT, Steurer J, ter Riet
it in a consistent and reproducible fashion in their daily G. Accuracy of Ottawa Ankle Rules to exclude frac-
practice. In one survey on the use and recall of the tures of the ankle and mid-foot: systematic review.
OAR among Canadian EPs, a majority of respondents BMJ. 2003; 326:417.
reported routine use of the rules and rated them as 8. Dowling S, Spooner CH, Liang Y, et al. Accuracy of
easy to learn and remember.6 While the survey data Ottawa Ankle Rules to exclude fractures of the ankle
suggested that the majority of physicians apply the and midfoot in children: a meta-analysis. Acad
rules from memory, less than one-third demonstrated Emerg Med. 2009; 16:277–87.
perfect recall of the specific elements of the rules. 9. Lohman M, Kivisaari A, Kallio P, Puntila J, Vehmas
Whether these results translate into routine misapplica- T, Kivisaari L. Acute paediatric ankle trauma: MRI
tion of the rules is uncertain, but similar findings were versus plain radiography. Skeletal Radiol. 2001;
demonstrated in another survey on the use and sponta- 30:504–11.
neous recall of other CDRs.12 Given these findings and 10. Simanovsky N, Hiller N, Leibner E, Simanovsky N.
the exponential increase in the proliferation of CDRs in Sonographic detection of radiographically occult
recent years, this is an area richly deserving of future fractures in paediatric ankle injuries. Pediatr Radiol.
investigation. 2005; 35:1062–5.
In summary, the OAR appear to be an excellent alter- 11. McGinn TG, Guyatt GH, Wyer PC, Naylor CD, Stiell
native to routine radiographs for detecting ankle and IG, Richardson WS. Users’ guides to the medical lit-
midfoot fractures in children 6 years or older after erature: XXII: how to use articles about clinical
blunt trauma. Clinicians can confidently assure patients decision rules. Evidence-Based Medicine Working
and their families that the sensitivity of this CDR is Group. JAMA. 2000; 284:79–84.
354 Runyon • COMMENTARY ON OAR IN CHILDREN

12. Runyon MS, Richman PB, Kline JA, Pulmonary with suspected pulmonary embolism: variations by
Embolism Research Consortium Study Group. practice setting and training level. Acad Emerg
Emergency medicine practitioner knowledge and Med. 2007; 14:53–7.
use of decision rules for the evaluation of patients