Early versus Late Profiles of Inflammatory Cytokines after Mild Traumatic Brain Injury and Their Association with Neuropsychological Outcomes

Abstract

Despite pre-clinical evidence for the role of inflammation in traumatic brain injury (TBI), there is limited data on inflammatory biomarkers in mild TBI (mTBI). In this study, we describe the profile of plasma inflammatory cytokines and explore associations between these cytokines and neuropsychological outcomes after mTBI. Patients with mTBI with negative computed tomography and orthopedic injury (OI) controls without mTBI were prospectively recruited from emergency rooms at three trauma centers. Plasma inflammatory cytokine levels were measured from venous whole–blood samples that were collected at enrollment (within 24 h of injury) and at 6 months after injury. Neuropsychological tests were performed at 1 week, 1 month, 3 months, and 6 months after the injury. Multivariate regression analysis was performed to identify associations between inflammatory cytokines and neuropsychological outcomes. A total of 53 mTBI and 24 OI controls were included in this study. The majority of patients were male (62.3%), and injured in motor vehicle accidents (37.7%). Plasma interleukin (IL)-2 (p = 0.01) and IL-6 (p = 0.01) within 24 h post-injury were significantly higher for mTBI patients compared with OI controls. Elevated plasma IL-2 at 24 h was associated with more severe 1-week post-concussive symptoms (p = 0.001). At 6 months, elevated plasma IL-10 was associated with greater depression scores (p = 0.004) and more severe post-traumatic stress disorder (PTSD) symptoms (p = 0.001). Plasma cytokine levels (within 24 h and at 6 months post-injury) were significantly associated with early and late post-concussive symptoms, PTSD, and depression scores after mTBI. These results highlight the potential role of inflammation in the pathophysiology of post-traumatic symptoms after mTBI.

Introduction

Mild traumatic brain injury (mTBI) constitutes the majority of annual TBI cases across the United States¹ and can result in persistent neuropsychological sequelae.² Prior studies have investigated anatomical^3-5 and functional changes^6,7 in the brain to explain cognitive deficits and post-traumatic symptoms associated with mTBI. Recently, there has been increased focus on blood biomarkers of neural injury to detect^8-10 and prognosticate mTBI.^11-13 The identification of blood biomarkers with prognostic value in mTBI may help us to better understand the pathological basis for persistent cognitive deficits and other symptoms, as well as explore therapeutic interventions.

Inflammation has been shown to have an important role in the pathophysiology of secondary brain injury after TBI.¹⁴ The majority of prior studies have evaluated blood and cerebrospinal fluid (CSF) inflammatory markers such as interleukins (ILs), and have shown a link between these markers and outcome after severe TBI.^15–19 There are limited studies on peripheral blood inflammatory markers in patients with mTBI.²⁰ Since the levels of pro- and anti-inflammatory markers in the blood change during early and delayed recovery after TBI,²¹ a description of the profile of inflammatory markers at separate time-points after mTBI is necessary. Additionally, there is limited data on the association between levels of blood inflammatory markers and clinical outcomes after mTBI.²¹

In this study, we describe the profile of plasma inflammatory markers as well as investigate associations between inflammatory markers and neuropsychological outcomes in a cohort of patients with mTBI.

Methods

This prospective study was approved by the local Institutional Review Board and by the sponsor, the Department of Defense. This investigation was part of a larger study conducted in two Level 1 trauma centers and one Level 3 community hospital. All potential mTBI patients in the emergency department were screened, and patients were able to provide informed consent if their Galveston Orientation and Amnesia Test was ≥75. All patient had scores ≥75 in this study. Patients with pre-existing substance abuse, alcohol dependence, and psychiatric disorders were excluded. Inclusion criteria were adult patients (age 18-50 years) with mTBI at admission to the emergency room, Glasgow Coma Scale score 13-15, presence of head injury, negative computed tomography imaging of the head, loss of consciousness less than 30 min, and post-traumatic amnesia less than 24 h. A comparison group of orthopedic injury (OI) patients without mTBI also were analyzed in this study. This control group of OI patients included adult patients (age 18-50 years) with extremity or pelvis injuries (Abbreviated Injury Scale <3) and without evidence of head injury. The OI group of patients reported no head trauma concomitant with their orthopedic trauma, did not have external signs of head trauma, and did not lose consciousness or memory after the injury. Demographic and clinical data related to the injury were collected at admission.

Inflammatory markers

Venous whole–blood samples were collected at enrollment (within 24 h of injury) and again at 6 months after injury. Samples were centrifuged at 4°C for 10 min at 1500 × g, the supernatant was removed to clean tubes, and the samples were further clarified by centrifugation 4°C for 10 min, 10,000 × g. The resulting platelet-depleted plasma fraction was divided into aliquots and stored at -80°C until assayed. A high-sensitivity multiplex panel (Thermofisher, cat# EPXS090-12199-901) was used to simultaneously assay IL-1β, IL-2, IL-4, IL-6, IL-10, IL12p70, IL-17a, interferon (IFN)γ, and tumor necrosis factor (TNF)α using a Luminex Magpix. This cytokine panel was chosen based upon previously published data on the relationship of multiple cytokines with different outcome measures after TBI,^22–27 and our past experience with multiplex assay panels.

Reference standards provided in the kit were run in tandem with the experimental samples to generate standard curves for each target and used to quantify the amount of each target in the patient samples. A mixture of magnetic antibody beads was added to the reaction plate, followed by 25 μL of plasma and 25 μL of universal assay buffer. The plate was incubated with shaking for 30 min at room temperature, then overnight at 4°C, followed by an additional 30 min at room temperature. The beads were washed two times with the buffer provided in the kit and a magnetic baseplate was used to retain the beads. The plates were then incubated for 30 min at room temperature with 25 μL of the biotinylated detection antibody mixture, washed twice, incubated with 50 μL streptavidin-phycoerythrin, then washed twice. The samples were then incubated at room temperature for 30 min with 25 μL amplification reagent 1, washed twice, incubated with 25 μL amplification reagent 2 for 30 min at room temperature, followed by two washes. After the final wash, the samples were resuspended in 120 μL reading buffer (provided with the kit) by 5 min vigorous shaking at room temperature and analyzed using a Luminex Magpix plate reader.

Neuropsychological outcomes

Neuropsychological outcomes for mTBI patients were assessed as described previously at 1 week, 1 month, 3 months, and 6 months after injury.^28,29 Health-related quality of life was recorded using the 12 item Short-Form Health Survey (SF-12), which measures physical and mental health.³⁰ The neuropsychological tests used for analysis in this study are described below.

Verbal Selective Reminding Test (VSRT)

The VSRT measures verbal episodic memory using 12 unrelated words over six trials as described previously.²⁹ The participant was asked to recall all 12 words on each trial; after the examiner orally presented the word list on trial 1, the examiner presented only those words that the participant failed to recall on the preceding trial. The variables analyzed in this study included the consistent long-term retrieval (total number of words recalled for each word from the first through sixth trial) and delayed recall (total number of words recalled at 30 min after the sixth trial). We also recorded cued recall (CR) by providing the first two or three letters of each word on the list, as well as recognition memory recall (REC) after cueing, which measured if participants could identify the words on the list amongst four alternative words.

Symbol-Digit Modalities Test (SDMT)

The SDMT³¹ detects reduced processing speed for timed substitution tasks when the participant pairs numbers with geometric figures in a fixed time limit. For this study, we analyzed the total number of correct responses for both the written and oral versions.

Rivermead Post-Concussion Symptoms Questionnaire (RPCSQ)

The RPCSQ^32-34 is a self-reported scale of post-concussion symptoms, including cognitive, emotional and somatic symptoms after mTBI compared with pre-injury; each symptom is rated by the subject on a 5-point scale (0 to 4) wherein a score of 1 denotes a symptom that is unchanged from pre-injury. The total score (0-64) was used for analysis in this study. A higher score indicated greater severity.

Post-traumatic Stress Checklist-Civilian Form (PCL-C)

The PCL-C is a self-reported measure of PTSD symptom severity for civilians.³⁵ Participants rate the severity of their symptoms on a 5-point scale with higher scores indicating greater severity. The total score (17-85) was used for analysis in this study.

Center for Epidemiologic Studies Depression Scale (CES-D)

The CES-D is a self-reported measure of depression related symptoms with higher scores associated with greater symptom severity.³⁶ The total score (0-60) was used as the primary variable in this study, and the baseline score (recorded as symptoms for the week prior to injury) was used as a covariate in the analysis.

Brief Visuospatial Memory Test-Revised (BVMT-R)

The BVMT-R³⁷ measures visuospatial learning and memory by asking participants to recall and draw figures in a 2 × 3 array after a learning trial. The scores summed across three learning trials and delayed recall were used in this study.

Delis-Kaplan Executive Functioning System Verbal Fluency Test (D-KEFS VFT)

D-KEFS VFT measures verbal fluency by asking the participant to generate as many words as possible beginning with specific letters according to rules presented beforehand to the participant; the total number of correct words summed over three 1-min trials is the key variable (letter fluency).³⁸ For category fluency, the participant generates as many words as possible in a particular category for two 1-min trials. For category-switching, the participant generates as many words as possible in two alternating categories. For this study, we analyzed total correct scores for letter fluency, category fluency, and category-switching. D-KEFS VFT measures executive function, switching strategies, initiation of response, and adherence to rules. Expressive language is also implicitly tested by this test.

Delis-Kaplan Executive Functioning System Color-Word Interference Test (D-KEFS CWIT)

The D-KEFS CWIT³⁹ assesses response time for naming the color of font in color words when it is dissonant with the color word (e.g., naming “red” when the color word “blue” is printed in red font) compared with the time to read the words printed in non-dissonant font). By subtracting response times for the conditions with and without interference, this test measures the executive function of inhibition. The raw scores for the inhibition (interference) and inhibition-switching (alternating naming of font color and reading the color word) conditions were analyzed for this study.

Return to class/work

At each follow-up interval, participants were asked if they returned to school or work and this was analyzed as an additional outcome variable.

Statistical analysis

All statistical analysis, visualizations, and summary tables were implemented with R (3.6.1).

Exploratory analysis

A total of 53 subjects were analyzed. Median and interquartile range were calculated for all numeric variables. Categorical variables were counted and calculated as a percent of total. Shapiro-Wilk tests and box plots with jittered points were used to assess the normality of each cytokine at the 24-h and 6-month collection points. Wilcoxon signed rank tests were used to assess significant differences between cytokine levels at 24 h and 6 months. The results from neuropsychological outcomes were assessed for normality using Shapiro-Wilk tests at each time-point (1 week, and 6 months). Wilcoxon signed-rank test was used to assess significant differences in outcome scores between 1 week and 6 months after injury. McNemar's test was used for the binary “return to class/work” outcome. Univariate models for each potential outcome:predictor pair were generated to identify significant linear associations.

Regression analysis

Least absolute shrinkage and selection operator (LASSO) was used to identify potentially associated cytokines and confounders (age, sex, race, socioeconomic index, Glasgow Coma Scale verbal score, and injury severity score) for each outcome. All outcomes were analyzed for cytokines measured at 24 h. Only 6-month outcomes were analyzed for cytokines measured at 6 months.

For each outcome with at least one cytokine predictor and one confounder, the relationship between the cytokine and confounder was assessed for interaction. If this interaction was significant, the interaction term was added to the model. Interaction was assessed by comparing predicted outcome values between the original study data and a modified data set that held the interaction term constant.

Multiple regression models were constructed for each outcome that had at least one potential predictor identified by LASSO. Transformations (e.g., log, square root) of the outcome measure were conducted if they improved the diagnostic fit. If transformations did not improve diagnostic fit, the model was fitted using restricted cubic splines on nonlinear predictors. The restricted cubic spline (RCS) fitted model was then compared with the original model using an analysis of variance chi squared test. All spline fitted models had p values >0.05 for the chi-squared test, and were not included in the final summary.

Significant associations between the outcome and cytokine predictors were adjusted using Bonferroni's correction, with the number of comparisons calculated by counting the number of models with at least one LASSO-selected cytokine (for tests administered at 24 h, n = 16 and p = 0.05/16 = 0.003; at 6 months, n = 6 and p = 0.05/6 = 0.008). These adjusted models were compared with the univariate models that were generated regardless of LASSO selection. If an outcome:cytokine pair exhibited a p value below the Bonferroni-adjusted alpha, the result was determined to be significant and reported in the list of significant models.

Results

Study sample

A total of 104 mTBI and 73 OI controls were enrolled in the original larger study. For this study, we included 53 mTBI and 24 OI controls for whom plasma cytokine data and 6-month follow-up was available. Baseline demographic and clinical data are shown in Table 1. The majority of mTBI patients were male (62.3%) and non-black (71.7%), with motor vehicle accidents (37.7%) being the most common cause of injury.

Table 1.

Baseline Demographic and Clinical Data for the 53 mTBI Patients included in This Study

	mTBI (n = 53)	OI (n = 24)	p Value
Age
Median (min-max)	26 (18-49)	28 (20-50)	0.41
Sex
Male	33 (62.3%)	16 (66.7%)	0.71
Female	20 (37.7%)	8 (33.3%)
GCS (Verbal)
4	3 (5.7%)	0 (0%)	0.55
5	50 (94.3%)	24 (100%)
GCS (Total)
14	3 (5.7%)	0 (0%)	0.55
15	50 (94.3%)	24 (100%)
Race
Non-Black	38 (71.7%)	17 (70.8%)	0.94
Black	15 (28.3%)	7 (29.2%)
Injury
Motor vehicle	20 (37.7%)	1 (4.2%)
Assault	10 (18.9%)	0 (0%)
Blow to head	6 (11.3%)	0 (0%)
Fall moving object	4 (7.5%)	1 (4.2%)
Fall standing	4 (7.5%)	1 (4.2%)
Motorcycle	3 (5.7%)	1 (4.2%)
Auto pedestrian	3 (5.7%)	2 (8.3%)
Sports-related	2 (3.8%)	0 (0%)
ATV	1 (1.9%)	0 (0%)
Laceration	0 (0%)	8 (33.3%)
Crush	0 (0%)	7 (29.2%)
Dislocation	0 (0%)	1 (4.2%)
Fall from raised surface	0 (0%)	2 (8.3%)
ISS
1	8 (15.1%)	17 (70.8%)
2	9 (17.0%)	0 (0%)
3	13 (24.5%)	0 (0%)
4	2 (3.8%)	6 (25%)
5	8 (15.1%)	1 (4.2%)
6	12 (22.6%)	0 (0%)
8	1 (1.9%)	0 (0%)
SCI
Median (min-max)	25.00 (16-65)	28.5 (16-49)	0.22

mTBI, mild traumatic brain injury; OI, orthopedic injury; min, minimum; max, maximum; GCS, Glasgow Coma Scale; ATV, all-terrain vehicle; ISS, injury severity score; SCI, Socio-economic Composite Index.

Inflammatory cytokines for mTBI patients

The median plasma cytokine levels for samples collected at 24 h and at 6 months after mTBI subjects are shown in Table 2. For mTBI subjects, a statistically significant decrease in cytokine levels at 6 months compared with cytokine levels at 24 h was seen for IL-1b (p = 0.03), IL-4 (p < 0.001), IL-6 (p < 0.001), and IFN-γ (p < 0.001). IL-2 levels were significantly increased (p < 0.001) at 6 months compared with the levels at 24 h after mTBI (Fig. 1).

Table 2.

Levels of Plasma Inflammatory Cytokines (picograms/mL) at 24 h and 6 months after Injury for 53 mTBI Patients and 24 OI Controls

Plasma cytokines	mTBI (n = 53)	OI controls (n = 24)	p Value
Within 24 h of injury
IL-1b	2.23 (0.29-33.72)	1.7 (0.12-6.62)	0.061
IL-2	5.24 (0.58-16.64)	3.45 (0.58-12.04)	0.014*
IL-4	48.67 (7.64-910.84)	31.64 (0.92-147.5)	0.106
IL-6	25.86 (6.03-159.29)	15.99 (2.51-57.81)	0.01*
IL-10	0.63 (0.03-5.4)	0.38 (0.03-2.44)	0.093
IL-12p70	2.86 (0.47-62.22)	2.49 (0.47-10.08)	0.464
IL-17a	9.84 (0.25-54.66)	5.46 (0.25-31.98)	0.099
IFN-γ	33.88 (1.01-315.21)	21.7 (1.57-292.28)	0.059
TNF-α	4.00 (0.26-85.99)	2.56 (0.26-14.44)	0.108
At 6 months after injury
IL-1b	1.77 (0.12-18.81)	2.53 (0.12-5.29)	0.792
IL-2	10.56 (1.26-22.94)	11.79 (3.03-22.94)	0.834
IL-4	33.00 (5.72-206.92)	34.36 (2.53-83.66)	0.93
IL-6	10.02 (2.85-101.57)	8.34 (1.53-24.64)	0.044*
IL-10	0.47 (0.0-2.96)	0.505 (0.0-8.5)	0.733
IL-12p70	3.92 (0.85-12.91)	4.72 (0.85-9.6)	0.613
IL-17a	8.76 (0.33-26.14)	10.04 (0.33-26.27)	0.52
IFN-γ	8.73 (0.0-114.92)	6.21 (0.0-68.13)	0.222
TNF-α	3.63 (0.27-33.18)	4.08 (0.27-13.24)	0.847

mTBI, mild traumatic brain injury; OI, orthopedic injury; IL, interleukin; IFN, interferon; TNF, tumor necrosis factor.

FIG. 1.

Graphical representation of cytokine levels that changed significantly at 6 months after injury in mild traumatic brain injury patients. Each line represents a subject and the median value is represented by a circle.

Comparison of cytokine levels between mTBI patients and OI controls

Plasma levels for inflammatory cytokines showed a significant elevation for IL-2 (p = 0.014) and IL-6 (p = 0.01) within 24 h post-injury for mTBI patients compared with OI patients. No statistically significant differences were noted for other inflammatory cytokines at this time-point (Table 2).

Plasma levels for inflammatory cytokines at 6 months post-injury showed significant elevations for IL-6 (p = 0.044) for mTBI patients compared with OI patients. No statistically significant differences were noted for other inflammatory cytokines at 6 months post-injury.

Neuropsychological outcomes and posttraumatic symptoms

Significant differences in post-traumatic symptoms between mTBI and OI groups at each time-point are shown in Table 3. Intra-group differences were analyzed for the mTBI and OI group to assess changes in symptoms over time from 1 week to 6 months after injury. Significant improvements in BVMT_DR (p < 0.001), BVMT_Recall (p < 0.001), CES-D (p = 0.002), fatigue (p < 0.001), RPCSQ (p < 0.001), SDMT_Oral (p < 0.001), SDMT_Written (p < 0.001), SF12_MH (p = 0.004), and SF12_PH (p < 0.001) scores were observed for the mTBI group. For the OI group, significant improvements were noted for BVMT- Learning (p = 0.02), BVMT_Recall (p = 0.02), RPCSQ (p = 0.006), SDMT_Oral (p = 0.004), SDMT_Written (p = 0.001), and SF12_PH (p < 0.001).

Table 3.

Comparison of Mean Neuropsychological and Symptom Scores between mTBI and OI group at Different Time-Points after Injury

Outcome	Baseline	1 Week	1 Month	3 Months	6 Months
BVMT_DR		-3.4314; 0.032	-1.4072; 0.831	-3.1262; 0.086	-3.0135; 0.224
BVMT_Learning		3.4979; 0.043	0.3198; 0.651	1.7496; 0.38	-5.5465; 0.034
BVMT_Recall		-2.8606; 0.08	-2.4974; 0.098	-3.6189; 0.01	-5.9383; 0.089
CES-D	4.8066; 0.048	7.7015; 0.001***	4.6508; 0.012	4.9863; 0.044	4.9863; 0.044
CLTR		-6.1461; 0.02	-1.0373; 0.716	-0.9123; 0.268	-8.1839; 0.052
CR		-0.2628; 0.14	0.8093; 0.35	0.4919; 0.676	-0.4434; 0.511
DEL		-0.911; 0.01	-0.0835; 0.96	-0.1711; 0.532	-0.4318; 0.3
Fatigue		1.0609; 0.024	0.801; 0.002***	0.5396; 0.527	0.5783; 0.034
LTR		-4.225; 0.063	-0.6936; 0.75	-0.4776; 0.591	-5.1521; 0.074
LTS		-3.5141; 0.081	-0.8964; 0.561	-0.4137; 0.659	-3.5963; 0.122
OCM_ISCN		-0.8428; 0.403	-1.2908; 0.604	-1.1816; 0.888	0.32; 0.38
OCM_ISWR		-0.9619; 0.028	-2.3033; 0.007***	-1.675; 0.376	0.2493; 0.71
OMC_FITC		-8.2873; 0.044	-7.0858; 0.079	-6.4467; 0.025	-2.0465; 0.049
OMC_SITC		-3.3996; 0.133	-5.1744; 0.002***	-3.8019; 0.165	-2.0907; 0.13
PCL				7.0288; 0.033	5.193; 0.001***
REC		-0.1512; 0.037	0.3941; 0.798	0.4442; 0.785	0.1183; 0.732
RLTR		1.9211; 0.038	0.3438; 0.355	0.4348; 0.127	3.0318; 0.168
RPCSQ		11.835; < 0.001***	8.3798; < 0.001***	6.8124; 0.013	6.5482; < 0.001***
SDMT_Oral		-5.8582; 0.035	-6.6735; 0.062	-4.3931; 0.051	-9.4949; 0.036
SDMT_Written		-2.7798; 0.054	-4.5794; 0.029	-2.7367; 0.025	-7.7262; 0.061
SF12_MH		-7.4235; 0.002***	-6.6266; 0.002***	-5.3447; 0.01	-5.8563; 0.155
SF12_PH		-0.0968; 0.389	-2.1904; 0.987	-0.4267; 0.23	-1.7037; 0.851
STR		1.2198; 0.606	0.6358; 0.591	0.8832; 0.733	1.2096; 0.242

Data are presented as mean score.

^***p < 0.01 after multiple comparisons adjustment; n = 5.

BVMT_DR, Brief Visuospatial Memory Test-Revised Delayed Recall; BVMT_R_Learning, Brief Visuospatial Memory Test-Revised Learning; BVMT_R_Recall, Brief Visuospatial Memory Test-Revised Recall; CES-D, Center for Epidemiological Studies Depression Scale; CLTR, Consistent long-term retrieval (Verbal selective reminding test); DEL, Delayed recall (Verbal selective reminding test); LTR, Long term retrieval (Verbal selective reminding test); LTS, Long term storage (Verbal selective reminding test); OCM_ISWR, Inhibition score, DKEFS-Color Word Interference Test; OMC_FITC, Total Correct-Letter Fluency, DKEFS-Verbal Fluency Test; OMC_SITC, Total correct-Category Switching condition of Category Fluency, DKEFS-Verbal Fluency Test; PCL, Post-Traumatic Stress Symptom Checklist-Civilian Form; RLTR, Random long term retrieval (Verbal selective reminding test); RPCSQ, Rivermead Post Concussion Symptoms Questionnaire; SDMT_Oral, Symbol-Digit Modalities Test Oral; SDMT_Written, Symbol-Digit Modalities Test Written; SF12_MH, Short-Form 12_ Mental Health; SF12_PH, Short-Form 12_Physical Health; STR, Short-term retrieval (Verbal selective reminding test).

Association between inflammatory cytokines and neuropsychological outcomes

Multivariate analysis was performed to evaluate associations between symptom measures, neuropsychological outcomes and plasma cytokines obtained within 24 h and at 6 months after injury (Supplementary Tables S1, S2, and S3). Higher plasma IL-2 levels at 24 h were associated with more severe RPCSQ at 1 week after injury (p = 0.001; Fig. 2). Lower plasma IL-6 (p = 0.035) and IL-17a (p = 0.007) levels within 24 h were associated with more severe RPCSQ at 1 week after injury (Table 4).

FIG. 2.

Scatter plot showing statistically significant associations between plasma cytokine levels at 24 h and the Rivermead Post-Concussion Symptoms Questionnaire total score at 1 week.

Table 4.

Regression Models Showing Statistically Significant Associations between Plasma Cytokines and Neuropsychological Outcomes in 53 mTBI Patients

Outcome	Coefficient	95% CI	p Value	R2
RPCSQ at 1 week			0.002	0.255
IL2_T1	2.35	(1.029, 3.67)	0.001
IL6_T1	-0.113	(-0.218, -0.009)	0.035
IL17a_T1	-0.54	(-0.928, -0.152)	0.007
GCS_Verbal	6.98	(-8.289, 22.244)	0.363
SCI_total	-0.196	(-0.454, 0.063)	0.134
PCL-C at 6 months (log)			0.032	0.14
IL10_T2	0.218	(0.073, 0.362)	0.004
IL6_T2	-0.005	(-0.011, 0)	0.067
GCS_Verbal	0.206	(-0.21, 0.622)	0.324
SCI_total	-0.006	(-0.013, 0.001)	0.113
Race	0.064	(-0.153, 0.281)	0.557
CESD at 6 months (sqrt)			0.002	0.194
IL10_T2	1.2	(0.529, 1.868)	0.001
SCI_total	-0.034	(-0.066, -0.001)	0.044
OCM_ISWR at 6 months			<0.001	0.5
IL17a_T2	-0.791	(-1.167, -0.415)	<0.001
Age	-0.094	(-0.189, 0)	0.05
GCS_Verbal	-1.75	(-4.045, 0.539)	0.128
SCI_total	-0.051	(-0.095, -0.007)	0.024
IL17a_T2:Age	0.023	(0.012, 0.034)	<0.001

T1- within 24 h post-injury; T2- at 6 months post-injury.

mTBI, mild traumatic brain injury; CI, confidence interval; RPCSQ, Rivermead Post-Concussion Symptoms Questionnaire; IL, interleukin; GCS, Glasgow Coma Scale; SCI, Socio-economic Composite Index; PCL, Post-traumatic Stress Symptom Checklist-Civilian Form; OCM_ISWR, Inhibition score; D-KEFS CWIT, Delis-Kaplan Color Word Interference Test.

At 6 months after injury, higher IL-10 at 6 months was significantly associated with more severe PTSD (p = 0.004; Fig. 3) symptoms at 6 months and worse mood as measured by CES-D (p = 0.001; Fig. 4) at 6 months. Higher plasma IL-17a at 6 months after injury was associated with poorer executive function as measured by DKEFS-CWIT (p < 0.001; Table 4).

FIG. 3.

Scatter plots showing statistically significant associations between plasma interleukin-10 levels at 6 months and Post-traumatic Stress Checklist-Civilian Form score at 6 months.

FIG. 4.

Scatter plots showing statistically significant associations between plasma IL-10 levels at 6 months and Center for Epidemiologic Studies Depression Scale score at 6 months.

Discussion

This prospective study describes specific early and late profiles of plasma inflammatory cytokines after mTBI and significant associations between plasma cytokine levels and neuropsychological measures, including both severity of symptoms and performance on an executive function test. The time-point–dependent associations of specific biomarkers with specific neuropsychological outcomes is also a novel aspect of these findings. Plasma IL-2 and IL-6 levels were significantly higher for mTBI patients compared with OI controls within 24 h post-injury. For mTBI subjects, the plasma levels of IL-1b, IL-4, IL-6 and IFN-γ were reduced at 6 months after injury in comparison to levels within 24 h after injury. Elevated IL-2 within 24 h after the injury was associated more severe early post-concussive symptoms, while elevated plasma IL-10 level at 6 months was associated with more severe PTSD and mood scores at 6 months, respectively.

Neuroinflammation plays an important role in the pathophysiology of secondary brain injury. The inflammatory response originates in the brain, but may be augmented by inflammatory cytokines that cross a compromised blood brain barrier.¹⁴ Pre-clinical studies describe a cascade of pro- and anti-inflammatory cytokines early after mTBI.⁴⁰ Importantly, these cytokines play distinct roles in the acute and chronic phase after TBI. While early neuro-inflammation may worsen neuronal injury after mTBI, delayed inflammation can stimulate neural regeneration as well as repair.²¹ The majority of data for neuroinflammation in mTBI is from animal studies, which vary considerably in their use of injury models, species and behavioral tests.⁴¹ Recent studies have shown that blood inflammatory markers are elevated after adult mTBI.⁴² The present study evaluates inflammatory plasma cytokines in a homogenous set of adult mTBI patients with strict eligibility criteria, including negative CT head findings, to help mitigate confounds. In addition, we compared the plasma levels of inflammatory cytokines between mTBI patients and control trauma subjects showing an early elevation in plasma IL-2 and IL-6 within 24 h of injury, and elevation in plasma IL-6 levels at 6 months after the injury. These results are supported by animal data showing upregulation of IL-2⁴³ and IL-6⁴⁴ after TBI. Importantly, we identify potential relationships between neuropsychological outcomes and the pro-inflammatory state after mTBI.

In this study, elevated plasma IL-2 levels within 24 h of the injury were associated with more severe post-concussive symptoms as measured by the RPCSQ at 1 week. IL-2 plays an important role in regulating immune responses by regulating T cell activity. A transient elevation in IL-2 levels can stimulate effector T cells and drive an immune response.⁴⁵ Prior studies have shown that TBI can produce upregulation of IL-2,⁴³ and this in turn could contribute to enhanced neuroinflammation. Higher IL-2 levels early after mTBI suggest an enhanced immune response and potentially more neuronal injury due to an increased immune response. These effects may explain the association we found between elevated plasma IL-2 within 24 h of injury and increased severity of early post-concussive symptoms.

In the present study, we also found that higher plasma IL-10 levels at 6 months was associated with more severe PTSD symptoms. Interleukin-10 is an anti-inflammatory cytokine produced by microglia and astrocytes in the central nervous system, as well as by lymphocytes in the peripheral blood. Csuka and colleagues¹⁵ showed that IL-10 is predominantly produced in the central nervous system after TBI; however, the role of IL-10 in the pathophysiology after TBI is unclear. In animal studies, the administration of systemic IL-10 has been shown to improve recovery from TBI.⁴⁶ Conversely, higher CSF IL-10 levels in children have been shown to correlate with worse outcomes after severe TBI.⁴⁷ Prior studies have shown elevated serum IL-6⁴⁸ and IL-10^48,49 levels in patients with PTSD compared with controls. Elevated plasma IL-10 levels suggest a robust anti-inflammatory response that attempts to counter pro-inflammatory changes after mTBI. A stronger pro-inflammatory process may be contributory to the more severe PTSD symptoms as well as mood scores seen in mTBI patients with significantly higher plasma IL-10 levels at 6 months.

In the present study, we noted that plasma IL-6 was significantly elevated in mTBI patients compared with OI controls at 6 months after the injury. Lower plasma IL-6 and IL-17a levels within 24 h were associated with poorer RPCSQ scores at 1 week. IL-6 is a regulator of acute inflammation and has both pro- and anti-inflammatory properties, while IL-17a is a pro-inflammatory cytokine involved in the T-cell inflammatory response.⁵⁰ Ojo and colleagues⁵¹ showed elevated plasma IL-17a levels in an mTBI mouse model, but not in a PTSD model suggesting unique systemic inflammatory responses. The plasma cytokine levels measured in this study reflect systemic inflammatory responses, and it is possible that distinct neuroinflammatory cascades contribute to mTBI and post-concussive symptoms. Further research is necessary to determine how systemic cytokine levels affect the brain or if these correspond to an equivalent neuroinflammatory response. Although the association of cytokine levels with measures of symptom severity predominated, we found that a single measure of executive function, cognitive interference (or inhibition) at 6 months, was associated with IL-17a measured at the same time-point.

Neuropsychological sequelae are commonly noted after mTBI, and may be attributable to chronic neuroinflammation.^14,50 Prior studies have shown statistically significant differences in plasma inflammatory cytokine levels in patients with PTSD compared with controls.^48,49 Neuronal-derived exosomal IL-10 has been shown to be elevated in PTSD after mTBI indicating an inflammatory state in the brain.⁵² Increased peripheral inflammation can also contribute to increased neuroinflammation and this may in turn influence neurotransmitter release in the brain and affect behavior. Additional stressors including chronic exposure to combat in the military, the aftermath of natural disasters, intimate partner violence, or physical impairments as a result of the injury or socio-economic factors may perpetuate inflammation and neurotransmitter imbalance via the hypothalamic-pituitary axis.⁵³ While the exact mechanism of these neuropsychological sequelae has not been elucidated, results from this study support the link between neuroinflammation and neuropsychological outcomes after mTBI.

The present study is limited by its small sample size. We only included subjects from the study with plasma samples available for analysis, and therefore there is a potential sampling error due to exclusion of subjects. However, no significant differences were seen in baseline demographics, clinical injury scores, or 1-week or 6-month outcomes when comparing mTBI subjects included in this analysis and mTBI that were not. The small sample size also limited the number of variables that could be included in the multivariate analysis; however, we were able to include important confounders such as age, sex and socioeconomic status in the models. At 6 months after the injury, patients were not re-evaluated for additional stressors that may have impacted the measurement of systemic inflammatory cytokines. Although we measured a panel of nine plasma cytokines in this study, future studies to examine additional cytokines and chemokines may further elucidate the inflammatory processes after mTBI.

Conclusion

This study demonstrates that patients with mTBI have elevated plasma inflammatory cytokine levels within 24 h and at 6 months after the injury compared with orthopedic injury control patients without mTBI. Additionally, elevated plasma cytokine levels were associated with poorer post-traumatic symptom and executive function measures at 24 h and at 6 months, suggesting a potential relationship between inflammation and neuropsychological outcomes after mTBI.

Funding Information

Department of Defense CDMRP grants W81XWH-08-2-0132 (CSR), W81XWH-08-2-0131 (JJM), W81XWH-08-2-0133 (HSL), W81XWH-08-2-0134 (PKD), and National Institutes of Health NINDS Grant #NS088298 (PKD).

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3