Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Ottawa Knee Rule in children

Three Part Question

In [children with blunt knee injuries] can [the Ottawa Knee Rule be applied] to [aid your decision to x-ray the knee?]

Clinical Scenario

An eight year old girl presents to the Emergency Department with a painful knee having fallen at gymnastics. She is not able to bend it to 90 degrees and says she cannot walk. The knee is not particularly swollen and there are no skin wounds present. You wonder whether the Ottawa Knee Rule would help you decide whether or not to x-ray her knee, which is what her mother wants.

Search Strategy

Ovid MEDLINE (R) 1966-2008 October Week 1 and Ovid EMBASE 1980-2008 October Week 1 [Ottawa] AND [knee] AND [rule] ti, ab LIMIT to children (<18 years) AND English Language. Cochrane Library (2002, issue 3): Ottawa AND knee. Pubmed: Ottawa Knee Rule LIMIT to children <18 years

MEDLINE: 21 hits, EMBASE: 26 hits, Cochrane: 1 hit, Pubmed: 20 hits.

Search Outcome

A total of 6 articles were found, 3 of which were discarded as they excluded children. One additional paper was found from the references.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Khine et al.
234 paediatric patients(<18yrs) presenting to 2 urban Emergency Departments with acute knee trauma. Normal standard of care applied then Ottawa Knee Rule (OKR) prospectively validated.Prospective consecutive case series (Clinical Decision Rule, level II)Sensitivity of OKR at detecting fracture in study group92% sensitivity (95% CI=66-99), 99% negative predictive value (95% CI=95-100) for fracture detection. Using the OKR would have missed 1 of the 13 fractures identified.Small sample size - underpowered. Level I Paediatric Emergency Departments - may reduce external validity.
Bulloch et al.
750 paediatric patients (2-16yrs) presenting to 5 urban paediatric Emergency Departments with a knee injury in the preceding 5 days. Exclusions: reduced GCS, multiple injuries, underlying bone disease, verified knee fracture, isolated soft tissue injury.Prospective multicenter validation study.Sensitivity and specificity of the OKR at detecting fracture in the study group.100% sensitivity at fracture detection (CI=94.9-100%) 42.8% specificity (CI=39.1-46.5%)34% of eligible children not enrolled. All centers were academic paediatric EDs. Two of the 5 centers enrolled the majority of cases. Insufficient power as enrolment stopped at 750 patients(needed 873-1600 pts to achieve a power of 80%) Poor inter-observer reliability in 2-5 year age group due to small numbers recruited. Telephone follow-up used as surrogate outcome to exclude fracture.
Moore et al.
Convenience sample of 146 consecutive 2-16yr olds, presenting to 3 paediatric Emergency Departments. All patients had an x-ray. Patients were evaluated as to whether they could weightbear immediately after the injury or in tne ED, whether they could actively or passively flex the knee to 90o, whether there was bony tenderness and evidence of an effusion.Blinded prospective cohort study.Unable to weight bear69 patients (47%), all fractures identified in this group. Sensitivity 100% (CI=82-100%), specificity 59% (CI=50-67%), PPV 22% (CI=13-34%), NPV 100% (CI=94-100%).Small sample size-underpowered. High incidence of trampoline-related injuries.
Presence of fracture on X ray15 out of 146.
Inability to flex to 90oOnly 6 fractures identified.
Bony tendernessOnly 11 fractures identified.
Cohen et al.
254 patients who had knee X rays for an acute injury in the ED of a Children's Hospital were evaluated for bony tenderness, the ability to weightbear and the ability to flex to 90o.Retrospective chart review.Bony tendernessSensitivity 92% (CI=60-100%), specificity 19% (CI=14-25%), PPV 6% (CI=3-10%), NPV 98% (CI=87-100%).Retrospective study. Not all acute knee injuries were X rayed. 54 patients with an acute knee injury did not have X rays as the clinician felt they were not needed. Outcomes for these is uncertain.
Inability to weightbearSensitivity 91% (CI=57-99%), specificity 93% (CI=88-95%), PPV 37% (CI=20-57%), NPV 100 (CI=97-100%).
Inability to flex to 90oSensitivity 100% (CI=72-100%), specificity 96% (CI=92-98%), PPV 52% (CI=30-74%), NPV 100% (CI=98-100%).


One article (Khine et al) suggested adding point tenderness at the anterior tibial tuberosity to the OKR to increase sensitivity and specificity. Another article (Moore et al) identified the ability to weightbear immediately and in the ED as 100% sensitive in detecting all fractures. if they had only X rayed patients unable to weightbear they could have reduced the number of X rays by 53% without missing any fractures. If Cohen et al had only X rayed those patients unable weightbear and those unable to flex to 90o they could have reduced their use of X rays by 73% with no missed fractues. However all the studies had significant weaknesses.

Clinical Bottom Line

Larger studies are needed to validate the Ottawa Knee Rule in a paediatric population.


  1. Khine H, Dorfman DH, Avner JR. Applicability of the Ottawa Knee Rule for knee injury in children Pediatric Emergency Care 2001: 17 (6): 401-4
  2. Bulloch B, Neto G, Plint A, Lim R, Lidman P, Reed M, Nijsse4n-Jordan C, Tenenbein M, Klassen TP. Validation of the Ottawa Knee Rule in children: A multicenter study. Annals of Emergency Medicine 2003: 42 (1); 48-55.
  3. Moore BR, Hampers LC, Clark KD Performance of a decision rule for radiographs of pediatric knee injuries The Journal of Emergency Medicine 2005: 28 (3); 257-261.
  4. Cohen DM, Jasser JW, Kean JR, Smith GA. Clinical criteria for using radiography for children with acute knee injuries. Pediatric Emergency Care 1998: 14 (3); 185-187.