Introduction
During the past 5–10 years, there has been a growing interest in sport-related concussion due to recent deaths in high-performance athletes (Kelly, 2001; O’Connor et al., 2017). The Centers for Disease Control and Prevention, National Center for Injury Prevention and Control have reported that there is an estimated between 1.7 and 3.8 million concussions per year in the USA alone (Rutland-Brown et al., 2006; Daneshvar et al., 2011), with approximately 30% being sport-related (Cassidy et al., 2004; Langlois et al., 2006; Gaw and Zonfrillo, 2016). Globally, it is estimated that 64–74 million individuals experience traumatic brain injury annually (Dewan et al., 2018).
Recently there has been a focus on the management of the athlete post-concussion, including when it is safe for the athlete to return-to-learn and return-to-play so that the likelihood of re-injury is decreased (Aubry et al., 2002; Collins et al., 2002; McCrory et al., 2005, 2009, 2013, 2017; Ellis et al., 2016; Terwilliger et al., 2016). Traditionally, “prescribed rest” has been used to ensure the safety of the athlete (Moser et al., 2012; Elbin et al., 2014; Thomas et al., 2015; Buckley et al., 2016). Although recent consensus statements recommend that concussed athletes must participate in a graduated, medically supervised exercise protocol before returning-to-play (Aubry et al., 2002; McCrory et al., 2005, 2009, 2013, 2017), less information is available on whether aerobic exercise can be performed immediately post-injury, i.e., with the first 72 h. However, mounting research is recommending that mild-to-moderate aerobic exercise can be performed safely within the days and weeks following injury (Leddy et al., 2016a, 2017, 2018; Bishop et al., 2017b; Dech et al., 2019), although most studies initiate exercise as a rehabilitation strategy after ~21 days post-injury (Dematteo et al., 2015; Gagnon et al., 2016; Morissette et al., 2019).
Most recently, exertional exercise testing protocols, such as the Buffalo Concussion Treadmill and Bike Test have proven to be safe and efficacious (Leddy et al., 2016a, 2018; Haider et al., 2019; Morissette et al., 2019) in evaluating the athletes post-concussion. These laboratory testing protocols have been designed with the intent to differentiate the type of concussion injury (i.e., physiologic, vestibulo-occular, cervicogenic; Ellis et al., 2014, 2015; Leddy et al., 2016a), and to discern the physiological mechanisms (cerebral autoregulation, cerebrovascular reactivity, neurovascular coupling; Ellis et al., 2015, 2016, 2018; Neary et al., 2019). This research has added to our understanding of the pathophysiological changes that occur following a concussion (Len et al., 2011, 2013; Giza and Hovda, 2014; Bishop and Neary, 2017; Bishop et al., 2017b; Worts et al., 2019), and has also provided a greater understanding of how we can apply exercise safely as a treatment strategy. Thus, it appears that evidence supports the therapeutic benefits of mild-to-moderate intensity physical exercise (Kozlowski, 2014; Hinds et al., 2016; Leddy et al., 2016a,b, 2018) and that prolonged rest may be detrimental during acute recovery post-concussion (Elbin et al., 2014; Thomas et al., 2015; Buckley et al., 2016; Wells et al., 2016).
This evolving research that has examined the pathophysiology of concussion, includes monitoring cerebral blood flow velocity using transcranial Doppler (Len et al., 2011; Wright et al., 2018), cerebral blood-oxygen-level-dependent (BOLD) changes using functional magnetic resonance imaging (fMRI; Mutch et al., 2014, 2018), blood pressure (La Fountaine et al., 2016, 2019; Bishop et al., 2017b; Wright et al., 2018), heart rate variability (HRV; Gall et al., 2004; La Fountaine et al., 2009; Bishop et al., 2017a, b), cortical activation (Pearce et al., 2019), and cerebral hemodynamics using near-infrared spectroscopy (Urban et al., 2015; Bishop and Neary, 2017). In particular, near-infrared spectroscopy (NIRS) is an optical technique that has been validated to show quantitative changes in blood volume (an indirect measure of blood flow velocity) and oxygenation in cerebral tissue (Ferrari et al., 2004; Bhambhani et al., 2006). This measurement is taken at the level of capillaries, arterioles and venules in cerebral tissue (Belardinelli et al., 1995). By measuring the concentration of both the oxygenated (HbO2) and deoxygenated haemoglobin (HHb), changes in total blood volume (i.e., total haemoglobin; tHb = HHb + HbO2) can be determined and has been suggested as a surrogate for cerebral blood flow (Wolf et al., 2007).
NIRS has been used to monitor cerebral hemodynamics in patients following concussion (Kontos et al., 2014; Urban et al., 2015; Bishop and Neary, 2017), and TBI (Bhambhani et al., 2006). The neurocognitive research by Kontos et al. (2014) and Urban et al. (2015) suggests clinical implications of using pre-frontal cortex NIRS to assess brain cognition and connectivity. Furthermore, the pre-frontal cortex has been suggested as a major area to measure these hemodynamic changes during a hypercapnic challenge (Bishop and Neary, 2017), but no studies to our knowledge are available that used NIRS during incremental exercise when assessing acute concussion within 24–48 h following injury. Collectively, these studies suggest that the vascular involvement of the pre-frontal cortex is important during an acute concussion.
Because there is mounting research focused on both physiologic changes and the importance of exercise post-concussion (Worts et al., 2019), the focus of our research was to monitor the pathophysiological changes in athletes prospectively in the days following sport-related concussion in response to acute incremental aerobic exercise. It was hypothesised that exercise could be used safely to assess the physiological status of the concussed athlete and that there would be differences in the cerebral hemodynamic response between healthy control vs. mTBI subjects. This study used an evidence-based approach to discern differences between groups to contribute to our understanding and insight into the physiological response to exercise post-mTBI.
Discussion
We demonstrated two important findings in this study: (1) that acute, short duration mild-to-intense exercise testing (HR = 110 bpm–170 bpm) can be performed safely in the days (D2, D4, D7) immediately following the diagnosis of concussion; and (2) that physiological measures can be easily collected and can be used to add to our understanding of the pathophysiological effects of concussion. This has important implications for treatment strategies, as current research suggests that sport-related concussion or mTBI is a physiological brain injury (McCrory et al., 2017; Leddy et al., 2018), and that concussion testing can potentially differentiate the type of mTBI (Ellis et al., 2015). Furthermore, our data support recent reports that physical exercise can be used safely when assessing recently concussed athletes within the first week post-injury in a controlled laboratory environment (Leddy et al., 2016a, 2017, 2018; Haider et al., 2019).
Basis of Physiological Concussion Testing
The primary objective of concussion testing is to characterise the physiologic response of the injured athlete. The effects of moderate-to-high intensity exercise on physiological function in the concussed state are unknown (Worts et al., 2019), and therefore we explored whether such exercise intensity could be safely maintained for a short duration (increments of 5-min workloads). To our knowledge this is the first study to systematically evaluate the physiological effects of repeated exercise tests using mild-to-intense exercise intensity (HR~ 110–170 bpm) performed within 2-days (D2) after the medical diagnosis of sport-related concussion in high-performance athletes. Based on the growing body of literature in this area, moderate-to-intense exercise testing can be performed in concussed athletes safely, but most studies were performed more than a week after injury (Baker et al., 2012; Ellis et al., 2015; Hinds et al., 2016; Leddy et al., 2016b, 2018; Haider et al., 2019; Morissette et al., 2019). Using an evidence-based approach, similar to our study, Leddy et al. (2011, 2017) developed the Buffalo Concussion Treadmill and Bike Tests to specifically evaluate symptom limitations in post-concussion syndrome (PCS) patients and showed that concussed athletes assessed on average of 33 weeks (6–36 weeks) post-injury could reliably differentiate physiological dysfunction and quantify exercise capacity in PCS patients. This test has now been used to assess the acute effects of acute concussion (Hinds et al., 2016; Haider et al., 2019). Furthermore, Ellis et al. (2015) conducted a review of the literature to describe and differentiate the type of concussion and suggested that concussed individuals can exercise to maximal volitional fatigue during supervised graded exercise testing. They classified the type of concussion as either physiologic, vestibulo-ocular, or cervicogenic (Ellis et al., 2015). It is well accepted that exercise increases the metabolic demand and stresses the autonomic nervous system function. Thus, dynamic cerebral autoregulation (Bailey et al., 2013), neurovascular coupling (Wright et al., 2017; Sharma et al., in press), heart rate (Haider et al., 2019) and HRV (Gall et al., 2004; La Fountaine et al., 2009), functional connectivity and brain activation (Kontos et al., 2014; Urban et al., 2015; Pearce et al., 2019) and blood pressure (La Fountaine et al., 2016, 2019) can be assessed to characterise the autonomic/physiological effects of acute mTBI. Therefore, this increasing body of literature demonstrates that objective physiologically based concussion testing using exercise is an important part of athlete management post-injury.
Comparison of NIRS Hemodynamics Between Control vs. mTBI Participants
Our results showed that athletes who sustained a sport-related concussion or mTBI can undertake an incremental exercise test in 5-min stages from mild-to-intense exercise intensity on D2, D4 and D7 post-injury. Although we hypothesised that there would be significant differences between the CON and mTBI groups, the statistical analysis revealed no significant differences in the physiological responses in HbO2, HHb and tHb using NIRS, between groups or testing days (Figures 1–3). Previous research, including our own (Bishop and Neary, 2017), has shown that significant differences exist between days post-injury. For example, under resting conditions, HbO2 was statistically different (p ≤ 0.05) between days of recovery in mTBI participants; i.e., Day 1–3 vs. D7–14 and D4–6 vs. D7–14 (Bishop and Neary, 2017). As well, there were significant differences in the SD values between days post-injury, with the mTBI group showing reduced SD. However, it is important to note that this was in response to a hypercapnic (breath-hold) challenge which is a different paradigm than our exercise protocol. Hypercapnia has demonstrated that cerebral blood flow velocity is altered leading to cerebrovascular reactivity changes and differences between healthy controls and mTBI (Len et al., 2011, 2013; Mutch et al., 2014). Based on these observations, we extended our hypercapnic protocol (Neary et al., 2019) in the current study to include incremental (5 min) exercise workloads to assess the physiological consequences of mTBI in these athletes accustomed to intense physical exercise. No statistical differences in our hemodynamic data suggest that cerebral blood flow and oxygenation variables were not different during the exercise test in the days following acute concussion. When transitioning from rest to exercise, to sustain exercise intensity, energy (ATP) production is needed and was maintained adequately following concussion and did not compromise our athletes’ ability to complete the testing protocol.
However, a word of caution is warranted, as our data showed that differences as much as 40% (p > 0.05) exist between groups. Closer observation of our hemodynamic results showed that there were 30–40% differences in the mean value of tHb and HHb variables, and 20–30% differences in mean HbO2 between the CON vs. D2 and D4 at particular exercise workloads (EX1, EX2; Figures 1–3). However, the standard deviations in our data were extremely large (2–5 times the mean) and eta values were low (0.01–0.02 for HbO2, 0.06–0.12 for HHb, 0.02–0.04 for tHb) making it difficult to reach statistical significance. Based on these statistical parameters, we have estimated that approximately 45 participants per group (CON, D2, D4 and D7) are required to demonstrate statistical differences. If the participants had baseline data collected first, a within-subject design based on this effect size will require 29 participants per group to demonstrate significance. This demonstrates the importance of including baseline testing in sport-related concussion research which has been documented previously (Len et al., 2013; Neary et al., 2019).
The high SD’s are most likely related to two factors, the continuous wave NIRS device used in this study gives a relative change value and therefore differences between individual participants is common, and the small number of CON (n = 5) and mTBI (n = 14) participants in the study limited our statistical power. Individual differences could potentially nullify any differences that exist between groups as there is more variability to the data.
Limitations of the Study
As suggested above, the major limitation of our study was the small number of participants, especially in the CON group. Future studies of this nature must increase the number of control participants, both for within- and between-group comparisons. Because of the limited number of participants in this study, the standard deviation of the oxygenation variables were quite large. This is an inherent limitation of continuous-wave NIRS technology in that it gives a relative change value in oxygenation concentration (μM). Typically, the best method to deal with relative differences between participants is to conduct a resting baseline and then zero the relative changes from the resting baseline (Bishop and Neary, 2017). We did this in our study to account for differences between participants but still, the SD values were high. Based on the SD and eta values we estimated that 45 subjects would be needed to show statistical significance, or based on a within-subject design, 29 participants per group. Other limitations included not collecting rating of perceived exertion or symptom scores (Sport Concussion Assessment Tool; SCAT) to see how each athlete felt after their test, and whether this changed from D2 to D7. Anecdotally, the players reported feeling that exercise on D2 was difficult for them but felt better by D7 post-injury. However, all athletes were able to complete all the testing days. Recent research by Morissette et al. (2019) showed that symptomatic subjects when assessed at day 45 post-injury reported significantly higher ratings of perceived exertion vs. controls, although metabolic demands were not different at the same workload. The inclusion of a symptom checklist before and after the exercise testing may have provided additional information about their rating of perceived exertion, and considered for future research. However, the reader is reminded that SCAT scores are subjective and cannot always be relied upon. Finally, ECG recordings instead of using an HR monitor would have allowed us to calculate HRV changes from day-to-day. Previous research has demonstrated that HRV is altered in acute concussion (Gall et al., 2004; Papaioannou et al., 2008). Based on our results, and because of these limitations, it was not possible to determine the extent to which autonomic dysfunction occurred in our mTBI participants (Bishop et al., 2017b), or to characterise the type of concussion (Ellis et al., 2015).