Application of carbon dioxide to the skin and muscle oxygenation of human lower-limb muscle sites during cold water immersion

Subjects’ characteristics

We recruited 15 healthy young male subjects (age 23.1 ± 0.4 yrs, weight 61.4 ± 1.8 kg, height 171 ± 1.4 cm, mean ± SE) for the study. The subjects’ body mass index (BMI) values and body fat percentage were 20.8 ± 0.5 and 15.7 ± 1.1%, respectively. Written informed consent was obtained from all subjects after they received a detailed explanation about all procedures, the purpose of the study, and the possible risks and benefits of their participation. This study conformed to the Declaration of Helsinki, and the Ethics Committee of Doshisha University approved the purpose and all procedures of the study (no. 15085). None of the subjects had cardiovascular abnormalities or skin lesions.

Water immersion protocol

All of the experiments were performed at the same time of the day in a climatic chamber with its ambient temperature and relative humidity maintained at 25.0 ± 0.5 °C and 55 ± 3%, respectively. After entering the climatic chamber, the subject assumed a sitting position on a chair for 10 min (rest stage) and then immersed his right lower leg up to the knee in a 47-L water container for 15 min in CO₂-rich water (1150 ppm, CCWI) or tap water (CWI) maintained at 20 °C (Fig. 1). After the 15-min immersion, the subject withdrew his lower leg from the CCWI or CWI and then rested for a 20-min recovery period. The subjects performed both a CWI and a CCWI and were not informed whether the water was CCWI or CWI in order to ensure a double-blind design.

The CO₂-rich water was prepared by dissolving CO₂ in tap water using a dual-chamber/dual-vortex high speed rotation system (DDHRS; type S1, Taikougiken, Kumamoto, Japan). Water was pumped into the DDHRS. Gas was injected into the DDHRS and was reduced in the DDHRS by the centrifugation effect. In this study, the flow rates of CO₂ and water were 0.3 L/min and 20 L/min, respectively. This DDHRS can produce a microbubble water flow in which CO₂ is dissolved very quickly (within 10 min) by a connection with a small CO₂ cartridge (74 g), and the temperature was maintained at 20 °C by an isothermal cooling system (BB310, Yamato, Shiga, Japan).

Measurements and data analysis

As the subjects’ core temperature (T_core), the sublingual temperature was measured with an electric thermometer (MC-652LC, Omron, Kyoto, Japan). Two skin temperatures were measured at the lower leg (i.e., the tibialis anterior [TA] muscle) under water immersion (T_sk-WI) and at the subject’s non-immersed control leg (T_sk-cont) with commercial thermistors (LT-8, Gram Co., Saitama, Japan). The thermistors for the lower legs were positioned at the point at one-third of the length of the lateral epicondyle of the femur and the malleolus lateralis line from the lateral epicondyle of the femur and the same length horizontally back.

At the same time as the temperature measurements, the subject’s systolic and diastolic blood pressure (SBP, DBP) and heart rate (HR) were recorded by an automated tonometer technique (HBP-1300, Omron) using the subject’s upper right arm which rested at heart level alongside his body. The mean arterial pressure (MAP) was automatically calculated with the formula MAP = DBP + $\frac{1}{3}$ (SBP − DBP).

The subject’s thermal sensation was reported by the subject every 5 min after the start of the pre-immersion and throughout the experiment. The thermal sensations were rated by the subject based on a conventional seven-point scale (Fanger, 1984) comprising cold (−3), cool (−2), slightly cool (−1), neutral (0), slightly warm (1), warm (2), and hot (3). In this study, we limited the thermal sensations to a narrow scale: below −3 to 0.

We measured each subject’s skin adipose tissue thickness (ATT) on two regions of the TA and gastrocnemius (GAS) muscles by using an ultrasonic device (SSD-3500SV, Hitachi-Aloka Medical, Tokyo), because the transmissive pulsed light in the NIRS device is influenced by the ATT when the ATT is >5 mm in both of the two muscle regions (Binzoni et al., 1998).

NIRS provides a noninvasive window into the microvascular and intramuscular oxygenation status of skeletal muscle under a wide range of conditions at rest (Adami et al., 2015; Binzoni et al., 2000) and during exercise (Koga et al., 2013). In healthy and clinical populations, the total heme concentration (total [Hb+Mb]) and the tissue O₂ saturation (StO₂) can be estimated (Yamabata et al., 2016). In the present study, the oxygenation (oxy[Hb+Mb]) and deoxygenation (deoxy[Hb+Mb]) concentration profiles and their sum (i.e., the total[Hb+Mb] concentration profile) on the two regions of the TA and GAS muscles were simultaneously recorded by two customized continuous-wave NIRS devices (BOM-L1 TR, Omega Wave Co., Tokyo). This system was able to continuously monitor the changes in the oxy[Hb+Mb], deoxy[Hb+Mb], and total[Hb+Mb] concentrations at two different sites (the skin and muscle). Data were collected as previously described in Yamabata et al. (2016). Specifically, we calculated from the light attenuation changes by using a modification of the Beer-Lambert law.

In this study, we defined the 7.5-mm layer as ‘superficial’ and the 15-mm layer as ‘muscle’. The TA and GAS muscles with the attached probe holder were then wrapped in a dark-colored elastic bandage to further secure the probes and to eliminate ambient light that might contaminate the NIRS signals. The NIRS data reflected the relative concentration changes in the hemoglobin chromophores and thus did not reflect the absolute tissue O₂ values. As we set the probe gain setting at 0 prior to testing the subjects at rest in the sitting position, the relative changes from the resting (Δoxy[Hb+Mb], Δdeoxy[Hb+Mb], and Δtotal[Hb+Mb]) values were calculated. With the use of two detectors at the lower penetration depth of 7.5 mm, the superficial Δoxy[Hb+Mb], Δdeoxy[Hb+Mb], and Δtotal[Hb+Mb] values were assumed to reflect the cutaneous blood flow (i.e., Δoxy[Hb+Mb]_superficial, Δdeoxy[Hb+Mb]_superficial, and Δtotal[Hb+Mb]_superficial).

We also measured the data from the deep penetration depth of 15 mm (i.e., the muscle Δoxy[Hb+Mb], Δdeoxy[Hb+Mb], and Δtotal[Hb+Mb]) values. The Δoxy[Hb+Mb]_muscle, Δdeoxy[Hb+Mb]_muscle, and Δtotal[Hb+Mb]_muscle were obtained by subtracting these superficial variables from the muscle variables (Δoxy[Hb+Mb]_muscle), Δdeoxy[Hb+Mb]_muscle, and Δtotal[Hb+Mb]_muscle according to the modified Beer-Lambert law (Ando et al., 2013; Horiuchi et al., 2016; Yamabata et al., 2016).

From these data, we calculated the S_tO₂ as oxy[Hb+Mb]/total[Hb+Mb] × 100% in the superficial and muscle regions. The NIRS-S_tO₂ measurements reflect a mixture of arteriole, capillary, and venous blood flows and do not separate venous from arterial saturations (Culver et al., 2003). NIRS-derived S_tO₂ values closely correspond to the actual microvascular oxygenation in skeletal muscle (S_mO₂) (Koga et al., 2013; Sperandio et al., 2009; Wüst et al., 2014).

We calculated the averaged data of each NIRS parameter for 30 s at each 3-min interval throughout the experiment: water immersion (WI)3 through WI15 at 3, 6, 12, and 15 min, and recovery (rec)3 through rec18 at 3, 6, 12, 15, and 18 min. The baseline values were averaged for the final 30 s of the subject’s 10-min rest in a sitting position. The data of the cardiovascular indexes were calculated at each 5-min interval, i.e., baseline; 10-min sitting position; the rest periods, WI for 15 min (WI5, WI10, and W15), and recovery for 20 min (rec5, rec10, rec15, and rec20), averaging the values for 30 s at each period.

Statistical analysis

All values are presented as the mean ± standard error (SE). The significance of differences in each variable (Δoxy[Hb+Mb], Δdeoxy[Hb+Mb], and Δtotal[Hb+Mb]) was determined by a two-way analysis of variance (ANOVA) comparing water conditions (CCWI and CWI) × time periods (rest, WI3–WI15, and recovery at rec3–rec18). A post hoc comparison was applied by Bonferroni test for the appropriate data sets when a significant F-value was obtained. A partial eta-square (η²) was also determined. Another post hoc comparison was applied by Dunnett’s test to determine multiple comparisons from baseline. All analyses were performed using SPSS software (Abacus Concepts, Berkeley, CA) with significance in all cases set at the 5% level.

Regarding the sample determination using PS Power and Sample Size Calculations (Dupont & Plummer, 1998), we are planning a study of a continuous response variable from matched pairs of 15 healthy young subjects. Prior data of ΔStO₂ of TA and GAS muscles indicate that the difference in the response of matched pairs is normally distributed with the standard deviation 3.486. If the true difference in the mean response of matched pairs is 3.9, we will need to study 8 pairs of subjects to be able to reject the null hypothesis that this response difference is zero with the probability (power) 0.8. The Type I error probability associated with this test of the null hypothesis is 0.05. Therefore, judging from the criteria of the power 0.8, we recognized that the sample size of our 15 pairs of subjects provides the power 0.941.

The ATT findings

The average ATT value in the TA region was 3.0 ± 0.2 mm, and that in the GAS region was 3.9 ± 0.3 mm.

Cardiovascular responses during water immersion

The 15 subjects’ MAP values remained unchanged during the immersion and recovery periods, with no significant difference between CCWI and CWI (interaction effect: F(9,126) = 0.892, p = 0.534) (Fig. 2A). As shown in Fig. 2B, the HR values during the CCWI remained constant throughout the experiment (74 ± 2.2 beats min⁻¹ at baseline and WI15; 73 ± 2.3 beats min⁻¹ at rec20) (interaction effect: F(9,126) = 0.446, p = 0.908), which supported the idea that the diving reflex in HR could not be induced with only one leg immersed in the water. These results indicate that both the cold CO₂-rich water and cold tap water immersions themselves did not influence the subjects’ cardiovascular responses during one-leg water immersion, irrespective of the CO₂ condition.

Core temperature and skin temperatures in the WI and non-WI legs during CO₂-rich immersion and recovery

The sublingual temperature (T_core) values during the CCWI remained constant at a narrow range between 36.4° and 36.6 °C and were not significantly lower than those during the CWI (Fig. 2C) (interaction effect: F(9,126) = 1.131, p = 0.346). The T_sk-WI values decreased abruptly during each water immersion (CCWI: 21.8 ± 0.1 °C, CWI: 21.8 ± 0.2 °C, at WI15) and returned at the end of the recovery, i.e., rec18 (CCWI: 25.8 ± 0.2 °C, CWI: 26.4 ± 0.2 °C) (Fig. 3A).

Compared to the baseline, there was significant lower T_sk-WI during the CCWI and CWI until the end of recovery (time effect: F(1,154) = 446.973, p < 0.001, η² = 0.970). Compared to the CWI, the significantly lower T_sk-WI values continued from rec9 to rec18 of the recovery period following the CCWI (interaction effect: F(11,154) = 1.901, p = 0.048, η² = 0.160). In addition, the T_sk-cont values gradually and significantly decreased from baseline to rec18 and were significantly lower from WI9 during the CCWI and from WI15 during the CWI to rec18 (time effect: F(1,154) = 19.445, p < 0.001, η² = 0.599, Fig. 3B).

It is very noteworthy that the control leg (T_sk-cont) values tended to be lower during the CCWI than during the CWI until the end of recovery, corresponding to the reduction in T_sk-WI. In addition, the thermal sensation was significantly decreased during the immersion and began increasing again during the recovery period (Fig. 2D) (time effect: F(1,126) = 8.985, p < 0.001, η² = 0.391), but it was slightly lower at a later recovery point in the CCWI compared to the CWI (condition effect: F(1,14) = 0.440, p = 0.518).

TA muscle

The Δoxy[Hb+Mb]_superficial increased significantly from baseline (by approx. 0.2 AU) during the CCWI (periods WI3 to rec3, p < 0.01) and the CWI (periods WI9 to WI12, p < 0.01), and this response returned to baseline following the water immersion (time effect: F(11, 143) = 14.076, p < 0.001, η² = 0.573, Fig. 4A). The Δdeoxy[Hb+Mb]_superficial during the CCWI remained constant until the end of the water immersion, and it was significantly increased at rec15 and rec18 from baseline due to vasodilation in venous veins during recovery (time effect: F(11, 143) = 5.087, p < 0.001, η² = 0.298, Fig. 4B). The Δtotal[Hb+Mb]_superficial increased by approx. 0.2 AU during water immersion and stayed higher during the recovery following immersion (time effect: F(11, 143) = 1.763, p = 0.066, Fig. 4C). Overall, the increase in Δtotal[Hb+Mb]_superficial was a result of the Δoxy[Hb+Mb]_superficial increase during water immersion, and during recovery this response was a result of the Δdeoxy[Hb+Mb]_superficial increase.

The Δoxy[Hb+Mb]_muscle increased gradually from baseline and stayed at a higher level during the recovery period with a moderate ES at all points between CCWI and CWI (ES: 0.67–0.80, interaction effect: F(11, 143) = 1.813, r = 0.060, Fig. 4D). By contrast, the Δdeoxy[Hb+Mb]_muscle was unaltered from baseline and increased slightly after the CCWI (Fig. 4E). The Δtotal[Hb+Mb]_muscle consequently and gradually increased until the end of recovery without a significant increase from baseline, and this trend contributed mostly to the Δoxy[Hb+Mb]_muscle response with a significant difference at some time points during the CCWI compared to the CWI (interaction effect: F(11, 143) = 1.983, p = 0.044, η²=0.161, Fig. 4F).

GAS muscle

In the CCWI protocol, the Δoxy[Hb+Mb]_superficial was increased from baseline and gradually decreased to below the baseline after immersion (Fig. 5A). Notably, the Δoxy[Hb+Mb]_superficial was significantly elevated from baseline (time effect: F(11, 143) = 9.258, p < 0.001, η² = 0.416). By contrast, the Δdeoxy[Hb+Mb]_superficial during the water immersion abruptly decreased to a significantly lower level from baseline in both the CCWI and CWI (time effect: F(11, 143) = 24.118, p < 0.001, η² = 0.650) and increased again during recovery (interaction effect: F(11,143) = 2.107, p = 0.023, η² = 0.139, Fig. 5B). During the CCWI, the Δtotal[Hb+Mb]_superficial tended to be lower than the baseline values, but this decrease was sustained below baseline until the end of recovery (Fig. 5C).

At the GAS muscle region, the Δoxy[Hb+Mb]_muscle was constant throughout the CCWI. After the CWI, Δoxy[Hb+Mb]_muscle gradually decreased until the end of recovery (time effect: F(11, 143) = 6.602, p < 0.001, η² = 0.355, Fig. 5D). The Δdeoxy[Hb+Mb]_muscle was gradually augmented until the end of recovery (time effect: F(11, 143) = 8.875, p < 0.001, η² = 0.425, Fig. 4E), which reflected mostly the Δtotal[Hb+Mb]_muscle response (Fig. 5F). By contrast, the Δtotal[Hb+Mb]_muscle showed the following: a decrease of Δoxy[Hb+Mb]_muscle and an increase of Δdeoxy[Hb+Mb]_muscle that offset each other.

Alternations in O₂ saturation inmuscle as measured by NIRS

The StO₂ values gradually increased during the water immersion and then returned to a decrease during the recovery period (Supplemental Information). At the 20-min recovery, the StO₂ values had fallen to below the baseline. Since the baseline values of StO₂ differed between the CCWI and CWI, we estimated the change in the StO₂ from baseline (ΔStO₂) to compare the values between the CCWI and CWI and between the TA and GAS muscles (Fig. 6). For the TA muscle, a significant increase from baseline in ΔStO_2muscle reached 7.5 ± 1.1% at WI9 in the CCWI (time effect: F(11,143) = 17.025, p < 0.001, η²=0.654), which was very similar to the ΔStO_2superfical value 8.0 ± 0.9%. It is therefore likely that the CO₂ diffusion arrives at deep muscle tissue in the TA. By contrast, for the GAS muscle, the ΔStO_2muscle was increased at 3.6 ± 1.5% at WI9, which is approx. one-half compared to the ΔStO_2superfical of 7.5 ± 0.9% at WI9; this suggests that it might be difficult for CO₂ to diffuse into the GAS. Apparently, CO₂ induced a greater ΔStO₂ in both superficial and muscle regions in both the TA and GAS (interaction effect: p = 0.044–0.002, η² = 0.61–0.251).

CO₂ promotes more the heat transfer from the body to the water and lower skin temperature during CCWI than CWI

Our analyses revealed that compared to the CWI, the CCWI induced a slightly lower T_sk-WI and a slightly higher thermal sensation at the recovery period, which indicates that CCWI facilitates not only a decline in T_sk-WI but also a less cold sensation following water immersion. These observations confirm that an increase in cutaneous blood flow during CCWI may elicit excess heat transfer even if the water temperature is cool. CO₂ applied topically can increase the cutaneous blood flow (Diji, 1959; Komoto et al., 1986; Ito, Moore & Koss, 1989; Hartmann, Bassenge & Pittler, 1997) due to cutaneous vasodilation (Ito, Moore & Koss, 1989). In the present investigation, this effect promoted not only the heat transfer from the body to the water and lower T_sk-WI values; it also produced lower CO₂-induced T_sk-cont values, which is one of the strong countermeasures of a hot environment in sports. Moreover, compared to the CWI, the CCWI facilitated the slightly lower cold sensation. It has been demonstrated that CO₂ inhibits the activity of cold receptors and facilitates the activity of warm receptors of the skin (Dodt, 1956). Such modifications in the activity of skin receptors by CO₂ can explain the elevation of the thermal sensations that we observed during the CCWI compared to the CWI, since thermal sensations are caused predominantly by the signals from skin receptors rather than central receptors (Hensel, 1981).

We set the water temperature at 20 °C because cold receptors are discharged most vigorously at skin temperatures at 25 °C (Tansey & Johnson, 2015) and cold immersion even at 20 °C in this study would have excited a greater number of stimulated cold receptors. In a review of CWI studies, a water temperature at ∼14 °C was most commonly selected with a ≤15-min immersion duration (Choo et al., 2018a). We selected 20 °C with a longer immersion duration (15 min) for the present experiment, and the T_sk-WI was reduced from 31.8 ± 0.2 °C to 21.8 ± 0.1 °C at the end of the immersion, indicating that the underlying muscles were probably cooled to a significant degree and to the same extent between CCWI and CWI (Ihsan et al., 2013).

We expected that in the CCWI, increased cutaneous blood flow due to cutaneous vasodilation via CO₂ would facilitate the heat transfer from the subject’s body to the water. Indeed, an increased Δoxy[Hb+Mb]_superficial was observed even during the CCWI. Consequently, the Δtotal[Hb+Mb]_superficial at the TA muscle tended to be higher than the baseline (time effect: p = 0.066), and fell significantly from the baseline in the GAS muscle (time effect: p = 0.027). Our NIRS results thus support our hypothesis that CCWI depresses the marked reduction in the cutaneous blood flow.

The impact of hydrostatic pressure during water immersion leads to inflow into muscle

Hydrostatic pressure might reduce edema and inflammation by increasing the pressure gradient between the interstitial and intravascular spaces, promoting the re-absorption of interstitial fluid in a manner similar to compression stockings (Partsch, Winiger & Lun, 2004). The combination of a cold stimulus and hydrostatic pressure could also act synergistically; decreased muscle temperature may reduce edema by reducing the muscle perfusion (cold-induced vasocontraction) and fluid diffusion into the interstitial space (Yanagisawa & Fukubayashi, 2010; Gregson et al., 2011), as well as through a reduced permeability of cellular, lymphatic, and capillary vessels (Coté et al., 1988)). This might complement any hydrostatic pressure effects on interstitial–intravascular fluid movement.

In the present study, the Δoxy[Hb+Mb]_muscle gradually increased from baseline and stayed at a relatively higher level during recovery. Overall, the increase in Δtotal[Hb+Mb]_muscle in the TA muscle was a result of the Δoxy[Hb+Mb]_muscle increase during the CCWI, and this response was a result of both the Δoxy[Hb+Mb]_muscle and Δdeoxy[Hb+Mb]_muscle during the recovery period. Choo et al. (2018b) reported that the skin blood flow had notably less influence on the total[Hb+Mb] signal during the recovery after cooling compared to a non-cooling recovery. However, our findings regarding the Δtotal[Hb+Mb]_muscle kinetics differed, and they indicate that CO₂ could influence the muscle blood flow increase.

By contrast, in the GAS muscle in the CCWI, the Δoxy[Hb+Mb]_muscle remained unchanged throughout the experiment. The gradually augmented Δdeoxy[Hb+Mb]_muscle values were mostly reflected by the Δtotal[Hb+Mb]_muscle response. The results of several intervention studies suggest that the deoxy[Hb+Mb]_muscle profile is likely to have different effects on the O₂ extraction-related changes and volume-related changes (Adami et al., 2015; Binzoni et al., 2000; Truijen et al., 2012). Thus, CCWI intervention without shivering would improve the redistribution of blood flow from the non-immersion site to the immersion site. If the 20 °C cold exposure reduces the metabolic demand, an abrupt decline in the Δdeoxy[Hb+Mb]_muscle from baseline would occur; however, the above result indicated no decline of the Δdeoxy[Hb+Mb]_muscle in the TA or GAS muscles.

Different oxygenation in the TA and GAS regions during water immersion

We observed significant increases in the ΔStO₂ in both the TA and GAS muscles during the CCWI (time effect: p < 0.001). Apparently, CO₂ induced a greater ΔStO₂ in both superficial and muscle regions in both the TA and GAS. However, for the TA muscle, there was a similar increase from the baseline in the ΔStO₂ between the skin layer and muscle layer during the CCWI. By contrast, for the GAS muscle, the ΔStO_2muscle was increased by one-half compared to the ΔStO_2superfical, suggesting that it might be difficult for CO₂ to diffuse into the GAS. It may be possible that specific characteristics of TA muscle account for muscle fiber-type-dependent muscle oxygenation responses to a combined cold and CO₂ stimulus, because >70% of the TA muscle in humans consists of only slow-twitch fibers (Dahmane et al., 2005). Slow-twitch fibers have more developed muscle capillaries than fast-twitch fibers, which means that the vasodilatory effect of CO₂ is further improved.

Underlying mechanism of vasodilation during CO₂ water immersion

A plausible mechanism of CO₂-induced vasodilation is associated with extracellular acidosis. Traditional studies demonstrated that acidosis might reduce the contractility of the vascular smooth muscle, leading to vasodilation (Tobian, Martin & Eilers, 1959; Vanhoutte & Clement, 1968). The reduction in smooth muscle contractility has been ascribed to a reduction in calcium influx or to the suppression of myofilament contractility (Breemen et al., 1972; Fabiato & Fabiato, 1978). An in vitro study examining the contractility of the rat aorta exposed to a small change in pH (from 7.4 to 7.0) demonstrated that even this small change in pH could reduce vascular smooth muscle contractility (Loutzenhiser et al., 1990). That study also indicated that H⁺-induced vasodilation is associated with an increase in the amount of calcium sequestered in the norepinephrine-sensitive intracellular calcium store.

Other investigations of the coronary, cerebral and aortic circulations have shown that nitric oxide (Fukuda et al., 1990; Gurevicius et al., 1995; Aalkjær & Peng, 1997) and the activation of potassium channels (Ishizaka & Kuo, 1996) may contribute to this acidosis-induced vasodilation. Both skin and muscle arterioles are known to be regulated by sympathetic outflow and vasodilatory substances such as nitric oxide (Hickner et al., 1997; Kellogg Jr, 2006; Bernjak et al., 2012). However, we did not measure the blood substances of pH, potassium, or nitric oxide or their underlying mechanisms in this study.

Study limitations

To avoid the shivering with which an involuntary muscle contraction of immersed muscles can increase metabolic demand (Proulx, Ducharme & Kenny, 2003) and the associated alteration of NIRS profiles, we set the water temperature of CWI and CCWI at 20 °C, a temperature at which subjects can easily tolerate the cooling muscle treatment. Although we did not directly measure our subjects’ shivering response, there is only a slight possibility that shivering occurred in light of the subjects’ thermal sensations.

Compared to the more global information provided by blood flow in a large supplying artery, the use of NIRS as a monitoring technique provides the ability to assess the specific effects on the tissue of interest. NIRS cannot discriminate between myoglobin (Mb) and hemoglobin (Hb) saturation or between arterial and venous compartments; thus, no unambiguous explanation can be provided. Irrespective of the underlying reasons, the increase in tissue oxygenation that we observed herein was a very consistent feature of the hemodynamic response to the compression of resting muscle.

Athletic application of CO₂-rich water immersion

Cold-water immersion is one of the beneficial recovery techniques that is commonly used by athletes post-exercise to promote the restoration of body systems to baseline conditions and return the physiological system to a pre-exercise state (Bleakley & Davison, 2010). Our present experiment was conducted with the subjects in a resting condition without a post-exercise state, and thus the initial muscle blood flow at the onset of water immersion was quite different from that in the above-cited study of athletes. We thus conducted a second experiment to explore whether this CCWI protocol at post-exercise is an effective intervention to maintain exercise performance, which is associated with a reduced blood lactate concentration and reduced heart rate (Supplemental File). Further research regarding the changes in oxygenation profiles following athletic exercise is of interest.