Factors related to adverse long-term outcomes after mild traumatic brain injury in children: a scoping review

Objective To identify demographic, premorbid and injury-related factors, or biomarkers associated with long-term (≥3 months) adverse outcomes in children after mild traumatic brain injury (mTBI). Design Scoping review of literature. Patients Children and adolescents with mTBI. Risk factors Any demographic, premorbid and injury-related factors, or biomarkers were included. We excluded genetic and treatment-related factors. Main outcome measures Postconcussion syndrome (PCS), recovery. Results Seventy-three publications were included, reporting 12 long-term adverse outcomes, including PCS in 12 studies and recovery in 29 studies. Additional outcomes studied were symptom scores/severity (n=22), quality of life (n=9) and cognitive function (n=9). Forty-nine risk factors were identified across studies. Risk factors most often assessed were sex (n=28), followed by age (n=23), injury mechanism = (n=22) and prior mTBI (n=18). The influence of these and other risk factors on outcomes of mTBI were inconsistent across the reviewed literature. Conclusions The most researched risk factors are sex, age and mechanism of injury, but their effects have been estimated inconsistently and did not show a clear pattern. The most studied outcomes are recovery patterns and symptom severity. However, these may not be the most important outcomes for clinicians and patients. Future primary studies in this area should focus on patient-important outcomes. Population-based prospective studies are needed that address prespecified hypotheses on the relationship of risk factors with given outcomes to enable reliable prediction of long-term adverse outcomes for childhood mTBI.


Eligibility criteria
We included observational studies that reported factors potentially associated with long-term adverse outcomes of mTBI. Using the World Health Organisation definition of childhood, we included studies reporting data on patients aged 0-18 years (14), and included any author-defined mTBI. Studies that assessed only children with clinically important TBI (ciTBI) with structural brain injury (15), or intentional/abusive TBI, were excluded. These injuries are often of greater severity with set management pathways, unlike unintentional mTBI which is expected to resolve spontaneously. Factors of interest were developed through consultation with clinicians who treat children with mTBI. These included demographics (e.g., socioeconomic status (SES), age, sex, ethnicity, parental education, family size, social services involvement, and any other parental or child characteristics), pre-morbid conditions (e.g., physical, sensory, behavioural, or learning difficulties prior to injury, or previous history of TBI), injury-related factors (e.g., mechanism, severity, Glasgow Coma Score (GCS), time and place of injury, and symptoms immediately post-injury), and biomarkers (blood biomarkers or imaging abnormalities). These factors were most relevant to the scope of our review, which aimed to create a list of factors to help clinicians pre-empt which children may need longer follow-up for PCS or other long-term adverse outcomes. We excluded genetic factors as these are hard to assess in emergency departments and difficult to modify, and treatment-related factors as pathways and options vary across healthcare systems and are better studied for effectiveness in intervention reviews.
Our primary outcome was presence of PCS three months or longer after mTBI. We included all PCS as defined by authors. We also included all author-defined recovery and other long-term adverse BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s) outcomes of relevance to patients and clinicians (physical, behavioural, psychological) reported at 3 months or longer post-mTBI (16). We included all author-defined outcomes due to the absence of agreed definitions, so as not to miss any studied risk factors across all reported domains.

Study selection and data charting
We assessed abstracts and relevant full-text articles for eligibility in duplicate. One author extracted data from included studies and another checked a 20% sample. Extracted data included study design, author, location, year of publication, participants, risk factors identified or studied, and outcome(s) (Supplementary material).
As we aimed to comprehensively identify risk factors for adverse outcomes rather than quantify associations or assess their validity, we did not extract estimates (e.g., odds ratios) of their associations with outcomes. Methodological quality of included studies was not assessed for the same reason.

Synthesis
We extracted data for all demographic, premorbid and injury-related factors, or biomarker factors affecting primary and secondary outcomes on a standard template. Data were charted, thematically categorised and analysed iteratively after discussion with subject experts (JM, ML, IW). Findings were summarised into a table of risk factors studied for each outcome, and in figures presenting long-term adverse outcomes and risk factors studied in literature.

Stakeholder consultation
We presented preliminary findings to stakeholders comprising two young people with a history of PCS and their parents, and three clinicians who treat children with mTBI. This dialogue helped refine the synthesis and add patients' perspectives to our findings.   (1) Prior mTBI (1) Child's IQ (1) Premorbid anxiety (2) Premorbid depression (2) Prior ill health/medical conditions (1) Prior headaches, pre-injury sleep problems, preinjury ability/ disability, learning disorder, preinjury academic achievement, parental education, and family history of migraine.