Age-Dependent Differences in Blood Levels of Glial Fibrillary Acidic Protein but Not Ubiquitin Carboxy-Terminal Hydrolase L1 in Children

Abstract

Objectives

Despite the growing evidence of the clinical utility of blood-brain biomarkers in adults with traumatic brain injury (TBI), less is known about the performance of these biomarkers in children. We characterize age-dependent differences in levels of ubiquitin carboxy-terminal hydrolase L1 (UCH-L1) and glial fibrillary acidic protein (GFAP) in children without TBI.

Methods

Plasma biobank specimens from children and adolescents aged 0–<19 years without TBI were obtained, and UCH-L1 and GFAP levels were quantified. The relationship between age and biomarker expression was determined using previously defined aged epochs (<3.5 years, 3.5 years to <11 years, 11 years and older), then biomarker levels were compared with established thresholds for ruling out the need for a head CT in adults with a mild TBI (mTBI) (UCH-L1 400 pg/mL, GFAP 35 pg/mL).

Results

The age range of the 366 control patients was 3 months–18 years. There was a significant negative association between age and GFAP but not UCH-L1. Only 1.4% of samples exceeded the UCH-L1 cutoff; however, 20% of samples exceeded the GFAP cutoff and 100% children younger than 3.5 years had values that exceeded the cutoff.

Discussion

Age seems to modify physiologic plasma GFAP levels. Diagnostic cutoffs for TBI based on GFAP but not UCH-L1 and may need to be adjusted in children younger than 11 years.

Introduction

Recently, the US Food and Drug Administration (FDA) approved the use of 2 blood biomarkers, ubiquitin carboxy-terminal hydrolase L1 ([UCH-L1], a neuronal marker involved in protein degradation) and glial fibrillary acidic protein ([GFAP], a structural protein in astroglial cells), for identifying adult patients with traumatic brain injury (TBI) likely to have negative findings on CT.¹ Less is known about the performance of these biomarkers in children, and neither biomarker has been approved for use in pediatric TBI.

There is a special need for blood-brain biomarkers in pediatric TBI because developmental differences make children difficult to assess using standard TBI assessments, yet children are also more sensitive to ionizing radiation such as CT imaging. For these reasons, blood biomarkers may offer particular utility in the evaluation of children with possible TBI. However, unlike adults where population norms for UCH-L1 and GFAP have been well established, normal levels for UCH-L1 and GFAP in children are not well described. This is especially important for TBI-related biomarkers because diagnostic thresholds in adults may not be applicable in children because of the marked developmental changes that occur in the brain, especially during early childhood.

Therefore, the objective of this study was to investigate age-dependent differences in GFAP and UCHL-1 in a population of children without TBI seeking care at a quaternary care facility, relevant to established diagnostic thresholds for TBI in adults.

Methods

Blood samples were obtained from the Boston Children's Hospital (BCH) PrecisionLink Biobank,² in which all BCH patients are eligible to enroll. The research study was approved by the Boston Children's Hospital Institutional Review Board. Research assistants obtain informed consent for all participants, which gives permission for (1) research use of electronic health record (EHR) data including demographics, dates and types of encounters (e.g., ED visit, hospitalization), clinic and hospital discharge notes, diagnoses, laboratory orders and values, physical examination findings, medications prescribed and procedures performed; (2) use of residual specimens collected in the process of routine clinical care, including blood and plasma; and (3) optional additional collection of a research tube during a future clinical laboratory draw.

Venous blood samples are collected in 2 ways:

1. For participants who consent to a research blood draw, the Biobank collects a 4-mL blood sample in an EDTA tube, which is centrifuged at 2,000×g for 10 minutes at room temperature. Plasma is then aliquoted into 0.5-mL microcentrifuge tubes and stored at −80C in the Biobank Core Lab facility.

2. For the use of residual samples, blood samples drawn in EDTA tubes are identified through automated reports run in the EHR system and collected by Biobank staff and then frozen at −80°C in the Biobank Core Lab facility.

Biobank specimens are tracked and available for deidentified browsing using a customized version of the STARLIMS laboratory information management system (STARLIMS. Computer Software, version 10.06; Abbott Laboratories: Abbott Park, IL). To identify control patients for this study, the Biobank database was queried for participants who met the following inclusion criteria:

1. Age younger than 19 years

2. No history of traumatic brain injury or concussion <12 months ago based on International Classification of Diseases, Tenth Revision (ICD-10) coding

3. No recent admission/ED visit for trauma in past 12 months

4. No concurrent infectious illness at the time of sample collection (ICD-10 coding)

5. No complex chronic condition based on International Classification of Diseases, Tenth Revision (ICD-10) coding³

Using this query, 366 samples banked between February 2017 and May 2023 were identified for shipment; coded samples were then shipped overnight on dry ice in a single batch.

Samples underwent one freeze-thaw cycle and then were analyzed using the Alinity i TBI test, which is a panel of chemiluminescent microparticle immunoassays for the measurement of GFAP and UCH-L1 in plasma and serum. Two samples were not analyzed because of inadequate volume. The analytical measuring interval is 6.1–42,000.0 pg/mL for GFAP and 26.3 to 25,000.0 pg/mL for UCH-L1. The coefficient of variation was 2.8%–5.3% CV for GFAP and 2.1%–5.6% CV for UCH-L1. All samples were tested without dilution and in singlicate. Technicians performing biomarker measurements were blinded to clinical outcome data.

Simple descriptive statistics were used to describe the demographics of the study population. Based on the age partitions derived in the Canadian Laboratory Initiative on Pediatric Reference Intervals (CALIPER) study (<3.5 years, 3.5 to <11 years, 11 to <19 years),⁴ we conducted a spline-based regression using robust (heteroskedastic) standard errors with biomarker concentration as the outcome and age as the predictor, controlling for sample storage duration (acquisition to analysis). In addition, cutoffs were set for GFAP and UCH-L1 of 35.0 pg/mL and 400.0 pg/mL, respectively, derived from studies in the TBI literature relevant to CT-positive TBI in adults.⁵ All analyses were performed using STATA (College Station, TX). The anonymized patient data are not being publicly shared as they are being used for the development of a clinical trial.

Results

The age range of control patients was 3 months–18 years (median 13 years, interquartile range 9 years–16 years), and 55% (n = 202) of participants were female.

There were no sex-based differences in biomarker levels for UCH-L1 and GFAP (p = 0.89 and p = 0.80, respectively). GFAP but not UCH-L1 levels decreased with increasing age (Figure). For UCH-L1, only 5 control patients exceeded the threshold of 400 pg/mL, of whom 3 were older than 11 years. Using a cutoff of 35 pg/mL for GFAP, 100% of children younger than 3.5 years (n = 18), 40% of children 3.5–11 years (n = 46), and 3% of children 11 years and older (n = 8) exceeded this cutoff.

Figure. Age-Related Trends in Serum Levels of UCH-L1 and GFAP, Controlling for Time From Sample Collection.

Panel A: Age related decreases in UCH-L1 levels were not significant across the 3 age epochs (B = −74.8 [95% CI −155.3 to 5.6] p = 0.07 for age <3.5 years; B = −7.8 [95% CI −16.4 to 0.9], p = 0.08 for 3.5 ≤ age < 11; B = −3.2 [95% CI −7.2 to 0.8], p = 0.11 for age ≥ 11). Panel B: Age-related decreases in GFAP levels were significant for children <3.5 (B = −16.7 [95% CI −31.0 to 2.5] p = 0.02) and 3.5 to less than 11 years (B = −5.8 [95% CI −7.7 to −3 0.9], p < 0.001 for but not children older than 11 (B = −0.4 [95% CI −1.2 to 0.3], p = 0.3).

Discussion

In this investigation, we found age-dependent differences in GFAP but not UCHL1 in a population of children without TBI enrolled in a hospital-based biobank. Similar age-dependent trends were reported in the CALIPER cohort, which used the Quanterix HD-1 GFAP Discovery Kit, which is not currently approved by the FDA for clinical use.⁴ Using a platform already approved by the FDA for adult patients with TBI, our study investigates UCHL-1 and GFAP levels in children without TBI relative to established thresholds of diagnosing TBI in adult populations. We found that diagnostic thresholds for TBI used in adult populations have the potential to misclassify children because of developmental differences in biomarker norms.

Our data are important to inform studies to refine thresholds for GFAP and UCH-L1 in excluding the need for CT in children with TBI. Such studies are especially important in pediatric TBI, where blood biomarkers may reduce the utilization of CT, which carries risk of exposure to ionizing radiation.⁶ Moreover, despite the availability of excellent decision rules for head CT in children,⁷ CT utilization remains fairly high in children presenting to the emergency department with minor head trauma.⁸ Special considerations in the pediatric population—including the possibility of abusive head trauma in preverbal children—warrant scrupulous attention to biomarker thresholds across diverse mechanisms of injury.

There are several limitations to this study. First, it is possible that patients had medical conditions not documented in our system. Second, blood samples were collected and processed per biobank protocols, not using protocols optimized for UCH-L1 and GFAP analyses. Finally, given the age distribution, the sample may not fully represent the potential variations in younger children.

Taken together, our study demonstrates age-dependent differences in plasma levels of GFAP but not UCH-L1, suggesting the need for careful interpretation of appropriate biomarker-specific cutoffs across the developmental spectrum both for TBI and for other neurologic diseases in which biomarkers are being evaluated.

Appendix. Authors

Name	Location	Contribution
Rebekah Mannix, MD, MPH	Division of Emergency Medicine, Boston Children's Hospital, Harvard Medical School, MA	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data
Erin Borglund, MPH	Computational Health Informatics Program (CHIP), Boston Children's Hospital, MA	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data
Alexandra Monashefsky, BA	Computational Health Informatics Program (CHIP), Boston Children's Hospital, MA	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data
Christina Master, MD	Division of Sports Medicine, Children's Hospital of Philadelphia, PA	Drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data
Daniel Corwin, MD, MSCE	Division of Emergency Medicine, Children's Hospital of Philadelphia, PA	Drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data
Mohamed Badawy, MD	Division of Emergency Medicine, UT Southwestern Medical Center, Dallas, TX	Drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data
Danny G. Thomas, MD, MPH	Division of Emergency Medicine, Medical College of Wisconsin, Milwaukee	Drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data
Andrew Reisner, MD	Department of Neurosurgery, Children's Hospital of Atlanta, GA	Drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data

Study Funding

The work is funded in part by a grant to Boston Children's Hospital (Dr. Mannix) from Abbott Laboratories. The authors acknowledge material and/or data support from the PrecisionLink Biobank for Health Discovery at Boston Children's Hospital. This work was supported in part by Cooperative Agreement U01TR002623 from the National Center for Advancing Translational Sciences/NIH and the PrecisionLink Project at Boston Children's Hospital.

Disclosure

R. Mannix reports funding from Abbott Laboratories, the National Football League, the Department of Defense, and the NIH. C. Master has received funding from the Centers for Disease Control, the NIH, the Department of Defense, Pennsylvania Department of Health, the American Medical Society for Sports Medicine, the Children's Hospital of Philadelphia Frontier Program, the Chuck Noll Foundation for Brain Injury Research, and the Toyota Way Forward Fund. M. Badawy has received funding from the Centers for Disease Control, the NIH, EMSC-Network Development Demonstration Project, and the Health Resources & Services Administration. D. Thomas has received funding from the Centers for Disease Control, the NIH, the Clinical and Translational Science Institute of Southeastern Wisconsin, and the Children's (Wisconsin) Research Institute. The other authors report no relevant disclosures. Go to Neurology.org/N for full disclosures.