- Email: [email protected]
Prediction of Persistent Postconcussion Symptoms in Youth Using a Neuroimaging Decision Rule
Accepted Manuscript Prediction of Persistent Post-concussion Symptoms in Youth using a Neuroimaging Decision Rule Gregory Faris, MD, Terri Byczkowski, PhD, MBA, Mona Ho, MS, Lynn Babcock, MD, MS PII:
S1876-2859(15)00336-8
DOI:
10.1016/j.acap.2015.10.007
Reference:
ACAP 774
To appear in:
Academic Pediatrics
Received Date: 6 May 2015 Revised Date:
21 October 2015
Accepted Date: 24 October 2015
Please cite this article as: Faris G, Byczkowski T, Ho M, Babcock L, Prediction of Persistent Postconcussion Symptoms in Youth using a Neuroimaging Decision Rule, Academic Pediatrics (2015), doi: 10.1016/j.acap.2015.10.007. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1 Title: Prediction of Persistent Post-concussion Symptoms in Youth using a Neuroimaging Decision Rule
RI PT
Authors: Gregory Faris, MDaI
Division of Emergency Medicine, Cincinnati Children’s Hospital Medical Center, 3333 Burnet
Avenue, ML 2008, Cincinnati, OH 45229 I
SC
a
Indianapolis, IN 46202
Terri Byczkowski, PhD, MBAa a
M AN U
Present Address: Indiana University Health Methodist Hospital, 1701 North Senate Avenue,
Division of Emergency Medicine, Cincinnati Children’s Hospital Medical Center, 3333 Burnet
Mona Ho, MSa
Division of Emergency Medicine, Cincinnati Children’s Hospital Medical Center, 3333 Burnet
EP
a
TE D
Avenue, ML 2008, Cincinnati, OH 45229, [email protected]
AC C
Avenue, ML 2008, Cincinnati, OH 45229, [email protected]
Lynn Babcock, MD, MSa (Corresponding Author) a
Division of Emergency Medicine, Cincinnati Children’s Hospital Medical Center
3333 Burnet Avenue, ML 2008, Cincinnati, OH 45229, Phone (513) 803-2952, Fax (513) 6367967, [email protected]
ACCEPTED MANUSCRIPT 2
Running Title: Predicting post-concussive symptoms with a neuroimaging rule
RI PT
Key Words: mild traumatic brain injury, concussion, post-concussion syndrome, emergency department, children
SC
Abstract word count: 248
M AN U
Main text word count: 3477
Acknowledgements. Funding to conduct this trial was provided in part by (1) Academic Pediatric Association Young Investigator Award Grant, and (2) Cincinnati Children’s Hospital Medical Center, Division of Emergency Medicine. We would like to acknowledge the
TE D
exemplary work of Jenna Dyas as the lead research coordinator for this project, as well as the recruitment efforts of all the other research coordinators in the emergency department of
EP
Cincinnati Children’s Hospital Medical Center.
Financial Disclosure:
AC C
The authors have no financial relationships relevant to this article to disclose. Conflict of Interest:
The authors have no conflicts of interest to disclose.
ACCEPTED MANUSCRIPT 3 Contributor’s Statement: Gregory Faris: Dr. Faris conceptualized and designed the study, conducted the study, carried out the initial analyses, drafted the initial manuscript, and approved the final manuscript as
RI PT
submitted. Terri Byczkowski: Dr. Byczkowski assisted with study design, data analysis and reviewed, revised and approved the final manuscript as submitted. Mona Ho: Ms. Ho assisted with data analysis and reviewed, revised and approved the final manuscript as submitted. Lynn
SC
Babcock: Dr. Babcock assisted with study design, study conduct and data analysis, and reviewed
AC C
EP
TE D
M AN U
and revised the manuscript, and approved the final manuscript as submitted.
ACCEPTED MANUSCRIPT 4 What’s New Persistence of post-concussive symptoms cannot be predicted by the risk of clinically important
AC C
EP
TE D
M AN U
SC
predict children who develop persistent post-concussive symptoms.
RI PT
traumatic brain injury. Multifactorial studies are needed to develop a clinical decision rule to
ACCEPTED MANUSCRIPT 5 ABSTRACT Objective To evaluate the ability of risk strata generated by a neuroimaging rule, developed to assess risk
youth with an acute mild traumatic brain injury (TBI). Methods
RI PT
of clinically important traumatic brain injury (ciTBI), to predict post-concussive symptoms in
SC
Prospective cohort study of youth ages 5-17 years presenting to an emergency department (ED) within 24 hours of mild TBI. Risk strata (very low, intermediate, at risk) of ciTBI were
M AN U
determined in ED by criteria set forth by the neuroimaging rule. Post-concussive symptoms were assessed using the Health and Behavior Inventory (HBI) in the ED and at one, two and four weeks post-injury. General linear models were used to examine the relationship between the HBI score at one week and risk strata. Repeated measures analysis was used to measure change in
Results
TE D
HBI over time.
Of the 120 participants, 46 were categorized by the PECARN rule as very low-risk, 39 as
EP
intermediate-risk and 35 as at-risk for ciTBI. Adjusted mean HBI scores with 95% confidence intervals at one week were: 18.0 (13.9, 22.2) for at-risk, 13.8 (9.9, 17.6) for intermediate-risk and
AC C
17.1 (13.4, 20.8) for very low-risk. Risk strata were not significantly associated with the adjusted HBI score at one week (p=0.17). While adjusted HBI scores declined significantly over time (p<0.0001), the trajectories of the HBI score over time did not differ significantly by risk strata (p=0.68). Conclusion
ACCEPTED MANUSCRIPT 6 Risk of ciTBI as determined by factors within a neuroimaging rule alone is insufficient to predict
AC C
EP
TE D
M AN U
SC
RI PT
children with persistent post-concussive symptoms.
ACCEPTED MANUSCRIPT 7 Abbreviations: AIS: Abbreviated Injury Score ciTBI: clinically important traumatic brain injury
RI PT
ED: emergency department EHR: electronic health record IQR: interquartile range
SC
ISS: Injury Severity Score
PCS: post-concussion syndrome PCSS: Post Concussion Symptom Scale
M AN U
OR: odds ratio
PECARN: Pediatric Emergency Care Applied Research Network SD: standard deviations
AC C
EP
TE D
TBI: traumatic brain injury
ACCEPTED MANUSCRIPT 8 INTRODUCTION Traumatic brain injury (TBI) results in over 650,000 emergency department (ED) visits annually for youth ages 0-19 years.1 Acutely, many youth experience physical, cognitive and
RI PT
behavior symptoms that limit their ability to function in everyday settings.2-6 The proportion of youth experiencing post-concussive symptoms typically decline over the initial weeks following injury, from 60% at one week to 10-50% at one month post-injury.2,7-10
SC
Acute management in the ED focuses on identifying patients with intracranial injury and managing symptoms. The Pediatric Emergency Care Applied Research Network (PECARN)
M AN U
developed and internally validated a six factor rule that predicts pediatric patients at very low risk of clinically important TBI (ciTBI), where ciTBI is defined as a patient requiring neurosurgical intervention, hospital admission for greater than two nights, intubation for greater than 24 hours or death.11 The intention of the rule was to decrease exposure to ionizing radiation
TE D
in patients by decreasing the number of computed tomography scans (CTs) in children at very low-risk of ciTBI. As outlined by Kuppermann et al , risk of ciTBI can be stratified into three categories – very low (<0.05%), intermediate (0.9% ), and at-risk (4.3%) – based on clinician’s
EP
assessment of the Glasgow Coma Scale (GCS) and the presence of the other signs and symptoms within 24 hours of head trauma as displayed in Table 1.11 Although, factors contained within the
AC C
rule such as loss of consciousness, headache and vomiting have been associated with postconcussive symptoms, repurposing the rule to predict post-concussive symptoms has not been assessed.9,12,13 Evidence suggests children with intracranial injury may have a higher risk of post-concussive symptoms; thus it seems reasonable to hypothesize that children at highest risk of ciTBI may have greater risk of post-concussive symptoms.14,15 Symptoms persisting for one month or longer generally connote the diagnosis of post-concussion syndrome and impose a
ACCEPTED MANUSCRIPT 9 significant public health burden due to impact on daily functioning. Patients with symptoms persisting greater than one week usually necessitates re-evaluation by a medical professional for additional guidance on symptom management and safe return to activities.
RI PT
The primary objective of this pilot study was to determine if the risk strata outlined in the PECARN neuroimaging rule for detecting ciTBI would also be prognostic of post-concussive symptoms at one week. We hypothesized that post-concussive symptom burden would vary
SC
significantly as a function of risk for ciTBI, with those in the highest risk stratum (at-risk)
reporting the highest symptom burden. The secondary objective was to determine if the risk
M AN U
strata would predict post-concussive symptom burden over a month duration among youth presenting to the ED with mild TBI. The ability to use one tool to both identify patients at risk of both ciTBI and post-concussive symptoms would streamline ED and follow-up care.
Study Population
TE D
MATERIALS AND METHODS
This prospective cohort study involved youth between the ages of 5 and 17 years who
EP
presented between August 2012 and August 2013 to a large urban tertiary care ED within 24 hours of a reported direct or indirect blow to the head and GCS ≥14. This study was approved
AC C
by the Institutional Review Board prior to patient enrollment. Patients were excluded if they had penetrating head trauma, pre-existing neurologic impairment (e.g. stroke, cerebrospinal fluid shunt or brain tumor), pre-existing significant psychological problems (e.g depression, anxiety, bipolar, oppositional defiant disorder, or conduct disorder), developmental delay, altered mental status due to ingestion of substances of abuse or alcohol, had been prescribed medication that
ACCEPTED MANUSCRIPT 10 impairs cognition, or non-English speaking families. All inclusion and exclusion criteria were verified by a combination of chart review, treating physician, patient and guardian/parent.
RI PT
Study Design
Trained research coordinators in the ED screened, assented and consented eligible
participants typically during the hours between 8 am and 12 am. Information collected included
SC
age, self-described race/ethnicity, presence of premorbid conditions associated with post-
concussive symptoms including history of headaches16, history of prior concussions17-19, and
M AN U
attention and learning disability (e.g. attention deficit hyperactivity disorder, special education classes, speech therapy, and other learning disabilities), mechanism of injury, physical exam findings, and the results of head CT (if obtained). Physicians answered questions in our electronic health record about the seven variables contained in the PECARN neuroimaging rule
TE D
for children > 2 years of age (Table 1). These responses were used to calculate the risk of ciTBI. The published rule lists six factors because GCS <15 and other signs of altered mental status were combined into a single factor; however, GCS and altered mental status are asked as
EP
separate questions in our electronic health record resulting in seven separate variables. Postconcussive symptoms were assessed using the 20-item version of Health and Behavior Inventory
AC C
(HBI) administered in the ED and at one, two and four weeks post-injury.20 The HBI is a validated inventory of cognitive, somatic, emotional, and behavioral symptoms rated on a 4point scale ranging from 0 (never) to 3 (often) present over the past week. A higher total score signifies a higher burden of symptoms, where HBI score greater than zero at follow-ups was defined as post-concussive symptoms. HBI somatic (9-items) and cognitive (11-items) subscores were also calculated.20,21 Forms for parents and participants differed in wording to
ACCEPTED MANUSCRIPT 11 reflect first versus third person viewpoints.20,22 During the ED visit, parents completed the HBI to assess the child’s premorbid level of functioning. Patients, ages 9-17, completed the HBI in the ED to measure current symptom burden. For participants under the age of 9 or those who
RI PT
were too impaired to complete the questionnaire, their parents completed a second survey
pertaining to their current symptom levels. At one, two and four weeks post-injury, participants, or their parents for participants 5-8 years old, completed the HBI via structured phone follow-up
SC
interviews conducted by research coordinators who were naïve to initial risk categorization. Patients that had no symptoms (i.e. a score of zero on the HBI) at the two week follow-up were
M AN U
not contacted for the four week assessment.
Typical ED care was provided to all participants and management was at the discretion of the treating physician. If performed, clinical interpretation of head CTs was conducted by the on duty neuroradiologists. Traumatic intracranial injury on CT included intracranial hemorrhage or
TE D
contusion, cerebral edema, traumatic infarction, diffuse axonal injury, shearing injury, sigmoid sinus thrombosis, midline shift of intracranial contents or signs of brain, herniation, diastasis of the skull, pneumocephalus, and skull fracture depressed by at least the width of the table of the
EP
skull.11 Participants were given standard discharge instructions for concussions specifying a plan for gradual return to activities based on the 2012 Zurich Consensus Statement on Concussion in
AC C
Sport.23
To achieve our primary objective, we estimated that 30 participants in each risk stratum
were needed to detect an approximate 8 point difference in the mean total Post-Concussive Symptom Scale24 (range 0-132) at one week with a power of 80%. In this pilot study, our primary endpoint was at one week due to feasibility issues. Prior to the start of the study, we amended the protocol to use the HBI given that it was specified as a core common data element
ACCEPTED MANUSCRIPT 12 for assessment of post-concussive symptoms in children and adolescents endorsed by the National Institute of Neurological Diseases and Stroke.25 Thirty children in each stratum would yield over 95% power to detect a 5 point difference in the total HBI score at one week between
RI PT
each risk stratum.
Statistical Analysis
SC
Frequency distributions and descriptive statistics were developed to describe the
demographic characteristics of the study population and participants. T-tests and chi-square tests
M AN U
were used to test for differences in demographic and health-related characteristics between study enrollees and those not enrolled and between risk strata for those enrolled. For our primary objective, general linear models were used to examine the relationship between the HBI score at one week post-injury and risk strata, adjusting for age, gender, pre-injury HBI, and any of the
TE D
premorbid conditions. A repeated measures analysis using a mixed models approach that included a risk stratum by time interaction term was used to evaluate the association of the trajectory of the HBI score over three time points – week 1, 2, and 4, with the risk strata,
EP
adjusting for age, gender, pre-injury HBI, and any of the premorbid conditions. The relationship between the HBI subscores and risk strata were also calculated. Post hoc, we used general linear
AC C
models to assess the relationship between the HBI score at one week and each of the seven individual factors, the total number of symptoms, and the presence of one of the premorbid conditions. We performed all analysis using SAS/STAT software (version 9.3, SAS Institute Inc. Cary, NC). P-values of <0.05 were considered statistically significant.
RESULTS
ACCEPTED MANUSCRIPT 13 Of the 127 patients enrolled, 120 patients completed study measures in the ED. Details of patient recruitment and retention are presented in Figure 1. Study patients were younger by approximately one year (11.0 vs. 11.9 years, p=0.05) and were more likely to be of white race
RI PT
(75.8 vs. 68.1%, p=0.02) compared to those who declined participation or were unable to be enrolled in the ED. Fourteen participants were lost to follow-up at one week, 27 at two weeks, and 34 at four weeks. Those lost to follow-up were similar in terms of age, race, sex, and history
SC
of headaches, history of concussions, or presence of attention or learning disability when
compared to the group that completed follow-up at one week. Average days to follow-up were
M AN U
6.4 (SD 1.4) at one week, 13.5 (SD 1.6) at two weeks, and 26.7 (SD 1.6) at four weeks. Frequency of youth as opposed to parent unique respondents of the HBI were 79/120 at one week, 60/106 at one week, 57/93 at two weeks, and 48/89 at 4 weeks. Characteristics of the study population are shown in Table 2. Mechanisms of injury
TE D
amongst the participants were as follows: sports (20.8%) fall from standing height (20.0%), bicycle collision (11.7%) and other mechanisms (47.5%). Of the 120 study participants, 46 participants were categorized by the PECARN rule as very low-risk, 39 as intermediate-risk and
EP
35 as at-risk (Figure 1). Across the three risk strata, there were no statistically significant differences in age, gender, race, history of headaches, history of concussion, or presence of
AC C
attention or learning disability. Distribution of each of the PECARN risk variables by risk category are presented in Table 3. Within the cohort, a total of 34 participants underwent CT imaging of the head, six had traumatic intracranial injury (all six had an intracranial hemorrhage or contusion, and one participant had concomitant cerebral edema), and none had a ciTBI. Unadjusted mean HBI score pre-injury for all participants was 12.8 (SD 8.9). Unadjusted mean HBI scores of all participants in the ED and at one, two and four weeks post-injury were
ACCEPTED MANUSCRIPT 14 18.2 (SD 11.9), 15.4 (SD 11.0), 12.0 (SD 10.7), and 9.6 (SD 9.6), respectively. Mean total HBI scores and subscores adjusted for age, gender, history of headaches, history of concussions, presence of an attention or learning disability, and pre-injury HBI by risk category are listed in
RI PT
Table 4.
With respect to the primary objective, there was no significant association of the HBI score at one week with risk strata for ciTBI (p=0.17) after adjusting for age, gender, history of
SC
headaches, history of concussions, presence of an attention or learning disability, and pre-injury HBI. Similarly, there was no significant association of the unadjusted HBI score at one week
M AN U
with risk strata for ciTBI (p=0.10). While the unadjusted (p=0.04) mean cognitive subscale score was significantly associated with risk strata at one week, the adjusted (p=0.09) cognitive subscale score showed a marginal association. Both adjusted (p=0.43) and unadjusted (p=0.39) mean somatic subscale scores were not significantly associated with risk strata at one week.
TE D
Counterintuitively, compared to the medium risk strata at one week, the low risk strata demonstrated higher mean scores for the total HBI score, as well as both cognitive and somatic subscale scores.
EP
The repeated measures analysis showed that there was a significant decline in the adjusted HBI score over the three follow-up time periods (1, 2, and 4 week) regardless of risk
AC C
stratum (p< 0.0001). The risk strata by time interaction was not significant indicating that the trajectories of the adjusted HBI score over time did not vary by risk strata (p=0.68). The results of the repeated measures analysis for the cognitive and somatic subscales were similar with significant declines over time (p< 0.0001 for each), but no significant differences in trajectories over time by risk strata (somatic, p=0.89, and cognitive, p=0.62).
ACCEPTED MANUSCRIPT 15 Since there was no association between the risk strata and outcome, post-hoc analysis was conducted to assess association between the unadjusted HBI score at one week and other potential factors. Of the 7 variables incorporated in the PECARN neuroimaging rule in our
RI PT
electronic health record, only one, severe mechanism of injury, was associated with the
unadjusted HBI score at one week (p=0.03). Amongst the participants, the total number of ciTBI risk factors documented ranged from 0 to a maximum of 3. There was no statistically significant
SC
difference between the unadjusted HBI scores at one week for participants who had 3 risk factors versus no risk factors (p=0.57). History of headaches was significantly associated with the
M AN U
unadjusted HBI score at one week (p=0.009). History of concussion (p=0.64) and or presence of attention or learning disabilities (p=0.33) were not associated with the unadjusted HBI score at one week.
TE D
DISCUSSION
Contrary to our hypothesis, the use of an existing clinical decision rule aimed at decreasing unnecessary CTs in children did not improve identification of patients presenting to the ED with
EP
blunt head trauma who will experience post-concussive symptoms. Neither symptom burden at one week post-injury nor over the one month follow-up period differed as a function of risk
AC C
categorization of ciTBI (very low, intermediate, and at-risk). Clinicians should not rely on the risk stratification of the PECARN neuroimaging rule to identify children at risk for persistent symptoms following mild TBI. Further research is needed to determine acute symptom measures or biomarkers that can assist with disposition planning for children with TBI who are at high risk of post-concussive symptoms.
ACCEPTED MANUSCRIPT 16 Since the majority of the variables in the PECARN rule have been associated with postconcussive symptoms, the lack of association between post-concussive symptoms and these acute presenting factors when categorized into risk strata of ciTBI or in isolation in this cohort
RI PT
suggests that there may be other prognostic factors. Despite over 16 variables assessed to derive this rule, only 7 variables were predictive of ciTBI that also met strict Standards for the
Reporting of Diagnostic Accuracy criteria including moderate inter-observer agreement (kappa
SC
>0.05) agreement. The limited number of variables in the decision rule and the hierarchal
stratification may limit its utility for identifying children at risk of ongoing symptoms. As a
M AN U
single factor, only severe mechanism of injury was associated with post-concussive symptom burden at one week, yet this has not been previously associated with persistent symptoms as it was defined in this study. Yeates et al found that children were more likely to have prolonged post-concussive symptoms if they (a) had a higher severity of injury based on modified injury
TE D
severity score, GCS <15 and signs of altered mental status, or (b) presented with a constellation of vomiting, loss of consciousness, and headache.9,15 Within the PECARN rule, the presence of either GCS of 14 or signs of altered mental status alone places a patient into the at-risk category.
EP
Likewise, the constellation of vomiting, loss of consciousness, and headache places participants into the intermediate-risk category in the rule. Stratifying these symptoms as delineated by the
AC C
rule did not equate to a greater post-concussive symptom burden at one week. The rule does not include many of acute signs and symptoms such as dizziness and amnesia that have been shown to predict post-concussive symptoms.9,12,26 Furthermore, prior literature suggests premorbid factors such as gender, age, prior concussion, prior headaches, baseline cognitive functioning, and coping strategies not captured in the rule may be more predictive of post-concussive symptoms than acute injury related signs and symptoms. 4,12,27-29 In the present study, history of
ACCEPTED MANUSCRIPT 17 headaches was the only premorbid factor associated with symptom burden at one week, as has been found in numerous other studies.7,12,14 An attempt to control for these premorbid issues was made by adjusting for those that were assessed in this study which included age, gender and
RI PT
history of headaches, prior concussions, and attention and learning disabilities. Although there was an association between the risk strata and the unadjusted HBI cognitive subscore, this
association was no longer statistically significant after adjusting for premorbid issues, suggesting
SC
that premorbid issues warrant inclusion in future studies to assess the effect on cognitive functioning.
M AN U
No single or set of acute signs, symptoms, neurocognitive tests, blood/urine/cerebrospinal biomarkers or radiological indices have demonstrated high negative or positive predictive value for detection of post-concussive symptoms in adults or children.30,31 Concussion screening tools such as the Acute Concussion Evaluation – Emergency Department (ACE-ED) or the Sport
TE D
Concussion Assessment Tool (SCAT3) are more comprehensive diagnostic tools which assess 18 and 22 acute symptoms, respectively, as well as injury characteristics, cognitive function (in SCAT3), and premorbid risk factors associated with post-concussive symptoms.32-34 As a result,
EP
these tools may have prognostic capacity, yet to our knowledge this has not been reported. Total acute symptom burden has been associated with post-concussive symptoms in most studies,
AC C
however, a few studies, including the present study, have failed to find such association.9,35,36 While computerized neurocognitive tests designed to assess recovery from concussion can detect deficits acutely in the ED, there is contradictory evidence about their prognostic ability.36-38 Currently, there is a prospective, multicenter study being conducted in Canadian pediatric EDs to derive a clinical prediction rule for the development of persistent post-concussive symptoms in
ACCEPTED MANUSCRIPT 18 children and adolescents following acute head injury using clinically available factors from the ED.39 This pilot study has several limitations. Due to the enrollment strategy of continuous
RI PT
enrollment regardless of risk strata, selection bias may have occurred because the very low-risk category completed enrollment prior to completion of the other risk strata. Misclassification bias is a second limitation of the study. Despite inter-rater reliability of at least 50% for each factor
SC
contained in the rule during its development, the factors within the rule are primarily subjective signs and symptoms leading to increased likelihood of misclassification. For instance, GCS, a
M AN U
key factor in the rule, is one of the most widely used indicators of mental status in TBI, yet interrater reliability of this factor in other cohorts has been reported to be less than 50%.40 Misclassification of another key variable, “other signs of altered mental status” may have occurred because each clinician may have applied their own definition of altered mental status,
TE D
as opposed to the way this variable was derived in the formation of the PECARN rule which asked four separate questions about the presence of agitation, slow to respond to questions, repetitive questioning or somnolence. In the electronic health record at our site, these factors
EP
were annotated in small print under the variable altered mental status. Misclassification of either one of these variables alone could have changed a participant’s risk categorization. Establishing
AC C
inter-rater reliability may have reduced the risk of misclassification. About halfway through the study in December 2012, clinical decision support specifying the risk category was available to the clinician in the electronic health record based on the responses to the rule. No overt disclosure about the risk strata were conveyed by study staff to the physician or the patient; however prior knowledge of the rule and risk strata, as well as obvious risk categorization in the electronic health record after implementation of decision support, may have led to altered care
ACCEPTED MANUSCRIPT 19 and guidance based on risk stratification. Effects of different post-injury treatments were not accounted for. Post-concussive symptoms were measured using the HBI administered via the phone or email, yet this tool has not been validated for this mode. The variation in respondent
RI PT
may have affected the rating of symptom severity, although modest parent-child agreement on the HBI has been demonstrated.22 Post-concussive symptoms were only assessed using the HBI and we did not utilize other measures of disability associated with post-concussion syndrome
SC
such as balance testing or computerized neurocognitive testing. Incomplete follow-up may have
M AN U
affected outcome measurements and ultimately affected results.
CONCLUSION
Stratification of the risk of ciTBI based on the PECARN neuroimaging rule is not predictive of post-concussive symptoms burden following acute mild TBI. Thus physicians
TE D
should be aware that patient’s risk of ciTBI is not helpful in identifying those patients who will develop post-concussive symptoms. This study indicates that a single rule is inadequate to serve the dual purposes of assisting with imaging decisions and stratifying risk of persistent symptoms.
EP
Further research is needed to develop a clinical decision rule to identify patients at risk of prolonged post-concussive symptoms during their acute visit for head trauma in order to
AC C
appropriately triage patients for who are in need of ongoing care, as well as stratify patients eligible for possible treatments.
ACCEPTED MANUSCRIPT 20 REFERENCES 1.
Faul M, Xu L, Wald MW, Coronado VG. Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations, and Deaths 2002-2006. Atlanta, GA:
RI PT
Centers for Disease Control and Prevetnion, National Center for Injury Prevention and Control;2010. 2.
Barlow KM, Crawford S, Stevenson A, Sandhu SS, Belanger F, Dewey D. Epidemiology
SC
of postconcussion syndrome in pediatric mild traumatic brain injury. Pediatrics. 2010;126(2):e374-381.
Mittenberg W, Wittner MS, Miller LJ. Postconcussion syndrome occurs in children.
M AN U
3.
Neurosyschol. 1997;11(3):447-452. 4.
Ponsford J, Willmott C, Rothwell A, et al. Cognitive and behavioral outcome following mild traumatic head injury in children. J Head Trauma Rehabil. 1999;14(4):360. Yeates KO, Luria J, Bartkowski H, Rusin J, Martin L, Bigler ED. Postconcussive
TE D
5.
symptoms in children with mild closed head injuries. J Head Trauma Rehabil. 1999;14(4):337-350.
Eisenberg MA, Meehan WP, 3rd, Mannix R. Duration and course of post-concussive
EP
6.
symptoms. Pediatrics. 2014;133(6):999-1006. Carroll LJ, Cassidy JD, Peloso PM, et al. Prognosis for mild traumatic brain injury:
AC C
7.
results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J
Rehabil Med. 2004(43 Suppl):84-105.
8.
Sroufe NS, Fuller DS, West BT, Singal BM, Warschausky SA, Maio RF. Postconcussive
symptoms and neurocognitive function after mild traumatic brain injury in children. Pediatrics. 2010;125(6):e1331-1339.
ACCEPTED MANUSCRIPT 21 9.
Yeates KO, Taylor HG, Rusin J, et al. Longitudinal trajectories of postconcussive symptoms in children with mild traumatic brain injuries and their relationship to acute clinical status. Pediatrics. 2009;123(3):735-743. Eisenberg MA, Andrea J, Meehan W, Mannix R. Time interval between concussions and
RI PT
10.
symptom duration. Pediatrics. 2013;132(1):8-17. 11.
Kuppermann N, Holmes JF, Dayan PS, et al. Identification of children at very low risk of
SC
clinically-important brain injuries after head trauma: a prospective cohort study. Lancet. 2009;374(9696):1160-1170.
Zemek RL, Farion KJ, Sampson M, McGahern C. Prognosticators of Persistent
M AN U
12.
Symptoms Following Pediatric Concussion: A Systematic Review. JAMA Pediatr. 2013:1-7. 13.
Babcock L, Byczkowski T, Wade SL, Ho M, Mookerjee S, Bazarian JJ. Predicting
TE D
postconcussion syndrome after mild traumatic brain injury in children and adolescents who present to the emergency department. JAMA Pediatr. 2013;167(2):156-161. 14.
Hung R, Carroll LJ, Cancelliere C, et al. Systematic review of the clinical course, natural
EP
history, and prognosis for pediatric mild traumatic brain injury: results of the International Collaboration on Mild Traumatic Brain Injury Prognosis. Arch Phys Med
15.
AC C
Rehabil. 2014;95(3 Suppl):S174-191. Taylor HG, Dietrich A, Nuss K, et al. Post-concussive symptoms in children with mild
traumatic brain injury. Neurosyschol. 2010;24(2):148-159.
16.
Collins MW, Field M, Lovell MR, et al. Relationship between postconcussion headache
and neuropsychological test performance in high school athletes. Am J Sports Med. 2003;31(2):168-173.
ACCEPTED MANUSCRIPT 22 17.
Pellman EJ, Viano DC, Casson IR, et al. Concussion in professional football: repeat injuries--part 4. Neurosurg. 2004;55(4):860-873; discussion 873-866.
18.
Guskiewicz KM, McCrea M, Marshall SW, et al. Cumulative effects associated with
RI PT
recurrent concussion in collegiate football players: the NCAA Concussion Study. JAMA. 2003;290(19):2549-2555. 19.
Collins MW, Lovell MR, Iverson GL, Cantu RC, Maroon JC, Field M. Cumulative
SC
effects of concussion in high school athletes. Neurosurg.51(5):1175-1179; discussion 1180-1171.
Ayr LK, Yeates KO, Taylor HG, Browne M. Dimensions of postconcussive symptoms in
M AN U
20.
children with mild traumatic brain injuries. J Int Neuropsychol Soc. 2009;15(1):19-30. 21.
Moran LM, Taylor HG, Rusin J, et al. Do postconcussive symptoms discriminate injury severity in pediatric mild traumatic brain injury? J Head Trauma Rehabil.
22.
TE D
2011;26(5):348-354.
Hajek CA, Yeates KO, Taylor HG, et al. Agreement between Parents and Children on Ratings of Post-Concussive Symptoms Following Mild Traumatic Brain Injury. Child
23.
EP
Neuropsychol. 2010;17(1):17-33.
McCrory P, Meeuwisse WH, Aubry M, et al. Consensus statement on concussion in
AC C
sport: the 4th International Conference on Concussion in Sport held in Zurich, November 2012. Br J Sports Med. 2013;47(5):250-258.
24.
Lovell MR, Iverson GL, Collins MW, et al. Measurement of symptoms following sports-
related concussion: reliability and normative data for the post-concussion scale. Appl Neuropsychol. 2006;13(3):166-174.
ACCEPTED MANUSCRIPT 23 25.
NINDS Commond Data Elements. Traumatic Brain Injury. 2015; http://www.commondataelements.ninds.nih.gov/tbi.aspx#tab=Data_Standards. Accessed 4/1/2015. Lau BC, Kontos AP, Collins MW, Mucha A, Lovell MR. Which on-field signs/symptoms
RI PT
26.
predict protracted recovery from sport-related concussion among high school football players? Am J Sports Med. 2011;39(11):2311-2318.
Fay TB, Yeates KO, Taylor HG, et al. Cognitive reserve as a moderator of
SC
27.
postconcussive symptoms in children with complicated and uncomplicated mild
28.
M AN U
traumatic brain injury. J Int Neuropsychol Soc. 2010;16(1):94-105.
Ganesalingam K, Yeates KO, Ginn MS, et al. Family burden and parental distress following mild traumatic brain injury in children and its relationship to post-concussive symptoms. J Pediatr Psychol. 2008;33(6):621-629.
Kelly A M, Barbara B, Ann D, et al. Injury versus noninjury factors as predictors of
TE D
29.
postconcussive symptoms following mild traumatic brain injury in children. Neurosyschol. 2013;27(1):1-12.
Zemek RL, Farion KJ, Sampson M, McGahern C. Prognosticators of Persistent
EP
30.
Symptoms Following Pediatric Concussion: A Systematic Review. JAMA Pediatr.
31.
AC C
2013;167(3):259-265.
Carroll LJ, Cassidy JD, Holm L, Kraus J, Coronado VG. Methodological issues and
research recommendations for mild traumatic brain injury: the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J Rehabil Med. 2004(43 Suppl):113125.
ACCEPTED MANUSCRIPT 24 32.
SCAT3. Br J Sports Med. 2013;47(5):259. http://www.ncbi.nlm.nih.gov/pubmed/?term=br+j+sports+med+2013%3B47%3A5+259.
33.
Child SCAT3. Br J Sports Med. 2013;47(5):263.
RI PT
Accessed 4/1/2015.
http://www.ncbi.nlm.nih.gov/pubmed/?term=br+j+sports+med+2013%3B47%3A5+263. Accessed 4/1/2015.
Zuckerbraun NS, Atabaki S, Collins MW, Thomas D, Gioia GA. Use of Modified Acute
SC
34.
2014;133(4):635-642. 35.
M AN U
Concussion Evaluation Tools in the Emergency Department. Pediatrics.
Grubenhoff JA, Deakyne SJ, Brou L, Bajaj L, Comstock RD, Kirkwood MW. Acute concussion symptom severity and delayed symptom resolution. Pediatrics. 2014;134(1):54-62.
Hang B, Babcock L, Hornung R, Ho M, Pomerantz WJ. Can Computerized
TE D
36.
Neuropsychological Testing in the Emergency Department Predict Recovery for Young Athletes With Concussions? Pediatr Emerg Care. 2015;31(10):688-693. Brooks BL, Khan S, Daya H, Mikrogianakis A, Barlow KM. Neurocognition in the
EP
37.
emergency department after a mild traumatic brain injury in youth. J Neurotrauma.
38.
AC C
2014;31(20):1744-1749.
Thomas DG, Collins MW, Saladino RA, Frank V, Raab J, Zuckerbraun NS. Identifying
neurocognitive deficits in adolescents following concussion. Acad Emerg Med. 2011;18(3):246-254.
ACCEPTED MANUSCRIPT 25 39.
Zemek R, Osmond MH, Barrowman N. Predicting and preventing postconcussive problems in paediatrics (5P) study: protocol for a prospective multicentre clinical prediction rule derivation study in children with concussion. BMJ open. 2013;3(8).
RI PT
Gill MR, Reiley DG, Green SM. Interrater reliability of Glasgow Coma Scale scores in
EP
TE D
M AN U
SC
the emergency department. Ann Emerg Med. 2004;43(2):215-223.
AC C
40.
ACCEPTED MANUSCRIPT 26
RI PT
Table 1. Factors and Risk Stratification of ciTBI in the PECARN Neuroimaging Rule Sign and Symptoms in the Risk Stratification PECARN Neuroimaging Rule11 Very low-risk Intermediate-risk At-risk of ciTBI of ciTBI of ciTBI (<0.05%) (0.9%) (4.3%)
AC C
EP
TE D
M AN U
SC
Glasgow Coma Scale <15* X Altered mental status* X Signs of basilar skull fracture X Loss of consciousness X Severe mechanism of injury X Vomiting X Headache X Legend: ciTBI: clinically important traumatic brain injury *Glasgow Coma Scale and altered mental status represent as a single factor “altered mental status” in the published neuroimaging rule; however, they are asked as two separate questions in the electronic health version of the rule implemented at our institution Altered mental status: agitation, sleepiness, slow responses, or repetitive question Signs of basilar skull fracture: retro-auricular bruising, periorbital bruising, hemotympanum, cerebral spinal fluid otorrhea, or cerebral spinal fluid rhinorrhea Severe mechanism of injury: motor vehicle crash with patient ejection, death of another passenger, or rollover; pedestrian or bicyclist without helmet struck by a motorized vehicle; falls of more than 1·5 meters (5 feet); or head struck by a high-impact object
ACCEPTED MANUSCRIPT 27 Table 2. Characteristics of Study Participants Participants (n=120) 11.0 ± 3.4 70.0
AC C
EP
TE D
M AN U
SC
RI PT
Age (Mean years ± SD) Gender (% male) Race White (%) 75.8 Black (%) 16.7 Other (%) 7.5 Premorbid conditions History of headaches (%) 22.5 History of concussion (%) 17.7 Attention and learning disabilities (%) 39.2 Legend: SD: standard deviation Attention and learning disabilities: attention deficit hyperactivity disorder, special education classes, and speech therapy, and other learning disabilities, repeated school year, autism, special education classes
ACCEPTED MANUSCRIPT 28 Table 3. Presence of PECARN ciTBI Factors by Risk Groups Very low-risk Intermediate-risk (n=39) (n= 46) (%)† (%)†
At-risk (n=35) (%)†
AC C
EP
TE D
M AN U
SC
RI PT
Loss of Consciousness 0.0 61.5 48.7 Vomiting 0.0 23.1 28.6 Mechanism of injury Mild 76.1 64.1 45.7 Neither mild nor severe 23.9 15.4 28.6 Severe 0.0 20.5 25.7 Current headache None 55.6 33.3 17.1 Mild 33.3 43.6 40.0 Moderate 11.1 15.4 22.9 Severe 0.0 7.7 11.4 Unable to obtain 0.0 0.0 8.6 Glasgow Coma Scale < 15 0.0 0.0 31.4 Other signs of altered mental status 0.0 0.0 85.7 Signs of basilar skull fracture 0.0 0.0 5.7 CT obtained 2.2 30.8 60.0 Presence of ciTBI, n=0 0 0 0 Presence of intracranial injury, n=6 0 0.017 0.033 Legend: † Column percentage ciTBI: clinically important traumatic brain injury CT: cranial computed tomography Mechanism of injury: mild: ground-level falls or running into stationary objects severe: motor vehicle crash with patient ejection, death of another passenger, or rollover; pedestrian or bicyclist without helmet struck by a motorized vehicle; falls of more than 1·5 meters (5 feet); or head struck by a high-impact object Signs of basilar skull fracture: retro-auricular bruising, periorbital bruising, hemotympanum, cerebral spinal fluid otorrhea, or cerebral spinal fluid rhinorrhea Other signs of altered mental status: agitation, sleepiness, slow responses, or repetitive questions
ACCEPTED MANUSCRIPT 29
AC C
EP
TE D
M AN U
SC
RI PT
Table 4. Adjusted* mean Health and Behavior Inventory Score by PECARN Risk Strata associated with ciTBI Mean (95% Confidence Interval) HBI Scores Very low-risk Intermediate-risk At-risk of ciTBI of ciTBI of ciTBI ED Total 12.8 (9.10, 16.5) 17.8 (14.1, 21.5) 23.6 (19.6, 27.7) Cognitive 7.8 (5.3, 10.3) 8.5 (6.0, 11.1) 11.3 (8.6, 14.1) Somatic 5.2 (3.2, 7.2) 9.4 (7.3, 11.4) 12.6 (10.4, 14.7) 1 week Total 17.1 (13.4, 20.8) 13.8 (9.9, 17.6) 18.1 (13.9, 22.2) Cognitive 11.0 (8.6, 13.5) 8.1 (5.5, 10.6) 11.0 (8.3, 13.7) Somatic 6.2 (4.4, 8.0) 5.8 (3.9, 7.7) 7.2 (5.2, 9.2) 2 weeks Total 13.6 (9.9, 17.4) 12.4 (8.6, 16.3) 16.3 (12.2, 20.5) Cognitive 9.0 (6.5, 11.5) 7.8 (5.3, 10.4) 9.9 (7.1, 12.6) Somatic 4.7 (3.0, 6.4) 4.7 (3.0, 6.5) 6.5 (4.6, 8.4) 4 weeks Total 10.1 (6.6, 13.6) 8.8 (5.1, 12.4) 12.9 (8.9, 16.9) Cognitive 7.0 (4.6, 9.5) 5.5 (2.9, 8.0) 7.8 (5.1, 10.5) Somatic 3.4 (1.8, 4.9) 3.6 (2.0, 5.2) 5.4 (3.7, 7.2) Legend: ciTBI: clinically important traumatic brain injury HBI: Health and Behavior Inventory *Mean HBI adjusted for age, gender, history of headaches, history of concussion, and presence of attention and learning disabilities.
ACCEPTED MANUSCRIPT
Figure 1. Study participant flow diagram.
RI PT
Assessed for eligibility (n=966)
SC
Not meeting inclusion/exclusion criteria (n=725) Eligible but not approached (n=35) Refused enrollment (n=79)
M AN U
Enrolled in study (n=127)
Intermediate-risk (n=39) 1 week f/u 87% (n=34) 2 week f/u 79% (n=31) 4 week f/u 69% (n=27)1
EP
Very low-risk (n=46) 1 week f/u 87% (n=40) 2 week f/u 76% (n=35) 4 week f/u 74% (n=34)1
TE D
Did not complete enrollment survey (n=6) Deemed ineligible after enrollment (n=1)
At-risk (n=35) 1 week f/u 91% (n=32) 2 week f/u 77% (n=27) 4 week f/u 71% (n=25)1
AC C
Legend: 1 10 patients had no symptoms at the 2 week follow-up and were not contacted for the 4 week follow-up visit. Their 4th week HBI values were carried over from week 2.
S1876-2859(15)00336-8
DOI:
10.1016/j.acap.2015.10.007
Reference:
ACAP 774
To appear in:
Academic Pediatrics
Received Date: 6 May 2015 Revised Date:
21 October 2015
Accepted Date: 24 October 2015
Please cite this article as: Faris G, Byczkowski T, Ho M, Babcock L, Prediction of Persistent Postconcussion Symptoms in Youth using a Neuroimaging Decision Rule, Academic Pediatrics (2015), doi: 10.1016/j.acap.2015.10.007. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1 Title: Prediction of Persistent Post-concussion Symptoms in Youth using a Neuroimaging Decision Rule
RI PT
Authors: Gregory Faris, MDaI
Division of Emergency Medicine, Cincinnati Children’s Hospital Medical Center, 3333 Burnet
Avenue, ML 2008, Cincinnati, OH 45229 I
SC
a
Indianapolis, IN 46202
Terri Byczkowski, PhD, MBAa a
M AN U
Present Address: Indiana University Health Methodist Hospital, 1701 North Senate Avenue,
Division of Emergency Medicine, Cincinnati Children’s Hospital Medical Center, 3333 Burnet
Mona Ho, MSa
Division of Emergency Medicine, Cincinnati Children’s Hospital Medical Center, 3333 Burnet
EP
a
TE D
Avenue, ML 2008, Cincinnati, OH 45229, [email protected]
AC C
Avenue, ML 2008, Cincinnati, OH 45229, [email protected]
Lynn Babcock, MD, MSa (Corresponding Author) a
Division of Emergency Medicine, Cincinnati Children’s Hospital Medical Center
3333 Burnet Avenue, ML 2008, Cincinnati, OH 45229, Phone (513) 803-2952, Fax (513) 6367967, [email protected]
ACCEPTED MANUSCRIPT 2
Running Title: Predicting post-concussive symptoms with a neuroimaging rule
RI PT
Key Words: mild traumatic brain injury, concussion, post-concussion syndrome, emergency department, children
SC
Abstract word count: 248
M AN U
Main text word count: 3477
Acknowledgements. Funding to conduct this trial was provided in part by (1) Academic Pediatric Association Young Investigator Award Grant, and (2) Cincinnati Children’s Hospital Medical Center, Division of Emergency Medicine. We would like to acknowledge the
TE D
exemplary work of Jenna Dyas as the lead research coordinator for this project, as well as the recruitment efforts of all the other research coordinators in the emergency department of
EP
Cincinnati Children’s Hospital Medical Center.
Financial Disclosure:
AC C
The authors have no financial relationships relevant to this article to disclose. Conflict of Interest:
The authors have no conflicts of interest to disclose.
ACCEPTED MANUSCRIPT 3 Contributor’s Statement: Gregory Faris: Dr. Faris conceptualized and designed the study, conducted the study, carried out the initial analyses, drafted the initial manuscript, and approved the final manuscript as
RI PT
submitted. Terri Byczkowski: Dr. Byczkowski assisted with study design, data analysis and reviewed, revised and approved the final manuscript as submitted. Mona Ho: Ms. Ho assisted with data analysis and reviewed, revised and approved the final manuscript as submitted. Lynn
SC
Babcock: Dr. Babcock assisted with study design, study conduct and data analysis, and reviewed
AC C
EP
TE D
M AN U
and revised the manuscript, and approved the final manuscript as submitted.
ACCEPTED MANUSCRIPT 4 What’s New Persistence of post-concussive symptoms cannot be predicted by the risk of clinically important
AC C
EP
TE D
M AN U
SC
predict children who develop persistent post-concussive symptoms.
RI PT
traumatic brain injury. Multifactorial studies are needed to develop a clinical decision rule to
ACCEPTED MANUSCRIPT 5 ABSTRACT Objective To evaluate the ability of risk strata generated by a neuroimaging rule, developed to assess risk
youth with an acute mild traumatic brain injury (TBI). Methods
RI PT
of clinically important traumatic brain injury (ciTBI), to predict post-concussive symptoms in
SC
Prospective cohort study of youth ages 5-17 years presenting to an emergency department (ED) within 24 hours of mild TBI. Risk strata (very low, intermediate, at risk) of ciTBI were
M AN U
determined in ED by criteria set forth by the neuroimaging rule. Post-concussive symptoms were assessed using the Health and Behavior Inventory (HBI) in the ED and at one, two and four weeks post-injury. General linear models were used to examine the relationship between the HBI score at one week and risk strata. Repeated measures analysis was used to measure change in
Results
TE D
HBI over time.
Of the 120 participants, 46 were categorized by the PECARN rule as very low-risk, 39 as
EP
intermediate-risk and 35 as at-risk for ciTBI. Adjusted mean HBI scores with 95% confidence intervals at one week were: 18.0 (13.9, 22.2) for at-risk, 13.8 (9.9, 17.6) for intermediate-risk and
AC C
17.1 (13.4, 20.8) for very low-risk. Risk strata were not significantly associated with the adjusted HBI score at one week (p=0.17). While adjusted HBI scores declined significantly over time (p<0.0001), the trajectories of the HBI score over time did not differ significantly by risk strata (p=0.68). Conclusion
ACCEPTED MANUSCRIPT 6 Risk of ciTBI as determined by factors within a neuroimaging rule alone is insufficient to predict
AC C
EP
TE D
M AN U
SC
RI PT
children with persistent post-concussive symptoms.
ACCEPTED MANUSCRIPT 7 Abbreviations: AIS: Abbreviated Injury Score ciTBI: clinically important traumatic brain injury
RI PT
ED: emergency department EHR: electronic health record IQR: interquartile range
SC
ISS: Injury Severity Score
PCS: post-concussion syndrome PCSS: Post Concussion Symptom Scale
M AN U
OR: odds ratio
PECARN: Pediatric Emergency Care Applied Research Network SD: standard deviations
AC C
EP
TE D
TBI: traumatic brain injury
ACCEPTED MANUSCRIPT 8 INTRODUCTION Traumatic brain injury (TBI) results in over 650,000 emergency department (ED) visits annually for youth ages 0-19 years.1 Acutely, many youth experience physical, cognitive and
RI PT
behavior symptoms that limit their ability to function in everyday settings.2-6 The proportion of youth experiencing post-concussive symptoms typically decline over the initial weeks following injury, from 60% at one week to 10-50% at one month post-injury.2,7-10
SC
Acute management in the ED focuses on identifying patients with intracranial injury and managing symptoms. The Pediatric Emergency Care Applied Research Network (PECARN)
M AN U
developed and internally validated a six factor rule that predicts pediatric patients at very low risk of clinically important TBI (ciTBI), where ciTBI is defined as a patient requiring neurosurgical intervention, hospital admission for greater than two nights, intubation for greater than 24 hours or death.11 The intention of the rule was to decrease exposure to ionizing radiation
TE D
in patients by decreasing the number of computed tomography scans (CTs) in children at very low-risk of ciTBI. As outlined by Kuppermann et al , risk of ciTBI can be stratified into three categories – very low (<0.05%), intermediate (0.9% ), and at-risk (4.3%) – based on clinician’s
EP
assessment of the Glasgow Coma Scale (GCS) and the presence of the other signs and symptoms within 24 hours of head trauma as displayed in Table 1.11 Although, factors contained within the
AC C
rule such as loss of consciousness, headache and vomiting have been associated with postconcussive symptoms, repurposing the rule to predict post-concussive symptoms has not been assessed.9,12,13 Evidence suggests children with intracranial injury may have a higher risk of post-concussive symptoms; thus it seems reasonable to hypothesize that children at highest risk of ciTBI may have greater risk of post-concussive symptoms.14,15 Symptoms persisting for one month or longer generally connote the diagnosis of post-concussion syndrome and impose a
ACCEPTED MANUSCRIPT 9 significant public health burden due to impact on daily functioning. Patients with symptoms persisting greater than one week usually necessitates re-evaluation by a medical professional for additional guidance on symptom management and safe return to activities.
RI PT
The primary objective of this pilot study was to determine if the risk strata outlined in the PECARN neuroimaging rule for detecting ciTBI would also be prognostic of post-concussive symptoms at one week. We hypothesized that post-concussive symptom burden would vary
SC
significantly as a function of risk for ciTBI, with those in the highest risk stratum (at-risk)
reporting the highest symptom burden. The secondary objective was to determine if the risk
M AN U
strata would predict post-concussive symptom burden over a month duration among youth presenting to the ED with mild TBI. The ability to use one tool to both identify patients at risk of both ciTBI and post-concussive symptoms would streamline ED and follow-up care.
Study Population
TE D
MATERIALS AND METHODS
This prospective cohort study involved youth between the ages of 5 and 17 years who
EP
presented between August 2012 and August 2013 to a large urban tertiary care ED within 24 hours of a reported direct or indirect blow to the head and GCS ≥14. This study was approved
AC C
by the Institutional Review Board prior to patient enrollment. Patients were excluded if they had penetrating head trauma, pre-existing neurologic impairment (e.g. stroke, cerebrospinal fluid shunt or brain tumor), pre-existing significant psychological problems (e.g depression, anxiety, bipolar, oppositional defiant disorder, or conduct disorder), developmental delay, altered mental status due to ingestion of substances of abuse or alcohol, had been prescribed medication that
ACCEPTED MANUSCRIPT 10 impairs cognition, or non-English speaking families. All inclusion and exclusion criteria were verified by a combination of chart review, treating physician, patient and guardian/parent.
RI PT
Study Design
Trained research coordinators in the ED screened, assented and consented eligible
participants typically during the hours between 8 am and 12 am. Information collected included
SC
age, self-described race/ethnicity, presence of premorbid conditions associated with post-
concussive symptoms including history of headaches16, history of prior concussions17-19, and
M AN U
attention and learning disability (e.g. attention deficit hyperactivity disorder, special education classes, speech therapy, and other learning disabilities), mechanism of injury, physical exam findings, and the results of head CT (if obtained). Physicians answered questions in our electronic health record about the seven variables contained in the PECARN neuroimaging rule
TE D
for children > 2 years of age (Table 1). These responses were used to calculate the risk of ciTBI. The published rule lists six factors because GCS <15 and other signs of altered mental status were combined into a single factor; however, GCS and altered mental status are asked as
EP
separate questions in our electronic health record resulting in seven separate variables. Postconcussive symptoms were assessed using the 20-item version of Health and Behavior Inventory
AC C
(HBI) administered in the ED and at one, two and four weeks post-injury.20 The HBI is a validated inventory of cognitive, somatic, emotional, and behavioral symptoms rated on a 4point scale ranging from 0 (never) to 3 (often) present over the past week. A higher total score signifies a higher burden of symptoms, where HBI score greater than zero at follow-ups was defined as post-concussive symptoms. HBI somatic (9-items) and cognitive (11-items) subscores were also calculated.20,21 Forms for parents and participants differed in wording to
ACCEPTED MANUSCRIPT 11 reflect first versus third person viewpoints.20,22 During the ED visit, parents completed the HBI to assess the child’s premorbid level of functioning. Patients, ages 9-17, completed the HBI in the ED to measure current symptom burden. For participants under the age of 9 or those who
RI PT
were too impaired to complete the questionnaire, their parents completed a second survey
pertaining to their current symptom levels. At one, two and four weeks post-injury, participants, or their parents for participants 5-8 years old, completed the HBI via structured phone follow-up
SC
interviews conducted by research coordinators who were naïve to initial risk categorization. Patients that had no symptoms (i.e. a score of zero on the HBI) at the two week follow-up were
M AN U
not contacted for the four week assessment.
Typical ED care was provided to all participants and management was at the discretion of the treating physician. If performed, clinical interpretation of head CTs was conducted by the on duty neuroradiologists. Traumatic intracranial injury on CT included intracranial hemorrhage or
TE D
contusion, cerebral edema, traumatic infarction, diffuse axonal injury, shearing injury, sigmoid sinus thrombosis, midline shift of intracranial contents or signs of brain, herniation, diastasis of the skull, pneumocephalus, and skull fracture depressed by at least the width of the table of the
EP
skull.11 Participants were given standard discharge instructions for concussions specifying a plan for gradual return to activities based on the 2012 Zurich Consensus Statement on Concussion in
AC C
Sport.23
To achieve our primary objective, we estimated that 30 participants in each risk stratum
were needed to detect an approximate 8 point difference in the mean total Post-Concussive Symptom Scale24 (range 0-132) at one week with a power of 80%. In this pilot study, our primary endpoint was at one week due to feasibility issues. Prior to the start of the study, we amended the protocol to use the HBI given that it was specified as a core common data element
ACCEPTED MANUSCRIPT 12 for assessment of post-concussive symptoms in children and adolescents endorsed by the National Institute of Neurological Diseases and Stroke.25 Thirty children in each stratum would yield over 95% power to detect a 5 point difference in the total HBI score at one week between
RI PT
each risk stratum.
Statistical Analysis
SC
Frequency distributions and descriptive statistics were developed to describe the
demographic characteristics of the study population and participants. T-tests and chi-square tests
M AN U
were used to test for differences in demographic and health-related characteristics between study enrollees and those not enrolled and between risk strata for those enrolled. For our primary objective, general linear models were used to examine the relationship between the HBI score at one week post-injury and risk strata, adjusting for age, gender, pre-injury HBI, and any of the
TE D
premorbid conditions. A repeated measures analysis using a mixed models approach that included a risk stratum by time interaction term was used to evaluate the association of the trajectory of the HBI score over three time points – week 1, 2, and 4, with the risk strata,
EP
adjusting for age, gender, pre-injury HBI, and any of the premorbid conditions. The relationship between the HBI subscores and risk strata were also calculated. Post hoc, we used general linear
AC C
models to assess the relationship between the HBI score at one week and each of the seven individual factors, the total number of symptoms, and the presence of one of the premorbid conditions. We performed all analysis using SAS/STAT software (version 9.3, SAS Institute Inc. Cary, NC). P-values of <0.05 were considered statistically significant.
RESULTS
ACCEPTED MANUSCRIPT 13 Of the 127 patients enrolled, 120 patients completed study measures in the ED. Details of patient recruitment and retention are presented in Figure 1. Study patients were younger by approximately one year (11.0 vs. 11.9 years, p=0.05) and were more likely to be of white race
RI PT
(75.8 vs. 68.1%, p=0.02) compared to those who declined participation or were unable to be enrolled in the ED. Fourteen participants were lost to follow-up at one week, 27 at two weeks, and 34 at four weeks. Those lost to follow-up were similar in terms of age, race, sex, and history
SC
of headaches, history of concussions, or presence of attention or learning disability when
compared to the group that completed follow-up at one week. Average days to follow-up were
M AN U
6.4 (SD 1.4) at one week, 13.5 (SD 1.6) at two weeks, and 26.7 (SD 1.6) at four weeks. Frequency of youth as opposed to parent unique respondents of the HBI were 79/120 at one week, 60/106 at one week, 57/93 at two weeks, and 48/89 at 4 weeks. Characteristics of the study population are shown in Table 2. Mechanisms of injury
TE D
amongst the participants were as follows: sports (20.8%) fall from standing height (20.0%), bicycle collision (11.7%) and other mechanisms (47.5%). Of the 120 study participants, 46 participants were categorized by the PECARN rule as very low-risk, 39 as intermediate-risk and
EP
35 as at-risk (Figure 1). Across the three risk strata, there were no statistically significant differences in age, gender, race, history of headaches, history of concussion, or presence of
AC C
attention or learning disability. Distribution of each of the PECARN risk variables by risk category are presented in Table 3. Within the cohort, a total of 34 participants underwent CT imaging of the head, six had traumatic intracranial injury (all six had an intracranial hemorrhage or contusion, and one participant had concomitant cerebral edema), and none had a ciTBI. Unadjusted mean HBI score pre-injury for all participants was 12.8 (SD 8.9). Unadjusted mean HBI scores of all participants in the ED and at one, two and four weeks post-injury were
ACCEPTED MANUSCRIPT 14 18.2 (SD 11.9), 15.4 (SD 11.0), 12.0 (SD 10.7), and 9.6 (SD 9.6), respectively. Mean total HBI scores and subscores adjusted for age, gender, history of headaches, history of concussions, presence of an attention or learning disability, and pre-injury HBI by risk category are listed in
RI PT
Table 4.
With respect to the primary objective, there was no significant association of the HBI score at one week with risk strata for ciTBI (p=0.17) after adjusting for age, gender, history of
SC
headaches, history of concussions, presence of an attention or learning disability, and pre-injury HBI. Similarly, there was no significant association of the unadjusted HBI score at one week
M AN U
with risk strata for ciTBI (p=0.10). While the unadjusted (p=0.04) mean cognitive subscale score was significantly associated with risk strata at one week, the adjusted (p=0.09) cognitive subscale score showed a marginal association. Both adjusted (p=0.43) and unadjusted (p=0.39) mean somatic subscale scores were not significantly associated with risk strata at one week.
TE D
Counterintuitively, compared to the medium risk strata at one week, the low risk strata demonstrated higher mean scores for the total HBI score, as well as both cognitive and somatic subscale scores.
EP
The repeated measures analysis showed that there was a significant decline in the adjusted HBI score over the three follow-up time periods (1, 2, and 4 week) regardless of risk
AC C
stratum (p< 0.0001). The risk strata by time interaction was not significant indicating that the trajectories of the adjusted HBI score over time did not vary by risk strata (p=0.68). The results of the repeated measures analysis for the cognitive and somatic subscales were similar with significant declines over time (p< 0.0001 for each), but no significant differences in trajectories over time by risk strata (somatic, p=0.89, and cognitive, p=0.62).
ACCEPTED MANUSCRIPT 15 Since there was no association between the risk strata and outcome, post-hoc analysis was conducted to assess association between the unadjusted HBI score at one week and other potential factors. Of the 7 variables incorporated in the PECARN neuroimaging rule in our
RI PT
electronic health record, only one, severe mechanism of injury, was associated with the
unadjusted HBI score at one week (p=0.03). Amongst the participants, the total number of ciTBI risk factors documented ranged from 0 to a maximum of 3. There was no statistically significant
SC
difference between the unadjusted HBI scores at one week for participants who had 3 risk factors versus no risk factors (p=0.57). History of headaches was significantly associated with the
M AN U
unadjusted HBI score at one week (p=0.009). History of concussion (p=0.64) and or presence of attention or learning disabilities (p=0.33) were not associated with the unadjusted HBI score at one week.
TE D
DISCUSSION
Contrary to our hypothesis, the use of an existing clinical decision rule aimed at decreasing unnecessary CTs in children did not improve identification of patients presenting to the ED with
EP
blunt head trauma who will experience post-concussive symptoms. Neither symptom burden at one week post-injury nor over the one month follow-up period differed as a function of risk
AC C
categorization of ciTBI (very low, intermediate, and at-risk). Clinicians should not rely on the risk stratification of the PECARN neuroimaging rule to identify children at risk for persistent symptoms following mild TBI. Further research is needed to determine acute symptom measures or biomarkers that can assist with disposition planning for children with TBI who are at high risk of post-concussive symptoms.
ACCEPTED MANUSCRIPT 16 Since the majority of the variables in the PECARN rule have been associated with postconcussive symptoms, the lack of association between post-concussive symptoms and these acute presenting factors when categorized into risk strata of ciTBI or in isolation in this cohort
RI PT
suggests that there may be other prognostic factors. Despite over 16 variables assessed to derive this rule, only 7 variables were predictive of ciTBI that also met strict Standards for the
Reporting of Diagnostic Accuracy criteria including moderate inter-observer agreement (kappa
SC
>0.05) agreement. The limited number of variables in the decision rule and the hierarchal
stratification may limit its utility for identifying children at risk of ongoing symptoms. As a
M AN U
single factor, only severe mechanism of injury was associated with post-concussive symptom burden at one week, yet this has not been previously associated with persistent symptoms as it was defined in this study. Yeates et al found that children were more likely to have prolonged post-concussive symptoms if they (a) had a higher severity of injury based on modified injury
TE D
severity score, GCS <15 and signs of altered mental status, or (b) presented with a constellation of vomiting, loss of consciousness, and headache.9,15 Within the PECARN rule, the presence of either GCS of 14 or signs of altered mental status alone places a patient into the at-risk category.
EP
Likewise, the constellation of vomiting, loss of consciousness, and headache places participants into the intermediate-risk category in the rule. Stratifying these symptoms as delineated by the
AC C
rule did not equate to a greater post-concussive symptom burden at one week. The rule does not include many of acute signs and symptoms such as dizziness and amnesia that have been shown to predict post-concussive symptoms.9,12,26 Furthermore, prior literature suggests premorbid factors such as gender, age, prior concussion, prior headaches, baseline cognitive functioning, and coping strategies not captured in the rule may be more predictive of post-concussive symptoms than acute injury related signs and symptoms. 4,12,27-29 In the present study, history of
ACCEPTED MANUSCRIPT 17 headaches was the only premorbid factor associated with symptom burden at one week, as has been found in numerous other studies.7,12,14 An attempt to control for these premorbid issues was made by adjusting for those that were assessed in this study which included age, gender and
RI PT
history of headaches, prior concussions, and attention and learning disabilities. Although there was an association between the risk strata and the unadjusted HBI cognitive subscore, this
association was no longer statistically significant after adjusting for premorbid issues, suggesting
SC
that premorbid issues warrant inclusion in future studies to assess the effect on cognitive functioning.
M AN U
No single or set of acute signs, symptoms, neurocognitive tests, blood/urine/cerebrospinal biomarkers or radiological indices have demonstrated high negative or positive predictive value for detection of post-concussive symptoms in adults or children.30,31 Concussion screening tools such as the Acute Concussion Evaluation – Emergency Department (ACE-ED) or the Sport
TE D
Concussion Assessment Tool (SCAT3) are more comprehensive diagnostic tools which assess 18 and 22 acute symptoms, respectively, as well as injury characteristics, cognitive function (in SCAT3), and premorbid risk factors associated with post-concussive symptoms.32-34 As a result,
EP
these tools may have prognostic capacity, yet to our knowledge this has not been reported. Total acute symptom burden has been associated with post-concussive symptoms in most studies,
AC C
however, a few studies, including the present study, have failed to find such association.9,35,36 While computerized neurocognitive tests designed to assess recovery from concussion can detect deficits acutely in the ED, there is contradictory evidence about their prognostic ability.36-38 Currently, there is a prospective, multicenter study being conducted in Canadian pediatric EDs to derive a clinical prediction rule for the development of persistent post-concussive symptoms in
ACCEPTED MANUSCRIPT 18 children and adolescents following acute head injury using clinically available factors from the ED.39 This pilot study has several limitations. Due to the enrollment strategy of continuous
RI PT
enrollment regardless of risk strata, selection bias may have occurred because the very low-risk category completed enrollment prior to completion of the other risk strata. Misclassification bias is a second limitation of the study. Despite inter-rater reliability of at least 50% for each factor
SC
contained in the rule during its development, the factors within the rule are primarily subjective signs and symptoms leading to increased likelihood of misclassification. For instance, GCS, a
M AN U
key factor in the rule, is one of the most widely used indicators of mental status in TBI, yet interrater reliability of this factor in other cohorts has been reported to be less than 50%.40 Misclassification of another key variable, “other signs of altered mental status” may have occurred because each clinician may have applied their own definition of altered mental status,
TE D
as opposed to the way this variable was derived in the formation of the PECARN rule which asked four separate questions about the presence of agitation, slow to respond to questions, repetitive questioning or somnolence. In the electronic health record at our site, these factors
EP
were annotated in small print under the variable altered mental status. Misclassification of either one of these variables alone could have changed a participant’s risk categorization. Establishing
AC C
inter-rater reliability may have reduced the risk of misclassification. About halfway through the study in December 2012, clinical decision support specifying the risk category was available to the clinician in the electronic health record based on the responses to the rule. No overt disclosure about the risk strata were conveyed by study staff to the physician or the patient; however prior knowledge of the rule and risk strata, as well as obvious risk categorization in the electronic health record after implementation of decision support, may have led to altered care
ACCEPTED MANUSCRIPT 19 and guidance based on risk stratification. Effects of different post-injury treatments were not accounted for. Post-concussive symptoms were measured using the HBI administered via the phone or email, yet this tool has not been validated for this mode. The variation in respondent
RI PT
may have affected the rating of symptom severity, although modest parent-child agreement on the HBI has been demonstrated.22 Post-concussive symptoms were only assessed using the HBI and we did not utilize other measures of disability associated with post-concussion syndrome
SC
such as balance testing or computerized neurocognitive testing. Incomplete follow-up may have
M AN U
affected outcome measurements and ultimately affected results.
CONCLUSION
Stratification of the risk of ciTBI based on the PECARN neuroimaging rule is not predictive of post-concussive symptoms burden following acute mild TBI. Thus physicians
TE D
should be aware that patient’s risk of ciTBI is not helpful in identifying those patients who will develop post-concussive symptoms. This study indicates that a single rule is inadequate to serve the dual purposes of assisting with imaging decisions and stratifying risk of persistent symptoms.
EP
Further research is needed to develop a clinical decision rule to identify patients at risk of prolonged post-concussive symptoms during their acute visit for head trauma in order to
AC C
appropriately triage patients for who are in need of ongoing care, as well as stratify patients eligible for possible treatments.
ACCEPTED MANUSCRIPT 20 REFERENCES 1.
Faul M, Xu L, Wald MW, Coronado VG. Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations, and Deaths 2002-2006. Atlanta, GA:
RI PT
Centers for Disease Control and Prevetnion, National Center for Injury Prevention and Control;2010. 2.
Barlow KM, Crawford S, Stevenson A, Sandhu SS, Belanger F, Dewey D. Epidemiology
SC
of postconcussion syndrome in pediatric mild traumatic brain injury. Pediatrics. 2010;126(2):e374-381.
Mittenberg W, Wittner MS, Miller LJ. Postconcussion syndrome occurs in children.
M AN U
3.
Neurosyschol. 1997;11(3):447-452. 4.
Ponsford J, Willmott C, Rothwell A, et al. Cognitive and behavioral outcome following mild traumatic head injury in children. J Head Trauma Rehabil. 1999;14(4):360. Yeates KO, Luria J, Bartkowski H, Rusin J, Martin L, Bigler ED. Postconcussive
TE D
5.
symptoms in children with mild closed head injuries. J Head Trauma Rehabil. 1999;14(4):337-350.
Eisenberg MA, Meehan WP, 3rd, Mannix R. Duration and course of post-concussive
EP
6.
symptoms. Pediatrics. 2014;133(6):999-1006. Carroll LJ, Cassidy JD, Peloso PM, et al. Prognosis for mild traumatic brain injury:
AC C
7.
results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J
Rehabil Med. 2004(43 Suppl):84-105.
8.
Sroufe NS, Fuller DS, West BT, Singal BM, Warschausky SA, Maio RF. Postconcussive
symptoms and neurocognitive function after mild traumatic brain injury in children. Pediatrics. 2010;125(6):e1331-1339.
ACCEPTED MANUSCRIPT 21 9.
Yeates KO, Taylor HG, Rusin J, et al. Longitudinal trajectories of postconcussive symptoms in children with mild traumatic brain injuries and their relationship to acute clinical status. Pediatrics. 2009;123(3):735-743. Eisenberg MA, Andrea J, Meehan W, Mannix R. Time interval between concussions and
RI PT
10.
symptom duration. Pediatrics. 2013;132(1):8-17. 11.
Kuppermann N, Holmes JF, Dayan PS, et al. Identification of children at very low risk of
SC
clinically-important brain injuries after head trauma: a prospective cohort study. Lancet. 2009;374(9696):1160-1170.
Zemek RL, Farion KJ, Sampson M, McGahern C. Prognosticators of Persistent
M AN U
12.
Symptoms Following Pediatric Concussion: A Systematic Review. JAMA Pediatr. 2013:1-7. 13.
Babcock L, Byczkowski T, Wade SL, Ho M, Mookerjee S, Bazarian JJ. Predicting
TE D
postconcussion syndrome after mild traumatic brain injury in children and adolescents who present to the emergency department. JAMA Pediatr. 2013;167(2):156-161. 14.
Hung R, Carroll LJ, Cancelliere C, et al. Systematic review of the clinical course, natural
EP
history, and prognosis for pediatric mild traumatic brain injury: results of the International Collaboration on Mild Traumatic Brain Injury Prognosis. Arch Phys Med
15.
AC C
Rehabil. 2014;95(3 Suppl):S174-191. Taylor HG, Dietrich A, Nuss K, et al. Post-concussive symptoms in children with mild
traumatic brain injury. Neurosyschol. 2010;24(2):148-159.
16.
Collins MW, Field M, Lovell MR, et al. Relationship between postconcussion headache
and neuropsychological test performance in high school athletes. Am J Sports Med. 2003;31(2):168-173.
ACCEPTED MANUSCRIPT 22 17.
Pellman EJ, Viano DC, Casson IR, et al. Concussion in professional football: repeat injuries--part 4. Neurosurg. 2004;55(4):860-873; discussion 873-866.
18.
Guskiewicz KM, McCrea M, Marshall SW, et al. Cumulative effects associated with
RI PT
recurrent concussion in collegiate football players: the NCAA Concussion Study. JAMA. 2003;290(19):2549-2555. 19.
Collins MW, Lovell MR, Iverson GL, Cantu RC, Maroon JC, Field M. Cumulative
SC
effects of concussion in high school athletes. Neurosurg.51(5):1175-1179; discussion 1180-1171.
Ayr LK, Yeates KO, Taylor HG, Browne M. Dimensions of postconcussive symptoms in
M AN U
20.
children with mild traumatic brain injuries. J Int Neuropsychol Soc. 2009;15(1):19-30. 21.
Moran LM, Taylor HG, Rusin J, et al. Do postconcussive symptoms discriminate injury severity in pediatric mild traumatic brain injury? J Head Trauma Rehabil.
22.
TE D
2011;26(5):348-354.
Hajek CA, Yeates KO, Taylor HG, et al. Agreement between Parents and Children on Ratings of Post-Concussive Symptoms Following Mild Traumatic Brain Injury. Child
23.
EP
Neuropsychol. 2010;17(1):17-33.
McCrory P, Meeuwisse WH, Aubry M, et al. Consensus statement on concussion in
AC C
sport: the 4th International Conference on Concussion in Sport held in Zurich, November 2012. Br J Sports Med. 2013;47(5):250-258.
24.
Lovell MR, Iverson GL, Collins MW, et al. Measurement of symptoms following sports-
related concussion: reliability and normative data for the post-concussion scale. Appl Neuropsychol. 2006;13(3):166-174.
ACCEPTED MANUSCRIPT 23 25.
NINDS Commond Data Elements. Traumatic Brain Injury. 2015; http://www.commondataelements.ninds.nih.gov/tbi.aspx#tab=Data_Standards. Accessed 4/1/2015. Lau BC, Kontos AP, Collins MW, Mucha A, Lovell MR. Which on-field signs/symptoms
RI PT
26.
predict protracted recovery from sport-related concussion among high school football players? Am J Sports Med. 2011;39(11):2311-2318.
Fay TB, Yeates KO, Taylor HG, et al. Cognitive reserve as a moderator of
SC
27.
postconcussive symptoms in children with complicated and uncomplicated mild
28.
M AN U
traumatic brain injury. J Int Neuropsychol Soc. 2010;16(1):94-105.
Ganesalingam K, Yeates KO, Ginn MS, et al. Family burden and parental distress following mild traumatic brain injury in children and its relationship to post-concussive symptoms. J Pediatr Psychol. 2008;33(6):621-629.
Kelly A M, Barbara B, Ann D, et al. Injury versus noninjury factors as predictors of
TE D
29.
postconcussive symptoms following mild traumatic brain injury in children. Neurosyschol. 2013;27(1):1-12.
Zemek RL, Farion KJ, Sampson M, McGahern C. Prognosticators of Persistent
EP
30.
Symptoms Following Pediatric Concussion: A Systematic Review. JAMA Pediatr.
31.
AC C
2013;167(3):259-265.
Carroll LJ, Cassidy JD, Holm L, Kraus J, Coronado VG. Methodological issues and
research recommendations for mild traumatic brain injury: the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J Rehabil Med. 2004(43 Suppl):113125.
ACCEPTED MANUSCRIPT 24 32.
SCAT3. Br J Sports Med. 2013;47(5):259. http://www.ncbi.nlm.nih.gov/pubmed/?term=br+j+sports+med+2013%3B47%3A5+259.
33.
Child SCAT3. Br J Sports Med. 2013;47(5):263.
RI PT
Accessed 4/1/2015.
http://www.ncbi.nlm.nih.gov/pubmed/?term=br+j+sports+med+2013%3B47%3A5+263. Accessed 4/1/2015.
Zuckerbraun NS, Atabaki S, Collins MW, Thomas D, Gioia GA. Use of Modified Acute
SC
34.
2014;133(4):635-642. 35.
M AN U
Concussion Evaluation Tools in the Emergency Department. Pediatrics.
Grubenhoff JA, Deakyne SJ, Brou L, Bajaj L, Comstock RD, Kirkwood MW. Acute concussion symptom severity and delayed symptom resolution. Pediatrics. 2014;134(1):54-62.
Hang B, Babcock L, Hornung R, Ho M, Pomerantz WJ. Can Computerized
TE D
36.
Neuropsychological Testing in the Emergency Department Predict Recovery for Young Athletes With Concussions? Pediatr Emerg Care. 2015;31(10):688-693. Brooks BL, Khan S, Daya H, Mikrogianakis A, Barlow KM. Neurocognition in the
EP
37.
emergency department after a mild traumatic brain injury in youth. J Neurotrauma.
38.
AC C
2014;31(20):1744-1749.
Thomas DG, Collins MW, Saladino RA, Frank V, Raab J, Zuckerbraun NS. Identifying
neurocognitive deficits in adolescents following concussion. Acad Emerg Med. 2011;18(3):246-254.
ACCEPTED MANUSCRIPT 25 39.
Zemek R, Osmond MH, Barrowman N. Predicting and preventing postconcussive problems in paediatrics (5P) study: protocol for a prospective multicentre clinical prediction rule derivation study in children with concussion. BMJ open. 2013;3(8).
RI PT
Gill MR, Reiley DG, Green SM. Interrater reliability of Glasgow Coma Scale scores in
EP
TE D
M AN U
SC
the emergency department. Ann Emerg Med. 2004;43(2):215-223.
AC C
40.
ACCEPTED MANUSCRIPT 26
RI PT
Table 1. Factors and Risk Stratification of ciTBI in the PECARN Neuroimaging Rule Sign and Symptoms in the Risk Stratification PECARN Neuroimaging Rule11 Very low-risk Intermediate-risk At-risk of ciTBI of ciTBI of ciTBI (<0.05%) (0.9%) (4.3%)
AC C
EP
TE D
M AN U
SC
Glasgow Coma Scale <15* X Altered mental status* X Signs of basilar skull fracture X Loss of consciousness X Severe mechanism of injury X Vomiting X Headache X Legend: ciTBI: clinically important traumatic brain injury *Glasgow Coma Scale and altered mental status represent as a single factor “altered mental status” in the published neuroimaging rule; however, they are asked as two separate questions in the electronic health version of the rule implemented at our institution Altered mental status: agitation, sleepiness, slow responses, or repetitive question Signs of basilar skull fracture: retro-auricular bruising, periorbital bruising, hemotympanum, cerebral spinal fluid otorrhea, or cerebral spinal fluid rhinorrhea Severe mechanism of injury: motor vehicle crash with patient ejection, death of another passenger, or rollover; pedestrian or bicyclist without helmet struck by a motorized vehicle; falls of more than 1·5 meters (5 feet); or head struck by a high-impact object
ACCEPTED MANUSCRIPT 27 Table 2. Characteristics of Study Participants Participants (n=120) 11.0 ± 3.4 70.0
AC C
EP
TE D
M AN U
SC
RI PT
Age (Mean years ± SD) Gender (% male) Race White (%) 75.8 Black (%) 16.7 Other (%) 7.5 Premorbid conditions History of headaches (%) 22.5 History of concussion (%) 17.7 Attention and learning disabilities (%) 39.2 Legend: SD: standard deviation Attention and learning disabilities: attention deficit hyperactivity disorder, special education classes, and speech therapy, and other learning disabilities, repeated school year, autism, special education classes
ACCEPTED MANUSCRIPT 28 Table 3. Presence of PECARN ciTBI Factors by Risk Groups Very low-risk Intermediate-risk (n=39) (n= 46) (%)† (%)†
At-risk (n=35) (%)†
AC C
EP
TE D
M AN U
SC
RI PT
Loss of Consciousness 0.0 61.5 48.7 Vomiting 0.0 23.1 28.6 Mechanism of injury Mild 76.1 64.1 45.7 Neither mild nor severe 23.9 15.4 28.6 Severe 0.0 20.5 25.7 Current headache None 55.6 33.3 17.1 Mild 33.3 43.6 40.0 Moderate 11.1 15.4 22.9 Severe 0.0 7.7 11.4 Unable to obtain 0.0 0.0 8.6 Glasgow Coma Scale < 15 0.0 0.0 31.4 Other signs of altered mental status 0.0 0.0 85.7 Signs of basilar skull fracture 0.0 0.0 5.7 CT obtained 2.2 30.8 60.0 Presence of ciTBI, n=0 0 0 0 Presence of intracranial injury, n=6 0 0.017 0.033 Legend: † Column percentage ciTBI: clinically important traumatic brain injury CT: cranial computed tomography Mechanism of injury: mild: ground-level falls or running into stationary objects severe: motor vehicle crash with patient ejection, death of another passenger, or rollover; pedestrian or bicyclist without helmet struck by a motorized vehicle; falls of more than 1·5 meters (5 feet); or head struck by a high-impact object Signs of basilar skull fracture: retro-auricular bruising, periorbital bruising, hemotympanum, cerebral spinal fluid otorrhea, or cerebral spinal fluid rhinorrhea Other signs of altered mental status: agitation, sleepiness, slow responses, or repetitive questions
ACCEPTED MANUSCRIPT 29
AC C
EP
TE D
M AN U
SC
RI PT
Table 4. Adjusted* mean Health and Behavior Inventory Score by PECARN Risk Strata associated with ciTBI Mean (95% Confidence Interval) HBI Scores Very low-risk Intermediate-risk At-risk of ciTBI of ciTBI of ciTBI ED Total 12.8 (9.10, 16.5) 17.8 (14.1, 21.5) 23.6 (19.6, 27.7) Cognitive 7.8 (5.3, 10.3) 8.5 (6.0, 11.1) 11.3 (8.6, 14.1) Somatic 5.2 (3.2, 7.2) 9.4 (7.3, 11.4) 12.6 (10.4, 14.7) 1 week Total 17.1 (13.4, 20.8) 13.8 (9.9, 17.6) 18.1 (13.9, 22.2) Cognitive 11.0 (8.6, 13.5) 8.1 (5.5, 10.6) 11.0 (8.3, 13.7) Somatic 6.2 (4.4, 8.0) 5.8 (3.9, 7.7) 7.2 (5.2, 9.2) 2 weeks Total 13.6 (9.9, 17.4) 12.4 (8.6, 16.3) 16.3 (12.2, 20.5) Cognitive 9.0 (6.5, 11.5) 7.8 (5.3, 10.4) 9.9 (7.1, 12.6) Somatic 4.7 (3.0, 6.4) 4.7 (3.0, 6.5) 6.5 (4.6, 8.4) 4 weeks Total 10.1 (6.6, 13.6) 8.8 (5.1, 12.4) 12.9 (8.9, 16.9) Cognitive 7.0 (4.6, 9.5) 5.5 (2.9, 8.0) 7.8 (5.1, 10.5) Somatic 3.4 (1.8, 4.9) 3.6 (2.0, 5.2) 5.4 (3.7, 7.2) Legend: ciTBI: clinically important traumatic brain injury HBI: Health and Behavior Inventory *Mean HBI adjusted for age, gender, history of headaches, history of concussion, and presence of attention and learning disabilities.
ACCEPTED MANUSCRIPT
Figure 1. Study participant flow diagram.
RI PT
Assessed for eligibility (n=966)
SC
Not meeting inclusion/exclusion criteria (n=725) Eligible but not approached (n=35) Refused enrollment (n=79)
M AN U
Enrolled in study (n=127)
Intermediate-risk (n=39) 1 week f/u 87% (n=34) 2 week f/u 79% (n=31) 4 week f/u 69% (n=27)1
EP
Very low-risk (n=46) 1 week f/u 87% (n=40) 2 week f/u 76% (n=35) 4 week f/u 74% (n=34)1
TE D
Did not complete enrollment survey (n=6) Deemed ineligible after enrollment (n=1)
At-risk (n=35) 1 week f/u 91% (n=32) 2 week f/u 77% (n=27) 4 week f/u 71% (n=25)1
AC C
Legend: 1 10 patients had no symptoms at the 2 week follow-up and were not contacted for the 4 week follow-up visit. Their 4th week HBI values were carried over from week 2.