The blood-brain barrier (BBB) is a semipermeable membrane that regulates the transport of selected chemical species into and out of the CNS. This is essential to ensure normal brain function and to protect the CNS from pathogens and other potentially harmful substances. While the BBB is largely resistant to changes in permeability, there are situations where the function of the BBB may be compromised (e.g., neuronal damage, infections, fever). A widespread increase in barrier permeability has been observed after 30 min of forced swimming exercise in rodents (25,26). These changes were found to be relatively acute, with normal BBB function restored 2 h after the end of exercise. Recent work has reported an increase in serum S100β concentration, which is used as a peripheral marker of BBB integrity, after prolonged exercise in a warm environment, but no such change was apparent when exercise was performed in temperate conditions (28). This finding has additional implications because serum S100β is sometimes employed as a peripheral marker of brain damage in individuals who suffer traumatic head injuries during sports. A change in BBB permeability during exercise may result in misleading results in exercising individuals, particularly under conditions of significant heat stress.
A number of factors may contribute to increased BBB permeability during exercise; these include the development of hyperthermia (29), increased circulating ammonia (14) and adrenaline (1) concentrations, changes to central serotonergic neurotransmission (25,26), and the production of proinflammatory cytokines (7). Additional stress on the BBB may come from disturbances to fluid homeostasis associated with the performance of prolonged exercise in a warm environment. Hyperosmolality of the extracellular fluid (ECF), caused by the loss of large volumes of hypotonic sweat, may cause a shift of fluids from the brain into the peripheral circulation (11). The net movement of fluid across the BBB, and shrinkage of the barrier's structural endothelial cells, can result in a transient opening of the tight junctions, causing increased exchange of substances between the periphery and the CNS.
Prolonged exercise in a warm environment results in a significant loss of hypotonic fluid, in the form of sweat secreted on to the skin surface, to assist in the dissipation of heat. The resulting plasma hyperosmolality causes a shift of fluids from interstitial and intracellular spaces to defend blood volume, thus helping to maintain blood flow to the working muscles and skin (18). It is also clear that an elevation in extracellular osmolality can exert a marked influence on brain volume (11). While there has been considerable clinical study of the effects of hyperosmolality on brain water homeostasis (5,6,23), the possibility that changes in extracellular tonicity occurring during exercise may influence on brain volume has received little attention. A recent report has provided some preliminary evidence that brain volume is altered during prolonged exercise in a warm environment (8), but the impact on BBB function was not explored.
The aim of the present study was to examine changes in serum concentrations of S100β, a peripheral marker of BBB permeability, in response to exercise in the heat with and without fluid ingestion. Large elevations in serum S100β are observed after head trauma, and this protein has been widely employed as a marker of CNS damage in clinical settings as well as in investigations of head injury in sports. Recent work reported increases in serum S100β of up to 0.34 μg·L−1 after osmotic opening of the BBB, despite the absence of neuronal damage, suggesting that S100β may be used as a peripheral marker of BBB integrity (12).
Eight physically active males (mean ± SD: age 25.8 ± 6.5 yr, height 1.77 ± 0.05 m, body mass 76.5 ± 6.2 kg, V˙O2peak 4.57 ± 0.42 L·min−1) were recruited to take part in this investigation. All subjects were physically active and familiar with the sensation of strenuous exercise, but none were accustomed to exercising in a warm environment at the time of the study. Individuals who had spent prolonged periods of time exposed to warm environments were excluded, because there is evidence that repeated exposure to heat stress may attenuate losses in barrier integrity associated with acute hyperthermia (27). Before volunteering, all subjects received written details outlining the nature and purpose of the study. After any questions regarding the protocol, a written statement of consent was signed. The protocol was approved by the Loughborough University ethical advisory committee.
All subjects completed a preliminary test, a familiarization trial, and two experimental trials. V˙O2peak was determined using a discontinuous, incremental exercise test to volitional exhaustion on a cycle ergometer (Gould Corival 300, Groningen, The Netherlands). The familiarization trial was undertaken to ensure the subjects were accustomed to the procedures employed during the investigation and to minimize any potential learning or anxiety effects. The experimental trials were administered in a randomized order, separated by at least 7 d to limit the development of heat acclimation. Subjects were instructed to record dietary intake and physical activity during the 2 d before the first trial, and to replicate this before the subsequent experimental trial. No exercise or alcohol consumption was permitted for 24 h before each trial.
The experimental trials took place in the morning after an overnight fast, other than the ingestion of 500 mL of plain water 90 min before entering the laboratory. On arrival, subjects first emptied their bladder and inserted a rectal thermistor 10 cm beyond the anal sphincter to permit the measurement of core temperature. Skin thermistors were attached to the skin at four sites (chest, triceps, thigh, and calf) to determine weighted mean skin temperature (22), and a heart rate telemetry band was positioned (Polar Beat, Kempele, Finland). Volunteers then sat in a comfortable environment (22-24°C) for 20 min. One hand was immersed in warm water (42°C) for 10 min before a 21-g cannula was introduced into a superficial forearm vein to allow the collection of serial blood samples. To ensure the cannula remained patent, it was flushed with a small volume of heparinized saline immediately after each collection. During the rest period, temperatures and heart rate were recorded at 5-min intervals, and a 7.5-mL blood sample was drawn at 20 min.
Subjects then entered a climatic chamber maintained at a temperature of 35.0 ± 0.5°C (relative humidity 56 ± 5%), and nude body mass, with the subject holding the temperature probes, was measured to the nearest 10 g (Adam CFW150 digital scale, Milton Keynes, UK). The subject then completed up to six 15-min periods of cycle exercise at a workload corresponding to 55% of V˙O2peak. A 5-min rest period separated successive blocks of exercise, during which the subject's dry, nude body mass was recorded. In the fluid (F) trial, sweat losses, determined for each period of exercise through changes in body mass, were retrospectively replaced by the ingestion of plain water (Evian, Danone, UK) during the first 5 min of the next block of exercise. The temperature of the water was approximately 20°C. Nothing was ingested by the subjects during the no-fluid (NF) trial. Core and skin temperatures and heart rate were recorded at 5-min intervals during exercise. Ratings of perceived exertion (3) and thermal stress (using a 21-point scale ranging from unbearable cold (−10) to unbearable heat (+10)) were obtained at the end of each block. Blood samples (7.5 mL) were drawn during the final minute of each exercise period and, if necessary, when the subject stopped exercise because of volitional exhaustion. At the end of exercise, subjects were weighed before returning to a comfortable environment to allow recovery to be monitored for a further 15 min.
Blood handling and analysis.
Blood samples were drawn directly into dry syringes. A 2.5-mL aliquot was dispensed into tubes containing K2EDTA, with the remaining 5 mL placed into a plain tube. Hemoglobin (cyanmethemoglobin method) and hematocrit (microcentrifugation) values were used to estimate percentage changes in blood, plasma, and red cell volumes relative to the resting sample (9). The 5 mL added to a plain tube was left to clot at room temperature for 60 min before being centrifuged to yield serum. This was divided into two equal aliquots; one was stored at −20°C for the analysis of S100β using a commercially available enzyme-linked immunosorbant assay (Sangtec Medical, Bromma, Sweden), and the other was kept at 4°C for the analysis of serum sodium and potassium concentrations (Corning 410C, New York, NY), chloride concentration (Jenway PCLM 3, Essex, UK), and osmolality (Gonotec Osmomat 030, YSI, Farnborough, UK). The calculated intraassay coefficient of variation (CV) for serum S100β measurements was 7.2%.
Data are presented as means ± standard deviation (SD) unless otherwise stated. To identify differences in normally distributed results, two-way (time by trial) repeated-measures analysis of variance (ANOVA) were employed. Where a significant interaction was apparent, pairwise differences were evaluated using paired t-tests with Holm-Bonferroni adjustment for multiple comparisons. Statistical significance was accepted at P < 0.05. Based on the results of a previous investigation (28), we estimated an 80% probability of detecting a difference in serum S100β concentration of 0.10 μg·L−1 with a sample size of eight subjects.
Not all subjects were able to complete the 90 min of intermittent exercise. Four (of eight) subjects completed the six blocks in the NF trial (mean exercise time, 80.7 ± 13.0 min), with six finishing the protocol when sweat losses were replaced (85.1 ± 9.5 min).
Core temperature data are presented in Figure 1. There was no difference between trials in core temperature before the start of exercise (P = 1.000). Exercise resulted in a progressive increase in core temperature, with a similar response observed during both trials until 30 min. After this point, a slowing in the rate of rise was apparent in the F trial. Core temperature at the end of exercise was 39.2 ± 0.2 and 38.9 ± 0.3°C in the NF and F trials, respectively (P = 0.008). Weighted mean skin temperature was elevated throughout exercise in both trials, but this response was not significantly influenced by fluid ingestion during exercise or recovery (P = 0.423).
Subjects consumed 2.19 ± 0.38 L of water during the F trial. With the ingestion of fluid at regular intervals throughout exercise, at no point did body mass differ by more than 0.43 ± 0.15 kg from the preexercise value. During the NF trial, cumulative body mass loss increased progressively, reaching 2.12 ± 0.42 kg at the end of the recovery period. This represented a 2.8 ± 0.6% loss of body mass, compared with a 0.3 ± 0.1% loss apparent at the end of the fluid trial (P < 0.001). Sweat rate was significantly greater during exercise when fluids were ingested (1.52 ± 0.20 L·h−1) than in the NF trial (1.42 ± 0.17 L·h−1; P = 0.019).
Perceived exertion and thermal stress increased progressively during exercise in both trials (Fig. 2; P < 0.001). The ingestion of fluid at regular intervals during exercise attenuated this increase, with lower RPE and thermal stress reported after 30 min of exercise (RPE, P = 0.002; thermal stress, P = 0.005). There was no difference in resting heart rate before the start of exercise (P = 0.160). Heart rate was similar during the first 50 min of exercise, but the ingestion of fluid resulted in lower values towards the end of exercise and throughout the postexercise recovery period compared with those recorded when no fluid was provided (P = 0.027). Values at the end of exercise were 187 ± 12 bpm in the NF trial and 180 ± 12 bpm when fluid was ingested.
Serum S100β concentrations are shown in Figure 3. There was no difference in serum S100β concentrations before exercise, with values of 0.08 ± 0.02 and 0.09 ± 0.02 μg·L−l observed in the NF and F trials, respectively (P = 0.320). An increase in serum S100β was apparent during the NF trial (P < 0.001), but this change was abolished when fluid was ingested during exercise. At the end of exercise, serum S100β concentrations were 0.20 ± 0.06 μg·L−l during the NF trial and 0.13 ± 0.03 μg·L−l in the F trial (P = 0.046).
Serum electrolyte concentrations and the change in plasma volume are presented in Table 1. There was a marked reduction in plasma volume during the first 15 min of exercise during both trials, with a tendency for plasma volume to continue to fall as exercise continued. This response was not significantly different between trials at any time (P = 0.664). Serum sodium concentrations were similar before exercise (P = 0.549), and no difference was apparent during the first 30 min of exercise, but thereafter a steady fall in sodium concentration was observed when water was ingested. At the end of exercise, serum sodium concentrations were 145 ± 2 mmol·L−l in the NF trial and 139 ± 2 mmol·L−l in the fluid trial (P < 0.001). Exercise resulted in a marked increase in serum potassium levels in both trials (P < 0.001), but fluid ingestion did not influence this response (P = 0.545). There was no difference in serum osmolality at rest (P = 0.156) or during the first 45 min of exercise (Fig. 4). After 60 min of exercise with fluid ingestion, there was a progressive fall in serum osmolality that continued into the recovery period, whereas osmolality continued to rise in the NF trial, reaching a peak of 294 ± 3 mOsm·kg−l at the end of exercise (P = 0.003). During the course of the trial, there was an increase in serum osmolality of 12 ± 3 mOsm·kg−l when no fluid was provided, whereas no change was apparent when sweat losses were replaced (+2 ± 3 mOsm·kg−l). There was a significant association between the change in serum S100β and the change in serum osmolality (Fig. 5: r = 0.662; P = 0.005).
The results of the present study are consistent with previous work demonstrating that exercise in a warm environment results in an increase in the circulating serum S100β concentration (28). The ingestion of water in a volume sufficient to maintain euhydration during exercise in the heat attenuated the increase in serum S100β concentration observed when no fluid was provided. This response occurred despite all subjects exercising for the same duration or longer when water was ingested. If the evidence that the appearance of this protein in the periphery reflects an increase in BBB permeability is accepted (12), these data indicate that the progressive loss of body fluids during exercise in the heat may affect the integrity of the BBB. As the BBB plays a vital role in the regulation of exchange between the CNS and the peripheral circulation, any change in its function is likely to alter brain function and may influence exercise performance. In addition, serum S100β is now being employed as an early indicator of neuronal damage caused by head injuries sustained during sport. A change in barrier function may result in a false-positive in exercising individuals, potentially leading to an inappropriate course of treatment being administered.
Evidence from both human and rodent studies suggests that strenuous exercise can increase BBB permeability. Animals subjected to 30 min of forced swimming displayed a marked breakdown in barrier function, a response that was reversed after a 2-h recovery period (25,26). We have previously shown that, in humans, exercise in a warm environment results in an elevation in serum concentrations of S100β; this response was not apparent when exercise of the same duration and intensity was performed in temperate conditions (28). Unlike this previous work, serum S100β concentration was elevated in all subjects in the present study after exercise, without the replacement of fluids. While there was some degree of intersubject variability in this response (+0.12 ± 0.06 μg·L−l), all but one subject displayed an at least twofold increase in serum concentrations at the end of exercise compared with preexercise values. The ingestion of water during exercise attenuated this increase in serum S100β concentrations at the end of exercise in all subjects.
Potential contributing factors that may be important in changes to BBB permeability during exercise in a warm environment include the development of hyperthermia (29), increased central serotonin release (25,26), hyperammonemia (14), increased circulating epinephrine concentrations (1), and an upregulation in proinflammatory cytokine production (7). The exercise test employed in this study resulted in a marked disturbance to fluid homeostasis, and it is also possible that changes in total body water may affect the BBB, perhaps in combination with some or all of the factors listed above.
Exercise in a warm environment results in a progressive loss of hypotonic fluid, in the form of sweat that is secreted on to the skin surface to dissipate heat through evaporation. The resulting plasma hyperosmolality causes a shift of fluids from interstitial and intracellular spaces to defend blood volume, thus helping to maintain blood flow to the working muscles and skin (18). It is also clear that an elevation in extracellular osmolality can exert a marked influence on brain volume (11). While the transport of nutrients across the BBB is tightly regulated through a number of selective transport systems, the movement of water is governed by osmotic and hydrostatic forces. The volume of the brain is therefore dictated largely by the solute content of the brain tissue relative to that of the extracerebral space, and a change in plasma osmolality can result in a net water flux across the BBB. This has been demonstrated in work conducted by Cserr and colleagues (6,7) that reported a marked loss of water from the brain after peripheral infusions of hypertonic solutions of sodium chloride, mannitol, or sucrose (5). The degree of water loss was not as great as predicted, assuming ideal osmotic behaviour, because of a net influx of electrolytes from the cerebrospinal fluid and plasma (6). This volume regulation occurs rapidly (within 30 min) and acts as a defense mechanism, limiting the degree of brain shrinkage (11).
In the present study, serum osmolality increased by 12 ± 3 mOsm·kg−l during exercise when no fluid was ingested. This change in osmolality was similar to that observed in the previous study investigating changes in serum S100β concentrations during exercise in the heat (12 ± 3 mOsm·kg−l) (28). While this change is relatively small compared with that caused by mannitol infusion, which has been reported to increase serum osmolality to over 300 mOsm·kg−l, the resulting osmotic gradient between the periphery and CNS may have caused a net loss of fluid from the brain to the circulation. The movement of water across the BBB, in addition to the shrinkage of barrier endothelial cells, will produce a transient opening of the tight junctions and a temporary loss of barrier integrity. This hypothesis is supported by the significant increase in serum S100β concentrations at the end of exercise when no fluid was provided. This magnitude of change was similar to that reported by Kapural and colleagues (12) when examining changes in serum S100β after mannitol infusion. The ingestion of water limited the progressive rise in serum osmolality apparent during exercise in the NF trial, consequently attenuating a potential osmotically driven shift of fluid from the brain and preserving BBB integrity. Hyperosmolality, induced through the peripheral infusion of mannitol, is commonly employed in a clinical setting to produce a marked but transient opening of the BBB, thus enabling the delivery of therapeutic agents into the CNS that would otherwise not cross the intact BBB (23).
At present, there are limited data on the effects of hypohydration on the CNS. Recent work has employed MRI to determine changes in brain and ventricular volume after dehydration induced by exercise in the heat (8). A marked change in brain volume was reported after exercise, with the magnitude and direction of change related to the degree of hypohydration induced. However, these data are limited by the methodologies employed, potentially confounding the results of this work. The MRI procedure took around 20 min to complete, but the rapid movement of osmolytes from the CSF and serum into the brain serves to defend brain volume and limit fluid losses. Water loss and electrolyte uptake are likely to occur almost simultaneously, and the magnitude of brain water losses has been demonstrated to be only 35% of that expected 30 min after osmotic opening of the BBB (6). Animal studies investigating the effect of dehydration by 10% of initial body mass also report no change in brain volume, but these measurements were taken once core temperature had returned to basal levels (19). Again, it is probable that there was a redistribution of fluids during this time to restore brain volume. Both of these studies are seriously hampered by the time lapsed between the end of exercise and the measurement of brain volume, making it difficult to accept their findings. The operation of a regulatory mechanism to defend brain volume is supported by the rapid return of serum S100β concentrations to preexercise levels 15 min after the cessation of exercise.
In line with previous reports, fluid ingestion attenuated the rise in core temperature, heart rate, and ratings of perceived exertion and thermal stress during the latter stages of exercise (16). Core temperature was 0.3 ± 0.2°C lower at the end of exercise when sweat losses had been replaced. Hyperthermia has been implicated in changes in BBB permeability, with whole-body hyperthermia demonstrated to impair BBB function in free-living rats (29). No such association during exercise was apparent in the previous work conducted by Watson et al. (28), but it is unlikely that a difference of the magnitude observed in the present study would have significantly altered this response. The effect of fluid ingestion on brain temperature is not clear at present.
The protocol employed in the present investigation resulted in the loss of approximately 2.8 ± 0.6% of the subjects' initial body mass during the NF trial (2.12 ± 0.42 kg). This level of hypohydration is sufficient to reduce exercise capacity, particularly when exercise is performed in warm ambient conditions (16). There is relatively little work to investigate the effects of exercise-induced dehydration on the brain, but evidence suggest that alertness, concentration, and performance of cognitive tasks are reduced at relatively mild levels of dehydration (15). These findings suggest a disturbance in brain function with dehydration, possibly related to changes in brain volume, although the underlying neurobiological mechanism for this response is not clear at present.
An elevation in serum S100β concentration, of a similar magnitude to that reported in the present study, has also been observed after soccer heading drills (17), bouts of boxing (20), and prolonged running (20), but the possibility that this response occurred through a change to BBB function was not explored. For a number of years, appearance of this protein in the peripheral circulation has been employed as an index of brain damage, but serum concentrations of up to 0.34 μg·L−1 are reported in the absence of neural damage and have been proposed to reflect changes in BBB permeability (12). While the beta subunit of the S100 protein family is highly specific to the CNS, it has been identified, albeit in minute quantities, in some peripheral tissues including bone, muscle, heart, and adipose tissue (2,21). Recent work suggests that S100β is liberated into the circulation from these tissues only when subjected to significant trauma, such as that experienced during surgery (10). Exercise-induced increases in serum S100β have been attributed to muscle damage, but the nature of the exercise task used in this study is unlikely to have resulted in significant loss of membrane integrity in the working muscles (13). Furthermore, fluid ingestion significantly attenuated the increase in serum S100β concentrations, but it seems improbable that this practice would influence the degree of muscle damage incurred during exercise.
As S100 proteins are cleared from the peripheral circulation by the kidneys, with a serum half-life of around 2 h, it is possible that the change observed after exercise without fluid ingestion may have been caused by a reduction in renal S100β clearance. Prolonged exercise, particularly under conditions of heat stress, is associated with a reduction in renal blood flow (24), but there is evidence that extraction of low-molecular weight proteins remains high during strenuous exercise (4), suggesting that the clearance of S100β may not be altered under these conditions and that this change does reflect the influx of S100β from the CNS into the periphery. Because the differences in serum S100β concentration observed at the end of exercise were not accompanied by any significant change in plasma volume, it also seems that volemia did not influence this response.
In conclusion, prolonged intermittent exercise in a warm environment resulted in a marked increase in serum S100β concentrations, suggesting that BBB permeability may be increased in these conditions. This response was abolished by the ingestion of water in a volume sufficient to maintain euhydration during exercise. A number of factors can influence BBB permeability during exercise, but the present data suggest that exercise-induced hyperosmolality may be important. While there was a progressive increase in serum osmolality when no water was consumed, ingestion of plain water maintained serum osmolality at preexercise levels, thus potentially limiting any osmotically driven movement of fluid across the BBB.
The authors would like to acknowledge the support of the Gatorade Sport Science Institute (GSSI), Barrington, IL. RJM is a member of the GSSI advisory board, but the results of the present study do not constitute an endorsement of any product.
1. Abdul-Rahman, A., N. Dahlgren, B. B. Johansson, and B. K. Siesjo. Increase in local cerebral blood flow induced by circulating adrenaline: involvement of blood-brain barrier dysfunction. Acta Physiol. Scand.
2. Anderson, R. E., L. O. Hansson, O. Nilsson, J. Liska, G. Settergren, and J. Vaage. Increase in serum S100A1-B and S100BB during cardiac surgery arises from extracerebral sources. Ann. Thorac. Surg.
3. Borg, G. A. Psychophysical bases of perceived exertion. Med. Sci. Sports Exerc.
4. Castenfors, J., F. Mossfeldt, and M. Piscator. Effect of prolonged heavy exercise on renal function and urinary protein excretion. Acta Physiol. Scand.
5. Cserr, H. F., M. DePasquale, and C. S. Patlak. Regulation of brain water and electrolytes during acute hyperosmolality
in rats. Am. J. Physiol.
6. Cserr, H. F., M. DePasquale, and C. S. Patlak. Volume regulatory influx of electrolytes from plasma to brain during acute hyperosmolality
. Am. J. Physiol.
7. de Boer, A. G., and D. D. Breimer. Cytokines and blood-brain barrier permeability. In: Progress in Brain Research: Brain Function in Hot Environment
, H. Sharma and J. Westman (Eds.). Amsterdam: Elsevier Science, pp. 425-451, 1998.
8. Dickson, J. M., H. M. Weavers, N. Mitchell, et al. The effects of dehydration on brain volume - preliminary results. Int. J. Sports Med.
9. Dill, D. B., and D. L. Costill. Calculation of percentage changes in volumes of blood, plasma, and red cells in dehydration. J. Appl. Physiol.
10. Fazio, V., S. K. Bhudia, N. Marchi, B. Aumayr, and D. Janigro. Peripheral detection of S100beta during cardiothoracic surgery: what are we really measuring? Ann. Thorac. Surg.
11. Gullans, S. R., and J. G. Verbalis. Control of brain volume during hyperosmolar and hypoosmolar conditions. Annu. Rev. Med.
12. Kapural, M., L. Krizanac-Bengez, G. Barnett, et al. Serum S-100beta as a possible marker of blood-brain barrier disruption. Brain Res.
13. Koller, A., J. Mair, W. Schobersberger, et al. Effects of prolonged strenuous endurance exercise on plasma myosin heavy chain fragments and other muscular proteins. Cycling vs running. J. Sports Med. Phys. Fitness
14. Mans, A. M., J. F. Biebuyck, and R. A. Hawkins. Ammonia selectively stimulates neutral amino acid transport across blood-brain barrier. Am. J. Physiol.
15. Maughan, R. J. Impact of mild dehydration on wellness and on exercise performance. Eur. J. Clin. Nutr.
57(Suppl 2):S19-S23, 2003.
16. Montain, S. J., and E. F. Coyle. Influence of graded dehydration on hyperthermia
and cardiovascular drift during exercise. J. Appl. Physiol.
17. Mussack, T., J. Dvorak, T. Graf-Baumann, and M. Jochum. Serum S-100B protein levels in young amateur soccer players after controlled heading and normal exercise. Eur. J. Med. Res.
18. Nose, H., G. W. Mack, X. R. Shi, and E. R. Nadel. Shift in body fluid compartments after dehydration in humans. J. Appl. Physiol.
19. Nose, H., T. Morimoto, and K. Ogura. Distribution of water losses among fluid compartments of tissues under thermal dehydration in the rat. Jpn. J. Physiol.
20. Otto, M., S. Holthusen, E. Bahn, et al. Boxing and running lead to a rise in serum levels of S-100B protein. Int. J. Sports Med.
21. Pelinka, L. E., L. Szalay, M. Jafarmadar, R. Schmidhammer, H. Redl, and S. Bahrami. Circulating S100B is increased after bilateral femur fracture without brain injury in the rat. Br. J. Anaesth.
22. Ramanathan, L. M. A new weighting system for mean surface temperature of the human body. J. Appl. Physiol.
23. Rapoport, S. I. Osmotic opening of the blood-brain barrier: principles, mechanism, and therapeutic applications. Cell Mol. Neurobiol.
24. Rowell, L. B. Human cardiovascular adjustments to exercise and thermal stress. Physiol. Rev.
25. Sharma, H. S., J. Cervos-Navarro, and P. K. Dey. Increased blood-brain barrier permeability following acute short-term swimming exercise in conscious normotensive young rats. Neurosci. Res.
26. Sharma, H. S., J. Westman, J. C. Navarro, P. K. Dey, and F. Nyberg. Probable involvement of serotonin in the increased permeability of the blood-brain barrier by forced swimming. An experimental study using Evans blue and 131I-sodium tracers in the rat. Behav. Brain Res.
27. Shivers, R. R., and J. A. Wijsman. Blood-brain barrier permeability during hyperthermia
. In: Progress in Brain Research: Brain Function in Hot Environment
, S. Sharma and J. Westman (Eds.). Amsterdam: Elsevier Science, pp. 413-424, 1998.
28. Watson, P., S. M. Shirreffs, and R. J. Maughan. Blood-brain barrier integrity may be threatened by exercise in a warm environment. Am. J. Physiol. Regul. Integr. Comp. Physiol.
29. Wijsman, J. A., and R. R. Shivers. Heat stress affects blood-brain barrier permeability to horseradish peroxidase in mice. Acta Neuropathol. (Berl.)