Lucas Oliveira Junqueira e Silva1,2 and Fernanda Bellolio1 1 Department of Emergency Medicine, Mayo Clinic, Rochester, MN, USA 2 Department of Emergency Medicine, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil Traumatic brain injury is a leading cause of morbidity and mortality in children and adolescents and accounts for a significant proportion of the >1 million annual emergency department (ED) visits and hospitalizations due to trauma‐related head injuries.1 The clinical challenge when evaluating a child with head trauma is to determine the presence of an intracranial injury. Because the radiation risks of computed tomography (CT) are greater in children, North American and European investigators have developed clinical decision rules (CDRs) for imaging children with blunt head trauma. The objectives of these head injury CDRs are to identify patients who are either at low risk for injury and can thus be evaluated without imaging, or at sufficient risk for injury to require imaging. Can elements of the history and physical exam identify children with minor head injury who are at low risk for intracranial injury and who therefore do not need CT imaging? In a systematic review, Pickering et al.2 found six CDRs that were developed to identify children who are at low risk of intracranial injury after a mild traumatic brain injury (mTBI). However, only four rules have been both derived and validated. The first rule to be validated and the most widely used is by Pediatric Emergency Care Applied Research Network (PECARN).3 They enrolled 42,412 children younger than 18 years old who presented within 24 hours of blunt trauma and had Glasgow coma scale (GCS) scores of 14 or 15, and 14,969 (35%) of them had head CTs (HCTs) performed. Patients were enrolled at 25 EDs, and their outcome of clinically important traumatic brain injury (ciTBI) was defined a priori as death from traumatic brain injury, neurosurgery, intubation for more than 24 hours, or hospital admission for two nights or more associated with traumatic brain injury on CT. Medical records were reviewed for the outcomes, and telephone follow‐up was performed between 7 and 90 days after ED visits to identify any missed injuries. The PECARN rule was derived in a population of 33,785 patients and then validated in 8627 patients by continued data collection at the same sites. The PECARN investigators analyzed preverbal (<2 years old) and verbal (≥2 years old) children separately to account for differences in communicative ability, mechanisms, and risks for intracranial injury. Of the 8502 preverbal children in the derivation cohort, there were 73 cases (0.86%) of intracranial injuries and, of the 25,283 verbal children, there were 215 (0.85%) cases of intracranial injuries. The PECARN decision rule for children <2 years old (Table 11.1) had a sensitivity of 99%, a specificity of 54%, a positive likelihood ratio (LR+) of 2.2, and a negative likelihood ratio (LR−) of 0.02. Similarly, the rule for children ≥2 years old (Table 11.2) had a sensitivity of 97%, a specificity of 59%, a LR+ of 2.4, and a LR− of 0.05. Test characteristics were similar in the validation cohorts (Tables 11.3 and 11.4). Table 11.1 The PECARN head CT rule for children <2 years old Source: Data from [3]. If none of these are present their risk of clinically significant injury is very low. Table 11.2 The PECARN head CT rule for children ≥2 years old Source: Data from [3]. If none of these are present their risk of clinically significant injury is very low. Table 11.3 Performance of the PECARN head CT rule in children <2 years old Source: Data from [3]. Table 11.4 Performance of the PECARN head CT rule in children ≥2 years old Source: Data from [3]. Aside from its original derivation and validation cohorts, several external validation and implementation studies have evaluated the performance of PECARN across different ED settings throughout the globe. In a large US‐based single‐center prospective cohort study, Easter et al.4 enrolled 1009 children with minor head injury presenting within 24 hours of their injuries, and compared the PECARN rule with physician judgment and two other CDRs (Canadian Assessment of Tomography for Childhood Head injury [CATCH] and Children’s Head injury ALgorithm for prediction of Clinically Important Events [CHALICE]). The authors found that only physician’s practice (as defined by actual CT ordering practice) and the PECARN rule identified all ciTBIs as well as all injuries requiring neurosurgical intervention, with sensitivity of 100%. Schonfeld et al.5 performed a cross‐sectional study to validate PECARN in two pediatric EDs in the United States and Italy. A total of 2439 children were included, of whom 373 (15%) had a CT performed, 69 (3%) had a positive CT, and 19 (0.8%) had a ciTBI. None of the patients with ciTBI were classified as very low risk by PECARN (overall sensitivity 100%, specificity 55%, negative predictive value 100%). In a prospective cohort study in 1179 children from Singapore, Thiam et al.6 evaluated the diagnostic performance of PECARN, CATCH, and CHALICE. There were 12 (1%) patients who underwent CT and only 6 (0.5%) who had positive findings. Although the PECARN originally used the outcome of ciTBI, this study evaluated its ability to predict a positive HCT. The sensitivity of the rule was 100% even though a different outcome was used from its original derivation and validation studies. Lorton et al.7 prospectively validated the PECARN rule in a French population. Among 1499 children from three EDs in France, there were 9 (0.6%) children with ciTBI, and none were classified as very low risk by the PECARN rule, obtaining a sensitivity of 100% in both cohorts of children <2 years and ≥2 years. The Pediatric Research in Emergency Departments International Collaborative (PREDICT) performed a multicenter prospective external validation study8 including 20,137 children in 10 Australian and New Zealand EDs, and evaluated the performance of PECARN, CATCH, and CHALICE. They enrolled all children (age < 18) with head injuries of any severity, and collected data on the inclusion and exclusion criteria of these three CDRs, as well as on their predictor variables and outcomes measures. In this cohort, 280 (1%) had ciTBI as defined by PECARN. PECARN did not miss any patients younger than 2 years but did miss one patient aged 2 years or older who did not require neurosurgery (sensitivity of 100% for <2 years and 99% for ≥2 years). Ide et al.9,10 validated the PECARN rule in Japan both retrospectively and prospectively. In their retrospective study,9 among 2208 children there were 24 (1.1%) patients with ciTBI. They found a reduced sensitivity and specificity when compared to the PECARN derivation and validation cohorts. The sensitivity was 85.7% in children <2 years. There were 16 cases of physically abused children in their cohort, 10 of which were in the <2 years group. There were two cases of abused children with ciTBI that were identified as very low risk and therefore missed by the PECARN rule. When the authors analyzed the cohort excluding these cases, however, the sensitivity of the rule improved to 100%. In the Japanese prospective study10, Ide et al. enrolled children younger than 16 with minor head trauma (GCS ≥ 14) who presented to six participating EDs within 24 hours of their injuries. Among 6585 patients, 463 (7%) underwent a HCT and 23 (0.35%) had ciTBI. Two patients with ciTBI were classified as very low risk by PECARN, yielding an overall sensitivity of 91.3% (95% confidence interval [CI] 72–98.9%) and a negative predictive value of 99.96% (95% CI 99.86–100%). Table 11.5 Performance of the PECARN head CT rule in external validation studies N = sample size; NPV = negative predictive value; PPV = positive predictive value; LR− = likelihood ratio for a negative test, LR+ = likelihood ratio for a positive test. * This study did not report the performance of PECARN separated by age. ** This study did not report the raw numbers for their calculations. The performance of PECARN across the different external validation cohorts is summarized in Table 11.5. Implementation studies of the PECARN rule were also performed and evaluated its impact on HCT rates. Three studies showed safe reductions (i.e., without missing ciTBI) with imaging rates ranging from 4% to 17%.11–13 This is in contrast with high rates of CT imaging shown by validation studies8 when the rule was applied alone without taking into consideration clinical judgment. This rule was not intended to be applied without clinician input to the decision‐making process. While the PECARN investigators suggest that patients at intermediate risk of ciTBI should either undergo observation or proceed with imaging, the use of shared decision‐making is key. Hess et al.14 performed a cluster randomized trial to evaluate the impact of a decision aid tool to improve parents‐centered outcomes. Its use was associated with significant improvement in parent knowledge, decisional conflict, and involvement in decision‐making. No changes in the rates of imaging or ciTBI were found. The 2009 PECARN rules were the first validated CT CDRs for pediatric blunt trauma and it is the most comprehensively studied, but there are three other CDRs that were derived before the PECARN. The first of these, a preplanned analysis of the National Emergency X‐Radiography Utilization Study II (NEXUS II) investigation by Oman et al.15 studied patients ≤18 years old to examine the performance of the NEXUS II low‐risk HCT rule on this population. The NEXUS II HCT rules are shown in Table 11.6. In an adaptation suited for the study of children, only seven of the eight variables were evaluated (the age > 65 criterion was dropped). NEXUS II enrolled 1666 children, all of whom underwent HCT. The outcomes evaluated were the same as for the larger NEXUS II study: clinically important intracranial injury requiring neurosurgical intervention or likely to lead to significant long‐term neurological impairment. The prevalence of clinically significant intracranial injury was 8.3% (138/1666). The performance of the adapted NEXUS II HCT rules is shown in Table 11.7
Chapter 11
Blunt Head Injury in Children
Background
Clinical question
CT can be avoided in children without any of the following:
CT can routinely be avoided in children without any of the following:
Derivation cohort
PECARN‐defined clinically important traumatic brain injury
Decision rule
Injury
No injury
Totals
Positive
72
3901
3973
Negative
1
4528
4529
Totals
73
8429
8502
Sensitivity (95% confidence interval [CI])
99% (93–100%)
Specificity (95% CI)
54% (53–55%)
Positive likelihood ratio (LR+)
2.2
Negative likelihood ratio (LR−)
0.02
Validation cohort
PECARN‐defined clinically important traumatic brain injury
Decision rule
Injury
No injury
Totals
Positive
25
1015
1040
Negative
0
1176
1176
Totals
25
2191
2216
Sensitivity (95% CI)
100% (86–100%)
Specificity (95% CI)
54% (52–56%)
Positive likelihood ratio (LR+)
2.2
Negative likelihood ratio (LR−)
0
Derivation cohort
PECARN‐defined clinically important traumatic brain injury
Decision rule
Injury
No injury
Totals
Positive
208
10,412
10,620
Negative
7
14,656
14,663
Totals
215
25,068
25,283
Sensitivity (95% confidence interval [CI])
97% (93–99%)
Specificity (95% CI)
59% (58–59%)
Positive likelihood ratio (LR+)
2.4
Negative likelihood ratio (LR−)
0.05
Validation cohort
PECARN‐defined clinically important traumatic brain injury
Decision rule
Injury
No injury
Totals
Positive
61
2550
2611
Negative
2
3798
3800
Totals
63
6348
6411
Sensitivity (95% CI)
97% (89–100%)
Specificity (95% CI)
60% (59–61%)
Positive likelihood ratio (LR+)
2.4
Negative likelihood ratio (LR−)
0.05
Rule
N
Sensitivity
Specificity
NPV
PPV
LR−
LR+
Easter et al.4 (the United States) – outcome of clinically important TBI
PECARN*
981
21/21, 100%
361/960, 62%
100%
5.5%
0.0
2.7
Easter et al.4 (the United States) – outcome of injury requiring neurosurgical intervention
PECARN*
981
4/4, 100%
599/977, 61%
100%
1%
0.0
2.6
Schonfeld et al.5 (the United States, Italy) – outcome of clinically important TBI
PECARN < 2 years
959
6/6, 100%
410/950, 43.2%
100%
1.7%
0.0
1.76
PECARN ≥ 2 years
1439
13/13, 100%
705,1459, 48.3%
100%
2%
0.0
1.93
Thiam et al.6 (Singapore) – outcome of positive head CT
PECARN*
1179
6/6, 100%
723/1173, 61.6%
100%
1.3%
0.0
2.6
Lorton et al.7 (France) – outcome of clinically important TBI
PECARN < 2 years
421
3/3, 100%
267/418, 64%
100%
2%
0.0
2.77
PECARN ≥ 2 years
1078
6/6, 100%
774/1072, 72.2%
100%
2%
0.0
3.59
Babl et al.8 (Australia, New Zealand) – outcome of clinically important TBI
PECARN < 2 years
4011
38/38, 100%
2139/3973, 53.8%
100%
2%
0.0
2.16
PECARN ≥ 2 years
11,152
97/98, 99%
5067/11054, 45.8%
100%
1.6%
0.02
1.83
Ide et al.9 (Japan) – outcome of clinically important TBI
PECARN < 2 years
792
12/14, 85.7%
572/778, 73.5%
99.7%
5.5%
0.19
3.23
PECARN ≥ 2 years
1416
10/10, 100%
1032/1406, 73.4%
100%
2.6%
0.0
3.76
Ide et al.10 (Japan) – outcome of clinically important TBI
PECARN** < 2 years
2237
86.7%
71.1%
99.9%
2%
0.19
3
PECARN ≥ 2 years
4348
100%
79.7%
100%
0.9%
0.0
4.9
Full access? Get Clinical Tree
