Hyperphosphorylated neurofilament NF-H as a predictor of mortality after brain injury in children

Objective

The aim of the study was to determine whether serum levels of hyperphosphorylated neurofilament NF-H correlate with severity of brain injury in children.

Introduction

Traumatic brain injury (TBI) is a major cause of death and disability worldwide, especially in children and young adults. TBI is difficult to assess and clinical examinations are of restricted value during the first hours and days after injury.

Methods

The study protocol and informed consent was approved by Ethics committee of the University Hospital Brno.
Forty-nine patients with TBI were enrolled into the prospective study. Venous blood samples were taken after admission and every 24 h for a maximum of 6 consecutive days. Serum pNF-H was measured by sandwich enzyme immunoassay for the quantitative measurement of human phosphorylated neurofilament H (BioVendor, Laboratorní medicína a.s., Czech Republic).
Clinical outcome was evaluated 6 months after TBI using Glasgow Outcome Scale (GOS) in all patients.

Results

GOS was taken as principal end-point for all predictive analyses, i.e. analyses that have been performed in order to assess the predictive potential of a set of patient’s and injury characteristics.
Limit of ELISA detection was 12 pg/ml.
Table 1 documents summary statistics of accessible initial characteristics of injured children. It is apparent that common descriptors like age, sex and weight are not related to the final GOS values.
Table 1 also indicates statistically significant relationship between some descriptors and a group of patients with GOS = 1:

• GCS ≤ 8 (all patients with GOS = 1 belong to this category)
• Marshall classification score > II (no patient with GOS=1 has Marshall score I or II)- chart 2
• Increased incidence of some type of injury, as diagnosed using CT examination:
  • intracranial bleeding, subdural hematoma, skull fracture, edema – chart 3
  • multiple CT findings, namely skull fracture in combination with some other injury type- chart 4

All the parameters which significantly distinguished patients with GOS=1 from the others (Table 1) serve as candidates for effective predictors.

The quantitative level of pNF-H remained significantly higher in the risk group (GOS = 1) in comparison with the other patients for the 2^nd – 4^nd day (Table 2). Ratio pNF-H in 2^nd – 6^nd day/initial pNF-H value was statistically significantly differentiated both compared group of patients (with GOS =1 and with GOS > 1) (Table 2– chart 1).

Levels of pNF-H was significantly higher in patients with diffuse axonal injury on initial CT scan (Table 3– chart 2). Increased pNF-H levels (2^nd day of hospitalization) was found in Marshal classification (Table 4– chart 3).
Single findings and multiple findings are significantly associated with risk score of GCS and Marshall classification score (Table 5). Except for this relationship, the GCS score appears to be rather independent on injury type.
Normal levels pNF-H in the pediatric population is unknown. Objective ROC analysis was identification of optimal cut-offs of proteins for prediction of GOS = 1. The initial level of pNF-H reached very low sensitivity and specificity, in 2nd day cut-off was 117.1 (pg/ml) (Table 6).

Conclusions

Although further, prospective study is warranted, these findings suggest, that levels of hyperphosphorylated neurofilament NF-H correlate with mortality and may be useful as predictors of outcome in children with TBI..

Key Words

children, brain injury, biomarkers, outcome, hyperphosphorylated neurofilament

Table 1:  Summary statistics of sample data ¹

Parameters	The whole sample	GOS > 1	GOS = 1	p2
Sample size	N=49	N=42	N=7
Age (months)	103 (8 – 204)	96 (11 – 204)	134 (8 – 171)	0.605
Weight (kg)	36 (4 – 120)	26 (4 – 90)	47 (8 – 120)	0.439
Sex (male)	N=26	N=23	N=3	0.559
GCS
GCS≤8	32	25	7
GCS=9	4	4	0
GCS≥10	13	13	0	0.037
Marshal classification
I+II	33	33	0
III+IV	6	3	3
V+VI	10	6	4	<0.001
Initial CT diagnostics
No findings	6	6	0	0.159
Single findings
Intracranial bleeding	23	17	6	0.021
Diffuse axonal injury	9	6	3	0.099
Subdural hematoma	16	11	5	0.022
Contusion	9	8	1	0.757
Skull fracture	20	14	6	0.008
Epidural hematoma	6	4	2	0.200
Subarachnoidal bleeding	27	21	6	0.063
Edema	23	16	7	0.001
Multiple findings
Scull fracture + other findings	18	12	6	0.004
Multiple bleeding findings	17	12	5	0.031
Other combinations	11	11	0	0.049
Days in hospital	6	6	5	0.492

¹ Categorical variables are summarized as N and percentage of a given category; continuous variables are described by median and by 5th – 95th percentile range (in parentheses)

² Statistical significance of difference between groups with GOS > 1 and GOS = 1 was analyzed by means of Mann-Whitney U test (continuous variables) and maximum likelihood chi-square (categorical variables)

Table 2. Summary statistics of NF-H values

Values of NF-H and sample size:	The whole sample	GOS > 1	GOS = 1	P 2
(patients:with GOS > 1 / with GOS=1)
Quantitative values of NF-H in pg/ml 1
Initial value (38 / 7)	12 (12 – 867)	12 (12 – 867)	36.1 (12 – 1 482)	0.246
2nd day (37 / 7)	12 (12 – 2 200)	12 (12 – 2 200)	151.9 (12 – 6 765)	0.027
3rd day (37 / 6)	13 (12 – 3 793)	12 (12 – 2 330)	221.6 (12 – 6 860)	0.019
4th day (34 / 6)	54.2 (12 – 2 761)	12 (12 – 979)	410 (12 – 7 845)	0.010
5th day (31 / 5)	106.7 (12 – 3 484)	94.5 (12 – 3 484)	729.8 (12 – 1 067)	0.091
6th day (30 / 5)	228.8 (12 – 3 607)	203.5 (12 – 3 289)	855.3 (12 – 3 607)	0.086

¹ Quantitative values are summarized by median and by 5^th – 95^th percentile range 

² Statistical significance of difference between GOS > 1 and GOS = 1 was analyzed by means of Mann-Whitney U test

Chart 1: Quantitative values of NF-H in pg/ml¹

Ratio pNF-H in 2nd – 6 nd day/initial pNF-H value was statistically significantly differentiated both compared group of patients (with GOS =1 and with GOS > 1)

Table 3. Initial CT diagnostics in relation to initial values of marker

CT examination	No of patients N 1	NF-H initial value (pg/ml) 2
Initial CT diagnostics
No findings	6	15.5    (12.0 – 90.9)
Single findings
Intracranial bleeding	23	12.0  (12.0 – 727.0)
Diffuse axonal injury	9	158.7  (12.0 – 867.0)*
Subdural hematoma	16	12.0  (12.0 – 1 482.0)
Contusion	9	12.0   (12.0 – 224.0)
Skull fracture	20	12.0 (12.0 – 727.0)
Epidural hematoma	6	12.0 (12.0 – 36.1)
Subarachnoidal bleeding	27	12.0 (12.0 – 1 442.0)
Edema	23	12.0 (12.0 – 867.0)
Multiple findings
Scull fracture + other findings	18	12.0 (12.0 – 1 482.0)
Multiple bleeding findings	17	12.0 (12.0 – 1 482.0)
Other combinations	11	12.0  (12.0 – 1 442.0)

¹ Diagnostic findings within the whole dataset

² Median and 5th – 95th percentile range

* statistically significant difference between groups of patients with and without given CT finding; analyzed by Mann-Whitney U test 

Chart 2: Initial CT diagnostics

Levels of pNF-H was significantly higher in patients with diffuse axonal injury on initial CT scan

Table 4: NF-H values stratified according to diagnostic kriteria

	NF-H initial value (μg/l) 1	NF-H 2nd day (µg/l) 1
GCS
≤ 8	12 (12.00 – 867.00)	12 (12.00 – 2 413.00)
> 8	12 (12.00 – 1 442.00)	12 (12.00 – 2 200.00)
p-value2	0.559	0.301
Marshal classification
I – II	12 (12.00 – 867.00)	12 (12.00 – 519.00)
III – VI	33.49 (12.00 – 1 482.00)	62.05 (12.00 – 6 765.00)
p-value2	0,062	0,004
Days in hospital
in hospital ≤  6 days	12 (12.00 – 1 442.00)	12 (12.00 – 2 200.00)
in hospital > 6 days	12 (12.00 – 867.00)	12 (12.00 – 2 413.00)
p-value2	0.607	0.688

¹ median and 5th – 95th percentile range

² statistical significance was analyzed by means of Mann-Whitney U test

Chart 3: NF-H values stratified according to diagnostic criteria

Increased pNF-H levels (2nd day of hospitalization) was found in Marshal classification.

Chart 4: Initial CT diagnostics in relation to GCS and Marshal classification

Table 5: Initial CT diagnostics in relation to GCS and Marshal classification

CT examination	No of patients	GCS ≤  8	Marshall classification III – VI
	N 1	N 2	N 2
Initial CT diagnostics
No findings	6	3	0
Single findings
Intracranial bleeding	23	17	11
Diffuse axonal injury	9	8	3
Subdural hematoma	16	11	9
Contusion	9	5	3
Skull fracture	20	11	11
Epidural hematoma	6	4	6
Subarachnoidal bleeding	27	18	12
Edema	23	17	12
Multiple findings
Scull fracture + other findings	18	9	11
Multiple bleeding findings	17	11	10
Other combinations	11	10	1

¹ Diagnostic findings within the whole dataset

² Diagnostic findings in the patients with GCS ≤ 8 or with Marshall classification > II

* statistically significant difference between groups of patients with and without given CT finding; analyzed by ML chi-square

Single findings and multiple findings are significantly associated with risk score of GCS and Marshall classification score.

Table 6: Identification of optimal cut-offs of proteins for prediction of GOS = 1

Potential predictors	Cut-off (µg/l)	Sensitivity	Specificity	AUC (95 % CI) 1	p
NF-H initial	134.0	42.9	84.2	0.641 (0.398 –0.884)	0.240
NF-H 2nd day	117.1	71.4	83.8	0.764 (0.541 – 0.988)	0.028

References

1. Kant R, Smith-Seemiller L, Isaac G, Duffy J. Tc-HMPAO SPECT in persistent postconcussion syndrome after mild head injury: comparison with MRI/CT. Brain Inj 1997;11: 115–124.
2. Schwartz SM, Duffy JY, Pearl JM, Nelson DP. Cellular and molecular aspects of myocardial dysfunction. Crit Care Med 2001;29:S214–219.
3. Haskins WE, Kobeissy FH, Wolper RA, Ottens AK. Rapid discovery of putative protein biomarkers of traumatic brain injury by SDS-PAGE-capillary liquid chromatography- tandem mass spectrometry. J. Neurotrauma 2005;22:629–644
4. Ingebrigtsen, T, Romner B. Biochemical serum markers for brain damage: a short review with emphasis on clinical utility in mild head injury. Restor Neurol Neurosci 2003;21:171–176.
5. Denslow N, Michel ME, Temple MD, Hsu CY. Application of proteomics technology to the field of neurotrauma. J. Neurotrauma 2003;20:401–407.
6. Žurek J, Bartlova L, Marek L, Fedora M. Serum S100B Protein as a molecular marker of severity in traumatic brain injury in children. Cesk Slov Neurol N 2010;1:37–44
7. Shaw G, Yang C, Zhang L, Cook P, Pike BR, Hill WD. Characterization of the bovine neurofilament NF-M protein and cDNA sequence and identification of in vitro and in vivo calpain cleavage sites, Biochem. Biophys. Res. Commun. 2004;325:619–625
8. Strong MJ, Strong WL, Jaffe H, Traggert B, Souper MM, Pant HC. Phosphorylation state of the native high-molecular-weight neurofilament subunit protein from cervical spinal cord in sporadic amyotrophic lateral sclerosis, J. Neurochem. 2001;76:1315–1325.
9. Teasdale G, Jennett B. Assessment of coma and impaired consciousness. A practical scale. Lancet 1974;2:81– 84.
10. Marshall LF, Marshall SB, Klauber MR, van Berkum -Clark M, Eisenberg HM, Jane JA, Luerssen TG, Marmarou A, Foulkes MA. A new classification of head injury based on computerized tomography. J.Neurosurg 1991; S14– S20.
11. Jennett B, Bond M. Assessment of outcome after severe brain damage. Lancet 1975;1:480– 484.
12. Anderson KJ, Scheff SW, Miller KM, Roberts KN, Gilmer LK, Yang C, Shaw G. The phosphorylated axonal form of the neurofilament subunit NF-H (pNF-H) as a blood biomarker of traumatic brain injury. J Neurotrauma 2008;1079–85.
13. Buki A, Povlishock JT. All roads lead to disconnection?–Traumatic axonal injury revisited. Acta Neurochir.(Wien) 2006;148:181–193.
14. Haqqani AS, Hutchison JS, Ward R, Stanimirovic DB. Biomarkers and diagnosis; protein biomarkers in serum of pediatric patients with severe traumatic brain injury identified by ICAT-LC-MS/MS. J Neurotrauma 2007;1:54–74
15. Stephen B Lewis, Regina A Wolper, Lynn Miralia, Cui Yang, Gerry Shaw. Detection of phosphorylated NF-H in the cerebrospinal fluid and blood of aneurysma subarachnoid hemorrhage patiens. Journal of Cerebral Blood Flow & Metabolism 2008;28:1261–1271
16. Anderson RE, Hansson LO, Nilsson O, Dijlai-Merzoug R, and Settergren G. High serum S100B levels for trauma patients without head injuries. Neurosurgery 2001;48:1255–1260.

Data sources

The data sources were taken from:

• Zurek J, Bartlová L, Fedora M. Hyperphosphorylated neurofilament NF-H as a predictor of mortality after brain injury in children. Brain Inj. 2011; 25(2): 221–6

Catalog Number	Species	Analyte	Assay	Regulatory	Format
RD191138300R	Human	Phosphorylated Neurofilament H	Sandwich ELISA, HRP-labelled antibody	RUO	96 wells (1 kit)

categories: Neural tissue damage markers