Traumatic brain injury blood biomarkers are associated with non-neurological organ dysfunction: a secondary analysis of ProTECT III and Bio-ProTECT

David J Barton; Erica Fan; Ian G Gober; Nabil Awan; Raj G Kumar; Jenna C Carlson; Michael R Frankel; David W Wright; Amy K Wagner

doi:10.1177/08977151251362234

. Author manuscript; available in PMC: 2026 Mar 24.

Published in final edited form as: J Neurotrauma. 2025 Aug 7;42(21-22):1947–1958. doi: 10.1177/08977151251362234

Traumatic brain injury blood biomarkers are associated with non-neurological organ dysfunction: a secondary analysis of ProTECT III and Bio-ProTECT

David J Barton ¹, Erica Fan ², Ian G Gober ², Nabil Awan ³, Raj G Kumar ⁴, Jenna C Carlson ^5,⁶, Michael R Frankel ^7,⁸, David W Wright ⁹, Amy K Wagner ^2,^10,^11,¹²

¹Department of Emergency Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, USA.

²Department of Physical Medicine & Rehabilitation, University of Pittsburgh, Pittsburgh, Pennsylvania, USA.

³Department of Statistics, University of Wisconsin, Madison, Wisconsin, USA.

⁴Department of Rehabilitation and Human Performance, Icahn School of Medicine at Mount Sinai, New York, New York, USA.

⁵Department of Human Genetics, University of Pittsburgh, Pittsburgh, Pennsylvania, USA.

⁶Department of Biostatistics, University of Pittsburgh, Pittsburgh, Pennsylvania, USA.

⁷Marcus Stroke and Neuroscience Center, Grady Memorial Hospital, Atlanta, Georgia, USA.

⁸Department of Neurology, Emory University, Atlanta, Georgia, USA.

⁹Department of Emergency Medicine, Emory University, Atlanta, Georgia, USA

¹⁰Department of Neuroscience, University of Pittsburgh, Pittsburgh, Pennsylvania, USA.

¹¹Safar Center for Resuscitation Research, University of Pittsburgh, Pittsburgh, Pennsylvania, USA.

¹²Clinical and Translational Science Institute, University of Pittsburgh, Pittsburgh, Pennsylvania, USA.

Authorship confirmation/contribution statement:

DJB: conceptualization, data curation, formal analysis, investigation, visualization, writing – original draft. EF: data curation, formal analysis, investigation, methodology, writing – review and editing. IGG: data curation, formal analysis, investigation, methodology, writing – review and editing. NA: conceptualization, funding acquisition, writing – review and editing. RGK: conceptualization, funding acquisition, writing – review and editing. JCC: investigation, formal analysis, methodology, writing – review and editing. MRF: conceptualization, funding acquisition, writing – review and editing. DWW: conceptualization, funding acquisition, writing – review and editing, AKW: conceptualization, funding acquisition, investigation, project administration, methodology, supervision, writing – review and editing.

^✉

Address correspondence to: David J. Barton, MD, Department of Emergency Medicine, University of Pittsburgh, Iroquois Building, Suite 400A, 3600 Forbes Ave, Pittsburgh, PA 15213, USA, bartond2@upmc.edu or Amy K. Wagner, MD, Department of Physical Medicine and Rehabilitation, University of Pittsburgh, 3471 Fifth Avenue, Suite 910, Pittsburgh, PA 15213, USA, wagnerak@upmc.edu

PMCID: PMC13007255 NIHMSID: NIHMS2135603 PMID: 40779423

Abstract

Non-neurological organ dysfunction (NNOD) is a prevalent complication and contributes to poor outcome after traumatic brain injury (TBI). Contributing factors to NNOD may include initial TBI severity, but this relationship has not been rigorously studied. The objectives of this study are to describe the frequency and timing of NNOD after TBI, evaluate the association between NNOD and outcome (mortality and Glasgow Outcome Score-Extended [GOSE] at 6-months post-injury), and examine the relationship between multimodal markers of initial TBI severity and NNOD. We performed a secondary analysis of data from participants in both the ProTECT-III clinical trial (progesterone versus placebo in participants with moderate-to-severe TBI) and the embedded Bio-ProTECT blood biomarker study (N=536 individuals). We reviewed laboratory and clinical data to determine prevalence of NNOD in renal, hematologic, hepatic, cardiovascular, and respiratory systems, based on the Sequential Organ Failure Assessment (SOFA) system. TBI severity was assessed using index Glasgow Coma Scale score (iGCS - first GCS post primary resuscitation), Rotterdam CT score, head-region abbreviated injury scale scores, and baseline TBI biomarkers (S100b, GFAP, UCHL1, SBDP150). NNOD frequencies by organ system were 72% (respiratory), 52% (cardiovascular), 45% (hematologic), 8% (renal), and 2% (hepatic). All TBI severity markers were positively correlated (using Spearman coefficients) with number of systems in dysfunction. To examine effects of NNOD on outcome independent of TBI severity, we used logistic regression and adjusted for age, sex, iGCS, Rotterdam CT score, and biomarker load score (mean biomarker quartile), wherein each additional system of dysfunction resulted in a 1.30x higher odds of unfavorable GOSE (95% CI 1.01–1.67; p=0.04). Stratification analyses revealed the relationship between greater NNOD and worse outcome was most pronounced among individuals with more severe Rotterdam CT and lower GCS scores. In conclusion, NNOD occurs frequently after moderate-to-severe TBI, is associated with higher odds of unfavorable GOSE at 6-months, and is positively associated with multimodal markers of TBI severity. This is the first study to demonstrate a relationship between TBI blood biomarker levels and NNOD. Future study is needed to determine mechanisms of NNOD and their relationships to subsequent neurological injury. While TBI research has historically focused on brain-centric measures and outcomes, this study builds on mounting evidence that non-neurological organ systems play an important role in injury response after TBI.

Keywords/search terms (4–6): traumatic brain injury, biomarkers, non-neurological organ dysfunction, outcome

Introduction

Non-neurological organ dysfunction (NNOD) is common after traumatic brain injury (TBI) and a significant contributor to patient outcome. Zygun et al. were the first to report on NNOD after TBI.¹ Of 209 patients with severe TBI, 89% developed dysfunction in at least one system (respiratory, cardiovascular, coagulation, renal, or hepatic), and adjusted analyses demonstrated 1.63x increased odds of in-hospital mortality if at least one system was in dysfunction.¹ In a similar study focused on death after severe TBI (n=135), NNOD was present in 81% of patients, contributed to in-hospital death in two-thirds of patients, and was the primary cause of death in 24% of patients.²

Despite this epidemiological evidence, the relationship between TBI and non-neurologic organ systems remains a largely understudied aspect of TBI associated pathophysiology. TBI may play a causal role in non-neurologic organ dysfunction (NNOD) after trauma, and conversely, NNOD may affect TBI pathophysiology.

More recent reports have worked to identify blood biomarkers associated with NNOD after TBI, which could aid in both pathophysiologic understanding of these processes and clinical prognostication. For example, Hendrickson et al. observed plasma levels of tissue inhibitor of matrix metalloproteinase-3 (TIMP-3), a soluble protein that regulates matrix metalloproteinase activity and inflammation systemically³, were higher upon hospital arrival among individuals who developed acute respiratory distress syndrome post-TBI compared to those who did not.⁴ Higher TIMP-3 levels also were associated with mortality.⁴ Lee et al. reported that higher plasma levels of interleukin (IL)-6, IL-8, and IL-10 (proinflammatory cytokines) were associated with development of multiple organ dysfunction syndrome (MODS) and worse Glasgow Outcome Scale (GOS) scores at six months post-TBI.⁵ Additionally, our group previously investigated serum estradiol (E2) and tumor necrosis factor-alpha (TNF-α) levels in a single site study involving participants with severe TBI.⁶ We found that higher levels of early E2 (<72 hours) and later TNF-α (72–144 hours) were associated with greater odds of death within 6 months post-TBI, and this effect was partially mediated by NNOD.⁶ Together, these data suggest systemic proinflammatory biomarkers may be indicative of NNOD and prognosticate outcome after TBI.

Our previous Bio-ProTECT work has also shown that the brain-specific blood biomarkers S100 calcium binding protein B (S100B), glial fibrillary acidic protein (GFAP), ubiquitin C-terminal hydrolase-L1 (UCHL1), and spectrin breakdown products (SBDP150) were independently associated with 6-month outcome and can enhance outcome prediction after TBI.⁷ However, the relationships between these prognostic brain-injury biomarkers and NNOD after TBI remain unstudied.

Further evidence is needed regarding how and when NNOD develops after TBI. It remains unknown how TBI severity, especially reflected by objective brain-specific biomarkers, is associated with NNOD risk and its related effects on mortality and neurologic outcome. In this study, we investigated how multimodal measures of TBI severity, including clinical, radiologic, and blood biomarker measures, interact with NNOD after trauma to affect global neurological outcome. We hypothesized 1) NNOD would be associated with higher rates of death and unfavorable Glasgow Outcome Scale-Extended (GOSE) score at 6 months after TBI; 2) NNOD is associated with measures of brain injury severity (GCS score, Rotterdam CT score, and brain-injury blood biomarkers) after TBI; and 3) brain injury severity affects the relationship between NNOD and outcome after TBI.

Methods

Study design

We conducted a secondary analysis of data collected in the ProTECT III trial, a randomized clinical trial testing the effect of progesterone vs. placebo for treatment of moderate-to-severe TBI,⁸ and the embedded BioProTECT study, in which investigators collected, stored, and analyzed blood specimens at 0, 24, and 48 hours after enrollment in a subset of consecutively enrolled participants in the ProTECT III trial.⁷ Study participants were enrolled within four hours of injury from 22 academic hubs that included 49 trauma centers involved in the Neurological Emergencies Treatment Trials (NETT) network within the United States.

The BioProTECT study cohort has been previously characterized.⁷ Blood samples were available from 574 participants; 285 participants were randomized into the placebo group, and 289 participants were randomized into the progesterone treatment group. Inclusion criteria were moderate-to-severe TBI (post-resuscitation GCS 4–12 or motor response 2–5 if intubated), age > 18 years, blunt TBI with altered mental status due to the TBI, and ability to initiate study drug infusion within 4 hours post-injury. Exclusion criteria were bilateral dilated unresponsive pupils, cardiopulmonary arrest, prisoner or incarcerated status, active status epilepticus, progesterone allergy, clotting disorder, and pregnancy.

In this secondary analysis of the BioProTECT cohort, we excluded 34 participants who did not have outcome information available at 6 months (due to loss to follow-up or early withdrawal from the trial), 3 participants who received the non-assigned trial drug treatment, and 1 participant with documented interruption in trial drug treatment. Removing these study participants from analysis resulted in a total of 536 participants for our current analysis (Figure 1).

Demographic and clinical variables:

All data were acquired from the ProTECT III clinical trial data set. Baseline patient demographics such as age, sex, and race were collected at enrollment. We used study data including injury characteristics, CT injury types, and Rotterdam CT score. Trained staff assessed index GCS (iGCS), representing a participant’s best GCS pre-enrollment and after initial resuscitation. Due to the clinical confounders associated with full GCS scoring in the emergency setting (i.e., endotracheal intubation, neuromuscular blockade, and sedation), index GCS motor subscore (iGCSm) was also recorded and primarily used in our analyses. Research assistants abstracted Abbreviated Injury Scale (AIS) and Injury Severity Scale (ISS) scores⁹, hospital complications including any infections, and mortality from the medical record within the trial protocol. We used recorded clinical data and laboratory values to ascertain NNOD status.

To account for non-head related injury severity, we also generated a non-head ISS score, which was the sum of squares of the highest three body region AIS scores excluding the head region, analogous to normal ISS score calculation⁹ and as previously reported.^10,11

NNOD score calculation:

We based NNOD categorization on the Sequential Organ Failure Assessment (SOFA) score, which is more strongly associated with outcomes after TBI compared to the Multiple Organ Dysfunction (MOD) scoring system.¹² We reviewed laboratory and clinical data to determine prevalence of NNOD in renal, hematologic, hepatic, cardiovascular, and respiratory systems, respectively defined as creatinine >1.2 mg/dL, platelet count <150,000/mm³, total bilirubin >1.2 mg/dL, mean arterial pressure <70 mmHg for ≥2 hours, and mechanical ventilation (Table 1). Dysfunction in each system was defined as abnormal parameters for ≥2 consecutive days, similar to prior work.⁶ When measuring time from injury to organ dysfunction, we assigned the day of dysfunction as the day it started. Due to data availability within the trial dataset, two modifications to SOFA categories were required for adjudicating NNOD status for each body system. First, for cardiovascular dysfunction, vasopressor data were not available. Thus, we used mean arterial pressure < 70 mmHg to define cardiovascular dysfunction. Second, arterial blood gas data were not available. Thus, we used endotracheal intubation status as a next-best available proxy for respiratory dysfunction. Of note, cardiovascular and respiratory dysfunction were exclusion criteria for trial enrollment. Thus, these systems could only become dysfunctional starting on day 1 post-injury in our data set, whereas renal, hematologic, and hepatic dysfunction were able to be categorized in dysfunction on day 0. We used the total sum of systems with NNOD (total NNOD) as an independent variable in analyses, with a range of 0–5.

Table 1:

Non-neurological organ dysfunction definitions.

System	Criteria
Renal	Creatinine levels > 1.2 mg/dL
Hematologic	Platelets <150,000/mm3
Hepatic	Total bilirubin levels >1.2 mg/dL
Cardiovascular	Mean arterial pressure <70 mmHg
Respiratory	Intubated & mechanically ventilated

Open in a new tab

Sample collection, storage, assays:

Biomarker data in our analyses were all previously measured as part of the Bio-ProTECT study without addition. Prior to initiation of study drug, blood was obtained from individuals, immediately centrifuged for 15 minutes, aliquoted, and frozen at −80 °C for later analysis of TBI markers. Additional blood was collected at 24 and 48 hours after randomization and processed similarly. Samples were run in duplicate at Banyan Laboratories at Banyan Biomarkers, Inc. Assay performance was previously reported.⁷ Laboratory staff were blinded to treatment group and outcome status. We did not impute missing data, which occurred with either biomarker values that were outside of the upper and lower limits of quantification, or if blood samples were unavailable at that time point. Consistent with the original Bio-ProTECT biomarker analysis⁷, we did not include biomarker values that were outside the upper and lower limits of quantification in our primary analyses. However, as a sensitivity analysis for cases where measured/detected values were estimated but were outside the limits of quantification, we re-performed our primary regression analyses that included these values.

Outcome variables:

The primary outcome is Glasgow Outcome Score-Extended (GOSE) score at 6 months post-injury.¹³ GOSE scores range from 1 (death) to 8 (good recovery). GOSE was dichotomized into unfavorable outcome (1–4) and favorable outcome (5–8). Mortality at 6 months was a secondary outcome.

Statistical analysis:

Descriptive statistics (including means and standard deviations [SD], medians and interquartile ranges [IQR], and frequencies) were computed to summarize variables of interest. Biomarker data was treated continuously, and outcomes of interest were dichotomous. Between-group comparisons were assessed using Student’s t-test for continuous variables (or Wilcoxon rank sum test for data with visually skewed distributions) and with Chi-squared tests for categorical variables (or Fisher’s exact tests when expected cell counts were <5).

A biomarker load score was generated to summarize the combined blood biomarker burden of S100b, GFAP, UCHL1, and SBDP150 from serum collected at enrollment and prior to trial drug administration. Individuals were assigned a quartile rank for each biomarker, and the mean of each individual’s quartiles was used as a biomarker load score (possible range of 1–4).

NNOD was evaluated in relation to measures of brain injury severity available within the study data. Spearman correlation coefficients were calculated to measure association between the total number of dysfunctional organ systems in the NNOD score (total NNOD) and 1) biomarker load score (defined above), 2) GCS motor score, 3) Rotterdam CT score, and 4) head AIS score. Additionally, Spearman correlation coefficients were calculated for relationships between total NNOD number of systems and all other AIS body region scores (including non-head ISS score) to examine effects of non-head injuries on NNOD.

Unadjusted and adjusted (multivariable) binary logistic regression models were used to assess the associations between neurological biomarkers and 6-month outcome (GOSE and mortality). We selected variables that could be potential confounders to include in the multivariable models a priori, including NNOD (number of systems in dysfunction), age, sex, iGCSm, Rotterdam CT score, and biomarker load score. We reported odds ratio (OR), 95% confidence interval (CI) and p values for each independent variable. Area under the receiver operating curve (AUC) of each primary logistic regression model was reported with 95% CI. For each primary multivariable regression model, the following assumptions and parameters were evaluated to ensure satisfactoriness: fit statistics and goodness-of-fit test, predicted probabilities vs. outcomes, standardized residuals, deviance, leverage, influence, linearity of each predictor, multicollinearity, and sensitivity analysis for high residuals.

We performed stratified analyses on clinically relevant groups to evaluate how initial injury severity (using either Rotterdam CT score or iGCSm) affected the strength of relationship between NNOD and outcome. Rotterdam CT score was grouped into a low group (score of 0–2) or a high group (score of 3–6). iGCSm scores were grouped into a low group (1–4) and high group (5–6). These cut points were chosen to reflect above and below the median for each variable and balance group sizes.

Analyses were performed using STATA version 17 (StataCorp LLC, College Station, TX, USA). P values < 0.05 were considered significant unless otherwise mentioned.

Results

Clinical characteristics

Cohort characteristics of the included 536 participants are summarized in Table 2. We tested all variables for differences by trial treatment group (progesterone vs. placebo), and no significant differences were observed.

Table 2:

Demographic and clinical characteristics.

Variable	Total N=536	Control (placebo) n=266	Treatment (progesterone) n=270	p-value
Age (years) – median (IQR)	35 (23–52)	35 (24–52)	34.5 (23–52)	0.75
Sex (male) - n (%)	402 (75)	202 (76)	200 (74)	0.62
Initial GCS – median (IQR)	8 (6–10)	8 (6–10)	8 (6–10)	0.69
Initial motor GCS – median (IQR)	4 (4–5)	4 (4–5)	4 (4–5)	0.45
Head AIS score - median (IQR)	4 (3–5)	4 (3–5)	4 (3–5)	0.64
ISS score – median (IQR)	25 (17–34)	26 (17–34)	24 (17–34)	0.59
Non-head ISS score – median (IQR)	9.5 (4–17)	9 (4–17)	10 (5–17)	0.33
Rotterdam CT Score - median (IQR)	3 (1)	3 (1)	3 (1)	0.44
In-hospital infection – n (%)	241 (45)	112 (42)	129 (48)	0.19
GOSE outcome – median (IQR)	5.5 (3–7)	6 (3–7)	5 (3–7)	0.49
GOSE unfavorable outcome (1–4) – n (%)	231 (43)	117 (44)	114 (42)	0.68
Mortality – n (%)	100 (19)	46 (17)	54 (20)	0.42

Open in a new tab

IQR, interquartile range; GCS, Glasgow Coma Scale; AIS, abbreviated injury scale; ISS, Injury Severity Score; CT, computed tomography; GOSE, Glasgow outcome scale-extended.

Non-neurological organ dysfunction

We determined the frequency of NNOD for each organ system using the definitions from Table 1, and we analyzed differences in NNOD frequency by key categorical variables, including sex, trial treatment group, and outcomes. Respiratory and cardiovascular function were most common organ systems effected and present in a majority of individuals (Table 3).

Table 3:

NNOD frequency by trial treatment group, GOSE outcome at 6 months, and mortality at 6 months.

By trial treatment group	Total N=536	Control (placebo) n=266	Treatment (progesterone) n=270	p-value
Hematologic dysfunction - n (%)	241 (45)	119 (45)	122 (45)	0.92
Hepatic dysfunction - n (%)	13 (2)	6 (2)	7 (3)	0.80
Renal dysfunction - n (%)	44 (8)	19 (7)	25 (9)	0.37
Respiratory dysfunction - n (%)	384 (72)	183 (69)	201 (74)	0.15
Cardiovascular dysfunction - n (%)	280 (52)	145 (55)	135 (50)	0.30
Total NNOD (mean (SD))	1.8 (1.1)	1.8 (1.1)	1.8 (1.2)	0.71

By GOSE outcome	Total N=536	Favorable GOSE (5–8) N=305	Unfavorable GOSE (1–4) N=231	p-value
Hematologic dysfunction - n (%)	241 (45)	127 (42)	114 (49)	0.08
Hepatic dysfunction - n (%)	13 (2)	5 (2)	8 (3)	0.17
Renal dysfunction - n (%)	44 (8)	19 (6)	25 (11)	0.06
Respiratory dysfunction - n (%)	384 (72)	181 (59)	203 (88)	<0.001
Cardiovascular dysfunction - n (%)	280 (52)	133 (44)	147 (64)	<0.001
Total NNOD (mean (SD))	1.8 (1.1)	1.5 (1.1)	2.2 (1.0)	<0.001

By mortality	Total, N=536	Alive N=436	Died N=100	p-value
Hematologic dysfunction - n (%)	241 (45)	195 (45)	46 (46)	0.82
Hepatic dysfunction - n (%)	13 (2)	8 (2)	5 (5)	0.08
Renal dysfunction - n (%)	44 (8)	27 (6)	17 (17)	<0.001
Respiratory dysfunction - n (%)	384 (72)	299 (69)	85 (85)	0.001
Cardiovascular dysfunction - n (%)	280 (52)	222 (51)	58 (58)	0.20
Total NNOD (mean (SD))	1.8 (1.1)	1.7 (1.1)	2.1 (1.1)	0.005

Open in a new tab

NNOD, non-neurological organ dysfunction; GOSE, Glasgow outcome scale-extended; SD, standard deviation. Bold font indicates p<0.05; italic indicates p<0.1.

Men had a significantly higher frequency than women of renal dysfunction (11% vs. 1%; p<0.001). Women tended to have higher frequencies of cardiovascular dysfunction (50% vs. 59%; p=0.07) and hematologic dysfunction (43% vs. 52%; p=0.05) than men, but these differences did not reach statistical significance. Men and women had similar frequencies of respiratory dysfunction (71% vs. 75%; p=0.38) and hepatic dysfunction (3% vs. 1%; p=0.20).

While no differences were observed between trial treatment groups (Table 3), those with unfavorable GOSE outcome had significantly higher frequency of respiratory dysfunction (88% vs. 59%; p<0.001) and cardiovascular dysfunction (64% vs. 44%; p<0.001) than those with favorable GOSE outcome (Table 3). Those who died by 6 months post-TBI had higher rates of renal dysfunction (17% vs. 6%; p<0.001) and respiratory dysfunction (85% vs. 69%; p=0.001) compared to survivors, and they had similar rates in other systems (Table 3).

We analyzed time-to-organ dysfunction (in days) during hospitalization and observed that NNOD typically occurred early in the hospitalization course, with the median time to any organ dysfunction being 1 day (IQR 1–1; range 0–3) (Supplemental Figure 1). Additionally, we report a histogram of individuals’ number of body systems in NNOD (Supplemental Figure 2); individuals most often had two systems in dysfunction.

Infections

Infections occurred in 241 individuals (45%), and the median time to infection was 5 days post-injury (IQR, 3–7 days). Similar to other of our studies¹⁴, infections occurred more frequently in those with vs. without renal dysfunction (66% vs. 43%; p<0.01), hematologic dysfunction (51% vs. 40%; p=0.01), respiratory dysfunction (95% vs. 5%; p<0.001), and cardiovascular dysfunction (64% vs. 36%; p<0.001). Individuals with infection had modestly higher biomarker load scores (mean 2.7 vs. 2.4; p<0.0001), head AIS score (mean 4.0 vs. 3.5, p<0.0001), Rotterdam CT score (mean 3.0 vs. 2.9; p<0.01), and lower iGCS (mean 7.6 vs. 8.8; p<0.0001). Individuals with any infection (compared to those without infection) had higher median ISS scores (29 vs. 22; p<0.0001), non-head ISS scores (12 vs. 8; p<0.001), and were more likely to have an unfavorable GOSE (52% vs. 36%; p<0.001) but had lower mortality (14% vs 22%; p=0.02).

TBI biomarkers

Mean serum levels of TBI biomarkers (GFAP, UCHL1, S100B, and SBDP150) were plotted by timepoint and treatment group (Figure 2); these data were reported independently in a prior secondary analysis by Korley et al. in a different format.¹⁵ For all biomarkers, levels are highest at baseline (time 0) and decline over 48 hours post-injury. No significant differences were observed between treatment groups at any time point.

We next compared serum TBI biomarker levels at enrollment between individuals with and without NNOD in each organ system (Figure 3). Respiratory, cardiovascular, and hematologic dysfunction were associated with higher levels of all biomarkers measured. Renal dysfunction was associated with higher UCHL1, S100B, and SBDP150 levels, but group differences did not reach statistical significance for GFAP. Hepatic dysfunction was not associated with differences in TBI biomarker levels.

NNOD and injury severity

Due to observed associations between NNOD systems and TBI biomarkers, we tested the correlations between total NNOD and other measures of brain injury severity available within the study data. Total NNOD was significantly correlated with all measures (biomarker load, GCS motor score, Rotterdam CT score, and head AIS score), where greater brain injury severity correlated with a higher number of NNOD-positive body systems (Table 4). Total ISS was also associated with total NNOD (Spearman’s r=0.42, p<0.001), as was the non-head ISS (Spearman’s r=0.032, p<0.0001). Other individual body regions were also associated with total NNOD, including neck, chest, abdomen, extremity, and external AIS categories (Supplemental Table 1), although all had smaller Spearman correlation coefficients than the head region.

Table 4.

Correlation coefficients between NNOD number of systems and TBI severity measures.

	Spearman’s r	P value
Biomarker load (mean quartile)	0.37	<0.0001
GCS (initial motor subscore)	−0.10	0.01
Rotterdam CT score	0.30	<0.0001
Head AIS score	0.34	<0.0001

Open in a new tab

NNOD, non-neurological organ dysfunction; TBI, traumatic brain injury; GCS, Glasgow Coma Scale; CT, computed tomography; AIS, abbreviated injury scale.

Multivariable regression models

We performed multivariable logistic regression to evaluate the effects of NNOD on outcomes (GOSE and mortality), while controlling for potential confounders of age, sex, iGCSm, Rotterdam CT score, and brain biomarker load. These results are shown in Table 5. Each additional NNOD system resulted in 1.26x higher odds of unfavorable GOSE (95% CI [1.02–1.55]; p=0.04). Sensitivity analyses performed showed that within the survivor group only (n=428), total NNOD remained significantly associated with unfavorable GOSE outcome in multivariable logistic regression (OR 1.36, 95% CI [1.07–1.73], p=0.01).

Table 5.

Multivariable logistic regression models

Unfavorable GOSE at 6 months	Odds ratio	95% CI	P value
Total NNOD	1.26	1.02–1.55	0.04
Age (y)	1.05	1.03–1.06	<0.001
Sex (male)	1.08	0.66–1.77	0.76
GCS (initial motor score)	0.75	0.68–0.84	<0.001
Rotterdam CT score	2.12	1.59–2.84	<0.001
Biomarker load score	2.24	1.60–3.14	<0.001
Overall model AUC: 0.85		0.81–0.88

Mortality at 6 months
Total NNOD	0.95	0.73–1.24	0.70
Age (y)	1.06	1.05–1.08	<0.001
Sex (male)	1.32	0.70–2.49	0.39
GCS (initial motor score)	0.66	0.52–0.85	0.001
Rotterdam CT score	2.04	1.54–2.70	<0.001
Biomarker load score	3.08	2.07–4.57	<0.001
Overall model AUC: 0.88		0.85–0.92

Survivors only – Unfavorable GOSE at 6 months	Odds ratio	95% CI	P value
Total NNOD	1.36	1.07–1.73	0.01
Age (y)	1.03	1.02–1.05	<0.001
Sex (male)	1.05	0.60–1.83	0.86
GCS (initial motor score)	0.68	0.54–0.84	<0.001
Rotterdam CT score	1.81	1.37–2.39	<0.001
Biomarker load score	1.63	1.18–2.26	0.003
Overall model AUC: 0.78		0.73–0.83

Open in a new tab

GOSE, Glasgow outcome scale-extended; NNOD, non-neurological organ dysfunction; CI, confidence interval; GCS, Glasgow Coma Scale; CT, computed tomography; AUC, area under the receiver operating curve. Bold font indicates p<0.05.

Non-head ISS score was not significantly associated in bivariate analyses with unfavorable GOSE (p=0.08) or mortality (p=0.86). When repeating the multivariable regression models and adding in non-head ISS score as an independent variable, non-head ISS score remained not significantly associated with unfavorable GOSE (OR 1.00, 95% CI 0.97–1.02; p=0.81) or mortality (OR 0.98, 95% CI 0.95–1.01; p=0.23).

Biomarker sensitivity analyses

Serum biomarker levels at enrollment were available in the majority of individuals included for analysis (S100B: n=509 [95%]; GFAP: n=498 [93%]; UCHL1: n=405 [76%]; SBDP150: n=521 [97%]). When including estimated biomarker levels whose raw values fell outside the limits of quantification, the number of available samples increased to: S100B, n=521 (96%); GFAP, n=525 (98%); UCHL1, n=596 (98%); SBDP, n=525 (98%). Biomarker load (mean quartile) updated with these values remained similarly associated with sum of NNOD systems (Spearman r=0.36, p<0.0001). Using the same multivariable logistic regression models as in Table 5 with updated biomarker load score (see Supplemental Table 2), biomarker load score remained a significant independent variable in the model for identifying unfavorable GOSE (OR 2.13, 95% CI [1.61–2.81], p<0.001) and mortality (OR 3.01, 95% CI [2.05–4.42], p<0.001) at 6 months post-injury. Total NNOD remained significant in the GOSE model (OR 1.27, 95% CI [1.02–1.57], p=0.03) and non-significant in the mortality model (OR 0.95, 95% CI [0.73–1.25], p=0.73).

Stratification analyses

Among participants with a Rotterdam CT score of 3–6, multivariable logistic regression controlling for age, sex, iGCSm, and biomarker load score revealed each system of NNOD increased the odds of unfavorable GOSE (OR 1.36, 95% CI [1.06–1.74], p=0.02). Among individuals with lower Rotterdam CT scores of 0–2, this relationship was not significant (OR 1.04, 95% CI [0.69–1.57], p=0.84). When modeling mortality as the outcome, total NNOD remained a nonsignificant independent variable in both strata; see Supplemental Tables 3 & 4.

Similarly, in the group of individuals with iGCSm ≤ 4, multivariable logistic regression controlling for age, sex, Rotterdam CT score, and biomarker load score showed each system of NNOD increased the odds of unfavorable GOSE (OR 1.34, 95% CI [1.02–1.78], p=0.04); this relationship was not significant among individuals with iGCSm >4 (OR 1.21, 95% CI [0.87–1.67], p=0.26). Total NNOD was not associated with mortality in either strata in these models; see Supplemental Tables 5 and 6.

Discussion

In this secondary analysis of data from the ProTECT III trial of progesterone vs. placebo in moderate-to-severe TBI and the Bio-ProTECT study, we observed that NNOD occurs frequently during hospitalization and is associated with poor neurological outcome and mortality at 6 months post-injury. Further, we show novel data that blood levels of central nervous system-specific biomarkers, Rotterdam CT score, iGCSm score, and head-region AIS score at time of hospital presentation are all worse in people with higher levels of NNOD. These data suggest that initial brain injury severity is linked to NNOD severity, which in turn has downstream consequences for global outcome and mortality.

One key observation is that multimodal markers of initial brain injury severity are related to NNOD after TBI. These are the first data demonstrating blood-based biomarkers of brain injury (including GFAP, UCHL1, SBDP, and S100b) on day of injury are higher in people who develop NNOD across various systems compared to those who do not. The evidence of this association is strengthened by the multiple biomarkers, from multiple brain cell sources, that are positively associated with multiple systems of dysfunction. A key question is whether NNOD is a function of brain injury or a function of total (including non-head) injury burden. Our data support a prominent role for brain injury severity. In addition to the multiple brain biomarker findings, the anatomic (Rotterdam CT score and head AIS score) and functional (GCS score) TBI measures also have consistent relationships with NNOD. This association was not explained by older age among people with NNOD. It remains possible that concomitant traumatic injuries may add to the NNOD burden – total ISS scores, non-head ISS scores, and other body region AIS scores were positively correlated with NNOD. However, the non-head ISS score was not significantly related to outcome in this cohort and had a smaller correlation with total NNOD than the head AIS score. As such, TBI rather than systemic trauma appears to be the main driver of total NNOD in this population.

It is also notable that each TBI classification category (blood biomarker, anatomic, and functional/clinical) adds prognostic value independently within multivariable regression models. This evidence supports using a more sophisticated TBI classification model, as proposed¹⁶ and recently worked on at the National Institute of Neurological Disorders and Stroke (NINDS) TBI Classification and Nomenclature Workshop.¹⁷

The underlying mechanisms for TBI-related contribution to NNOD are unclear but are suspected to involve systemic inflammatory cytokines^18–20, hormone modulators^6,21–23, or autonomic nervous system dysfunction.²⁴ Prior literature has recognized the important effects of brain injury on other organ systems. For example, acute brain injuries are suspected to predispose to acute lung injury,^25–28 and cardiac injury such as Takutsubo’s cardiomyopathy is a well-known complication of acute brain injury, including TBI.²⁹ Additionally, endocrine, renal, hepatic, immune, musculoskeletal, and blood/coagulation systems have varying degrees of dysfunction after TBI although with less clear pathophysiology.^24,30 We have also considered that hospital-acquired infections may play a role in the relationship between TBI, NNOD, and outcome. TBI likely induces a state of immune suppression and increased vulnerability for infection.^31–33 Infections can independently, and perhaps additively or synergistically with TBI, lead to NNOD. For example, ventilator-associated pneumonia in people hospitalized with TBI has been observed to increase rates of NNOD and mortality.^34,35 We assessed if infections may play a role in the present cohort. Infection status was related to injury severity, NNOD, and outcomes. However, infections occurred a median of 5 days after injury (consistent with prior data³³), which was temporally later than most instances of NNOD. Thus, infections were likely not a driving force of NNOD in the present study, but may be another type of secondary insult to brain and/or body that can contribute to unfavorable outcome.

Another key observation is that NNOD is associated with unfavorable GOSE after TBI. This finding corroborates prior literature reporting NNOD is frequent and associated with poor outcomes after TBI.^1,2 Interestingly, our multivariable regressions showed that GOSE was significantly related to the number of NNOD systems, whereas mortality was not, after controlling for covariates. One possible reason for this difference in findings between outcomes is that this cohort includes more moderate TBI cases (and more limited mortality) compared to prior literature demonstrating NNOD’s effect on mortality primarily in severe TBI cohorts.^1,6,12 As such, the mortality effect may be attenuated or underpowered due to a lesser frequency of death. Another possible reason is NNOD may primarily affect survivor-based outcomes in this population through downstream effects and associations with subsequent blood biomarker levels (which we did not analyze beyond the enrollment time point in this study). NNOD effect sizes on unfavorable GOSE outcome tended to be larger in the survivor-only models compared to whole sample models. We also observed larger NNOD effect sizes on GOSE within more severe injury groups when stratifying by TBI severity. Taken together, these findings suggest that NNOD remains an important contributor to global outcome, especially among survivors with greater injury severity. While research on TBI outcomes has historically focused on brain-centric measures, this study builds on mounting evidence that non-neurological organ systems have an important role in injury response after TBI, and their associated biomarker patterns may uniquely inform outcome and risk for secondary conditions after TBI.

The timing of NNOD after TBI may also have important implications for recovery. Some systems tend to occur earlier in the post-injury time course (e.g., renal and respiratory) whereas others tend to take on a more delayed onset (e.g. hepatic). These observations are corroborated by prior work.³³ It remains unknown how the timing of these systems’ dysfunction affects vulnerable brain tissue and injury recovery. While NNOD may impact acute mortality directly in some circumstances, it is also possible that NNOD triggers additional inflammation or neurodegenerative mechanisms resulting in further secondary brain injury. Relatedly, the trial data set did not contain primary causes of death, including withdrawal of life support, which can result in underestimated hospital complications such as NNOD.

Optimal methods for combining biomarker data are not clear. We performed a mean quartile calculation of the four brain biomarkers measured in this study, but other methods could also be reasonable depending on context of use. In prior work, an empirically derived cutoff threshold was used to maximize prognostic value.⁷ Our group has used similar quartile score systems^20,36 as well as weighted scoring systems³⁷ as a dimension reduction technique to quantify inflammatory cytokine levels after TBI. Longitudinal biomarker level areas under the curve have also been reported as a measure in neurotrauma from blast exposure.³⁸ The use of multiple biomarker levels in a quantitative fashion remains an underexamined aspect despite the field’s movement towards using multiple biomarker assays and panels, and further consideration of this in future work is warranted.

There are limitations associated with this study. This study relies on retrospective adjudication of NNOD using available trial data. This research design could potentially lead to unknown missing data for NNOD variables, as some lab abnormalities or vital sign changes may not have been completely reflected in the trial documentation. Due to the trial exclusion criteria, we also had to adjust our NNOD definitions (see Methods) such that they were grounded in, but not identical to, the standard SOFA criteria.

Our findings present multiple directions for future study. The interplay between NNOD and brain injury remains not fully understood. These data intriguingly show initial brain injury severity is related to subsequent early NNOD development. It remains unknown whether NNOD further fuels secondary brain injury and how this is reflected in longitudinal biomarker levels. Cross-lag analyses over time intervals may offer additional insight into longitudinal effects of biomarker levels, NNOD, and outcome after TBI.⁶ Additionally, we remain interested in systemic blood biomarkers that are not brain-derived, and the potential role of these biomarkers in elucidating both brain and non-brain injury responses after TBI. The data from this clinical trial present special opportunities to interrogate the effects of progesterone, other hormones, and inflammation profiles on injury response.

Conclusions:

Acute moderate-to-severe TBI frequently results in NNOD. Greater brain injury severity at presentation, as measured by biochemical (blood biomarkers), functional (GCS score), and anatomic (Rotterdam CT score, head-region AIS score) measures, is associated with higher levels of NNOD. NNOD is in turn associated with unfavorable outcome at 6-months after TBI primarily in the severely injured.

Transparency, Rigor, and Reproducibility Summary for a Clinical Study:

Study registration took place for the underlying clinical trial from which this analysis was based (NCT00822900). This study’s secondary analysis plan was not formally registered. Sample size was limited to the underlying trial population. A study flow diagram is included as Figure 1. Due to this being a secondary analysis of existing data, the data analyses were performed by investigators who had access to relevant characteristics of the participants and were not blinded to treatment group. All data were contained within the ProTECT trial database. Data were analyzed using STATA version 17 (StataCorp LLC, College Station, TX, USA). The primary clinical outcome measure (GOSE) is an established standard in the field.³⁹ Statistical tests and assumptions are described in the methods section. Effect sizes and confidence intervals have been reported in the main text for all outcomes. Methods that do not require correction for multiple comparisons were used, including multivariable binary logistic regression. No replication or external validation studies have been performed or are planned/ongoing at this time to our knowledge. De-identified data from this study are available in a protected archive: FITBIR (https://fitbir.nih.gov), DOI: 10.23718/FITBIR/1503320, as of 10/16/2024. Data can be obtained by completing a data access request with FITBIR. Analytic code used to conduct the analyses presented in this study are not available in a public repository but may be available by emailing the corresponding authors. The authors agree to provide the full content of the manuscript on request by contacting the corresponding authors.

Supplementary Material

NIHMS2135603-supplement-1.pdf^{(174.5KB, pdf)}

Acknowledgments

The authors would like to thank Jessa Darwin for editorial assistance on the manuscript.

Funding:

This work was supported by NIH R21NS111063, NIH T32HL134615, NIH U01NS062778, and NIH R01NS071867.

Footnotes

Authors’ disclosure (conflict of interest) statement:

The authors have no conflicts of interest to disclose.

This work was presented, in part, as an abstract at the National Neurotrauma Society Annual Symposium, virtually, in June 2021.

References

1.Zygun DA, Kortbeek JB, Fick GH, Laupland KB, Doig CJ. Non-neurologic organ dysfunction in severe traumatic brain injury. Crit Care Med. 2005;33(3):654–660. doi: 10.1097/01.ccm.0000155911.01844.54 [DOI] [PubMed] [Google Scholar]
2.Kemp CD, Johnson JC, Riordan WP, Cotton BA. How We Die: The Impact of Nonneurologic Organ Dysfunction after Severe Traumatic Brain Injury. Am Surg. 2008;74(9):866–872. doi: 10.1177/000313480807400921 [DOI] [PubMed] [Google Scholar]
3.Brew K, Nagase H. The tissue inhibitors of metalloproteinases (TIMPs): An ancient family with structural and functional diversity. Biochim Biophys Acta. 2010;1803(1):55. doi: 10.1016/j.bbamcr.2010.01.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Hendrickson CM, Gibb SL, Miyazawa BY, et al. Elevated plasma levels of TIMP-3 are associated with a higher risk of acute respiratory distress syndrome and death following severe isolated traumatic brain injury. Trauma Surg Acute Care Open. 2018;3(1):e000171. doi: 10.1136/tsaco-2018-000171 [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Lee S, Hwang H, Yamal JM, et al. IMPACT probability of poor outcome and plasma cytokine concentrations are associated with multiple organ dysfunction syndrome following traumatic brain injury. J Neurosurg. 2019;131(6):1931–1937. doi: 10.3171/2018.8.JNS18676 [DOI] [PubMed] [Google Scholar]
6.Kumar RG, DiSanto D, Awan N, et al. Temporal Acute Serum Estradiol and Tumor Necrosis Factor-α Associations and Risk of Death after Severe Traumatic Brain Injury. J Neurotrauma. 2020;37(20):2198–2210. doi: 10.1089/neu.2019.6577 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Frankel M, Fan L, Yeatts SD, et al. Association of Very Early Serum Levels of S100B, Glial Fibrillary Acidic Protein, Ubiquitin C-Terminal Hydrolase-L1, and Spectrin Breakdown Product with Outcome in ProTECT III. J Neurotrauma. 2019;36(20):2863–2871. doi: 10.1089/neu.2018.5809 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Wright DW, Yeatts SD, Silbergleit R, et al. Very Early Administration of Progesterone for Acute Traumatic Brain Injury. N Engl J Med. 2014;371(26):2457–2466. doi: 10.1056/NEJMoa1404304 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Baker SP, O’Neill B, Haddon W, Long WB. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma. 1974;14(3):187–196. [PubMed] [Google Scholar]
10.Chien DK, Hwang HF, Lin MR. Injury severity measures for predicting return-to-work after a traumatic brain injury. Accid Anal Prev. 2017;98:101–107. doi: 10.1016/j.aap.2016.09.025 [DOI] [PubMed] [Google Scholar]
11.Rubenstein R, McQuillan L, Wang KKW, et al. Temporal Profiles of P-Tau, T-Tau, and P-Tau:Tau Ratios in Cerebrospinal Fluid and Blood from Moderate-Severe Traumatic Brain Injury Patients and Relationship to 6–12 Month Global Outcomes. J Neurotrauma. 2024;41(3–4):369–392. doi: 10.1089/neu.2022.0479 [DOI] [PubMed] [Google Scholar]
12.Zygun D, Berthiaume L, Laupland K, Kortbeek J, Doig C. SOFA is superior to MOD score for the determination of non-neurologic organ dysfunction in patients with severe traumatic brain injury: a cohort study. Crit Care. 2006;10(4):R115. doi: 10.1186/cc5007 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Wilson JT, Pettigrew LE, Teasdale G. Structured interviews for the Glasgow Outcome Scale and the extended Glasgow Outcome Scale: guidelines for their use. J Neurotrauma. 1998;15(8):573–585. [DOI] [PubMed] [Google Scholar]
14.Wagner A, Sjobeck G, Maczuzak J, et al. Characterizing Acute Infections and their Associations with Injury Severity, Hospital Course and Seizure Risk after Moderate-to-Severe TBI. In: Poster presentation at 2024 Military Health System Research Symposium, August 26-29, 2024, Kissimmee, Florida. [Google Scholar]
15.Korley F, Pauls Q, Yeatts SD, et al. Progesterone Treatment Does Not Decrease Serum Levels of Biomarkers of Glial and Neuronal Cell Injury in Moderate and Severe Traumatic Brain Injury Subjects: A Secondary Analysis of the Progesterone for Traumatic Brain Injury, Experimental Clinical Treatment (ProTECT) III Trial. J Neurotrauma. 2021;38(14):1953–1960. doi: 10.1089/neu.2020.7072 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.National Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Health Care Services; Board on Health Sciences Policy; Committee on Accelerating Progress in Traumatic Brain Injury Research and Care. Traumatic Brain Injury: A Roadmap for Accelerating Progress. (Matney C, Bowman, Berwick D, eds.). National Academies Press (US); 2022. Accessed January 22, 2025. http://www.ncbi.nlm.nih.gov/books/NBK580081/ [PubMed] [Google Scholar]
17.NINDS TBI Classification and Nomenclature Workshop | National Institute of Neurological Disorders and Stroke. Accessed January 22, 2025. https://www.ninds.nih.gov/news-events/events/ninds-tbi-classification-and-nomenclature-workshop
18.Rowland B, Savarraj JPJ, Karri J, et al. Acute Inflammation in Traumatic Brain Injury and Polytrauma Patients Using Network Analysis. Shock Augusta Ga. 2020;53(1):24–34. doi: 10.1097/SHK.0000000000001349 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Lewis CT, Savarraj JPJ, McGuire MF, Hergenroeder GW, Alex Choi H, Kitagawa RS. Elevated inflammation and decreased platelet activity is associated with poor outcomes after traumatic brain injury. J Clin Neurosci Off J Neurosurg Soc Australas. 2019;70:37–41. doi: 10.1016/j.jocn.2019.09.004 [DOI] [PubMed] [Google Scholar]
20.Santarsieri M, Kumar RG, Kochanek PM, Berga S, Wagner AK. Variable neuroendocrine-immune dysfunction in individuals with unfavorable outcome after severe traumatic brain injury. Brain Behav Immun. 2015;45:15–27. doi: 10.1016/j.bbi.2014.09.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Rakholia MV, Kumar RG, Oh BM, et al. Systemic Estrone Production and Injury-Induced Sex Hormone Steroidogenesis after Severe Traumatic Brain Injury: A Prognostic Indicator of Traumatic Brain Injury-Related Mortality. J Neurotrauma. 2019;36(7):1156–1167. doi: 10.1089/neu.2018.5782 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Ranganathan P, Kumar RG, Oh BM, Rakholia MV, Berga SL, Wagner AK. Estradiol to Androstenedione Ratios Moderate the Relationship between Neurological Injury Severity and Mortality Risk after Severe Traumatic Brain Injury. J Neurotrauma. 2019;36(4):538–547. doi: 10.1089/neu.2018.5677 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Wagner AK, McCullough EH, Niyonkuru C, et al. Acute serum hormone levels: characterization and prognosis after severe traumatic brain injury. J Neurotrauma. 2011;28(6):871–888. doi: 10.1089/neu.2010.1586 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.McDonald SJ, Sharkey JM, Sun M, et al. Beyond the Brain: Peripheral Interactions after Traumatic Brain Injury. J Neurotrauma. 2020;37(5):770–781. doi: 10.1089/neu.2019.6885 [DOI] [PubMed] [Google Scholar]
25.Ziaka M, Exadaktylos A. Brain-lung interactions and mechanical ventilation in patients with isolated brain injury. Crit Care Lond Engl. 2021;25(1):358. doi: 10.1186/s13054-021-03778-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Finsterer J Neurological Perspectives of Neurogenic Pulmonary Edema. Eur Neurol. 2019;81(1–2):94–102. doi: 10.1159/000500139 [DOI] [PubMed] [Google Scholar]
27.Komisarow JM, Chen F, Vavilala MS, Laskowitz D, James ML, Krishnamoorthy V. Epidemiology and Outcomes of Acute Respiratory Distress Syndrome Following Isolated Severe Traumatic Brain Injury. J Intensive Care Med. 2022;37(1):68–74. doi: 10.1177/0885066620972001 [DOI] [PubMed] [Google Scholar]
28.Ziaka M, Exadaktylos A. Pathophysiology of acute lung injury in patients with acute brain injury: the triple-hit hypothesis. Crit Care Lond Engl. 2024;28(1):71. doi: 10.1186/s13054-024-04855-w [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Coppalini G, Salvagno M, Peluso L, et al. Cardiac Injury After Traumatic Brain Injury: Clinical Consequences and Management. Neurocrit Care. 2024;40(2):477–485. doi: 10.1007/s12028-023-01777-3 [DOI] [PubMed] [Google Scholar]
30.Robba C, Bonatti G, Pelosi P, Citerio G. Extracranial complications after traumatic brain injury: targeting the brain and the body. Curr Opin Crit Care. 2020;26(2):137–146. doi: 10.1097/MCC.0000000000000707 [DOI] [PubMed] [Google Scholar]
31.Gandasasmita N, Li J, Loane DJ, Semple BD. Experimental Models of Hospital-Acquired Infections After Traumatic Brain Injury: Challenges and Opportunities. J Neurotrauma. 2024;41(7–8):752–770. doi: 10.1089/neu.2023.0453 [DOI] [PubMed] [Google Scholar]
32.Sharma R, Shultz SR, Robinson MJ, et al. Infections after a traumatic brain injury: The complex interplay between the immune and neurological systems. Brain Behav Immun. 2019;79:63–74. doi: 10.1016/j.bbi.2019.04.034 [DOI] [PubMed] [Google Scholar]
33.Levochkina M, McQuillan L, Awan N, et al. Neutrophil-to-Lymphocyte Ratios and Infections after Traumatic Brain Injury: Associations with Hospital Resource Utilization and Long-Term Outcome. J Clin Med. 2021;10(19):4365. doi: 10.3390/jcm10194365 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Zygun DA, Zuege DJ, Boiteau PJE, et al. Ventilator-associated pneumonia in severe traumatic brain injury. Neurocrit Care. 2006;5(2):108–114. doi: 10.1385/ncc:5:2:108 [DOI] [PubMed] [Google Scholar]
35.Kumar RG, Kesinger MR, Juengst SB, et al. Effects of Hospital-Acquired Pneumonia on Long-Term Recovery and Hospital Resource Utilization Following Moderate to Severe Traumatic Brain Injury. J Trauma Acute Care Surg. 2020;88(4):491–500. doi: 10.1097/TA.0000000000002562 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Kumar RG, Boles JA, Wagner AK. Chronic Inflammation After Severe Traumatic Brain Injury: Characterization and Associations With Outcome at 6 and 12 Months Postinjury. J Head Trauma Rehabil. 2015;30(6):369–381. doi: 10.1097/HTR.0000000000000067 [DOI] [PubMed] [Google Scholar]
37.Milleville KA, Awan N, Disanto D, Kumar RG, Wagner AK. Early chronic systemic inflammation and associations with cognitive performance after moderate to severe TBI. Brain Behav Immun - Health. 2021;11:100185. doi: 10.1016/j.bbih.2020.100185 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Tate CM, Wang KKW, Eonta S, et al. Serum Brain Biomarker Level, Neurocognitive Performance, and Self-Reported Symptom Changes in Soldiers Repeatedly Exposed to Low-Level Blast: A Breacher Pilot Study. J Neurotrauma. 2013;30(19):1620–1630. doi: 10.1089/neu.2012.2683 [DOI] [PubMed] [Google Scholar]
39.Teasdale G, Maas A, Lecky F, Manley G, Stocchetti N, Murray G. The Glasgow Coma Scale at 40 years: standing the test of time. Lancet Neurol. 2014;13(8):844–854. doi: 10.1016/S1474-4422(14)70120-6 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS2135603-supplement-1.pdf^{(174.5KB, pdf)}

[R1] 1.Zygun DA, Kortbeek JB, Fick GH, Laupland KB, Doig CJ. Non-neurologic organ dysfunction in severe traumatic brain injury. Crit Care Med. 2005;33(3):654–660. doi: 10.1097/01.ccm.0000155911.01844.54 [DOI] [PubMed] [Google Scholar]

[R2] 2.Kemp CD, Johnson JC, Riordan WP, Cotton BA. How We Die: The Impact of Nonneurologic Organ Dysfunction after Severe Traumatic Brain Injury. Am Surg. 2008;74(9):866–872. doi: 10.1177/000313480807400921 [DOI] [PubMed] [Google Scholar]

[R3] 3.Brew K, Nagase H. The tissue inhibitors of metalloproteinases (TIMPs): An ancient family with structural and functional diversity. Biochim Biophys Acta. 2010;1803(1):55. doi: 10.1016/j.bbamcr.2010.01.003 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R4] 4.Hendrickson CM, Gibb SL, Miyazawa BY, et al. Elevated plasma levels of TIMP-3 are associated with a higher risk of acute respiratory distress syndrome and death following severe isolated traumatic brain injury. Trauma Surg Acute Care Open. 2018;3(1):e000171. doi: 10.1136/tsaco-2018-000171 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R5] 5.Lee S, Hwang H, Yamal JM, et al. IMPACT probability of poor outcome and plasma cytokine concentrations are associated with multiple organ dysfunction syndrome following traumatic brain injury. J Neurosurg. 2019;131(6):1931–1937. doi: 10.3171/2018.8.JNS18676 [DOI] [PubMed] [Google Scholar]

[R6] 6.Kumar RG, DiSanto D, Awan N, et al. Temporal Acute Serum Estradiol and Tumor Necrosis Factor-α Associations and Risk of Death after Severe Traumatic Brain Injury. J Neurotrauma. 2020;37(20):2198–2210. doi: 10.1089/neu.2019.6577 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] 7.Frankel M, Fan L, Yeatts SD, et al. Association of Very Early Serum Levels of S100B, Glial Fibrillary Acidic Protein, Ubiquitin C-Terminal Hydrolase-L1, and Spectrin Breakdown Product with Outcome in ProTECT III. J Neurotrauma. 2019;36(20):2863–2871. doi: 10.1089/neu.2018.5809 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Wright DW, Yeatts SD, Silbergleit R, et al. Very Early Administration of Progesterone for Acute Traumatic Brain Injury. N Engl J Med. 2014;371(26):2457–2466. doi: 10.1056/NEJMoa1404304 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.Baker SP, O’Neill B, Haddon W, Long WB. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma. 1974;14(3):187–196. [PubMed] [Google Scholar]

[R10] 10.Chien DK, Hwang HF, Lin MR. Injury severity measures for predicting return-to-work after a traumatic brain injury. Accid Anal Prev. 2017;98:101–107. doi: 10.1016/j.aap.2016.09.025 [DOI] [PubMed] [Google Scholar]

[R11] 11.Rubenstein R, McQuillan L, Wang KKW, et al. Temporal Profiles of P-Tau, T-Tau, and P-Tau:Tau Ratios in Cerebrospinal Fluid and Blood from Moderate-Severe Traumatic Brain Injury Patients and Relationship to 6–12 Month Global Outcomes. J Neurotrauma. 2024;41(3–4):369–392. doi: 10.1089/neu.2022.0479 [DOI] [PubMed] [Google Scholar]

[R12] 12.Zygun D, Berthiaume L, Laupland K, Kortbeek J, Doig C. SOFA is superior to MOD score for the determination of non-neurologic organ dysfunction in patients with severe traumatic brain injury: a cohort study. Crit Care. 2006;10(4):R115. doi: 10.1186/cc5007 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] 13.Wilson JT, Pettigrew LE, Teasdale G. Structured interviews for the Glasgow Outcome Scale and the extended Glasgow Outcome Scale: guidelines for their use. J Neurotrauma. 1998;15(8):573–585. [DOI] [PubMed] [Google Scholar]

[R14] 14.Wagner A, Sjobeck G, Maczuzak J, et al. Characterizing Acute Infections and their Associations with Injury Severity, Hospital Course and Seizure Risk after Moderate-to-Severe TBI. In: Poster presentation at 2024 Military Health System Research Symposium, August 26-29, 2024, Kissimmee, Florida. [Google Scholar]

[R15] 15.Korley F, Pauls Q, Yeatts SD, et al. Progesterone Treatment Does Not Decrease Serum Levels of Biomarkers of Glial and Neuronal Cell Injury in Moderate and Severe Traumatic Brain Injury Subjects: A Secondary Analysis of the Progesterone for Traumatic Brain Injury, Experimental Clinical Treatment (ProTECT) III Trial. J Neurotrauma. 2021;38(14):1953–1960. doi: 10.1089/neu.2020.7072 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.National Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Health Care Services; Board on Health Sciences Policy; Committee on Accelerating Progress in Traumatic Brain Injury Research and Care. Traumatic Brain Injury: A Roadmap for Accelerating Progress. (Matney C, Bowman, Berwick D, eds.). National Academies Press (US); 2022. Accessed January 22, 2025. http://www.ncbi.nlm.nih.gov/books/NBK580081/ [PubMed] [Google Scholar]

[R17] 17.NINDS TBI Classification and Nomenclature Workshop | National Institute of Neurological Disorders and Stroke. Accessed January 22, 2025. https://www.ninds.nih.gov/news-events/events/ninds-tbi-classification-and-nomenclature-workshop

[R18] 18.Rowland B, Savarraj JPJ, Karri J, et al. Acute Inflammation in Traumatic Brain Injury and Polytrauma Patients Using Network Analysis. Shock Augusta Ga. 2020;53(1):24–34. doi: 10.1097/SHK.0000000000001349 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Lewis CT, Savarraj JPJ, McGuire MF, Hergenroeder GW, Alex Choi H, Kitagawa RS. Elevated inflammation and decreased platelet activity is associated with poor outcomes after traumatic brain injury. J Clin Neurosci Off J Neurosurg Soc Australas. 2019;70:37–41. doi: 10.1016/j.jocn.2019.09.004 [DOI] [PubMed] [Google Scholar]

[R20] 20.Santarsieri M, Kumar RG, Kochanek PM, Berga S, Wagner AK. Variable neuroendocrine-immune dysfunction in individuals with unfavorable outcome after severe traumatic brain injury. Brain Behav Immun. 2015;45:15–27. doi: 10.1016/j.bbi.2014.09.003 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] 21.Rakholia MV, Kumar RG, Oh BM, et al. Systemic Estrone Production and Injury-Induced Sex Hormone Steroidogenesis after Severe Traumatic Brain Injury: A Prognostic Indicator of Traumatic Brain Injury-Related Mortality. J Neurotrauma. 2019;36(7):1156–1167. doi: 10.1089/neu.2018.5782 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] 22.Ranganathan P, Kumar RG, Oh BM, Rakholia MV, Berga SL, Wagner AK. Estradiol to Androstenedione Ratios Moderate the Relationship between Neurological Injury Severity and Mortality Risk after Severe Traumatic Brain Injury. J Neurotrauma. 2019;36(4):538–547. doi: 10.1089/neu.2018.5677 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R23] 23.Wagner AK, McCullough EH, Niyonkuru C, et al. Acute serum hormone levels: characterization and prognosis after severe traumatic brain injury. J Neurotrauma. 2011;28(6):871–888. doi: 10.1089/neu.2010.1586 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] 24.McDonald SJ, Sharkey JM, Sun M, et al. Beyond the Brain: Peripheral Interactions after Traumatic Brain Injury. J Neurotrauma. 2020;37(5):770–781. doi: 10.1089/neu.2019.6885 [DOI] [PubMed] [Google Scholar]

[R25] 25.Ziaka M, Exadaktylos A. Brain-lung interactions and mechanical ventilation in patients with isolated brain injury. Crit Care Lond Engl. 2021;25(1):358. doi: 10.1186/s13054-021-03778-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] 26.Finsterer J Neurological Perspectives of Neurogenic Pulmonary Edema. Eur Neurol. 2019;81(1–2):94–102. doi: 10.1159/000500139 [DOI] [PubMed] [Google Scholar]

[R27] 27.Komisarow JM, Chen F, Vavilala MS, Laskowitz D, James ML, Krishnamoorthy V. Epidemiology and Outcomes of Acute Respiratory Distress Syndrome Following Isolated Severe Traumatic Brain Injury. J Intensive Care Med. 2022;37(1):68–74. doi: 10.1177/0885066620972001 [DOI] [PubMed] [Google Scholar]

[R28] 28.Ziaka M, Exadaktylos A. Pathophysiology of acute lung injury in patients with acute brain injury: the triple-hit hypothesis. Crit Care Lond Engl. 2024;28(1):71. doi: 10.1186/s13054-024-04855-w [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] 29.Coppalini G, Salvagno M, Peluso L, et al. Cardiac Injury After Traumatic Brain Injury: Clinical Consequences and Management. Neurocrit Care. 2024;40(2):477–485. doi: 10.1007/s12028-023-01777-3 [DOI] [PubMed] [Google Scholar]

[R30] 30.Robba C, Bonatti G, Pelosi P, Citerio G. Extracranial complications after traumatic brain injury: targeting the brain and the body. Curr Opin Crit Care. 2020;26(2):137–146. doi: 10.1097/MCC.0000000000000707 [DOI] [PubMed] [Google Scholar]

[R31] 31.Gandasasmita N, Li J, Loane DJ, Semple BD. Experimental Models of Hospital-Acquired Infections After Traumatic Brain Injury: Challenges and Opportunities. J Neurotrauma. 2024;41(7–8):752–770. doi: 10.1089/neu.2023.0453 [DOI] [PubMed] [Google Scholar]

[R32] 32.Sharma R, Shultz SR, Robinson MJ, et al. Infections after a traumatic brain injury: The complex interplay between the immune and neurological systems. Brain Behav Immun. 2019;79:63–74. doi: 10.1016/j.bbi.2019.04.034 [DOI] [PubMed] [Google Scholar]

[R33] 33.Levochkina M, McQuillan L, Awan N, et al. Neutrophil-to-Lymphocyte Ratios and Infections after Traumatic Brain Injury: Associations with Hospital Resource Utilization and Long-Term Outcome. J Clin Med. 2021;10(19):4365. doi: 10.3390/jcm10194365 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R34] 34.Zygun DA, Zuege DJ, Boiteau PJE, et al. Ventilator-associated pneumonia in severe traumatic brain injury. Neurocrit Care. 2006;5(2):108–114. doi: 10.1385/ncc:5:2:108 [DOI] [PubMed] [Google Scholar]

[R35] 35.Kumar RG, Kesinger MR, Juengst SB, et al. Effects of Hospital-Acquired Pneumonia on Long-Term Recovery and Hospital Resource Utilization Following Moderate to Severe Traumatic Brain Injury. J Trauma Acute Care Surg. 2020;88(4):491–500. doi: 10.1097/TA.0000000000002562 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] 36.Kumar RG, Boles JA, Wagner AK. Chronic Inflammation After Severe Traumatic Brain Injury: Characterization and Associations With Outcome at 6 and 12 Months Postinjury. J Head Trauma Rehabil. 2015;30(6):369–381. doi: 10.1097/HTR.0000000000000067 [DOI] [PubMed] [Google Scholar]

[R37] 37.Milleville KA, Awan N, Disanto D, Kumar RG, Wagner AK. Early chronic systemic inflammation and associations with cognitive performance after moderate to severe TBI. Brain Behav Immun - Health. 2021;11:100185. doi: 10.1016/j.bbih.2020.100185 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] 38.Tate CM, Wang KKW, Eonta S, et al. Serum Brain Biomarker Level, Neurocognitive Performance, and Self-Reported Symptom Changes in Soldiers Repeatedly Exposed to Low-Level Blast: A Breacher Pilot Study. J Neurotrauma. 2013;30(19):1620–1630. doi: 10.1089/neu.2012.2683 [DOI] [PubMed] [Google Scholar]

[R39] 39.Teasdale G, Maas A, Lecky F, Manley G, Stocchetti N, Murray G. The Glasgow Coma Scale at 40 years: standing the test of time. Lancet Neurol. 2014;13(8):844–854. doi: 10.1016/S1474-4422(14)70120-6 [DOI] [PubMed] [Google Scholar]

PERMALINK

Traumatic brain injury blood biomarkers are associated with non-neurological organ dysfunction: a secondary analysis of ProTECT III and Bio-ProTECT

David J Barton

Erica Fan

Ian G Gober

Nabil Awan

Raj G Kumar

Jenna C Carlson

Michael R Frankel

David W Wright

Amy K Wagner

Abstract

Introduction

Methods

Study design

Figure 1.

Demographic and clinical variables:

NNOD score calculation:

Table 1:

Sample collection, storage, assays:

Outcome variables:

Statistical analysis:

Results

Clinical characteristics

Table 2:

Non-neurological organ dysfunction

Table 3:

Infections

TBI biomarkers

Figure 2.

Figure 3.

NNOD and injury severity

Table 4.

Multivariable regression models

Table 5.

Biomarker sensitivity analyses

Stratification analyses

Discussion

Conclusions:

Transparency, Rigor, and Reproducibility Summary for a Clinical Study:

Supplementary Material

Acknowledgments

Funding:

Footnotes

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases