A Randomized Placebo-Controlled Trial of Mild Hyperbaric Oxygen on Serum Biomarkers in Persistent Post-Concussive Symptoms: Analysis at 13-Week Follow-Up

Abstract

Background: The management of persistent post-concussive symptoms (PPCS) is limited by the absence of objective biomarkers to guide treatment. We examined the early effects of a mild hyperbaric oxygen protocol on serum biomarkers of neuronal injury (neurofilament light chain, NfL), astrogliosis (glial fibrillary acidic protein, GFAP), acute neuronal injury (ubiquitin C-terminal hydrolase L1, UCH-L1), and axonal stability (total tau) in patients with PPCS. Methods: In this single-center, randomized, placebo-controlled trial, we enrolled adults with PPCS lasting from 3 months to 5 years after mild traumatic brain injury. Participants received 40 sessions of either active treatment (≥99% O₂ at 1.5 atmospheres absolute, ATA) or a true chamber placebo (21% O₂ with simulated pressure changes). Serum samples were collected at baseline and 13 weeks after treatment. The primary outcome was the difference between groups in serum NfL levels. Analysis was performed on an intention-to-treat basis using a two-way ANOVA with Šídák’s multiple comparison test. Findings: Of 84 individuals assessed, 20 were randomized (Placebo, n = 9; Intervention, n = 11). Eight from each group received their respective interventions. At 13 weeks, one participant from each group was lost to follow-up, leaving seven per group for analysis. We found no significant differences in serum levels of GFAP, NfL, total tau, or UCH-L1 between the intervention and placebo groups from baseline to 13 weeks. Conclusions: A 40-session mild hyperbaric oxygen protocol at 1.5 ATA did not significantly change serum biomarkers of neuronal injury, astrogliosis, or acute neuronal damage at 13 weeks post-treatment in individuals with PPCS. This early-phase analysis, at the highest point of participant retention, provides no evidence of a treatment effect on these pathophysiological markers.

1. Introduction

Mild traumatic brain injury (mTBI) is a major public health concern, with a significant subset of patients experiencing persistent post-concussive symptoms (PPCS) that endure for months to years [1]. The subjective nature of PPCS complicates both diagnosis and the objective assessment of novel therapies. Blood-based biomarkers offer a quantifiable window into the neuropathological sequelae of mTBI. Serum neurofilament light chain (NfL), a marker of axonal injury, can remain elevated chronically [2]. In contrast, glial fibrillary acidic protein (GFAP) and ubiquitin C-terminal hydrolase L1 (UCH-L1), markers of astrogliosis and acute neuronal injury, respectively, typically peak shortly after injury [3,4,5].

Hyperbaric oxygen treatment (HBOT) has been explored as a potential therapy for PPCS, with proposed mechanisms including enhanced cerebral oxygenation, reduced inflammation, and support for neuroplasticity [6,7]. A critical challenge in interpreting the HBOT literature has been heterogeneity in treatment parameters and patient populations. Furthermore, the therapeutic definition of HBOT has evolved [8].

We hypothesized that 40 sessions of hyperbaric oxygen at 1.5 ATA would reduce serum NfL, suggesting mitigation of ongoing axonal degeneration, and conducted a parallel assessment of GFAP, tau, and UCH-L. While the original study design included follow-ups to 18 months, substantial attrition at later time points rendered those comparisons statistically unreliable. To provide a robust and interpretable analysis, we prioritized the 13-week post-intervention time point, where participant retention was the highest, and the risk of spurious findings due to a small sample size was minimal. This analysis tests the early biological effect of the intervention on a panel of relevant serum biomarkers.

2. Materials and Methods

2.1. Study Design and Participants

We conducted a randomized, placebo-controlled trial at a single center. Adults (≥18 years) with a neurologist-confirmed diagnosis of PPCS following non-penetrating mTBI, with symptoms persisting for 3 to 60 months, were eligible. Exclusion criteria included prior HBOT, untreated pneumothorax, or unstable psychiatric conditions. The Essentia Health Institutional Review Board approved the protocol (code #21714) on 24 September 2021, and all participants provided written informed consent. The trial was registered at ClinicalTrials.gov (NCT05173818) before participant enrollment.

Individuals with persistent post-concussive symptoms were identified through referrals from outpatient clinics and hospitals across North Dakota and western Minnesota to the Essentia Health Fargo neurology clinic, where diagnostic evaluation was conducted by the study neurologist. Hyperbaric research nurses performed preliminary eligibility screening. Individuals deemed potentially eligible completed an initial screening consent during their neurology evaluation, at which point study procedures were explained in detail. Those who expressed willingness to participate underwent further assessment by the hyperbaric research staff, including fitness to receive treatment, and written informed consent was obtained before formal enrollment. Subsequently, participants were randomized using the REDCap platform, and each participant completed baseline outcome measures, laboratory, and imaging studies as indicated.

When this study was designed and initiated, the Undersea and Hyperbaric Medical Society (UHMS) defined therapeutic HBOT as oxygen delivery at pressures ≥ 1.4 ATA [8]. Our protocol of 1.5 ATA with 100% oxygen thus met the contemporary standard. The 1.5 ATA with 100% oxygen for 60 min protocol was chosen in keeping with published studies at the time of designing this study [9]. However, the UHMS has since revised its definition, specifying pressures ≥ 2.0 ATA with longer session durations for therapeutic HBOT [10]. Consequently, our intervention is now more accurately classified as a mild hyperbaric oxygen exposure. This reclassification is essential for contextualizing our findings within the current therapeutic landscape.

2.2. Procedures

Participants were randomly assigned using concealed allocation.

• Intervention Group: Received 40 daily sessions in a monoplace chamber, breathing ≥99% oxygen at 1.5 ATA for 60 min. Daily sessions in this context denote once a day, five days a week (weekdays).
• Placebo-Control Group: Underwent a valid placebo procedure breathing 21% oxygen (room air). A protocol involving a 5 min brief compression to 1.2 ATA with ambient noise, followed by 55 min at 1.0 ATA in silence, and a final noisy “ascent” simulated the sensory experience of active hyperbaric oxygen treatment.

2.3. Biomarker Analysis

Serum was collected at baseline and 13 weeks (±2 weeks) after the final treatment session. Concentrations of GFAP, NfL, total tau, and UCH-L1 were quantified in a single batch using the Quanterix Simoa Neurology 4-Plex B kit (Quanterix, Billerica, MA, USA) according to the manufacturer’s protocol. Laboratory personnel were blinded to treatment allocation and clinical outcomes until all sample analyses were completed.

2.4. Statistical Analysis

Data were analyzed using GraphPad Prism 10 (GraphPad Software, Boston, MA, USA). Biomarker concentrations were compared between groups at the two time points (Baseline and 13-Week) using a two-way repeated-measures ANOVA with Šídák’s multiple-comparisons test for post hoc analysis. Baseline characteristics were compared using unpaired t-tests or Fisher’s exact tests. The primary analysis was intention-to-treat. Statistical significance was set at p ≤ 0.05.

2.5. Outcome Measures

All participants completed the Neurobehavioral Symptom Inventory (NSI), reporting each of the 22 symptoms in the two weeks before baseline evaluation. The NSI total score is a sum of the 22 items (range 0 to 88) and includes three subscales: somatic/sensory (11 items), affective (7 items), and cognitive (4 items). The NSI was administered at baseline only and was not repeated post-treatment. Neuropsychological performance was assessed at baseline and post-treatment using the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS); these findings are the subject of a companion manuscript currently in preparation.

3. Results

Between June 2021 and January 2023, 84 individuals were assessed for eligibility, and 20 were randomized (Placebo: n = 9; Intervention: n = 11) (Figure 1). Sixty individuals were excluded because they did not meet the inclusion criteria for different reasons, such as being less than 18 years old, a concussive episode occurring either too close to the enrollment date (less than 3 months) or remote from the enrollment date (5 years or more), and moderate or severe TBI. Two participants withdrew due to severe claustrophobia, one participant withdrew due to ongoing TBI-related symptoms, and one participant was involuntarily withdrawn due to safety concerns in the hyperbaric chamber. Baseline demographics and clinical characteristics were balanced between groups (

Table 1. (a) Baseline Participant Demographics. Data presented as Mean ± SD where applicable; (b) Baseline Neurobehavioral Symptom Inventory (NSI) Scores by Group. Data presented as Mean ± SD. Higher scores indicate greater symptom burden (NSI range 0 to 88).

(a)
Characteristic		Intervention	Placebo	p-Value
Age, years, Mean ± SD		42.1 ± 10.6	48.9 ± 14.3	0.301 †
Sex, n (%)				1.000 ‡
Male		2 (25.0)	1 (12.5)
Female		6 (75.0)	7 (87.5)
Time since injury, months, Mean ± SD		15.1 ± 10.0	20.6 ± 14.0	0.382 †
Mechanism of Injury, n (%)				0.467 ‡
Non-sport related		8 (100.0)	6 (75.0)
Sport related		0 (0.0)	2 (25.0)
(b)
Scale/Subscale	All Participants Mean ± SD	Intervention Mean ± SD	Placebo Mean ± SD	p-Value (95% CI)
NSI Total	40.75 ± 16.00	45.75 ± 12.61	35.75 ± 18.23	0.22 (−6.9 to 26.9)
Somatic/Sensory	14.25 ± 8.42	15.25 ± 7.29	13.25 ± 9.82	0.65 (−7.2 to 11.2)
Affective	9.94 ± 3.61	11.25 ± 2.61	8.63 ± 4.14	0.16 (−1.2 to 6.4)
Cognitive	16.56 ± 6.67	19.25 ± 6.90	13.88 ± 5.59	0.11 (−1.5 to 12.2)

^† Unpaired t-test, ^‡ Fisher’s exact test.

). Most participants (87.5%, n = 14) sustained their concussive episode(s) from non-sport-related mechanisms, with motor vehicle collisions representing the predominant cause. Two participants (12.5%) reported sport-related concussive episodes, both in the placebo group (Table 1). No statistically significant difference in mechanism of injury was observed between groups (p = 0.467, Fisher’s exact test). At the 13-week follow-up, one participant from each group was lost to follow-up, resulting in an analysis sample of n = 7 per group. Baseline differences between groups were not statistically significant; however, confidence intervals were wide and consistently favored higher symptom burden in the intervention group, suggesting potential baseline imbalance (Table 1b).

Analysis of serum biomarkers from baseline to 13 weeks post-intervention revealed no statistically significant differences between the placebo and intervention groups for GFAP, NfL, total tau, or UCH-L1 (Figure 2). The two-way ANOVA showed no significant main effect of treatment and no significant time-by-treatment interaction for any of the four biomarkers at this early time point (

Table 2. Analysis of Mixed Effects (Baseline to 13 Weeks). p values for the two-way ANOVA mixed effects analysis.

Effect	GFAP	NfL	Tau	UCH-L1
Time	0.1682	0.1996	0.4090	0.6757
Treatment	0.2874	0.5837	0.9192	0.1351
Time × Treatment	0.2874	0.5837	0.9192	0.1351

GFAP = glial fibrillary acidic protein; NfL = neurofilament light chain; Tau = total tau; UCH-L1 = ubiquitin C-terminal hydrolase L1. Each cell reports the p-value for the corresponding sources of variation from a two-way repeated-measure ANOVA mixed-effects model. ‘Time’ = main effect of assessment point (Baseline vs. 13 weeks), pooled across both groups. ‘Treatment’ = main effect of group assignment (Intervention vs. Placebo), pooled across both time points. ‘Time × Treatment’ = interaction term testing whether the change in biomarker level from baseline to 13 weeks differed between groups; this is the primary test of intervention effect. Significance threshold: p ≤ 0.05. Post hoc comparison used Šídák’s correction.

). The coefficient of variation for baseline vs. intervention is in

Table 3. Coefficient of variation (Baseline to 13 Weeks).

Time–Treatment	GFAP	NfL	Tau	UCH-L1
Baseline-Sham	83.3%	37.0%	46.6%	27.5%
13 week Sham	77.2%	78.0%	62.0%	79.5%
Base-Intervention	74.4%	66.2%	27.5%	97.7%
13 week-Intervention	51.7%	75.1%	34.2%	75.2%

. Since this pilot study had a small sample size and we wanted to rule out a Type II error or a false negative due to low statistical power, we performed a Wilcoxon Signed Rank Test (

Table 4. Secondary within treatment group analysis using the Wilcoxon Signed Rank Test. p values between baseline and 13 weeks.

Treatment	GFAP	NfL	Tau	UCH-L1
Sham	0.9658	0.5195	0.9658	0.7002
Intervention	0.1937	0.5416	0.6221	0.1909

). The Wilcoxon test did not lead to a significant effect between treatment groups for any of the biomarkers.

4. Discussion

In this randomized placebo-controlled trial, we found that a 40-session protocol of mild hyperbaric oxygen at 1.5 ATA did not produce a significant early modulation of serum biomarkers of neuronal injury (NfL), astrogliosis (GFAP), acute neuronal injury (UCH-L1), or axonal stability (total tau) in patients with PPCS when assessed at 13 weeks post-treatment [3,4,5]. This focused analysis, conducted at the time of maximal cohort retention, provides a reliable null result concerning the short-term biological impact of this specific intervention.

The evolution of the therapeutic definition for HBOT is a crucial context for our findings. Our 1.5 ATA protocol was designed in alignment with the UHMS guidelines at the study’s inception [8]. Its subsequent reclassification as a mild hyperbaric exposure underscores the dynamic nature of this therapeutic field [10]. These results do not speak to the efficacy of current therapeutic HBOT (≥2.0 ATA) but rather indicate that this lower-pressure regimen did not induce a detectable biomarker response in the short term. This aligns with the updated UHMS perspective that higher pressures may be necessary for a therapeutic biological effect.

The absence of biomarker change at 13 weeks suggests several possibilities. First, the pathological processes in chronic PPCS, as reflected by these specific serum markers, may be relatively inert or slowly progressive and unresponsive to this level of hyperoxic stimulus. Second, the chosen biomarkers, particularly GFAP and UCH-L1, which are associated with acute injury phases, may lack sensitivity to detect subtle treatment effects in a chronic, stable population [4,5]. NfL, a marker of ongoing axonal degeneration, also remained stable, suggesting that the intervention did not alter this underlying process within the observed timeframe [3].

A significant strength of this analysis is its focus on a time point with preserved statistical integrity. By halting the primary comparison at 13 weeks, we avoid the profound limitations associated with extreme attrition, which can lead to unstable and potentially spurious estimates [11]. This approach prioritizes scientific rigor over longitudinal completeness, providing a clear and interpretable answer to a focused question.

Limitations

This study has important limitations. The sample size, though adequate for this early-phase comparison, remains modest, limiting generalizability and power to detect small effects [11]. Furthermore, we cannot rule out the possibility of later biomarker changes; however, any such analysis in our cohort would be critically underpowered.

An additional limitation is the absence of a post-treatment symptom reassessment. Although the NSI characterized baseline symptom burden at study entry, it was not readministered following the intervention. This precludes any determination of subjective symptoms change and means the present analysis cannot delineate biological from subjective neurobehavioral symptomatic changes, including potential placebo responses. However, neuropsychological outcomes were assessed pre- and post-treatment using the RBANS and will be reported in a companion manuscript currently in preparation.

5. Conclusions

This early-phase analysis finds no evidence that a mild hyperbaric oxygen protocol (1.5 ATA) alters serum levels of key neuropathological biomarkers at 13 weeks in individuals with PPCS. These findings highlight the challenge of modifying established biomarker profiles in chronic brain injury and emphasize the importance of using standardized, currently defined therapeutic protocols in future HBOT research [8]. Future trials should employ higher-pressure HBOT regimens (≥2.0 ATA) against truly inert control conditions and continue to explore the utility of biomarker panels in guiding and assessing therapy for PPCS [2,12].