Why is this study needed?
Cervical spine injuries are a rare but serious finding following blunt trauma, affecting around 1-2% of presentations. In adult trauma cases, there has been significant research to help decide which patients require X-ray or CT imaging for possible cervical spine injuries, leading to the development of the National Emergency X-Ray (NEXUS) prediction rule and the Canadian C-Spine Rule (CCR).
However, there is a lack of large-scale paediatric studies, and so no validated, accurate prediction rule currently exists.
What is the problem?
Without an accurate prediction tool for detecting paediatric cervical spine injuries, children with significant injuries may be missed or receive unnecessary imaging.
What did previous studies show?
A 2024 Cochrane systematic review compared available clinical decision rules (CDRs) (NEXUS, Canadian C-Spine Rule, PECARN retrospective criteria, NICE guidelines CG56 and CG176, Leonard de novo model, and PEDSPINE) for the evaluation of potential cervical spine injury following blunt trauma in children. Limited evidence supported the diagnostic accuracy of the CDRs, especially in children under 8 years old. Well-designed multicentre large studies in children were warranted; cue Leonard et al. (2024).
What were the aims of the study?
To derive and validate a clinical prediction rule to identify the risk of cervical spine injury in children aged 0 – 17 years presenting following blunt trauma and guide decision-making around imaging.
A derivation cohort was used to select variables in the clinical prediction rule. A validation cohort was then used to determine the potential effect of the CDR on reducing CT and X-ray use in paediatric patients with blunt trauma.
What were the results?
They recruited 22,430 children with suspected or known blunt trauma (11,857 children in the derivation cohort and 10,573 in the validation cohort).
Children in the validation cohort were admitted to the intensive care unit or operating room less frequently than those in the derivation cohort. Cervical spine injury was more common in the derivation cohort (274/11,857 children) than in the validation cohort (159/10,573 children).
The authors then created a tiered decision rule with high-risk and medium-risk factors to guide imaging.
High-risk factors include
- Glasgow Coma Score (GCS) of 3-8 or unresponsive on the AVPU scale
- Abnormal airway, breathing, or circulation
- Focal neurological deficit on examination
Medium risk factors include:
- GCS of 9-14 or responsive to Voice / Pain only on the AVPU scale
- Self-reported neck pain or neck tenderness on examination
- Significant head or torso injuries requiring admission for observation or surgical intervention
12.1% (187/1549) of those with at least one high-risk factor had a c-spine injury. The CDR recommends performing CT scans in these patients.
2.8% (217/7676) of those with at least one medium-risk factor had a c-spine injury. The CDR recommends performing X-rays in these patients.
Only 0.2% (29/13,205) of those without risk factors had a c-spine injury. The CDR recommends no imaging for these patients.
Their model reported an overall sensitivity of 94.3% (95% CI 90.7-97.9), specificity of 60.4% (95% CI 59.4-61.3) and Negative Predictive Value (proportion of children with no risk actors who do not have a cervical spine injury) of 99.9% (95% CI 99.8-100).
What were the limitations of the study?
The study was conducted in Level 1 paediatric trauma centres. This may limit the generalisability of the results in regional/community hospital settings.
They included children who had already received cervical spine imaging and had injuries warranting transfer to a paediatric trauma centre, so there is potential for reporting bias.
A subgroup of included patients were victims of abuse with unclear time and mechanism of trauma. The authors performed subgroup analysis excluding these specific groups and yielded similar test characteristics.
In practice, questions remain about who applies the CDR and the inter-rater reliability. A proportion of the patients who were initially missed using the CDR were found to actually have risk factors documented in EMS reports or the medical record.
CASP checklist for Clinical Prediction Rule (CPR) studies
Is CPR clearly defined?
Yes
Did the population from which the rule was derived include an appropriate spectrum of patients?
Yes – it is important to note that this clinical prediction model did not exclude children who may have had a non-accidental injury with an unclear mechanism of injury. Penetrating trauma was the only exclusion.
Was the rule validated in a different group of patients?
Yes – They had different derivation and validation groups.
Were the predictor variables and the outcome evaluated in a blinded fashion?
A proportion of participants’ imaging results were viewed by the treating clinician prior to their completing the questionnaire.
Were the predictor variables and the outcome evaluated in the initial sample?
A proportion of participants were missed because the ED provider refused enrollment or said “Other,” but this is not well described. Some participants did not receive any imaging and received follow-up phone calls, but it was not reported if they reached all these patients via phone.
Are the statistical methods used to construct and validate the rule clearly described?
Yes
Can the performance of the rule be calculated?
Yes
How precise was the estimate of the treatment effect?
Did they try to refine the rule with other variables to see whether the precision could be improved or the rule simplified?
Yes. Four subanalyses were performed on the combined derivation and validation cohorts, and the prediction rule test characteristics were similar to the overall test characteristics.
Would the prediction rule be reliable and the results interpretable if used for your patient?
Yes
Is the rule acceptable in your case?
Yes
Would the rule’s results modify your decision about managing the patient or the information you can give them?
Yes
What did the authors conclude, and what does it mean for current practice?
The authors have devised a three-tiered prediction rule to detect and guide imaging in children with cervical spine injury. It can potentially reduce radiation exposure from neck imaging amongst children presenting following trauma.
What do I do for the child who falls under the medium risk category and has abnormal X-rays OR normal X-rays with persistent neck pain?
This question is not addressed in this manuscript.
Practice will vary based on institutional guidelines and clinician judgement. Possible options include further imaging with CT or MRI or consultation with specialists.
Stay tuned for additional research on the utility of CT after negative X-rays in this population.
A note from the author, Dr Julie Leonard
“If we had followed the approach of the NEXUS or Canadian C-spine Rules with an algorithm that merely recommends imaging or not imaging, we would unlikely change practice. It is the tiered aspect of our clinical prediction rule, triaging patients to risk-based imaging, that holds the potential to significantly improve practice.
I was very excited to see that applying this rule will reduce CT imaging by 50%! This significant reduction will decrease radiation exposure and unnecessary tests. I am hopeful it will also improve the flow in the ED.”
