Hypoxia often causes body water deficits (hypohydration, HYPO); however, the effects of HYPO on aerobic exercise performance and prevalence of acute mountain sickness (AMS) at high altitude (ALT) have not been reported. We hypothesized that 1) HYPO and ALT would each degrade aerobic performance relative to sea level (SL)-euhydrated (EUH) conditions, and combining HYPO and ALT would further degrade performance more than one stressor alone; and 2) HYPO would increase the prevalence and severity of AMS symptoms. Seven lowlander men (25 ± 7 yr old; 82 ± 11 kg; mean ± SD) completed four separate experimental trials. Trials were 1) SL-EUH, 2) SL-HYPO, 3) ALT-EUH, and 4) ALT-HYPO. In HYPO, subjects were dehydrated by 4% of body mass. Subjects maintained hydration status overnight and the following morning entered a hypobaric chamber (at SL or 3,048 m, 27°C) where they completed 30 min of submaximal exercise immediately followed by a 30-min performance time trial (TT). AMS was measured with the Environmental Symptoms Questionnaire-Cerebral Score (AMS-C) and the Lake Louise Scoring System (LLS). The percent change in TT performance, relative to SL-EUH, was −19 ± 12% (334 ± 64 to 278 ± 87 kJ), −11 ± 10% (334 ± 64 to 293 ± 33 kJ), and −34 ± 22% (334 ± 64 to 227 ± 95 kJ), for SL-HYPO, ALT-EUH, and ALT-HYPO, respectively. AMS symptom prevalence was 2/7 subjects at ALT-EUH for AMS-C and LLS and 5/7 and 4/7 at ALT-HYPO for AMS-C and LLS, respectively. The AMS-C symptom severity score (AMS-C score) tended to increase from ALT-EUH to ALT-HYPO but was not significant (P = 0.07). In conclusion, hypohydration at 3,048 m 1) degrades aerobic performance in an additive manner with that induced by ALT; and 2) did not appear to increase the prevalence/severity of AMS symptoms.
- time trial
during acute altitude exposure (<24 h), a decline in arterial oxygen pressure, oxygen saturation of hemoglobin, and arterial oxygen content reduces aerobic exercise performance and causes acute mountain sickness (AMS) in proportion to the altitude exposed (26, 32, 34). In addition, high-altitude exposure increases fluid losses (23, 28), which result in body water deficits (hypohydration, HYPO). Therefore, it is not unusual for civilian and military personnel to be performing stressful exercise at high altitude while hypohydrated (7).
The impact of hypohydration on aggravating AMS and reducing aerobic exercise performance at high altitude has not been studied experimentally. Epidemiological and field studies (7, 14) suggest an association between hypohydration and increased AMS, but this is not well supported by controlled studies. Hypoxia and hypohydration each reduce aerobic exercise performance independently (26, 23a). For example, maximal aerobic power (V̇o2max) is reduced by 15% at an altitude of 3,000 m (18) and by 5% with moderate hypohydration (∼3% body mass) at sea level (23a). Likewise, aerobic performance tests requiring ∼30 min to complete are degraded by 10–15% at an altitude of 3,000 m (18) and by 6–8% during moderate (3%) hypohydration at sea level (2, 12, 23a). It is unknown whether combining altitude exposure with hypohydration will exacerbate the prevalence and severity of AMS symptoms and/or further reduce aerobic exercise performance.
The purpose of this study was to determine the effects of hypohydration (∼4% body mass loss, hyperosmotic, hypovolemic) on aerobic exercise performance at high altitude and on the prevalence and severity of AMS symptoms. We hypothesized that hypohydration and altitude exposure would each degrade exercise performance and that the combination of these stressors would exacerbate the performance impairment. We also hypothesized that hypohydration would increase the prevalence and severity of AMS symptoms. Our approach was to use a hyperosmotic, hypovolemic model of hypohydration (−4%), caused by exercise-induced sweating, to mimic the type and magnitude of hypohydration commonly observed during recreational activities (climbers, trekkers) and military operations. To understand the potential effect of hypohydration on cerebral edema, S100β was measured as a marker of blood-brain barrier function (3, 4). Furthermore, since decreased diuresis has been noted in individuals who are susceptible to AMS (5, 6), fluid-regulatory hormones were measured to provide insight into possible mechanisms of increased AMS symptoms. Finally, we used a warm ambient environmental temperature to replicate the conditions in many high-altitude regions within 30° of the equator.
Seven male lowlander volunteers provided written informed consent to participate in this study, which was approved by the Scientific and Human Use Review Boards of the U.S. Army Research Institute of Environmental Medicine and the U.S. Army Medical Research and Materiel Command. The subjects volunteered after being fully informed of the requirements and risks associated with the research. Investigators adhered to Army Regulation (AR) 70–25 and U.S. Army Medical Research and Materiel Command Regulation 70–25 on the use of volunteers in research. Subject characteristics (mean, SD) were age, 25 ± 7 yr old; height, 176 ± 5 cm; body mass, 82 ± 11 kg; peak oxygen uptake (V̇o2peak) at sea level, 44.1 ± 4.9 ml·kg−1·min−1; V̇o2peak at 3,048 m, 38.6 ± 5.5 ml·kg−1·min−1; and percent body fat, 18.7 ± 3.9%.
Several weeks of preliminary tests were completed before beginning the experimental trials. The first week of preliminary testing included the measurement of baseline body mass using a calibrated scale (model WSI-600, Mettler-Toldeo, Columbus, OH). Subject body mass was measured each morning before breakfast and after voiding to establish a reliable baseline for euhydration trials. Body composition and percent body fat were determined from dual-energy x-ray absorptiometry (DEXA, model DPX-L, Lunar, Madison, WI). V̇o2peak was measured, using an incremental test, on an electronically braked cycle ergometer (Lode Excalibur Sport, Groningen, The Netherlands) and a metabolic measurement system (True-Max, ParvoMedics, Sandy, UT). Peak power output (in W) was obtained during this test while the subjects were euhydrated. The V̇o2peak test was conducted twice, once at sea level (SL) and once at an altitude of 3,048 m (ALT).
On a separate day, following completion of the V̇o2peak tests, volunteers cycled, at sea level, for 5–8 min at 100 and 140 W to determine the oxygen uptake (V̇o2) at these power outputs. The relative exercise intensity (relative to V̇o2peak at 3,048 m) was determined for these two power outputs (46% V̇o2peak at 100 W and 62% V̇o2peak at 140 W). The subjects then exercised at higher power outputs to determine the wattage needed at sea level so that the relative exercise intensity at sea level could be matched to the relative exercise intensity at 3,048 m. The power output was also determined that elicited 50% V̇o2peak at SL.
Subjects performed three to four exposure trials (walking at 1.34 m/s in 40°C air temperature using two 50/10 min work/rest cycles) to partially acclimate (∼60%) to the heat, enhance cardiovascular stability, and reduce between-subject variability. Subjects also received a 30-min familiarization flight in the altitude chamber to go over safety issues within the chamber and to get an appreciation of ascending and descending from 3,048-m simulated altitude. Before the experimental trials all volunteers practiced the cycle time trial three times to establish a baseline performance time. This ergometer allows for pedal rate-independent (hyperbolic) and -dependent (linear) cycling modes. Linear factors and pedal cadences were determined for each individual as previously described (11). The coefficient of variation for these three familiarization time trials was 3.1 ± 1.8%.
Four separate, counterbalanced experimental trials were conducted with subjects completing all four trials. The volunteer entered a 27.5 ± 0.7°C dry bulb, 6.9 ± 3.4°C dew point environment (27% relative humidity) at either SL or ALT (3,048 m; 10,000 ft) to complete each trial, which was 8 h in duration. SL was defined as an altitude of 152 m (chamber altitude was increased slightly to allow for environmental control). Subjects were blinded to the elevations. Altitude exposure was always followed by a sea-level exposure, and sea-level exposure was always followed by altitude exposure. At least three full days separated each experimental trial (with at least 6 days between altitude trials). Subjects reported the day before each experiment and underwent a hypohydration/euhydration session in the late afternoon hours. These sessions consisted of walking at 1.34 m/s in 50°C air (using 25/5 min work/rest cycles) for 2.5–3 h with and without fluid replacement. Subjects either replaced all the sweat secreted (measured by body mass losses where 1 ml = 1 g) or were given no fluid. The target mass loss in the no-fluid condition was 4%. Subjects slept and then after a light breakfast completed one of the four experimental trials. Two trials (SL and ALT) were performed while euhydrated (EUH) and two trials (SL and ALT) while hypohydrated (HYPO; 4% loss of body mass). Trials are denoted as SL-EUH, SL-HYPO, ALT-EUH, and ALT-HYPO. Figure 1 presents the experimental timeline.
Steady-state exercise and exercise performance test.
Two phases (steady state and time trial) were conducted during the exercise portion (60 min total). First subjects completed a 30-min steady-state phase. During this phase, constant-load submaximal exercise was performed in the hyperbolic mode setting that imposes a constant work rate independent of pedaling speed. The 30-min steady-state phase was broken into two workloads: the first workload was 20 min in duration and the second workload was 10 min in duration. Resistance for these two workloads was 100 W (46% V̇o2peak) and 140 W (62% V̇o2peak) during exercise at 3,048 m. At SL, subjects exercised at the same %V̇o2peak, based on the SL V̇o2peak results. Exercise was performed with a fan (3–5 mph) blowing on the subject. The steady-state exercise phase was performed after ∼1.25 h of exposure to SL or ALT conditions in the chamber.
After a short break, subjects then began the second phase, a 30-min exercise performance time trial (TT, linear mode). During the TT subjects were blinded to everything except elapsed time, and no motivation was given to the subjects. After completion of the time trial, the subjects were told the amount of work completed (kJ). In the linear mode, work rate increases as pedal rate increases [W = L × (rpm)2] similar to a mechanically braked cycle ergometer. Individual linear factors (L) were calculated for each volunteer to equal the SL 50% V̇o2peak exercise intensity at a pedal cadence of 60 rpm. The linear factor was kept at the sea level value to maintain the pedal cadence to workload relationship. This avoided the potential problem of dissimilar pedal cadences at equal workloads while preserving the required metabolic cost at any given workload. In the time trial, performance was assessed as the total amount of work completed in 30 min (kJ). Only time was visually displayed for subjects to monitor their progress. Percentage changes in performance, relative to SL-EUH, were calculated to determine the independent and combined stressors of HYPO and ALT. Pacing strategy was also examined. Individual pacing strategy for each trial was calculated by comparing the difference in actual work performed in each of ten 3-min work blocks normalized to the arithmetic mean (total work in 30 min divided by 10). A negative number indicates a slower-than-average pace for that block; a positive number, a faster than average pace.
AMS assessments were taken in the environmental chamber after 5 and 8 h of exposure during each experimental trial. Subject severity scores that reached the AMS threshold value at either time point were reported as positive for AMS. The prevalence and severity of AMS symptoms was determined from information gathered using a shortened version (8) of the Environmental Symptoms Questionnaire [ESQ (35)] presented on a hand-held personal digital assistant (iPAQ, Hewlett Packard). A weighted average of scores from 11 items (headache, lightheaded, dizzy, etc.) designated AMS-C was calculated. AMS-C scores equal to or greater than 0.7 are defined as diagnostic of AMS. AMS symptoms were also evaluated using the Lake Louise AMS Scoring System (LLS). The LLS consists of a five-question, self-reported assessment of AMS symptoms (33). Total LLS scores ≥ 3 (range, 0–15) that include headache at any time point are diagnostic of AMS.
At the completion of the PDA-based questionnaire, the volunteer's resting arterial oxygen saturation (SaO2) was measured (Dolphin Medical, Voyager Pulse Oximeter, Hawthorne, CA) for 1 min.
Measurements taken during the initial steady-state exercise phase included heart rate (HR, Polar E30, Polar USA, Lake Success, NY), core temperature (esophageal and pill, YSI, Yellow Springs, OH; Philips/Respironics, Bend, OR), skin temperature (YSI; forearm, tricep, chest, thigh, calf; mean skin temperature calculated according to Ref. 29), arterial oxygen saturation, oxygen uptake, local sweat rate, skin blood flow (SBF), blood pressure, ratings of perceived exertion (RPE), and thermal sensation. HR and SaO2 were also collected during the time trial. During the break period between the steady-state exercise and time trial, body mass was determined and water given in the euhydrated condition so that body mass was equal to the mass measured before the beginning of submaximal exercise. No fluid was given in the hypohydrated condition, except following the TT (to match sweat losses over 60 min of exercise). Esophageal and pill temperatures were not uniformly obtained in every subject on every trial; some subjects who could not insert an esophageal probe on a particular day were given a temperature pill to insert as a suppository.
SaO2 via noninvasive finger pulse oximetry (model 8600, Nonin Medical, Plymouth, MN) was monitored continuously during the V̇o2peak tests at sea level and 3,048 m and also during exercise in both the steady-state and performance phases. SaO2 was also determined immediately on finishing the performance portion of the time trial.
SBF (Peri-Med, PeriFlux System 5000) was measured on the ventral forearm (opposite arm of the blood pressure cuff, forearm kept at heart level) using laser-Doppler flowmetry. The flow probe (sampling at 780 nm) was held in place with adhesive, and the temperature underneath the probe was maintained at 33°C. Output (mV) was divided by mean arterial pressure (automated cuff, Cycle model, SunTech Medical, Morrisville, NC) to give an index of cutaneous vascular conductance (CVC, mV/mmHg). Following the time trial, maximal CVC was assessed by locally heating the forearm to 42°C for 30 min. Exercise CVC values were then expressed as a percentage of the maximal CVC, relative to each specific trial.
Local chest sweat rate was measured during the initial 30 min of exercise using a commercially available (Bi-Tronics, model BI102LS, Guilford, CT) ventilated (room air, 1 l/min) dew-point hygrometry system (22). The sensor was held in place with a strap placed around the chest.
Blood and body water.
Blood was sampled from an indwelling catheter inserted into a superficial forearm vein and was obtained at two time periods [before altitude exposure (0 h, outside the chamber, 21°C), and at 8 h (8 h) of exposure] after the subject had been sitting for a minimum of 20 min. Blood was analyzed for hemoglobin concentration, hematocrit, and plasma osmolality (freezing-point depression). Hemoglobin and hematocrit were used to calculate %plasma volume (PV) changes (15). Absolute PV was calculated from fat-free mass from the equation of Sawka et al. (38) for both EUH trials. To calculate the absolute PV for the HYPO trials, the %ΔPV at 0 h, relative to EUH, was determined using the Dill and Costill equation (15) and absolute PV adjusted for the HYPO trials for this change. Initial total body water (TBW) was calculated as 0.74 × lean body mass + 0.10 × fat mass as this relationship is constant in adults (23a). The change in TBW was calculated by the change in body mass for each trial from the initial pre-HYPO body mass.
Plasma renin activity (PRA), aldosterone (ALD), arginine vasopressin (AVP), and atrial natriuretic peptide (ANP) were measured before exposure (0 h) and after 8 h of exposure. PRA was measured with a commercially available two-site immunoradiometric assay (DSL-25100 ACTIVE Renin IRMA kit, Diagnostic Systems Laboratories, Webster, TX). The %CV for PRA was 4.5% (intra-assay) and 6.8% (interassay). ALD was measured with a commercially available radioimmunoassay (RIA, DSL-8600 ACTIVE kit, Diagnostic Laboratories, Webster, TX). The %CV for ALD was 5.9% (intra-assay) and 3.7% (interassay). AVP was measured with a commercially available double-antibody RIA (Buhlmann Laboratories AG, Schonenbuch, Switzerland). The intra-assay %CV for AVP was 17.9%. ANP was measured using a commercially available nonequilibrium RIA (Euro-Diagnostica, Malmo, Sweden). The %CV for ANP was 18.2% (intra-assay) and 12.9% (interassay). S100β was measured before and after 8 h of exposure as a marker of blood-brain barrier integrity (24). S100β was measured using a commercially available enzyme-linked immunosorbent assay (ALPCO Diagnostics, Salem, NH). The %CV for S100β was 4.3% (intra-assay) and 6.8% (interassay). S100β was used as a marker of possible blood-brain barrier disruption and has been previously measured in hypoxia studies to potentially explain cerebral edema (3, 4).
Total work, HR, SaO2, body mass, RPE, SBF, and blood measurements were analyzed using a three-way (altitude × hydration × time) repeated-measures analysis of variance (RMANOVA). AMS severity scores were analyzed using two-way (hydration × altitude) RMANOVA. The Newman-Keuls procedure was used when significant differences among means were identified using RMANOVA. Data are presented as means ± SD. The relative risk of developing AMS during altitude exposure was determined from the prevalence rates in the EUH and HYPO trials (prevalence rate in HYPO divided by prevalence rate in EUH).
For the time-trial performance, the practical importance of these differences was examined by comparing the mean and 95% confidence intervals against an a priori zone of indifference, defined as the %CV observed during performance training (calculated from familiarization trials 2 and 3 and equal to 3.1%). This procedure affords evaluation against a standard other than zero, is similar to equivalence testing in the clinical sciences (16), and affords determination of whether a performance change is of practical or meaningful importance. As well, the %change from SL-EUH for the other three trials was analyzed using a one-way RMANOVA.
The onset of chest sweating rate was determined by segmented linear regression, a method of regression analysis that identified the intersection of two line segments formed when a breakpoint occurred in the chest sweating rate as a function of time (10).
The primary outcome variable in this study was the amount of work completed during the performance time trial. A power analysis at α < 0.05 and β = 0.20 indicated that seven to eight subjects would provide sufficient power to detect a meaningful difference (6% change in time-trial performance; ∼20 kJ) among trials assuming a mean total work of ∼330 kJ and a CV of 3.1%. The desire to detect a twofold change from the %CV was chosen based on the likelihood of experimental perturbations (HYPO, ALT) producing practical or meaningful performance changes from SL-EUH (11).
Markers of hydration status are presented in Table 1. Volunteers were −4.0 ± 0.4% hypohydrated in the HYPO trials and −0.6 ± 0.5% hypohydrated in the EUH trials. Body masses were maintained at these levels throughout the 8-h exposure. Measurements of TBW, hematocrit, hemoglobin, PV, and plasma osmolality were consistent with hypohydration. HYPO resulted in a reduction in TBW (6–7%) and PV (10–12%), and an increase in plasma osmolality (Table 1). No differences were observed between SL and ALT during the HYPO trials.
Table 2 presents individual and mean time-trial performance data. Total work was lower in SL-HYPO compared with SL-EUH, and ALT-HYPO was significantly lower vs. the other three trials. One subject was unable to complete the entire 30-min bout during both HYPO trials. ALT-EUH performance was lower than SL-EUH but did not reach statistical significance using a RMANOVA (P = 0.06). Individual performance changes, relative to SL-EUH, demonstrated that 6/7 volunteers did worse during ALT-EUH, while 7/7 subjects did worse during SL-HYPO and ALT-HYPO. Figure 2 presents the percent change in performance, relative to SL-EUH, for ALT-EUH, SL-HYPO, and ALT-HYPO. Time-trial performance was −11.0 ± 10.4% worse for ALT-EUH, −18.6 ± 12.2% worse for SL-HYPO, and −33.7 ± 22.3% worse for ALT-HYPO. The confidence interval for SL-HYPO and ALT-HYPO falls entirely outside the a priori zone of indifference, which provides evidence that the negative effect of HYPO and ALT on performance in warm air is also of practical importance.
To evaluate pacing strategies, the average power output (PO, in W) over the entire 30-min time-trial period (n = 6) was calculated. Average PO was SL-EUH, 193.3 ± 57.0; SL-HYPO, 168.3 ± 64.5; ALT-EUH, 166.7 ± 44.5; and ALT-HYPO, 145.0 ± 53.2. Analysis of the pacing strategy (Fig. 3) indicated that during the last 3 min of the time trial (end spurt where subjects are trying to get in as much work as possible), subjects' pace was above the normalized average, but that this pace was lower at ALT vs. SL (main effect, P < 0.02). There was also a tendency throughout the time trial to have a lower normalized pace average when HYPO compared with EUH, but this was not statistically significant (P = 0.07).
Individual AMS-C and LLS scores are shown in Fig. 4. In general, both the severity of AMS-C and LLS scores and the number of subjects reporting AMS tended to increase in ALT-HYPO. However, there was no statistically significant difference between ALT-EUH and ALT-HYPO for severity scores using the AMS-C score (P = 0.07). For both the AMS-C and LLS, the symptom severity score increased ∼2.5 times from ALT-EUH to ALT-HYPO. Using the AMS-C score, in ALT-EUH, two subjects (29%) reached the AMS symptom threshold, whereas in the ALT-HYPO trial, five subjects (71%) met the threshold value of 0.7 for AMS symptoms (2.5 times increase in relative risk). For the AMS-C score, headache was identified in 2/7, 3/7, 2/7, and 2/7 volunteers for SL-EUH, SL-HYPO, ALT-EUH, and ALT-HYPO, whereas for the LLS, headache identification and AMS prevalence was 1/7, 3/7, 2/7, and 4/7. SaO2 levels after 8 h of exposure were 96.9 ± 0.5%, 96.5 ± 0.7%, 89.9 ± 2.2%, and 89.9 ± 2.1% for SL-EUH, SL-HYPO, ALT-EUH, and ALT-HYPO, respectively. ALT trials had a lower SaO2 than SL (P < 0.001), with no difference between hydration levels.
Physiological and perceptual responses to submaximal and time-trial bouts.
HR values during exercise were higher during HYPO vs. EUH (main effect, P < 0001), both during submaximal and time-trial exercise, but there were no differences between SL and ALT (Fig. 5). HR values at the end of the time trial, when volunteers were working to complete as much work as possible, were 171.3 ± 20, 175.3 ± 24.4, 165.8 ± 20.7, and 172.3 ± 19.7 beats/min for SL-EUH, SL-HYPO, ALT-EUH, and ALT-HYPO, respectively (P > 0.05). SaO2 levels during exercise were lower (P < 0.0001) in both ALT trials (81.0–82.5%) compared with SL (94.6–95.8%), but there were no differences between EUH and HYPO. (Fig. 5)
RPE was not different between SL and ALT during submaximal exercise but was significantly higher during HYPO vs. EUH (P < 0.05). At the end of the time trial, RPE was the same among trials (ranged from 18.7 ± 1.4 to 19.0 ± 1.3). Thermal sensation increased with time, but there were no differences between hydration levels or altitude conditions.
Thermal variables are presented in Table 3. Core temperature at the end of the submaximal exercise was ∼0.4°C higher in both HYPO trials, compared with EUH, but there was no difference between altitudes. Statistics were not run on core temperature due to the different measurement modes obtained on the subjects. Mean skin temperature decreased from minutes 0 to 10 and plateaued throughout the rest of the trials. There was a significant time-by-hydration interaction (P < 0.02) where HYPO trials had a 0.25–0.40°C higher mean skin temperature compared with EUH trials throughout submaximal exercise. The onset of chest sweating (in min, n = 5) was later (P < 0.05) in ALT-HYPO (9.5 ± 3.9 min) compared with SL-EUH (3.7 ± 2.6 min), SL-HYPO (5.9 ± 1.9 min), and ALT-EUH (5.0 ± 1.2 min). There were no differences among the latter three trials. CVC, expressed across time, was not different among the four trials.
Total fluid intakes (ml) over the 8-h exposure were 893 ± 242, 743 ± 146, 853 ± 418, and 679 ± 250 and total urine volumes (ml) over the 8-h exposure were 416 ± 419, 93 ± 126, 450 ± 452, and 131 ± 180 for SL-EUH, SL-HYPO, ALT-EUH, and ALT-HYPO, respectively, with HYPO values lower (P = 0.06) than EUH for urine volume.
Plasma fluid regulatory hormone responses are shown in Fig. 6. Results primarily indicate that there was a significant hydration effect, i.e., values were higher in HYPO vs. EUH trials for PRA, ALD, and AVP, with no differences between SL and ALT, whereas values in the HYPO trials were lower than EUH for ANP. Analysis of PRA also showed a significant altitude-by-hydration interaction, primarily caused by a higher 8-h value in SL-HYPO compared with the other trials. S100β was not different between hydration levels or altitude exposures (Fig. 7). Values for S100β before exposure ranged from 0.10 to 0.16 μg/l and after 8 h of exposure ranged from 0.18 to 0.22 μg/l.
This study was the first to experimentally evaluate the effect of hypohydration (hyperosmotic-hypovolemic) on aerobic exercise performance and the prevalence and severity of AMS symptoms during high-altitude exposure. Hyperosmotic hypovolemia was used as the model of hypohydration, caused by exercise-induced sweating, to mimic the type of hypohydration commonly observed in recreational and military activities. A moderate hypohydration level was employed, and a warm ambient temperature was selected to simulate a typical summer day at moderate altitude in equatorial regions. At sea level, moderate hypohydration will degrade aerobic performance in temperate, warm, and hot ambient conditions (25). The performance trial was selected to accentuate cardiovascular limitations (25), and a longer exercise bout would likely result in larger performance decrements by accentuating glycogen depletion and thermal strain (13). The principal findings from this study were 1) the combination of hypohydration and altitude exposure to 3,048 m reduces aerobic exercise performance in an additive manner over hypohydration and altitude exposure alone; and 2) hypohydration did not appear to increase the prevalence and severity of symptoms associated with AMS.
Hypohydration combined with 3,048-m altitude exposure caused a 34% decline in aerobic exercise performance while exercising in warm air. The combination of the two stressors was additive; independently hypohydration caused a 19% degradation and 3,048-m altitude exposure caused an 11% degradation in performance. Figure 2 clearly demonstrates that performance, relative to the SL-EUH trial, was significantly impaired in the other three trials. Further evidence for a practical importance for these findings is supported by the data showing that the 95% confidence limits for the SL-HYPO and ALT-HYPO trials lie outside the zone of indifference (defined as the coefficient of variation of the cycle ergometer time-trial test) and that the ALT-EUH trial confidence limits fall almost entirely outside the CV, synonymous with a significant difference, although not necessarily important. Furthermore, these data demonstrate that performance degraded to a similar extent at moderate HYPO as that observed at 3,048 m (10,000 ft). Many individuals quickly understand that altitude exposure degrades aerobic performance but fail to appreciate the same magnitude of performance degradation when hypohydrated.
Several mechanisms can be postulated for the large degradation in aerobic exercise performance during ALT-HYPO. ALT, independently, reduces submaximal exercise performance at 3,048 m, compared with SL, by 15–20% during events ranging from 20–30 min (18, 26). SaO2 is reduced at 3,048 m (to 83% during the time-trial performance) compared with sea level (95%), lowering the muscle capillary-to-mitochondrial oxygen gradient and contributing to a decline in the rate of oxygen delivery from the blood to the muscle (17). Furthermore, since subjects began the time trial at a linear factor set at 50% of the sea-level V̇o2peak, maintaining this power output at ALT to match the power output at SL would require subjects to work at a higher relative exercise intensity. Rather, subjects chose to reduce their power output at ALT, relative to the SL trials, by ∼20 W.
Hypohydration degrades aerobic exercise performance at sea level in temperate/warm-hot conditions (11, 25). Mechanisms for this include increased cardiovascular and thermoregulatory strain, hypovolemia, and increased perceptual strain (13). Subjects' HRs were ∼15–20 beats/min higher during submaximal exercise and 5–10 beats/min higher during the time trial when hypohydrated. Likewise, subjects' core temperatures were ∼0.4°C higher during HYPO vs. EUH and mean skin temperatures were higher in HYPO. These changes, combined with the plasma volume reduction, likely reduce cardiac filling and cardiac output. Gonzalez-Alonso et al. (20, 21) have shown that increased HR, hyperthermia, and lower plasma volume reduce cardiac output in HYPO compared with EUH. Furthermore, it has been demonstrated that V̇o2max is reduced by hypovolemia (27, 36). We observed a 9.5% decline in PV before exposure in HYPO compared with EUH. In temperate environments, a 4–8% decline in maximal aerobic power occurs after HYPO by 3% (23a). Thus during submaximal exercise in a HYPO state, subjects work at a higher relative exercise intensity and either become fatigued sooner or have to reduce their power output, as observed here. Subjects also perceived that they were working harder during HYPO, which may have also contributed to a reduction in performance. These data suggest that the large degradation in aerobic performance when exposed to 3,048-m altitude and hypohydrated is due to the additive effects of a reduction in arterial oxygen content, stroke volume, cardiac output, and V̇o2peak and an increase in cardiovascular, thermal, and perceptual strain.
This study was the first to prospectively examine the prevalence and severity of symptoms of AMS at 3,048 m over an 8-h time period with hyperosmotic hypohydration. When hypohydrated, AMS prevalence increased 2.5 times and severity scores by 2.5 times, but these increases were not statistically significant. There was greater variability in the altitude sickness scores when the subjects were hypohydrated and the number of subjects was not sufficient to observe significance.
Our findings are supported by Aoki and Robinson (1). In the only study to dehydrate subjects before hypobaric hypoxia before a prolonged altitude exposure (≥ 8 h), Aoki and Robinson (1) found no deleterious effect of 2.5% dehydration on AMS, using furosemide to induce an isosmotic hypovolemia (compared with hyperosmotic hypovolemia in present study). In contrast, other studies (7, 14, 30, 31) found that dehydration may worsen the symptoms of AMS. Two of these reports were epidemiological/field studies (7, 14) and the others examined AMS symptoms after only 100 min of hypoxic exposure. Although our findings were similar to Aoki and Robinson (1), we did observe a tendency for higher symptom severity scores with HYPO and the volunteers reported feeling sicker, which has important applications to field environments, although we acknowledge that we cannot distinguish between symptoms associated with AMS and HYPO (see Fig. 4B, where 2 subjects were “diagnosed” with AMS during SL-HYPO using the LLS; one subject had AMS in all trials). Further studies with larger numbers of subjects appear warranted to determine if hyperosmotic hypohydration indeed does increase AMS. Perhaps the nonsignificant elevation we observed in AMS severity scores was due to the type of dehydration induced. Using a diuretic, as Aoki and Robinson (1) used, will cause intravascular and interstitial volumes to decrease proportionately, causing no change in plasma solute concentrations, whereas dehydration induced by exercise in the heat increases plasma solutes and decreases intracellular fluid (osmotic gradient pulling fluids from tissues) to a greater extent than isosmotic hypovolemia. Specific studies comparing these two types of dehydration are warranted.
In summary, 4% hypohydration and 3,048-m exposure combine to additively reduce aerobic exercise performance. The performance degradation when these two stressors were combined (−33%) was equivalent to the sum of the performance degradations induced independently by hypohydration and altitude exposure. Hypohydration tended to increase the prevalence and severity of AMS symptoms, but not statistically. These findings strongly suggest that dehydration be avoided before ascending to high altitude such that further physical performance decrements are minimized.
No conflicts of interest, financial or otherwise, are declared by the author(s).
We give special thanks to the volunteers who endured the heat, dehydration, and altitude exposures. The expert technical assistance of Guy Tatum, Myra Reese, SPC Robert Hollins, and Jeff Staab is gratefully acknowledged.
This study is approved for public release; distribution is unlimited. The opinions or assertions contained herein are the private views of the author(s) and are not to be construed as official or reflecting the views of the U.S. Army or the Department of Defense. The investigators have adhered to the policies for protection of human subjects as prescribed in Army Regulation (AR) 70–25, and the research was conducted in adherence with the provisions of 32 CFR Part 219. Human subjects participated in these studies after giving their free and informed voluntary consent. Investigators adhered to AR 70–25 and U.S. Army Medical Research and Materiel Command Regulation 70–25 on the use of volunteers in research. Any citations of commercial organizations and trade names in this report do not constitute an official U.S. Department of the Army endorsement of approval of the products or services of these organizations.