Academia.eduAcademia.edu
LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 J Head Trauma Rehabil Copyright  c 2015 Wolters Kluwer Health, Inc. All rights reserved. Cerebral Blood Flow During Treadmill Exercise Is a Marker of Physiological Postconcussion Syndrome in Female Athletes Mary Clausen, MS; David R. Pendergast, EdD; Barry Willer, PhD; John Leddy, MD Objective: Some patients with postconcussion syndrome (PCS) have reduced exercise capacity that may reflect altered central cardiorespiratory control. The purpose of this study was to evaluate control of cerebral blood flow (CBF) during exercise in females with PCS. Setting: University Concussion Clinic. Participants: Nine female Division 1 collegiate team athletes with PCS (23 ± 6 years) and a reference group of 13 healthy female recreational aerobic athletes (21 ± 3 years). Design: A prospective experimental study. All PCS athletes were compared with the reference group at the beginning of the study. Six of the PCS athletes were subsequently measured before and after a subsymptom threshold aerobic exercise treatment program. Main Measures: Exercise treadmill test during which blood pressure (BP), minute ventilation (V˙ E), end-tidal CO2 (PETCO2 ), and CBF velocity (CBFV, by transcranial Doppler) were measured. Results: Participants with PCS had significantly lower V˙ E (by 18%) and greater PETCO2 (5%) and CBFV (14%) versus the reference group at similar workloads in association with appearance of symptoms and premature exercise cessation. Subthreshold exercise normalized V˙ E, PETCO2 , CBFV and exercise tolerance. Before treatment, PCS had low CO2 sensitivity that blunted their exercise ventilation. CO2 sensitivity and ventilation improved after exercise treatment. Conclusion: Some PCS patients have exercise intolerance due to abnormal CBF regulation that may be the result of concussion-induced altered sensitivity to CO2 . Return of normal CBF control and exercise tolerance may be physiological markers of recovery from concussion. Key words: carbon dioxide sensitivity, cerebral blood flow, exercise, postconcussion syndrome, ventilation A LTHOUGH THE MAJORITY OF PATIENTS with concussion recover within 7 to 10 days,1 a significant minority continues to experience symptoms is considered to be present in athletes with symptoms persisting several weeks to months.4 Some of these patients have symptom exacerbation during exercise5 called postconcussion syndrome (PCS).1,2 Postcon- that could be related to physiological changes resulting cussion syndrome is defined by the World Health from concussion, in this case, failure to appropriately Organization as persistence of 3 or more of the follow- regulate cerebral blood flow (CBF).6 Dysregulation of ing symptoms after head injury: headache, dizziness, resting CBF has been shown in patients soon after fatigue, irritability, insomnia, concentration difficulty, concussion by MRI measures of global CBF7 and by or memory difficulty.3 In sport-related concussion, PCS functional MRI in local brain regions acutely after concussion 8 and in patients with PCS.9,10 To our knowledge, dynamic CBF during exercise has not been Author Affiliations: Department of Physiology, Center for Research and evaluated in concussed patients, which has implications Education in Special Environments, State University of New York, for return to sport or military action after concussion. Buffalo (Ms Clausen and Dr Pendergast); SUNY Buffalo, Orthopaedics, University Sports Medicine, Buffalo, New York (Dr Leddy); and Cerebral autoregulation (CA) allows the brain to Department of Psychiatry, SUNY Buffalo (Dr Willer). maintain an almost constant perfusion pressure in the Supplemental digital content is available for this article. Direct URL citation face of varying levels of systemic blood pressure (BP), appears in the printed text and is provided in the HTML and PDF versions protecting the brain from changes in sympathetic nerve of this article on the journal’s Web site (www.headtraumarehab.com). activity, mean arterial blood pressure (MAP), and arte- The authors declare no conflicts of interest. rial CO2 levels (PaCO2 ).11,12 Concussion has effects on Corresponding Author: John Leddy, MD, SUNY Buffalo, Orthopaedics, systemic physiological systems.6 When compared with University Sports Medicine, 160 Farber Hall, Buffalo, NY 14214 controls, for example, patients with concussion and mild (leddy@buffalo.edu). TBI have altered heart rates (HRs)13 and HR variabil- DOI: 10.1097/HTR.0000000000000145 ity when moving from supine to standing14 and during 1 Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 2 JOURNAL OF HEAD TRAUMA REHABILITATION steady-state exercise,15 which reflects altered autonomic exercise in female patients with PCS. We focused on fe- nervous system (ANS) function.14,16 Cerebral autoregu- males because they may be at greater risk of concussion lation may also be disturbed after concussion such that and of developing PCS.24 Prior studies of the regulation changes in systemic BP may induce excessive changes of CBF during exercise have been conducted on men in cerebral perfusion pressure.17 We have shown that during cycle ergometry,11,12,20 but most exercise that patients with PCS exercising on a treadmill had abnor- is symptom-producing in PCS is running and the test mally increased BP at low exercise intensities associated proposed to evaluate the physiology of PCS patients with the appearance or exacerbation of symptoms5,18 is a walk/run on a treadmill.25 We hypothesized that and we hypothesized that this reflects a failure of female patients with PCS but not healthy patients would CA, with resulting hyperperfusion of the brain during have elevated PaCO2 levels at low exercise intensities exercise. and thus greater CBF leading to symptom exacerbation One of the primary influences on CBF is the arterial and early exercise cessation. We hypothesized that the CO2 tension (PaCO2 ). Depressed PaCO2 reduces CBF elevated PaCO2 would be the result of a pulmonary hy- and increased PaCO2 increases CBF.19 For a given car- poventilation (ie, reduced V˙ E ) during exercise because bon dioxide production (V˙ CO2 ), the PaCO2 is inversely of reduced CO2 sensitivity. Finally, we hypothesized proportional to the pulmonary ventilation (V˙ E ): as V˙ E that a program of subsymptom threshold aerobic ex- increases PaCO2 decreases and vice versa. Normally, as ercise treatment would normalize the physiology in oxygen consumption (V˙ O2 ) and V˙ CO2 increase with ex- patients with PCS and restore their exercise tolerance. ercise intensity, PaCO2 increases until the onset of ex- cess blood lactic acid accumulation, at which point METHODS there is a hyperventilation, that is, a respiratory com- The study protocol was reviewed and approved by pensation for the metabolic acidosis that reduces PaCO2 the Human Subjects Institutional Review Board at the and CBF.11,12,20 The onset of this compensatory hy- University at Buffalo and the participants signed an perventilation is called the “Ventilatory Threshold.”21 informed consent form. The PaCO2 where this hyperventilation begins varies among individuals; that is, some are more sensitive Subjects to the effects of CO2 than others. This is known as “CO2 sensitivity.”22 There is no reason to expect that Nine female athletes with PCS were recruited from the patients with concussion would have a different V˙ CO2 University at Buffalo Concussion Clinic (see Table 1). than healthy patients as the V˙ O2 for a given exercise in- All satisfied World Health Organization criteria for PCS tensity is not different.18 It is, however, possible that they and were symptomatic for more than 6 weeks but less have a different sensitivity to the effects of CO2 if con- than 12 weeks (mean duration of symptoms 9 weeks). cussion damaged the control centers for the ANS and Sports included soccer, basketball, and volleyball. A ref- V˙ E located in the brainstem. There is evidence that some erence group of 13 female age-matched, nonconcussed, patients with concussion have injury to the brainstem.23 aerobically trained recreational athletes was recruited If central damage altered V˙ E in patients with concussion, from the University community (see Table 1). They were abnormal control of CBF could produce symptoms training 3 to 5 days per week (30-60 min/day) for at least such as headache or dizziness that limit their exercise 2 years. They continued their regular training through- tolerance. out the study; thus, on the basis of prior work,26,27 we The purpose of this study was to examine the rela- assumed that their CO2 sensitivity would be stable and tionships between V˙ E , PaCO2 , and CBF during treadmill so they were only tested at the beginning of the study. TABLE 1 Physical characteristics Height, ˙ 2 peak, Vo PCSS Group Age, y cm Weight, kg mL/kg/min severity score Reference 21 ± 3 165 ± 9 60.95 ± 9.23 34.40 ± 4.51 5.5 PCS-Pre 23 ± 6 167 ± 3 62.23 ± 4.75 23.50 ± 2.82a 19a PCS-Post 23 ± 6 167 ± 3 62.23 ± 4.75 40.65 ± 1.62b 6.5b Abbreviations: PCS-Pre, concussed subjects before exercise treatment; PCS-Post, concussed subjects after exercise treatment; PCSS, postconcussion symptom scale, symptom severity score. a = significantly different from reference (<.001). b = significantly different from PCS-Pre (<.001). Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 CBF During Treadmill Exercise Is a Marker of Physiological PCS in Female Athletes 3 Design spite increasing exercise intensity and the respiratory exchange ratio exceeded 1.1.30 Heart rate was continu- All athletes with PCS were compared with the ref- ally monitored using a Polar heart rate monitor (Polar erence group at the beginning of the study. After this, 810i T61, Kempele, Finland). Blood pressure was taken 6 of the athletes with PCS were measured before and during the last minute of each stage from the left arm us- after a 12-week subsymptom threshold aerobic exercise ing a sphygmomanometer (Tango, SunTech Medical In- treatment program previously demonstrated to improve struments, Morrisville, North Carolina). As an estimate outcome in patients with PCS.5 Three athletes with PCS of PaCO2 , end-tidal CO2 (PET CO2 )31 was measured by did not complete the program because they were away mass spectrometer and recorded by a BIOPAC system from the university for an extended period. The refer- (MP150CE, Goleta, California). ence group was assessed only at the beginning of the study as we assumed that their data would not change CO2 sensitivity test significantly as they kept their training constant during the study period. The procedure used was a modified version of the re- breathing method described by Read and Leigh.32 Par- Physiological tests ticipants were seated in a chair with a curtain masking the equipment, except for the mouthpiece they breathed All participants completed a graded exercise test dur- from. They wore a nose clip throughout the test. They ing which physiological variables were measured as well were instructed to relax and breathe into and out of the as a standardized CO2 sensitivity test. Resting concus- mouthpiece as naturally as possible. The mouthpiece sion symptoms were recorded in the reference group was connected to a spirometer (Ohio Medical Products, and for subjects with PCS before and after exercise treat- Houston, Texas). The spirometer was filled to a volume ment, prior to exercise testing, using the Post-concussion of 10 L with a gas mixture containing approximately 5% Symptom Scale, a validated assessment instrument CO2 , 50% O2 , and 45% N2 . Participants breathed room that includes 22 symptoms of concussion (headache, air for the first 4 minutes after which they rebreathed dizziness, photophobia, etc) with sound psychomet- the gas mixture for 8 minutes until inhaled CO2 reached ric properties and normative data for men and women 9.0% or the participant voluntarily stopped. Heart rate (see Table 1).28 was continually recorded using a Polar heart rate mon- itor. Blood pressure was taken every 2 minutes us- Treadmill test ing a BioZ system sphygmomanometer (BioZ, Cardio Dynamics, San Diego, California). PET CO2 was mea- Participants performed a graded exercise test on sured by mass spectrometer (Perkin-Elmer, Pomona, a treadmill (Trackmaster, Full Vision Inc, Newton, California) and recorded by a BIOPAC system. Kansas). Participants walked initially at a speed of 3.0 mph at 0% grade. After 3 minutes, the grade was raised Cerebral blood flow velocity by 2.0% every 2 minutes until participants reached vol- untary exhaustion. This protocol allowed participants Cerebral blood flow velocity (CBFV) was measured to reach a “physiological steady state” for each intensity continuously during both the CO2 sensitivity and tread- level.21 The participants in both groups were instructed mill tests. Cerebral blood flow velocity was measured bi- to notify the administrator of increased or new concus- laterally from the middle cerebral artery (MCA) and con- sion symptoms. When they reported exacerbation of tinually recorded using the Transcranial Doppler (TCD). symptoms, the test was terminated. The test was termi- The MCA was observed through the transtemporal win- nated for reference subjects when the subject reported dow. The MCA was accurately identified by the depth of symptom exacerbation or reached exhaustion (defined the sample volume, direction of flow with relation to the as a rating of perceived exertion of ≥ 18).29 probe, and the relationship and depth of the M-mode Expired gas was collected in Douglas bags for the color bands to each other. Doppler probes were held last minute of each stage. The volume of the expired at a constant angle of insonation. Measurements were gas was measured using a dry gas meter (Harvard Dry taken at a depth of 35 to 55 mm. The Doppler power Gas Meter, Model # AH-50-6164, Harvard Apparatus was set between 80% and 100%, always using the lowest Inc, Holliston, Massachusetts) and the temperature de- possible power. The Doppler probes were fixated in the termined (Yellow Springs Instrument Co, Inc, Yellow appropriate position on the head using a secure head Springs, Ohio). Gas concentrations were analyzed by set (see Figure 1). The same trained, experienced techni- the mass spectrometer (Perkin Elmer 1100), which was cian recorded all TCD signals and used quality control calibrated before each test. V˙ O2 and V˙ CO2 were calcu- measures in setting up the experiments throughout the lated using standard procedures. V˙ O2 was considered study. TCD values from both sides of the head revealed maximum if it plateaued (<150 ml/min increase) de- mean values of 72 ± 13 cm/s for the right side and www.headtraumarehab.com Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 4 JOURNAL OF HEAD TRAUMA REHABILITATION bation or when their prescribed exercise duration was reached, whichever came first. Participants returned for treadmill tests every 3 weeks until their symptoms were no longer exacerbated on the treadmill. Participants kept daily reports of symptoms, which were reviewed weekly for compliance with the program. Participants saw the study physician for a repeat evaluation to ensure that they were symptom free after completion of the final treadmill test (to voluntary exhaustion). Physiologic res- olution of PCS was defined as the ability to exercise to voluntary exhaustion without exacerbation of concus- sion symptoms.5,29 Statistical analysis Figure 1. Representative example of the Transcranial Doppler setup during treadmill testing in a patient with postconcussion All statistics were calculated using SigmaPlot 10.0 by syndrome. Doppler probes are held at a constant angle of SigmaStat 3.5 (Systat Software Inc, Chicago, Illinois). insonation during exercise using the secure head set. All descriptive data were expressed as means ± standard deviation. A statistically significant result was defined 73 ± 11 cm/s for the left side. The correlation coeffi- as P ≤ .05. Data were first tested to see whether they cient between sides was 0.93 with a slope of 1.04. displayed a normal distribution. If the data passed the Inferring changes in CBF from CBFV requires the as- normality test, parametric statistics were used. If statis- sumption that the diameter of the carotid blood vessels tically significant differences were identified, the Holm- is constant.33 Although MCA diameters may be age- Sidak method was used for multiple comparisons to iso- related, the subjects of this study were of similar age and late the groups that differed. If the data did not pass the there is no evidence that the MCA diameter changes normality test, nonparametric statistics were used, that in women during exercise.34 TCD has been shown in is, the Friedman repeated-measures ANOVA on ranks many studies to be a valid indicator of CBF during was used to determine statistical significance. If the data exercise12,35-44 and our reference group data (at similar were statistically significant, the Tukey test was used V˙ O2 s) are in agreement with the CBFV values reported for multiple comparisons to isolate the groups that dif- in these studies. fered (see the Supplemental Digital Content, available The CO2 in the exhaled air was monitored at: http://links.lww.com/JHTR/A141). and recorded continuously by a mass spectrometer (MGA1100, Perkin-Elmer, Pomona, California) for de- RESULTS termination of PET CO2 . Ventilation was determined by Nine patients with PCS and 13 reference subjects a flow meter. Data were recorded and analyzed using a completed the exercise test at the beginning of the study. BIOPAC system (MP150CE, Goleta, California). V˙ E was Of the 9 patients with PCS, 6 of them performed an plotted as a function of the corresponding PET CO2 . A exercise treatment program and were retested after com- regression was determined for normal breathing and that pleting the program. Based on established criteria,45 all influenced by the CO2 inhalation. The CO2 threshold of the reference subjects reached their maximal V˙ O2 was the PET CO2 point at which the ventilation increased and had reliable CBFV, V˙ E and PET CO2 data, in spite above normal breathing. The slope of the V˙ E (L/min) in- of reaching only 95% of their age predicted maximal crease versus PET CO2 (mmHg) was defined as the CO2 HR. Twelve of the 13 reference subjects followed the sensitivity.22 same pattern in their exercise V˙ E , PET CO2 , and CBFV. They had similar CO2 sensitivity and their data agreed Exercise treatment program with previous studies.11,12,35,40 One subject increased Six participants with PCS underwent an aerobic exer- her V˙ E with V˙ CO2 such that her PaCO2 (PETCO2 ) was cise treatment program that has previously been shown maintained in the normal range (nominally 40 mmHg) to be safe and reliable.25 Participants performed aero- and thus her CBFV did not increase with submaximal bic exercise using an HR monitor 20 minutes per day, exercise. Her data therefore were not included in the 5 to 6 days per week, at an intensity of 80% of the reference group analysis. maximum HR achieved during the first treadmill test. A representative example of the TCD setup is pre- They were required to have someone present during sented in Figure 1. A representative color Doppler CBFV exercise for safety monitoring and were instructed to signal obtained during the treadmill test in a patient terminate exercise at the first sign of symptom exacer- with PCS is depicted in Figure 2. Prior to subthreshold Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 CBF During Treadmill Exercise Is a Marker of Physiological PCS in Female Athletes 5 TABLE 2 Physiological responses to exercise testing Maximum Exercise exercise Group Variable onset tolerance Reference HR (b/min) 82 ± 12 194 ± 8 PCS-Pre HR (b/min) 117 ± 17a 144 ± 20a PCS-Post HR (b/min) 115 ± 12a 173 ± 11b Reference SBP (mmHg) 116 ± 8 163 ± 16 PCS-Pre SBP (mmHg) 120 ± 6 138 ± 6a PCS-Post SBP (mmHg) 111 ± 11 160 ± 14b Reference DBP (mmHg) 77 ± 9 75 ± 8 PCS-Pre DBP (mmHg) 79 ± 10 74 ± 12 Figure 2. Representative raw data tracing from the screen of PCS-Post DBP (mmHg) 78 ± 6 81 ± 5 the Transcranial Doppler unit during treadmill walking in a patient with postconcussion syndrome. X-axes represent time. Abbreviations: b/min, beats per minute; DBP, diastolic blood pres- Top panel y-axis: depth of the artery in mm. Bottom panel sure; HR, heart rate; PCS-Pre, patients with PCS before exercise y-axis: blood flow speed in centimeters per second. treatment; PCS-Post, patients with PCS after exercise treatment; SBP, systolic blood pressure. a Significantly different from reference (<.001). exercise treatment, patients with PCS had limited exer- b Significantly different from PCS-Pre (<.001). cise tolerance because of symptom exacerbation dur- ing the treadmill test. After exercise treatment, they Expired ventilation (V˙ E ) was significantly depressed had normal exercise tolerance without symptom ex- and therefore PET CO2 was significantly elevated in the acerbation on the treadmill. Normative data from the patients with PCS compared with the reference group at Post-concussion Symptom Scale indicate that healthy all V˙ O2 s (see Figures 4 and 5, respectively). The CO2 sen- females can report up to a symptom severity score of sitivity of patients with PCS was significantly depressed 9.28 Table 1 shows that prior to exercise treatment, pa- (1.35 ± 2.11 L/min/mmHg CO2 ) versus the reference tients with PCS had an elevated level of resting symp- group (2.88 ± 0.60 L/min/mmHg CO2 ) before treat- toms when compared with the reference group. Resting ment (see Figure 6). symptoms in patients with PCS returned to reference group levels after exercise treatment. Before exercise treatment Up until the point of exercise cessation, there were no significant differences in the V˙ O2 values at all treadmill settings between patients with PCS and the reference group and between patients with PCS pre- and postex- ercise treatment (data not shown). Patients with PCS before treatment had greater HRs at exercise onset ver- sus the reference group (see Table 2). Patients with PCS stopped exercise on the treadmill at a V˙ O2 of 23.50 ± 2.82 mL/kg/min (see Table 1), an HR of 144 ± 20 beats per minute (see Table 2), and a systolic BP (SBP) of 138 ± 6 mmHg, with a nor- Figure 3. Mean data for cerebral blood flow velocity are plot- ted as a function of V˙ O2 for patients with postconcussion syn- mal DBP response (see Table 2), well below their re- drome (PCS) pre- (PCS-Pre, n = 6) and postexercise treatment spective predicted maximal values. They stopped exer- (PCS-Post, n = 6) as well as for reference subjects (Control, cise because of significant exacerbation of symptoms of n = 12). ∗ A significant difference (P ≤ .05) between values for headache and dizziness, which defined their symptom- PCS patients and reference subjects; + a significant difference limited threshold.5 As can be seen in Figure 3, patients (P ≤ .05) between PCS patients post- compared with preexer- with PCS had significantly higher CBFV at all V˙ O2 val- cise treatment. The coefficient of variation (CV) for the PCS ues. Their maximal V˙ O2 achieved was significantly lower patients preexercise treatment ranged from 4% to 14%, for at the point where symptoms increased (at their limit of postexercise treatment from 11% to 14%, and for reference exercise tolerance) versus the reference group. subjects from 30% to 33%. www.headtraumarehab.com Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 6 JOURNAL OF HEAD TRAUMA REHABILITATION Figure 6. Mean data ± standard deviation for CO2 sensitivity Figure 4. Mean data for expired ventilation are plotted as a are shown for patients pre- (n = 6) and postexercise treatment function of V˙ O2 for PCS patients pre- (PCS-Pre, n = 6) and (n = 6) as well as for reference subjects (Control, n = 12). postexercise treatment (PCS-Post, n = 6) as well as for refer- ∗ A significant difference (P ≤ .05) between values for PCS ence subjects (Control, n = 12). ∗ A significant difference (P ≤ patients and reference subjects; + a significant difference (P ≤ .05) between values for PCS patients and reference subjects; .05) between PCS patients post- compared with preexercise + a significant difference (P ≤ .05) between PCS patients pos- treatment. texercise treatment compared with preexercise treatment. The coefficient of variation (CV) for the PCS patients preexercise exhaustion without symptom exacerbation, achieving treatment ranged from 19% to 24%, for postexercise treatment 85% of their age predicted maximum HR (173 ± 11 from 14% to 19%, and for reference subjects from 18% to 20%. beats per minute), an SBP of 160 ± 14 mmHg, and a V˙ O2 of 40.65 ± 1.62 mL/kg/min (see Table 1). It should be noted that even after exercise treatment, pa- tients with PCS achieved only 90% of their predicted V˙ O2 max (based on HR data) and did not have the respi- ratory compensation for metabolic acidosis seen in the reference group (see Figure 4); thus, they did not have a reduction in CBFV at their highest V˙ O2 (see Figure 3). As can be seen in Figures 4 and 5, after exercise treat- ment V˙ E increased significantly and PET CO2 decreased significantly in patients with PCS versus pretreatment values and were not different from the reference group values. The increased V˙ E was associated with a signif- icant increase in CO2 sensitivity versus before treat- ment (see Figure 6). In addition, CBFV was significantly Figure 5. Mean data for end-tidal CO2 (an estimate of arterial lower after when compared with before treatment (see CO2 ) are plotted as a function of V˙ O2 for PCS patients pre- Figure 3). (PCS-Pre, n = 6) and postexercise treatment (PCS-Post, n = 6) as well as for reference subjects (Control, n = 12). ∗ A signifi- cant difference (P ≤ .05) between values for PCS patients and DISCUSSION reference subjects; + a significant difference (P ≤ .05) between This study shows that Division 1 female college ath- PCS patients postexercise treatment compared with preexer- letes with PCS had a physiological disturbance affect- cise treatment. The coefficient of variation (CV) for the PCS ing control of CBF in association with limited exer- patients preexercise treatment ranged from 14% to 15%, for cise tolerance. When compared at equivalent workloads postexercise treatment from 11% to 19%, and for reference ˙ 2 s) with a reference group of aerobically trained subjects from 19% to 20%. (ie, VO recreational female athletes, patients with PCS had ex- acerbation of symptoms of headache and dizziness that After exercise treatment occurred well below their maximal exercise capacity. The The HR and BP responses during the exercise test after physiological data show that patients with PCS had a rel- subthreshold exercise treatment are shown in Table 2. ative hypoventilation (lower V˙ E ) that disproportionately Patients with PCS after treatment still had greater HRs at increased their PaCO2 and CBFV, which suggests that exercise onset versus the reference group. In agreement CBF was similarly increased. The relative hypoventila- with a previous study,5 patients with PCS exercised to tion appears to have been a consequence of reduced Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 CBF During Treadmill Exercise Is a Marker of Physiological PCS in Female Athletes 7 sensitivity to the effects of CO2 . We hypothesize that sitivity would have changed had it been measured 12 this reflects altered central (medullary) and/or peripheral weeks later.26,27 Furthermore, endurance training lowers (carotid body) respiratory control that to our knowledge CO2 sensitivity in normal subjects. A 12-week program has not been reported previously in patients with con- of subthreshold aerobic exercise treatment increased CO2 cussion or patients with PCS. CO2 sensitivity in patients sensitivity in our PCS patients, which is consistent with with PCS normalized after a program of subthreshold an effect of concussion resolution rather than an effect aerobic exercise treatment in association with restora- of physical training. tion of V˙ E , PET CO2 , CBFV, and exercise tolerance to Cerebral artery resistance is very sensitive to changes reference group levels. These data suggest that concus- in PaCO2 .21 Breathing 7% CO2 can increase CBF ap- sion can alter the physiological response to arterial CO2 . proximately 2-fold in both men19 and women,50 which Reduced CO2 sensitivity resulted in a blunted ventila- is similar to what we observed in this study during exer- tory response to the increase in V˙ CO2 (CO2 production) cise. CO2 sensitivity is a measure of the level of ventila- during exercise. This elevated the PaCO2 that in turn ele- tion for a given arterial CO2 tension (PaCO2 ). Some vated exercise CBFV, which suggests that CBF increased. people are very sensitive; that is, they breathe a lot The elevated CBFV was out of proportion to exercise in- at a particular PaCO2 . Others, however, hypoventilate tensity (V˙ O2 ) and produced symptoms of headache and at the same PaCO2 ; that is, they have low CO2 sensi- dizziness that limited exercise tolerance. Subthreshold tivity and tolerate higher levels of PaCO2 chronically aerobic exercise treatment normalized the CO2 physi- (hypercapnia).51 Reduced CO2 response, hypoventila- ology and exercise capacity of our patients with PCS. tion, increased PET CO2 , and ensuing hypercapnia in Our data suggest that return of normal CO2 sensitivity response to exercise have been observed in trained un- and of exercise tolerance can be objective markers of derwater swimmers52 and in nonimmersed divers.53 The recovery from concussion. locus of altered CO2 sensitivity after concussion is not Some athletes do not breathe as much as do others known. One possibility is that concussion damages the in response to increasing CO2 levels; that is, they have brainstem, or communication with the brainstem, thus reduced CO2 sensitivity. In addition, their ventilatory interfering with the action of the central chemorecep- response to the increased CO2 production during ex- tors that are highly sensitive to changes in PaCO2 .54 We ercise is also blunted.46 The subjects with PCS in the have identified brainstem white matter abnormalities by present study had low CO2 sensitivity and a blunted diffusion tensor imaging in patients with PCS.23 Others ventilatory response to the increased V˙ CO2 during ex- have identified abnormal regional cerebrovascular re- ercise. We propose that the low CO2 sensitivity in our sponsiveness during provocative CO2 challenge in pa- athletes with PCS prior to subthreshold exercise training tients with PCS not present in healthy controls.55 Thus, was not a consequence of their training status. Subjects altered CO2 physiology may be a biological marker of with PCS in this study played anaerobic sports such as concussion. soccer, basketball, and volleyball. These types of athletes The physiological responses of the reference group do not have reduced CO2 sensitivity.47 Rather, reduced followed the pattern typically observed during treadmill CO2 sensitivity has been reported only in endurance- exercise.21,30 Many recent studies using TCD show that trained athletes.46,48 In addition, even if the subjects in CBF increases gradually during exercise. 36-38,40-44,56-66 this study had low CO2 sensitivity to begin with, inac- Ogoh et al,40 for example, found that both CBFV and tivity eliminates the depressed CO2 sensitivity acquired MAP increased by similar degrees during exercise, which during active training27,47,49 and our subjects with PCS is consistent with our results at low and medium levels were inactive for an average of 9 weeks prior to ex- of exercise. At greater workloads beyond the ventilatory ercise intervention. Therefore, even if they had expe- threshold, however, this relationship changed. Cerebral rienced some level of a training effect to lower CO2 blood flow velocity significantly decreased following the sensitivity, it would likely have been lost during their hyperventilatory reduction in PaCO2 despite increasing prolonged period of inactivity. Thus, it is reasonable to MAP. This suggests that PaCO2 is the primary driving conclude that the low CO2 sensitivity prior to subthresh- force to CBFV during exercise. Not everyone has, how- old training in the athletes with PCS was the result of ever, shown this relationship during exercise. Cerebral concussion. blood flow velocity has been reported to increase with Could the effects on CO2 sensitivity we observed increasing MAP without changes in PaCO2 .64 This may with subthreshold exercise training in the PCS subjects be the case at low exercise levels below the ventilatory have been observed with a similar training program in threshold but appears not to be at the higher inten- nonconcussed subjects? We think this is unlikely. The sities where hyperventilation normally reduces PaCO2 reference sample was composed of aerobically trained and CBFV. recreational athletes who continued their regular train- Our patients with PCS stopped exercise because of ing program and thus it is unlikely that their CO2 sen- the development of symptoms at a lower SBP than the www.headtraumarehab.com Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 8 JOURNAL OF HEAD TRAUMA REHABILITATION reference group but at levels that are typically within movement artifact. As shown in Figure 2, however, the the autoregulatory range.21 Despite this, CBFV was signals we obtained were of high quality and an expe- significantly elevated at all V˙ O2 s versus controls within rienced technician performed all of the measurements, the range of blood pressures where CA would normally thus removing any concern about operator dependence. prevent increased CBF,21 leading to speculation that CA We did not measure CA directly or CBFV in cerebral was not intact in patients with PCS. Laboratory67 and arteries other than the MCA. The CBF response to exer- clinical studies68 have shown abnormal CA in moderate cise may represent an increase in brain activation42 from to severe TBI that could be related to the reported alter- increased metabolites in neurons or as a result of neural ation in ANS activity in patients with brain injury.14,69 control.41,42,57 Linkis et al,57 for example, reported that Our results are consistent with these reports as patients during exercise of different body parts CBFV increased with PCS demonstrated altered HR responses versus the in both women and men in the artery supplying the reference group. The response of patients with PCS to cortical representation of the muscles used during the exercise treatment is consistent with a normalization of exercise. Jorgensen et al41 found that CBFV increased ANS function since HRs during exercise returned to the in the MCA but not the anterior cerebral artery in cy- reference group levels, which has been demonstrated cling women and men. We did not measure flow in in patients having recovered from moderate to severe the anterior cerebral artery so it is possible that the in- TBI.70 creased activation (ie, cerebral metabolism) in the cor- What could explain the resolution of abnormal con- tical representation of the muscles being used during cussion physiology with subthreshold exercise treat- exercise was responsible for the increased CBFV. Even ment? Exercise training is well known to increase if that were the case, it does not invalidate the integrated parasympathetic activity at rest71 and may have restored physiological response we observed in our patients ANS balance in patients with PCS. Progressive stepwise with PCS. aerobic training may improve CA by conditioning the brain to gradually adapt to repetitive mild elevations of CONCLUSION systolic BP.36 With respect to ventilation, it has been shown that a controlled progressive breathing training This study shows that some patients with PCS have program can increase CO2 sensitivity in persons with abnormal physiological function that limits their exer- low CO2 sensitivity to begin with.22 The progressive cise tolerance and that subthreshold aerobic exercise subthreshold exercise program performed by the pa- treatment can improve the physiology and restore ex- tients with PCS may have improved CO2 sensitivity ercise tolerance to normal. A key physiological mecha- by simulating such a breathing pattern. nism appears to be a reduced sensitivity to the effects Limitations of the study include the use of a refer- of arterial CO2 in patients after concussion that blunts ence group as opposed to employing an identical train- exercise ventilation, increasing arterial CO2 levels, and ing regimen in the healthy subjects and measuring CO2 therefore CBFV out of proportion to exercise intensity sensitivity afterward in them. The PCS patients were (V˙ O2 ). We believe that this disproportionate CBFV re- elite athletes from various team sports while the refer- sponse produces symptoms such as headache and dizzi- ence group included trained but recreational athletes. ness that limit exercise tolerance after concussion. These In future studies, a control group composed of seden- data are important for clinicians. They support the hy- tary subjects or subjects with fitness levels matched to pothesis that patients with PCS with exercise intolerance the athletes during their PCS would add to the under- have a physiological source for their symptoms that may standing of the relationships among CO2 sensitivity, be amenable to individualized subthreshold aerobic ex- V˙ E, V˙ CO2 , PaCO2 , and CBFV. Even though the differ- ercise treatment or to other approaches that improve ences between the PCS and the reference groups were autonomic function, such as ventilatory biofeedback,72 . significant, because of the potential heterogeneity of re- Of perhaps greater significance, the data suggest that re- sponses from PCS patients, further studies with more turn of normal control of exercise CBF and of exercise subjects would be desirable to confirm these results. Al- tolerance could be objective physiological markers of though TCD is widely used in similar studies and has recovery from concussion, which has implications for been shown to be valid and reliable by many authors, establishing prognosis and preventing premature return there are potential errors due to probe placement and to sport, activity, or military duty. REFERENCES 1. McCrory P, Meeuwisse W, Aubry M, et al. Consensus state- Concussion in Sport held in Zurich, November 2012. Clin J Sport ment on concussion in sport—the 4th International Conference on Med. 2013;23(2):89-117. Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 CBF During Treadmill Exercise Is a Marker of Physiological PCS in Female Athletes 9 2. Leddy JJ, Sandhu H, Sodhi V, Baker JG, Willer B. Rehabilita- 23. Polak P, Leddy JJ, Dwyer MG, Willer B, Zivadinov R. Diffusion tion of concussion and postconcussion syndrome. Sports Health. tensor imaging alterations in patients with postconcussion syn- 2012;4(2):147-154. drome undergoing exercise treatment: a pilot longitudinal study. 3. World Health Organization. International Statistical Classification of J Head Trauma Rehabil. 2015;30(2):E32-E42. Diseases and Related Health Problems. 10th ed. Geneva, Switzerland: 24. Covassin T, Elbin RJ. The female athlete: the role of gender in World Health Organization; 1992. the assessment and management of sport-related concussion. Clin 4. Harmon KG, Drezner JA, Gammons M, et al. American Medical Sports Med. 2011;30(1):125-131, x. Society for Sports Medicine position statement: concussion in 25. Leddy JJ, Baker JG, Kozlowski K, Bisson L, Willer B. Reliability sport. Br J Sports Med. 2013;47(1):15-26. of a graded exercise test for assessing recovery from concussion. 5. Leddy JJ, Kozlowski K, Donnelly JP, Pendergast DR, Epstein LH, Clin J Sport Med. 2011;21(2):89-94. Willer B. A preliminary study of subsymptom threshold exercise 26. Miyamura M, Hiruta S, Sakurai S, Ishida K, Saito M. Effects of training for refractory postconcussion syndrome. Clin J Sport Med. prolonged physical training on ventilatory response to hypercap- 2010;20(1):21-27. nia. Tohoku J Exp Med. 1988;156(Suppl.):125-135. 6. Leddy JJ, Kozlowski K, Fung M, Pendergast DR, Willer B. Regu- 27. Miyamura M, Ishida K. Adaptive changes in hypercapnic venti- latory and autoregulatory physiological dysfunction as a primary latory response during training and detraining. Eur J Appl Physiol characteristic of postconcussion syndrome: implications for treat- Occup Physiol. 1990;60(5):353-359. ment. NeuroRehabilitation. 2007;22(3):199-205. 28. Lovell MR, Iverson GL, Collins MW, et al. Measurement of 7. Maugans TA, Farley C, Altaye M, Leach J, Cecil KM. Pediatric symptoms following sports-related concussion: reliability and sports-related concussion produces cerebral blood flow alterations. normative data for the postconcussion scale. Appl Neuropsychol. Pediatrics. 2012;129(1):28-37. 2006;13(3):166-174. 8. McAllister TW, Saykin AJ, Flashman LA, et al. Brain activation 29. Leddy JJ, Willer B. Use of graded exercise testing in concus- during working memory 1 month after mild traumatic brain injury: sion and return-to-activity management. Curr Sports Med Rep. a functional MRI study. Neurology. 1999;53(6):1300-1308. 2013;12(6):370-376. 9. Leddy JJ, Cox JL, Baker JG, et al. Exercise treatment for postcon- 30. Whaley M. ACSM’s Guidelines for Exercise Testing and Prescrip- cussion syndrome: a pilot study of changes in functional magnetic tion. 7th ed. Philadelphia, PA: Lippincott Williams and Wilkins; resonance imaging activation, physiology, and symptoms. J Head 2005. Trauma Rehabil. 2013;28(4):241-249. 31. Yosefy C, Hay E, Nasri Y, Magen E, Reisin L. End tidal carbon 10. Chen JK, Johnston KM, Frey S, Petrides M, Worsley K, Ptito A. dioxide as a predictor of the arterial PCO2 in the emergency de- Functional abnormalities in symptomatic concussed athletes: an partment setting. Emerg Med J. 2004;21(5):557-559. fMRI study. Neuroimage. 2004;22(1):68-82. 32. Read DJ, Leigh J. Blood-brain tissue PCO2 relationships and ven- 11. Ogoh S, Dalsgaard MK, Yoshiga CC, et al. Dynamic cerebral tilation during rebreathing. J Appl Physiol. 1967;23(1):53-70. autoregulation during exhaustive exercise in humans. Am J Physiol 33. Serrador JM, Picot PA, Rutt BK, Shoemaker JK, Bondar RL. MRI Heart Circ Physiol. 2005;288(3):H1461-H1467. measures of middle cerebral artery diameter in conscious humans 12. Querido JS, Sheel AW. Regulation of cerebral blood flow during during simulated orthostasis. Stroke. 2000;31(7):1672-1678. exercise. Sports Med. 2007;37(9):765-782. 34. Tarasow E, Abdulwahed Saleh Ali A, Lewszuk A, Walecki J. 13. Gall B, Parkhouse WS, Goodman D. Exercise following a sport Measurements of the middle cerebral artery in digital subtrac- induced concussion. Br J Sports Med. 2004;38(6):773-777. tion angiography and MR angiography. Med Sci Monit. 2007;13 14. Hilz MJ, DeFina PA, Anders S, et al. Frequency analysis unveils (Suppl. 1):65-72. cardiac autonomic dysfunction after mild traumatic brain injury. 35. Ogoh S, Ainslie PN, Miyamoto T. Onset responses of ventila- J Neurotrauma. 2011;28(9):1727-1738. tion and cerebral blood flow to hypercapnia in humans: rest and 15. Gall B, Parkhouse W, Goodman D. Heart rate variability of re- exercise. J Appl Physiol. 2009;106(3):880-886. cently concussed athletes at rest and exercise. Med Sci Sports Exerc. 36. Brys M, Brown CM, Marthol H, Franta R, Hilz MJ. Dy- 2004;36(8):1269-1274. namic cerebral autoregulation remains stable during physical chal- 16. King ML, Lichtman SW, Seliger G, Ehert FA, Steinberg JS. lenge in healthy persons. Am J Physiol Heart Circ Physiol. 2003; Heart-rate variability in chronic traumatic brain injury. Brain Inj. 285(3):H1048-H1054. 1997;11(6):445-453. 37. Fisher JP, Ogoh S, Young CN, Raven PB, Fadel PJ. Regulation of 17. DeWitt DS, Prough DS. Traumatic cerebral vascular injury: the middle cerebral artery blood velocity during dynamic exercise in effects of concussive brain injury on the cerebral vasculature. humans: influence of aging. J Appl Physiol. 2008;105(1):266-273. J Neurotrauma. 2003;20(9):795-825. 38. Hellstrom G, Fischer-Colbrie W, Wahlgren NG, Jogestrand T. 18. Kozlowski KF, Graham J, Leddy JJ, Devinney-Boymel L, Willer BS. Carotid artery blood flow and middle cerebral artery blood Exercise intolerance in individuals with postconcussion syndrome. flow velocity during physical exercise. J Appl Physiol. 1996;81(1): J Athl Train. 2013;48(5):627-635. 413-418. 19. Imray CH, Walsh S, Clarke T, et al. Effects of breathing air con- 39. Hellstrom G, Wahlgren NG. Physical exercise increases middle taining 3% carbon dioxide, 35% oxygen or a mixture of 3% car- cerebral artery blood flow velocity. Neurosurg Rev. 1993;16(2): bon dioxide/35% oxygen on cerebral and peripheral oxygenation 151-156. at 150 m and 3459 m. Clin Sci (Lond). 2003;104(3):203-210. 40. Ogoh S, Fadel PJ, Zhang R, et al. Middle cerebral artery flow 20. Ogoh S, Ainslie PN. Regulatory mechanisms of cerebral blood velocity and pulse pressure during dynamic exercise in humans. flow during exercise: new concepts. Exerc Sport Sci Rev. 2009; Am J Physiol Heart Circ Physiol. 2005;288(4):H1526-H1531. 37(3):123-129. 41. Jorgensen LG, Perko G, Secher NH. Regional cerebral artery mean 21. Levy M, Koeppen B, Stanton B, eds. Berne and Levy Principles of flow velocity and blood flow during dynamic exercise in humans. Physiology. 4th ed. Philadelphia, PA: Elsevier Mosby; 2006. J Appl Physiol. 1992;73(5):1825-1830. 22. Pendergast DR, Lindholm P, Wylegala J, Warkander D, Lundgren 42. Jorgensen LG, Perko M, Hanel B, Schroeder TV, Secher NH. CE. Effects of respiratory muscle training on respiratory CO2 Middle cerebral artery flow velocity and blood flow during exercise sensitivity in SCUBA divers. Undersea Hyperb Med. 2006;33(6): and muscle ischemia in humans. J Appl Physiol. 1992;72(3):1123- 447-453. 1132. www.headtraumarehab.com Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. LWW/JHTR JHTR-D-14-00197 May 14, 2015 21:10 10 JOURNAL OF HEAD TRAUMA REHABILITATION 43. Moraine JJ, Lamotte M, Berre J, Niset G, Leduc A, Naeije R. Rela- 59. Ide K, Pott F, Van Lieshout JJ, Secher NH. Middle cerebral tionship of middle cerebral artery blood flow velocity to intensity artery blood velocity depends on cardiac output during exer- during dynamic exercise in normal subjects. Eur J Appl Physiol cise with a large muscle mass. Acta Physiol Scand. 1998;162(1): Occup Physiol. 1993;67(1):35-38. 13-20. 44. Thomas SN, Schroeder T, Secher NH, Mitchell JH. Cerebral blood 60. Giller CA, Giller AM, Cooper CR, Hatab MR. Evaluation of flow during submaximal and maximal dynamic exercise in hu- the cerebral hemodynamic response to rhythmic handgrip. J Appl mans. J Appl Physiol. 1989;67(2):744-748. Physiol. 2000;88(6):2205-2213. 45. ACSM’s Guidelines for Exercise Testing and Prescription. 7th ed. 61. Nybo L, Nielsen B. Middle cerebral artery blood velocity is re- Philadelphia, PA: Lippincott Williams & Wilkins; 2006. duced with hyperthermia during prolonged exercise in humans. 46. McConnell AK, Semple ES. Ventilatory sensitivity to carbon diox- J Physiol (Lond). 2001;534(Pt 1):279-286. ide: the influence of exercise and athleticism. Med Sci Sports Exerc. 62. Imray CHE, Myers SD, Pattinson KTS, et al. Effect of exercise 1996;28(6):685-691. on cerebral perfusion in humans at high altitude. J Appl Physiol. 47. Kelley MA, Laufe MD, Millman RP, Peterson DD. Ventilatory 2005;99(2):699-706. response to hypercapnia before and after athletic training. Respir 63. Querido JS, Rupert JL, McKenzie DC, Sheel AW. Effects of Physiol. 1984;55(3):393-400. intermittent hypoxia on the cerebrovascular responses to sub- 48. Miyamoto T, Inagaki M, Takaki H, et al. Adaptation of the maximal exercise in humans. Eur J Appl Physiol. 2009;105(3): respiratory controller contributes to the attenuation of exer- 403-409. cise hyperpnea in endurance-trained athletes. Eur J Appl Physiol. 64. Ide K, Horn A, Secher NH. Cerebral metabolic response to sub- 2012;112(1):237-251. maximal exercise. J Appl Physiol. 1999;87(5):1604-1608. 49. Katayama K, Sato Y, Morotome Y, et al. Ventilatory chemosensi- 65. Heckmann JG, Brown CM, Cheregi M, Hilz MJ, Neundorfer ¨ tive adaptations to intermittent hypoxic exposure with endurance B. Delayed cerebrovascular autoregulatory response to ergome- training and detraining. J Appl Physiol. 1999;86(6):1805-1811. ter exercise in normotensive elderly humans. Cerebrovasc Dis. 50. Hetzel A, Braune S, Guschlbauer B, Dohms K. CO2 reactivity 2003;16(4):423-429. testing without blood pressure monitoring? Stroke. 1999;30(2): 66. Doering TJ, Resch KL, Steuernagel B, Brix J, Schneider B, Fis- 398-401. cher GC. Passive and active exercises increase cerebral blood flow 51. Kerem D, Melamed Y, Moran A. Alveolar PCO2 during rest and velocity in young, healthy individuals. Am J Phys Med Rehabil. exercise in divers and nondivers breathing O2 at 1 ATA. Undersea 1998;77(6):490-493. Biomed Res. 1980;7(1):17-26. 67. Engelborghs K, Haseldonckx M, Van Reempts J, et al. Im- 52. Goff LG, Bartlett RG Jr. Elevated end-tidal CO2 in trained under- paired autoregulation of cerebral blood flow in an experimen- water swimmers. J Appl Physiol. 1957;10(2):203-206. tal model of traumatic brain injury. J Neurotrauma. 2000;17(8): 53. Lally DA, Zechman FW, Tracy RA. Ventilatory responses to exer- 667-677. cise in divers and nondivers. Respir Physiol. 1974;20(2):117-129. 68. Ter Minassian A, Dube L, Guilleux AM, Wehrmann N, Ursino M, 54. Nattie E. CO2 , brainstem chemoreceptors and breathing. Prog Beydon L. Changes in intracranial pressure and cerebral autoregu- Neurobiol. 1999;59(4):299-331. lation in patients with severe traumatic brain injury. Crit Care Med. 55. Mutch WA, Ellis MJ, Graham MR, et al. Brain MRI CO2 2002;30(7):1616-1622. stress testing: a pilot study in patients with concussion. PloS one. 69. Toledo E, Lebel A, Becerra L, et al. The young brain and concus- 2014;9(7):e102181. sion: imaging as a biomarker for diagnosis and prognosis. Neurosci 56. Pott F, Ray CA, Olesen HL, Ide K, Secher NH. Middle cerebral Biobehav Rev. 2012;36(6):1510-1531. artery blood velocity, arterial diameter and muscle sympathetic 70. Sviri GE, Aaslid R, Douville CM, Moore A, Newell DW. Time nerve activity during postexercise muscle ischaemia. Acta Physiol course for autoregulation recovery following severe traumatic Scand. 1997;160(1):43-47. brain injury. J Neurosurg. 2009;111(4):695-700. 57. Linkis P, Jorgensen LG, Olesen HL, Madsen PL, Lassen NA, Secher 71. Carter JB, Banister EW, Blaber AP. The effect of age and gender NH. Dynamic exercise enhances regional cerebral artery mean flow on heart rate variability after endurance training. Med Sci Sports velocity. J Appl Physiol. 1995;78(1):12-16. Exerc. 2003;35(8):1333-1340. 58. Pott F, Knudsen L, Nowak M, Nielsen HB, Hanel B, Secher NH. 72. Conder RL, Conder AA. Heart rate variability interven- Middle cerebral artery blood velocity during rowing. Acta Physiol tions for concussion and rehabilitation. Front Psychol. 2014;5: Scand. 1997;160(3):251-255. 1-7. Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.