Impaired Cerebral Vasoreactivity Despite Symptom Resolution in Sports-Related Concussion

Abstract

Traumatic brain injury (TBI) is associated with increased risk of later-life neurodegeneration and dementia. However, the underpinning mechanisms are poorly understood, and secondary injury resulting from perturbed physiological processes plays a significant role. Cerebral vasoreactivity (CVR), a measure of hemodynamic reserve, is known to be impaired in TBI. However, the temporal course of this physiological perturbation is not established. We examined CVR and clinical symptoms on day 3 (T1), day 21 (T2), and day 90 (T3) after concussion in collegiate athletes and cross-sectionally in non-injured controls. Changes in middle cerebral artery blood flow velocity (MCAV; transcranial Doppler ultrasonography) were measured during changes in end-tidal CO₂ (PetCO₂) at normocapnia, hypercapnia (inspiring 8% CO₂), and hypocapnia (hyperventilation). CVR was determined as the slope of the linear relationship and expressed as percent change in MCAV per mmHg change in PetCO₂. CVR was attenuated during the acute phase T1 (1.8 ± 0.4U; p = 0.0001), subacute phases T2 (2.0 ± 0.4U; p = 0.0017), and T3 (1.9 ± 0.6U; p = 0.023) post-concussion compared to the controls (2.3 ± 0.3U). Concussed athletes exhibited higher symptom number (2.5 ± 3.0 vs. 12.1 ± 7.0; p < 0.0001) and severity (4.2 ± 6.0 vs. 29.5 ± 23.0; p < 0.0001), higher Patient Health Questionnaire-9 score (2.2 ± 2.0 vs. 9.1 ± 6.0; p = 0.0003) at T1. However, by T2, symptoms had resolved. We show that CVR is impaired as early as 4 days and remains impaired up to 3 months post-injury despite symptom resolution. Persistent perturbations in CVR may therefore be involved in secondary injury. Future studies with a larger sample size and longer follow-up period are needed to validate this finding and delineate the duration of this vulnerable period.

Introduction

Concussion, a clinical syndrome resulting from mild traumatic brain injury (mTBI), is a global public health concern, with an annual incidence of approximately 1.6–3.8 million in the United States alone.¹ Recurrent mTBI may also be associated with a higher risk of neurodegeneration and dementia.^2–4 Although the mechanisms underpinning this observed association are poorly understood, a vulnerable window with secondary injury is likely at play in some settings. Secondary injury evolves over minutes to months after the primary traumatic injury and results in cascades of metabolic, cellular, and molecular events, including glutamate excitotoxicity, disruption of cellular calcium homeostasis, increases in free radical generation, inflammation, ischemia, and microvascular injury.^5,6 In sports-related concussion, return-to-play protocols rely on clinical symptom resolution to approximate the vulnerability window and minimize secondary injuries. However, clinical symptoms resolution after mTBI may not always coincide with physiological recovery, therefore generating a hidden window of cerebral vulnerability with heightened risk for secondary injuries.^3,7 As such, there is an urgent need to identify physiological, cellular, and molecular perturbations which may contribute to cumulative effects post-injury. These perturbations can be monitored as part of the return-to-play protocols to minimize secondary injury in high-risk athletes, but also they may be targeted for therapeutic interventions to prevent injury.

Cross-sectional and longitudinal studies investigating physiological perturbations have reported reductions in global cerebral blood flow (CBF) during acute, subacute, and chronic phases after mTBI.^7–10 CBF is regulated by the cerebrovascular bed, which maintains the cerebral microenvironment by changing cerebral vascular resistance in response to a number of stimuli, including cerebral perfusion pressure, arterial CO₂, O₂, pH, blood viscosity, as well as cerebral metabolic demand.¹¹ Cerebral vasoreactivity (CVR), a key measure of cerebral hemodynamic reserve, may be particularly altered in mTBI.¹² CVR, which is measured as the change in CBF velocity in response to changes in end-tidal CO₂ (PetCO₂), is thought to be mediated by nitric oxide and may represent endothelial function in the cerebral vasculature.^13,14 CVR helps regulate and maintain central pH.¹⁵ Increased partial pressure of arterial CO₂ (PaCO₂) is associated with arteriolar vasodilatation and subsequent increases in CBF in the upstream larger cerebral arteries to “wash out” CO₂ from the brain tissue. On the other hand, decreased PaCO₂ is associated with arteriolar vasoconstriction and subsequent decreases in CBF to attenuate the fall in CO₂ within the brain tissue.^12,15,16 A few earlier studies have examined CVR in the mTBI population. Using the breath-hold technique, CVR impairment was observed in athletes 2 days after concussion with normalization of CVR 5 days post-injury.¹⁷ In another study, CVR in response to inspired CO₂ was chronically impaired in active professional boxers compared to their age-matched controls.¹⁰ Specifically, CVR impairment was observed in boxers with greatest exposure to repetitive subconcussive TBIs. CVR impairment has also been linked to exacerbation of gray matter atrophy, reduced cerebral tissue oxygenation, and marked orthostatic hypotension.^10,18 In a cross-sectional study of individuals with a history of concussion, CVR was strongly associated with headache score and cognitive deficit.¹⁹ In a small longitudinal study utilizing magnetic resonance imaging, sports-related concussion was associated with patient-specific alterations in CVR, which persisted beyond clinical recovery.²⁰

Although available studies provide preliminary evidence for CVR perturbations in mTBI and sports-related concussion, the findings are undermined by inclusion of participants with a wide age range, variable time from injury, heterogeneous clinical syndromes including acute concussion as well as post-concussive syndromes, inadequate sample size, and absence of control groups for comparison.^17–20 Moreover, methods for CVR assessment have been heterogeneous and include vasoactive stimuli, such as breath-hold and rebreathing techniques, that may elicit significant cardiovascular responses in addition to CVR.^15,21 Inspired CO₂ under normoxia has been recommended as the most suitable vasoactive stimulus to assess CVR.²¹

In order to better understand the time course of CVR perturbation post-concussion and to examine the relationship between CVR and clinical symptoms, we used a longitudinal case-control study design in a homogenous population of collegiate athletes. We assessed clinical symptoms and measured CVR during acute (approximately day 3 [T1]) and subacute phases (day 21 [T2] and day 90 [T3]) after a concussion and compared them with non-injured age- and sports-matched control athletes. We hypothesized that CVR will be impaired during the acute phase (T1), but will normalize during the subacute phase (T2, T3), after injury. We also hypothesized that symptom resolution will coincide with normalization of CVR in the subacute phase. The three time points are in accord with the National Institute of Neurological Disorders and Stroke guidelines for sports-related concussion and coincided with their definition of acute injury phase (day 3), subacute phase with typical symptom resolution (day 21), and persistent chronic phase with post-concussion symptoms (day 90), if any.

Methods

Participants

A total of 56 collegiate athletes, both male and female, participating in National Collegiate Athletic Association (NCAA) Division I and recreational contact-collision sports (football, soccer, basketball, rugby, and cheerleading), were enrolled in the study. Concussion was diagnosed in accord with the NCAA criteria.²² Twenty-eight athletes were assessed on day 3 (T1) post-concussion. Twenty-four from that group returned for a second visit on day 21 (T2) post-concussion, with 18 athletes completing the final follow-up visit for day 90 (T3) post-concussion. Twenty-eight age- and sports-matched non-injured controls were included. Controls were required to be at least 1 year out from their most recent concussion. All the athletes were actively engaged in their sport at the time of participation and were healthy and free of cardiovascular and respiratory diseases or any self-reported learning disabilities.

Study protocol

All screening and experimental procedures were approved by the Institutional Review Board at the Southern Methodist University (SMU), Dallas, Texas (Protocol #2015-001-PURS) and were in accord with the guidelines of the Declaration of Helsinki. All subjects were informed of the study objectives and protocol verbally before obtaining their written consent for participation. The study is registered in clinicaltrial.gov as NCT02754206.

All experiments were performed in the Cerebrovascular Research Laboratory in the Department of Applied Physiology & Wellness at SMU. Subjects were asked to abstain from alcohol or caffeinated beverages 24 h before the laboratory visit. Subjects completed a health history questionnaire. Symptoms were evaluated with the Sport Concussion Assessment Tool 3rd Edition (SCAT3). Cognition was assessed with the Standardized Assessment of Concussion (SAC), a component of the SCAT3. Mood was assessed with the Patient Health Questionnaire (PHQ-9), and sleep quality was monitored utilizing the Insomnia Severity Index (ISI).

Heart rate was acquired with a standard three-lead electrocardiogram (ECG; Solar 8000i patient monitor; GE Healthcare, Chicago, IL). Beat-to-beat mean arterial blood pressure (MAP) was recorded continuously using finger photoplethysmography (Finometer Pro; Finapres Medical Systems, Amsterdam, Netherlands). Arterial blood pressure was also measured using an automated ambulatory arm cuff (Tango; SunTech Medical, Morrisville, NC). Beat-to-beat mean middle cerebral artery blood flow velocity (MCAV) was obtained with transcranial Doppler ultrasonography (Doppler BoxX; DWL USA, San Juan Capistrano, CA). A 2-MHz probe was placed bilaterally over the temporal bone to insonate the middle cerebral artery¹⁶ An adjustable silicon headband was used to stabilize the Doppler probe for the duration of the study. End-tidal CO₂ (PetCO₂) was continuously obtained from an infrared CO₂ analyzer (Capnocheck Plus; Smith Medical PM Inc, Waukesha, WI) connected to a nasal cannula. Subjects were instructed to breathe in and out only through the nose for the entire duration of the study.

Data analysis

MCAV was measured continuously while subjects were seated upright breathing room air (normocapnia) for 2 min. Through a non-rebreather mask, subjects then inspired a gas mixture of 8% CO₂, 21% O₂, with balanced nitrogen (hypercapnia) for 2 min. Subjects then mildly hyperventilated in order to obtain an end-tidal CO₂ of approximately 22 mmHg (hypocapnia) for 2 min. Percent change in MCAV from normocapnia was plotted against PetCO₂, and CVR was estimated from the slope of this relationship and expressed as percent change in MCAV per mmHg change in PetCO₂. Analog signals for ECG, MAP, MCAV, and PetCO₂ were sampled at 500 Hz and subsequently stored on a personal computer using the Windaq data acquisition system (Windaq DI720; DATAQ Instruments, Akron, OH) for offline analyses.

Statistical analysis

Medians and interquartile ranges were obtained for all continuous measures of clinical and behavioral symptoms and hemodynamic variables. Controls were compared to concussed athletes at three different time points (T1, T2, and, T3) using the non-parametric Mann-Whitney U test. A non-parametric Friedman's test was performed to examine longitudinal changes over time among the concussed athletes. A post-hoc Wilcoxon signed-ranks test was used to compare pairs of time points (T1, T2, and, T3). Significance for the post-hoc Wilcoxon tests was evaluated at p < 0.0167 to correct for multiple testing.

In a separate post-hoc analysis, subjects were divided into two groups based on median CVR at T1 (low CVR, <1.7024U; high CVR, >1.7024U). A non-parametric Wilcoxon rank-sum test was performed to examine whether low CVR at T1 was correlated with higher headache score (from SCAT3) at T3. IBM SPSS software (Version 25; IBM Corp., Armonk, NY) was used for all analyses, and, except where noted, statistical significance was evaluated using p < 0.05.

Results

Symptoms and hemodynamics

For concussed subjects, assessments were completed on 4 ± 1 days for T1, 22 ± 4 days for T2, and 95 ± 11 days for the T3 visit. Table 1 summarizes the baseline subject characteristics. Clinical symptoms and hemodynamic parameters are summarized in Table 2. As anticipated, compared to the controls, the concussed athletes on T1 reported higher symptom number (p < 0.0001) and symptom severity (p < 0.0001). Across the three time points in the concussed athletes, symptom number was reduced over time (p < 0.0001), with T2 (p < 0.0001) and T3 (p = 0.0003) significantly lower than T1. A similar trend in symptom severity was observed over time (p < 0.0001) with T2 (p < 0.0001) and T3 (p = 0.0003) significantly lower than T1. Cognition score (SAC) was lower (p = 0.0025) in the concussed athletes at T1, but normalized at T2 and T3, when compared to the controls. PHQ-9 was higher at T1 (p = 0.0003) compared to the controls with improvements over time in the concussed athletes (p < 0.0003) with lower scores at T2 (p = 0.0013) and T3 (p = 0.0026), which were similar to controls. ISI score was worse on T1 compared to T3 (p = 0.0058) in the concussed group. There were no other differences observed in the hemodynamic variables between the controls and concussed athletes at the three time points.

Table 1.

Baseline Subject Characteristics

	Control (N = 28)	Concussion (N = 28)
Demographics
Age (years)	20 ± 1	20 ± 1
Male (N)	16	18
Caucasian (N)	17	13
Characteristics
Height (cm)	179 ± 11	180 ± 12
Weight (kg)	81 ± 22	84 ± 23
BMI (kg/m2)	25 ± 4	26 ± 5
Education (years)	15 ± 1	15 ± 1
Previous concussions (N)	1 ± 1	1 ± 1
Sports
Football	10	10
Soccer	10	10
Other	8	8

BMI, body mass index.

Table 2.

Clinical Symptoms and Hemodynamic Parameters

		Concussion
	Control	T1 (Day 3)	T2 (Day 21)	T3 (Day 90)
Symptoms (SCAT-3rd Edition)
Total symptom number	1.00 [0.00–3.50]	11.00 [8.00–18.0]a	1.00 [0.00–5.50]b	0.00 [0.00–4.00]b
Symptom severity	2.00 [0.00–4.50]	23.0 [11.0–50.0]a	1.00 [0.00–6.50]b	0.00 [0.00–6.50]b
SAC (SCAT-3rd edition)	29.0 [27.0–29.0]	27.0 [26.0–28.0]a	27.5 [25.0–29.0]	27.0 [27.0–28.0]
Behavioral
Insomnia serverity index	3.00 [1.00–5.00]	6.00 [2.00–12.0]	3.00 [1.00–9.50]	1.00 [1.00–5.50]b
Patient health questionnaire 9	1.00 [1.00–3.00]	8.00 [4.00–14.0]a	2.00 [0.25–3.75]b	1.00 [0.00–2.00]b
Hemodynamic
Heart rate (bpm)	68 [61–72]	66 [62–74]	68 [60–77]	70 [65–74]
Mean arterial pressure (mmHg)	84 [78–86]	82.6 [77.9–89.2]	85.9 [78–92]	84.7 [80–88.9]
Middle cerebral artery velocity (cm/s)	59.3 [55.5–64.4]	64 [49.4–71.4]	66.5 [57.6–75.9]	61 [54.2–79.2]
Respiratory rate (BPM)	16 [14–19]	16 [15–17]	16 [14–18]	14.8 [13.2–16.3]
End tidal CO2 (mmHg)	40.8 [39–41.8]	41.7 [39.1–43.4]	40.2 [38.6–42.7]	39.8 [38–40.8]

Values are medians [1st–3rd quartile]. ^aSignificantly differently from the control. ^bWithin the concussed groups, significantly differently from T1.

bpm, beats per minute; BPM, breaths per minute.

Cerebral vasoreactivity

Average PetCO₂ during the CVR maneuver was 40.2 ± 2.6 mmHg (normocapnia), 22.8 ± 2.9 mmHg (hypocapnia), and 57.6 ± 3.3 mmHg (hypercapnia). Compared to the controls, CVR was attenuated on T1 (p = 0.0001) post-concussion (Fig. 1). CVR continued to be blunted during subacute phases T2 (p = 0.0017) and T3 (p = 0.02) post-injury. No significant difference in CVR was observed in the concussed groups across the three time points.

FIG. 1.

Cerebral vasoreactivity in controls and concussed athletes. Individual cerebral vasoreactivity (CVR) data from healthy non-injured controls and from the concussed athletes at three time points (T1, day 3; T2, day 21; and T3, day 90) after injury. Horizontal bar indicates median CVR score for each group and at each time point.

In addition, when concussed athletes were stratified into low- and high-CVR groups based the on the median CVR on T1, a trend in headache score on T3 was observed between the low- and the high-CVR groups (p = 0.06; Fig. 2). Athletes in the low-CVR group reported headache approximately 90 days post-injury, whereas the high-CVR group experienced no headache symptom during the chronic recovery phase.

FIG. 2.

Headache score in high versus low CVR group. A trend in headache score (p = 0.06) between the high- and low-CVR group was observed. Subjects in the low-CVR group reported headache compared to no headache reported in the high-CVR group on T3 (day 90). CVR, cerebral vasoreactivity.

Discussion

In this study, we report on the trajectory of CVR and clinical symptoms in collegiate athletes after sports-related concussion. First, CVR was impaired in the acute phase (T1) after a concussion compared to the controls. Second, CVR impairment persisted in the subacute phases T2 and T3 after concussion despite resolution of clinical symptoms. Third, lower CVR during the acute phase (T1) related to worse headache score during subacute phase (T3), although it did not reach significance. In summary, our data show that CVR is impaired as early as 4 days after concussion and remains impaired up to 3 months post-injury, suggesting that physiological recovery may lag clinical recovery as indicated by symptom resolution.^7,23 Moreover, lower CVR may be associated with worse headache scores 3 months after a sports-related injury.

Although mechanisms underpinning CVR are not completely understood, several lines of inquiry support an endothelial-mediated process. First, CVR, but not pressure-dependent autoregulation, is impaired in patients with endothelial dysfunction.¹³ Second, sodium nitroprusside infusion, which is an exogenous nitric oxide donor, is associated with enhanced CVR, but not pressure-dependent autoregulation.²⁴ In animal studies, l-arginine methyl ester, a nitric oxide synthase inhibitor, blunted hypercapnia-related vasodilatation response, suggesting the role of nitric oxide in CVR.¹⁴ Similar studies in humans are limited because of the relatively low concentrations of nitric oxide synthase inhibitors permitted for human experimental use, which thereby underestimates the role of nitric oxide in humans.¹⁴ Alterations in CVR are also speculated to result from direct impact on the blood vessels, disruption of the endothelium, or as a compensatory mechanism in response to the neurometabolic crisis post-injury.²⁵ Disruption of the endothelial cell layer, inflammation, diminished CBF, and focal tissue hypoxia are witnessed in animal models of mTBI.^25,26 Endothelial dysfunction may disrupt cerebral hemodynamics by reducing the bioavailability of nitric oxide and increasing the production of reactive oxygen species.^13–15 In healthy individuals, an increase in sympathetic neural activity is associated with diminished CVR sensitivity.²⁷ Autonomic dysfunction, which is implicated in mTBI, could also result in alterations in CVR, although autonomic function was not examined in this study.²⁸

Intact CVR and preserved hemodynamic reserve is essential for maintaining brain perfusion during exercise. Mild-to-moderate intensity exercise is associated with increases in PaCO₂ and subsequent rise in CBF.²⁹ On the other hand, heavy-intensity exercise is accompanied by exponential increase in ventilation and subsequent changes in CBF.³⁰ Physical exertion after return to play is associated with dynamic alteration in PaCO₂ and subsequent changes in CBF. Persistent CVR impairment during this phase could contribute to prolonged secondary injury in physically active athletes. Despite being asymptomatic at rest, recurrence of symptoms and exercise intolerance upon return to play is associated with incomplete physiological recovery in athletes with resultant suggestions not to return to play.³¹ CVR assessment in addition to clinical symptom evaluation as part of the return-to-play protocols could identify athletes in the vulnerable window and at high risk for secondary injury that are not ready to return to play.

Headache is one of the most common symptoms in post-concussion syndrome.^32,33 Although the mechanisms underpinning headache are not well understood, mismatch between increased metabolic demand and decreased CBF may be one mechanism.³³ We observed an inverse trend between CVR and headache score, but others have shown CVR to be associated with increased headache score and PCS severity.¹⁹ Differences in the time from concussion among participants limits comparison between these studies. In fact, patients with chronic migraines have been shown to have higher CVR during the headache-free periods, possibly manifesting a chronic disturbance in the autonomic control of CVR.³⁴ There is a clear need for longitudinal studies of CVR trajectory after concussion in patients with and without PCS extending over a longer period to better assess the temporal course of clinical symptoms and physiological perturbation in those with a protracted recovery.

It is important to note that several vasoactive stimuli have been utilized to elicit changes in PetCO₂ in order to assess CVR in other studies of mTBI and concussion. CVR evoked with repeated breath-holding was blunted in ice-hockey athletes 2 days after a concussion, with normalization of CVR 5 days post-injury.¹⁷ However, it is important to note that PetCO₂ between rest and the breath-hold protocol was not different in the study. Therefore, it could be argued that the applied CO₂ stimulus was not physiologically challenging, thereby resulting in CVR normalization within 5 days post-injury. In another study utilizing a rebreathing technique, a positive linear relationship between CVR and clinical symptoms was observed during the subacute phase after concussion.¹⁹ Again, it is important to note that techniques such as breath-hold and rebreathing elicit significant cardiovascular responses and subsequent sympathetic neural activation through reductions in venous return, MAP, and increases in tissue hypoxia in addition to CVR. Therefore, inspired CO₂ under normoxic conditions, the method utilized in our study, has been recommended as a more appropriate and effective stimulus to assess CVR.^15,21

Study limitations

Our study was limited by a relatively small sample size and only a 90-day follow-up. However, using a case-control design and limiting our sample to a well-characterized homogenous sample across age, type of injury, and time from injury, we are able to add to the existing knowledge in this field and demonstrate persistent physiological perturbation, and hence risk for prolonged secondary injury, despite symptoms resolution in collegiate athletes after a sports-related concussion. We did not utilize extensive neuropsychological and behavioral batteries to assess symptoms. Future studies with more comprehensive testing batteries are warranted to detect more subtle symptom persistence.

In summary, our results indicate that CVR is impaired as early as 4 days after a sports-related concussion, and the impairment persisted 3 months post-injury despite resolution of symptom. Therefore, CVR may be a useful biomarker for tracking physiological recovery, and it may aid in improving return-to-play decision-making protocols to reduce the risk of recurrent injuries. Future studies with a large sample size and longer follow-up period are needed to validate this preliminary finding.

Author Disclosure Statement

No competing financial interests exist.