Comparison of cardiovascular and biomechanical parameters of supine lower body negative pressure and upright lower body positive pressure to simulate activity in 1/6 G and 3/8 G

Abstract

For future space exploration missions, it is important to determine the best method of simulating on Earth cardiovascular and biomechanical conditions for lunar and Martian gravities. For this purpose, we compared exercise performed within a lower body negative pressure (LBNP) and a lower body positive pressure (LBPP) chamber. Twelve subjects underwent a protocol of resting and walking (0.25 Froude) within supine LBNP and upright LBPP simulation. Each protocol was performed in simulated 1/6 G and 3/8 G. We assessed heart rate (HR), mean arterial blood pressure, oxygen consumption (V̇o₂), normalized stride length, normalized vertical peak ground reaction force, duty factor, cadence, perceived exertion (Borg), and comfort of the subject. A mixed linear model was employed to determine effects of the simulation on the respective parameters. Furthermore, parameters were compared with predicted values for lunar and Martian gravities to determine the method that showed the best agreement. During walking, all cardiovascular and biomechanical parameters were unaffected by the simulation used for lunar and Martian gravities. During rest, HR and V̇o₂ were lower in supine LBNP compared with upright LBPP. HR, V̇o₂, and normalized vertical peak ground reaction force obtained with supine LBNP and upright LBPP showed good agreement with predicted values. Since supine LBNP and upright LBPP are lacking significant differences, we conclude that both simulations are suited to simulate the cardiovascular and biomechanical conditions during activity in lunar and Martian gravities. Operational characteristics and the intended application should be considered when choosing either supine LBNP or upright LBPP to simulate partial gravities on Earth.

missions of astronauts to moon and Mars are proposed in the future. From over 40 yr of manned spaceflight, it is known that the human body experiences cardiovascular and musculoskeletal losses and decreased aerobic fitness while exposed to reduced gravity (6, 7, 11). Mission durations of up to 2 yr or more in a reduced gravity environment are probable. A mission to Mars, for instance, will expose the human body to approximately 6 mo of microgravity during transition from Earth to Mars, followed by a stay on Mars for about 1 yr in 3/8 G and then transition back to Earth for another 6 mo of microgravity. Deconditioning of their musculoskeletal and cardiovascular systems during prolonged microgravity lowers exercise capacity and orthostatic tolerance, as well as decreases their overall crew performance, especially on arrival in Martian 3/8 G. Exposures to spaceflight so far document that the transition from microgravity to a higher gravity level is a critical phase for a crewmember (19, 27). Therefore, one important goal of space-related research is the development of countermeasures to prevent the aforementioned deconditioning. Countermeasure development requires Earth-based simulations of partial gravity to induce the aforementioned effects on the human body, and to test the efficacy of different countermeasures.

To advance the field of partial gravity research, further studies are needed to improve Earth-based simulations of reduced gravity environments. To date, several simulation methods have been tested. Assigning subjects to long-term bed rest in various tilt angles is a viable procedure to simulate the long-term effects of reduced gravity. Pavy Le Traon and coworkers (22), for example, recommend a 10° head-up tilt model to simulate the acute effects of 1/6 G (lunar gravity) on the cardiovascular system. Simulation of the deconditioning effects of microgravity on the musculoskeletal system is provided by supine bed rest, which nullifies the effective body weight along the body axis. Furthermore, the 6° head-down tilt model is usually employed during most bed-rest studies to simulate the fluid shift effect of microgravity on the cardiovascular system (13). To study the efficacy of countermeasures, it is, however, necessary that subjects can perform exercise within a simulated fractional-gravity environment.

Currently, three methods are well-suited to simulate exercise at fractional-gravity levels between 0 G and 1 G on Earth: supine lower body negative pressure (LBNP), upright lower body positive pressure (LBPP), and the partial gravity simulator (POGO). The POGO is a harness suspension system developed by the National Aeronautics and Space Administration (NASA) to simulate suited extravehicular activity (EVA) in reduced gravity environments, such as the moon and Mars. It achieves an unloading of the musculoskeletal system and, therefore, permits a good simulation of the dynamic aspects of exercise in reduced gravity environments (21). However, the POGO system does not alter the hydrostatic pressure gradients along the body axis and, therefore, imposes only a minimal impact on the cardiovascular system. In contrast to the POGO system, supine LBNP and upright LBPP use pressure differentials on the lower body to achieve an unloading (upright LBPP) or loading (supine LBNP) to the corresponding fractional body weight of a subject in reduced gravity (Fig. 1). Moreover, these pressure differentials change the hydrostatic pressure gradient along the body axis and thus reflect the changes in the cardiovascular system occurring in reduced gravity. Supine LBNP and upright LBPP are extensively studied in terms of ground reaction forces (GRFs), cardiovascular safety, and gait analyses (2, 8, 16). Also, some previous papers have described the principles of body loading or unloading in detail for LBNP and LBPP, respectively (2, 5, 8). Nevertheless, no previous study has directly compared these two methods with respect to both the biomechanical and cardiovascular responses of performing activity in lunar (1/6 G) and Martian (3/8 G) gravities. Therefore, the first aim of this study was to investigate two methods of simulating walking exercise in lunar (1/6 G) and Martian (3/8 G) gravities with regard to cardiovascular responses and biomechanics of gait. Second, we evaluated which method is better for simulating astronauts' activity in lunar and Martian gravity environments on Earth. Because actual physiological data are limited for moon exercise and lacking for Mars exercise, a prediction model was utilized to calculate comparable cardiovascular and biomechanical data. The better simulation method would then be defined by the following characteristics: 1) the produced data show a high level of agreement with both the predicted cardiovascular and biomechanical data; 2) the method shows no adverse effects; 3) the procedure is comfortable; and 4) the simulation is Earth based, inexpensive, and easily accessible to allow testing of a large number of subjects. We hypothesized that exercise performed in supine position within a LBNP chamber better simulates both the biomechanical and cardiovascular conditions occurring in lunar (1/6 G) and Martian gravity (3/8 G) compared with exercise performed in upright posture within a LBPP chamber. To formulate this hypothesis, we assumed that the posture of subjects in a 10° head-up-tilt bed rest is a partially validated method to simulate some effects of lunar gravity on the human body (22). Thus we assumed that supine LBNP better approximates the 10° head-up tilt compared with upright posture in LBPP. Moreover, during supine posture, application of negative pressure to the lower body can alter the blood pressure gradient along the vertical body axis without having to overcome the specific weight of the blood column. In contrast to that, while standing upright in LBPP, Earth's gravity vector creates a hydrostatic blood pressure gradient that must be somehow counteracted by the positive pressure applied to the lower body. As far as gait data during exercise are concerned, we found that such data are unavailable to compare these two simulation methods. Nevertheless, Boda and coworkers (2) compared gait patterns in supine LBNP with those of 1 G, whereas Cutuk and coworkers (8) analyzed gait during partial body unloading with ambulation in 1 G. Both found no significant differences in gait patterns. To determine the best procedure for an Earth-based simulation of moon and Mars exercise, both cardiovascular and biomechanical responses to exercise are needed.

Fig. 1.

Pictured is a schematic view of the principles of body unloading [upright lower body positive pressure (LBPP)] and loading [supine lower body negative pressure (LBNP)] in a lower body pressure chamber. During upright LBPP, a pressure differential (P1 > P2) across the waist seal creates a buoyancy force and ultimately reduces the effective ground reaction force toward the horizontal treadmill. Supine LBNP works the opposite way with a pressure differential (P1 > P2) across the waist seal, creating a suction force that generates a ground reaction force toward the vertical treadmill. (Modified schematic drawing by courtesy of H. Ruckstuhl.)

METHODS

Subjects

We studied 12 healthy human subjects, six women and six men. Anthropometrical data of the subjects are presented in Table 1. The sample size calculation was based on a priori power analysis using literature data for heart rate (HR) and mean arterial blood pressure (MAP) obtained in supine LBNP and upright LBPP at comparable pressure levels (8, 18). All subjects gave their written, informed consent before participation in the study. Subjects were excluded with history of chronic lower extremity pain, injuries, surgery of the lower extremities, cardiovascular or neurological diseases, or inguinal or abdominal hernia. This study was approved by the Institutional Review Board of the University of California, San Diego.

Table 1.

Anthropometric data of the subjects

	Men	Women	Total
n	6	6	12
Age, yr	25 ± 5	23 ± 3	24 ± 4
Height, m	1.84 ± 0.03	1.73 ± 0.09	1.78 ± 0.09
Weight, kg	92.8 ± 12.2	63.7 ± 8.7	78.2 ± 18.3
Body mass index, kg/m2	27.5 ± 3.9	21.4 ± 3.1	24.4 ± 4.6
Leg length, m	0.98 ± 0.04	0.93 ± 0.06	0.95 ± 0.06

Values are means ± SD; n, no. of subjects.

Measurement of Cardiovascular Responses

Overall we monitored three cardiovascular parameters continuously during the experiment: HR, MAP, and oxygen consumption (V̇o₂). HR and MAP were measured beat by beat employing the Finometer Model 1 (Finapres Medical Systems, Amsterdam, The Netherlands). The Finometer measures the arterial blood pressure (systolic, diastolic, and mean) and HR by pulse-wave detection and photoplethysmography at the finger. An automatic height-correcting system allowed compensating vertical arm movements during exercise and, therefore, calculation of brachial arterial blood pressure at heart level. Moreover, the instrumented arm was kept in a horizontal position to minimize the impact of the steps on the blood pressure measurement while walking. V̇o₂ during rest and exercise was measured using the Cosmed K4b system (Cosmed, Chicago, IL), a mobile system to monitor pulmonary gas exchange. For V̇o₂, however, we were only able to analyze 11 subjects because data from one subject had systematic errors and had to be excluded from further analysis.

Measurement of Gait Parameters

Four gait parameters were analyzed using force-sensitive shoe insoles (E.Q., Chalfont, PA). Force signals were sampled at 1,000 Hz and, due to high signal noise, were low-pass filtered using a windowed sinc-filter with Blackman window (cutoff frequency 20 Hz, length of the filter kernel 800). Normalized stride length, normalized vertical peak GRF (nVGRF), duty factor, and cadence (number of steps per second) were calculated from 20 consecutive right steps at the end of 4 min of walking. Stride length was calculated out of the stride time multiplied by the velocity of the treadmill and then normalized with respect to the subject's leg length, as suggested by Hof (12), with a stride length defined as the distance between two subsequent right foot touch downs. Leg length was measured from the greater trochanter of the femur to the ground (15). Duty factor was defined as stance time with respect to stride time (17). Calibration of the force-sensitive shoes insoles for vertical peak GRF was performed by having subjects alternately load their full body weight onto one leg. We then calculated the nVGRF referenced to each subject's body weight. Due to signal noise, however, some individual force data were not analyzed at low loading conditions (lunar gravity). In one subject, only 19 consecutive steps were analyzed during one trial. In addition, only in five subjects could a complete set of biomechanical parameters be used for statistical analysis, since some individual steps of the experimental protocol were not available for the remaining subjects. This was due to systematic errors as a cause of faulty shoe insoles.

Assessment of Subjective Parameters

We assessed exertion level while resting and walking in supine LBNP and upright LBPP, employing a Borg scale for rated perceived exertion (3). Subjects were asked to quantify their rated perceived exertion on a scale from 6 (no exertion at all) to 20 (maximum exertion). Moreover, subjects were asked to rate their level of comfort during rest and walking in supine LBNP and upright LBPP. Therefore, we used a modified visual analog scale in a range from 0 to 10, as suggested by Mundermann and coworkers (20) for assessing footwear comfort. Both subjective parameters were assessed in the fifth minute of each experimental step to allow time for the subject to adapt.

Experimental Protocol

All subjects completed a protocol consisting of resting and exercising in both supine LBNP and upright LBPP to simulate lunar (1/6 G) and Martian (3/8 G) gravities in random order. The waist seal for supine LBNP and upright LBPP was positioned at the level of the iliac crest in all subjects. Resting LBNP data were obtained while the subject rested in supine position with knees extended and feet in neutral position within the LBNP chamber. Comparable data were obtained in upright LBPP while the subject remained quiet and relaxed in upright position within the LBPP chamber. The walking exercise was performed at a Froude (Fr) number of 0.25 in both LBNP and LBPP, reflecting a comfortable walking speed (24). To obtain a Fr number of 0.25, treadmill speed was adjusted to the subject's leg length and the respective gravity condition. In other words, subjects walk more comfortably on the moon and Mars at a slower absolute speed because of reduced gravity (due to the concept of dynamic similarity between motions) (1). Treadmill speed was 3.4 ± 0.1 mph (mean ± SD) at 1 G, 2.1 ± 0.1 mph at 3/8 G, and 1.4 ± 0.1 mph at 1/6 G. Resting and walking at normal gravity (1 G) were performed for 5 min in supine LBNP and upright LBPP simulations, as well as during both gravity analogs (1/6 G and 3/8 G). Each protocol condition was preceded by a 10-min baseline recording without any gravity simulation or exercise. All conditions were randomized to avoid order effects of habituation and exhaustion. The pressures that were needed to load or unload the subject to 1/6 or 3/8 of body weight in supine LBNP and upright LBPP were individually prescribed using a weighing scale between treadmill and the subjects' feet. This was done before the experimental steps. The scale was then removed from the chamber. The mean pressure applied in supine LBNP was −11 ± 1 mmHg (mean ± SD) for lunar and −24 ± 3 mmHg for Martian simulation. For upright LBPP, 42 ± 4 mmHg (mean ± SD) was used for lunar and 30 ± 3 mmHg for Martian simulation.

Prediction of Lunar and Martian Parameters

Because actual cardiovascular and biomechanical data on Mars were not available to date and the Apollo missions only provided a limited number of data (26), a prediction model was utilized to produce comparable data. Previous literature provided evidence that GRF, in vivo tibial force, and knee range of motion all represent linear functions of partial gravity within the interval of 0 ≤ G ≤ 1 (9, 16). Pilot studies performed by our laboratory also indicated that both HR and V̇o₂ follow similar patterns within the interval of 0 ≤ G ≤ 1. Therefore, we obtained HR, MAP, V̇o₂, and nVGRF under 1 G conditions (upright standing and walking) and 0 G (resting in supine posture and movement of the horizontally suspended arms and legs with one swing per second). Subsequently, we performed a linear regression and computed the respective parameter as a function of gravity within the interval of 0 ≤ G ≤ 1. This allowed calculation of predicted values for our parameters for lunar and Martian conditions.

Statistics

All statistical analyses were employed using the software package IBM SPSS (20) (IBM SPSS, Munich, Germany). Descriptive statistics are reported as means and standard deviations (SD). A linear mixed model was fit to all outcome measures (25). Sex, activity, simulation, and gravity condition, and their interactions were entered as fixed effects, with activity, simulation and gravity condition treated as repeated effects. Covariance matrices were estimated by a restricted maximum likelihood method. In addition to compound symmetry (constant patterns of covariances), all models were also reanalyzed with unstructured covariance matrices (no assumption about the form of the covariance matrix). Finally, to account for subject variation, random effects (intercepts and slopes) were also considered for improving model fit. Penalized likelihood methods (Akaike information criterion) were used to determine model adequacy and tested for significant differences using χ² tests. Any significant main or interaction effects were followed up using Holm-Bonferroni-corrected t-tests. The level of significance was set to α = 0.05 for all testing. Furthermore, the parameters HR, V̇o₂, and nVGRF obtained with supine LBNP and upright LBPP were correlated with predicted parameters for 1/6 G and 3/8 G by computing Pearson's correlation coefficients.

RESULTS

We observed no adverse events, such as syncope, inguinal herniation, or other adverse events, during our study. Analyses of comfort data documented that subjects felt comfortable in both simulations using supine LBNP and upright LBPP. A complete overview of subjective, cardiovascular, and biomechanical parameters is presented in Table 2.

Table 2.

Overview of cardiovascular, biomechanical, and subjective parameters obtained during supine LBNP and upright LBPP simulations

		Supine LBNP				Upright LBPP
		Moon		Mars		Moon		Mars
	n (Men/Women)	Rest	Walking	Rest	Walking	Rest	Walking	Rest	Walking
HR, l/min	12 (6/6)	62 ± 11	72 ± 10	68 ± 12	77 ± 10	67 ± 12	74 ± 11	67 ± 12	80 ± 11
MAP, mmHg	10 (5/5)	105 ± 18	105 ± 13	103 ± 12	108 ± 14	105 ± 8	105 ± 9	104 ± 13	107 ± 13
V̇o2, ml·kg−1·min−1	11 (5/6)	2.75 ± 0.48	4.66 ± 1.41	3.01 ± 0.52	5.66 ± 1.39	3.29 ± 0.48	4.60 ± 0.77	3.25 ± 0.51	6.17 ± 1.32
Stride length normalized, AU	10 (5/5)*		1.06 ± 0.14		1.36 ± 0.12		1.02 ± 0.13		1.32 ± 0.13
Vertical peak GRF normalized, AU	10 (5/5)*		0.37 ± 0.23		0.73 ± 0.36		0.27 ± 0.16		0.54 ± 0.25
Duty factor, %	10 (5/5)*		62.5 ± 3.6		63.5 ± 2.9		60.8 ± 4.9		62.6 ± 3.0
Cadence, Hz	10 (5/5)*		1.23 ± 0.17		1.46 ± 0.13		1.24 ± 0.17		1.51 ± 0.14
Borg (6–20)	12 (6/6)	6.4 ± 0.7	7.8 ± 1.5	6.8 ± 1.5	8.3 ± 1.6	6.7 ± 1.2	7.7 ± 1.2	6.8 ± 1.7	8.3 ± 1.8
Comfort (VAS 1–10)	12 (6/6)	8.3 ± 1.1	7.9 ± 1.2	7.9 ± 1.3	7.3 ± 1.2	8.3 ± 1.0	8.0 ± 1.1	8.3 ± 0.9	8.0 ± 1.1

Values are means ± SD; n, no. of subjects.

LBNP, lower body negative pressure; LBPP, lower body positive pressure; HR, heart rate; MAP, mean arterial blood pressure; V̇o₂, oxygen consumption; AU, arbitrary units; GRF, ground reaction force; VAS, visual analog scale.

^*During lunar gravity, data were collected in four men and three women for supine LBNP and in four men and four women for upright LBPP, respectively.

Subjective Ratings

Borg (6–20).

Borg rating was significantly increased during walking compared with resting conditions (P < 0.001) (Fig. 2). Furthermore, a significant interaction effect between activity and sex was observed (P < 0.05), which was, however, due to a clinically negligible lower rating for men at rest. Simulation condition did not affect perceived exertion (P = 0.229–0.893 for main and interaction effects). Minimally, but significantly, higher ratings were also observed for Martian compared with lunar gravity (P < 0.05), substantiating the objective responses of V̇o₂ and HR which were higher for Martian gravity as well. There was no significant nor clinically relevant main or interaction effect between any conditions. Despite being a categorical variable, Borg ratings are often treated as a continuous variable. However, typically Borg ratings display a much stronger variation than in the present study. To minimize a type II error, Borg ratings were reanalyzed using Wilcoxon signed rank tests. The difference between rest and walking condition was clearly confirmed (P < 0.05 for all comparisons). Furthermore, the random effect of simulation condition was confirmed (P = 0.375–0.830). However, in contrast to parametric statistics, the comparison between gravity levels did not reach statistical significance (P = 0.094–1.0). Irrespective of this discrepancy, the differences in rating between lunar and Martian gravity (0.1 to 0.6 arbitrary unit) were so marginal that the different test statistics did not substantially confound the interpretation of data.

Fig. 2.

Subjective outcome measures during rest and walking [Froude (Fr) = 0.25], supine LBNP and upright LBPP, and lunar and Martian gravities are depicted. Comfort, rating of subjective comfort using a visual analog scale (1–10); Borg, rated perceived exertion using the Borg scale (6–20). AU, arbitrary units. All values are means ± SD. *Significant difference between indicated factors, i.e., between lunar vs. Martian gravity, and supine LBNP and upright LBPP, respectively (P < 0.05). #Significant difference compared with rest (P < 0.05).

Comfort (0–10).

Comfort ratings substantiated the findings for perceived exertion. Accordingly, significantly lower ratings were observed for walking compared with rest (P < 0.05) (Fig. 2). Furthermore, there was a minimal trend for higher comfort ratings during lunar gravity compared with Martian gravity (P = 0.075). This tendency was also confirmed by the significant interaction between simulation and gravity conditions during supine LBNP (P < 0.05), indicating higher comfort ratings for upright LBPP compared with supine LBNP during Martian gravity. During lunar gravity, however, ratings remained constant between supine LBNP and upright LBPP. While this result is contradictory to the higher stress levels observed during LBPP, this interaction effect was of little clinical relevance, as its difference amounted only to about 0.5 arbitrary unit. There were no significant nor clinically relevant main or interaction effects between any other condition (P = 0.415–0.870).

Cardiovascular Data

HR.

No significant sex effects were found for HR (P = 0.488). As expected, HR was significantly increased during walking compared with rest by ∼7–13 beats/min (P < 0.001). HR was also significantly increased under Martian gravity compared with lunar gravity (P < 0.001). Except for upright LBPP at rest, this difference ranged between 5 and 6 beats/min. HR also increased more during walking under Martian gravity compared with lunar gravity (P = 0.077 for interaction between gravity condition and activity). This effect was primarily due to increased HR from rest to walking during Martian gravity with upright LBPP. In addition, HRs were slightly, but significantly, elevated during upright LBPP compared with supine LBNP at rest (P < 0.05). Considering the additional sex interaction with the simulation condition (P < 0.05), this effect was predominantly due to higher increases in HR from supine LBNP to upright LBPP in women compared with men. In addition, the significant interaction between simulation and gravity condition (P < 0.05) showed that the effect of simulation condition was slightly more pronounced under lunar gravity compared with Martian gravity. The stronger impact of LBPP on the cardiovascular system was also indicated by the significant interaction effect between activity, gravity, and simulation condition (P < 0.05), showing that walking induced more pronounced HR increases relative to resting conditions in upright LBPP during Martian gravity compared with supine LBNP during Martian gravity (Fig. 3).

Fig. 3.

Diagrammed are cardiovascular outcome measures during rest and walking (Fr = 0.25), supine LBNP and upright LBPP, and during lunar and Martian gravity. HR, heart rate; MAP, mean arterial blood pressure at heart level; V̇o₂, oxygen consumption. All values are means ± SD. #Significant difference compared with rest (P < 0.05). *Significant difference between indicated factors, i.e., between lunar vs. Martian gravity, and supine LBNP and upright LBPP (P < 0.05).

MAP.

MAP was increased during walking compared with resting conditions under Martian gravity, but this difference was not significant (P = 0.171–0.808 for main and interaction effects) (Fig. 3). Neither gravity nor simulation conditions affected MAP (P = 0.186–0.992 for main and interaction effects). Only sex affected MAP, with higher MAP during upright LBPP compared with supine LBNP in men, and lower MAP during upright LBPP compared with supine LBNP in women (P < 0.001).

V̇o₂.

Changes in V̇o₂ are aligned well with our cardiovascular results. Hence, V̇o₂ was significantly increased during walking compared with resting conditions (P < 0.001) (Fig. 3). This increase was more pronounced in men compared with women (P < 0.001). Under Martian gravity, both men and women had a higher V̇o₂ (P < 0.001) compared with lunar gravity, except for upright LBPP at rest, which remained nearly unchanged. Furthermore, Martian gravity yielded a greater increase in V̇o₂ from rest to walking compared with lunar gravity (P < 0.001). Also, in line with cardiovascular responses, V̇o₂ was increased under upright LBPP by ∼300 ml, on average, except for walking during lunar gravity, where it remained unchanged. However, these differences in V̇o₂ were not significant (P = 0.104–0.938 for main and interaction effects). Again, similar to HR responses, a significant interaction between sex and simulation condition was observed, with significantly higher increases in V̇o₂ during upright LBPP compared with supine LBNP and in women compared with men (P < 0.05). Finally, a significant interaction between sex, activity, and simulation condition was found (P < 0.05), indicating that V̇o₂ was higher during upright LBPP compared with supine LBNP, except for men during walking, where V̇o₂ was lower during upright LBPP compared with supine LBNP.

Walking Biomechanics (Fr = 0.25)

Cadence.

Cadence was significantly increased during Martian compared with lunar gravities (P < 0.001). Differences between simulation conditions were minimal and not significant (P = 0.521). Moreover, any interaction effects including sex did not reach statistical significance (P = 0.440–0.704) (Fig. 4).

Fig. 4.

Illustrated are four biomechanical outcome measures during walking (Fr = 0.25), supine LBNP and upright LBPP, and lunar and Martian gravities. Stride length, normalized stride length; vertical peak GRF, normalized vertical peak ground reaction force. All values are means ± SD. *Significant difference between lunar vs. Martian gravity (P < 0.05).

Duty factor.

Data for duty factor only minimally differed between and within subjects. No significant main or interaction effects were observed (P = 0.152–0.878) (Fig. 4).

nVGRF.

nVGRF was significantly increased during Martian compared with lunar gravity (P < 0.001), irrespective of sex (P = 0.894 for interaction between sex and gravity condition). However, the difference was small, and Holm-Bonferroni-corrected post hoc testing failed to confirm these differences for both supine LBNP (P = 0.08) and upright LBPP (P = 0.06). As indicated in Fig. 4, nVGRF was slightly increased, but not significantly higher, under supine LBNP compared with upright LBPP (P = 0.089). While there was neither a significant main effect for sex (P = 0.350), nor an interaction between sex and simulation condition (P = 0.054), the higher data for nVGRF in supine LBNP vs. upright LBPP was predominantly due to women showing more pronounced differences in nVGRF between supine LBNP and upright LBPP compared with men. Neither the interaction between simulation and gravity condition (P = 0.503) nor between sex, simulation, and gravity condition (P = 0.720) were significant (Fig. 4).

Stride length.

Stride length was significantly increased during Martian gravity compared with lunar gravity (P < 0.001). Differences in stride length between simulation conditions were negligible and not significant (P = 0.560). Any interaction effects, including sex, also remained insignificant (P = 0.207–0.827) (Fig. 4).

Predicted parameters and comparison with supine LBNP and upright LBPP.

Predicted parameters for lunar and Martian gravites during walking (Fr = 0.25) are presented in Table 3. As previously mentioned, we found no statistically significant differences between supine LBNP and upright LBPP for walking in lunar and Martian gravity analogs. These findings were supported by linear regression analyses, which we subsequently performed for HR, V̇o₂, and nVGRF. There was good agreement between supine LBNP and upright LBPP in these parameters. Moreover, the high coefficients of determination (R² > 0.96) supported our a priori assumption of a linear correlation between HR, V̇o₂, and nVGRF and gravity within the interval of 0 ≤ G ≤ 1 (Fig. 5). Also, computing Pearson's correlation coefficients revealed a significant correlation between the simulated and predicted parameter HR for both supine LBNP and upright LBPP. We found this correlation existent for lunar and Martian gravities (Fig. 6).

Table 3.

Predicted parameters for walking (Froude = 0.25) in lunar and Martian gravity

		G Level		Linear Equation (y = mx + n)		Predicted Parameters For
Condition	n	0	1	m	n	1/6 G	3/8 G
HR, 1 × min
Rest	12	62	86	24	62	66	71
Walking	12	71	105	33	71	77	84
V̇o2, ml·kg−1·min−1
Rest	11	2.77	3.48	0.71	2.77	2.89	3.04
Walking	11	4.01	12.36	8.35	4.01	5.40	7.14
Vertical peak GRF normalized, AU
Rest
Walking	10	0.0	1.73	1.73	0.0	0.29	0.65

Values are means. AU, arbitrary units; G, gravity.

Fig. 5.

Illustrated are the parameters HR, V̇o₂, and normalized vertical peak GRF (nVGRF) obtained with either supine LBNP or upright LBPP during walking (Fr = 0.25), dependent on the gravity level. Linear regression analysis reveals 1) a strong linear relation of HR, V̇o₂, and nVGRF and the simulated gravity level in the interval 0 ≤ G ≤ 1; and 2) a good level of agreement between supine LBNP and upright LBPP. R² coefficient of determination is shown. All values are means ± SE.

Fig. 6.

Diagrammed are scatter plots of the parameter HR obtained with supine LBNP and upright LBPP during walking, dependent on the predicted HR under lunar and Martian gravity conditions. Pearson's correlation coefficient (r) shows significant high correlation between the simulated and predicted parameter HR for both supine LBNP and upright LBPP.

DISCUSSION

Our study compares two lower body pressure methods for simulating lunar and Martian gravity activities in terms of cardiovascular and biomechanical parameters, namely supine LBNP and upright LBPP. We hypothesized that exercise performed in supine position within a LBNP chamber better simulates both the biomechanical and cardiovascular responses for lunar (1/6 G) and Martian (3/8 G) gravities compared with exercise performed in upright posture within a LBPP chamber. The results of our study, however, did not support our hypothesis. Overall, our data suggest that exercise in both simulations, supine LBNP and upright LBPP, may simulate the cardiovascular and biomechanical conditions during activity in lunar and Martian gravities. This conclusion is supported by a lack of significant differences between supine LBNP and upright LBPP during walking; in fact, both simulations show good agreement for HR, V̇o₂, and nVGRF (Fig. 5).

The Cardiovascular System During Supine LBNP and Upright LBPP

Interestingly, we found that the simulation procedure (supine LBNP vs. upright LBPP) has an effect on V̇o₂ and HR during rest (Fig. 3 and Table 2). Subjects showed a higher V̇o₂ in upright LBPP compared with supine LBNP. This finding may reflect the higher activity of antigravity muscles of a subject in upright position to keep the body in balance without suspension. On the other hand, most skeletal muscles are not used while one is resting in supine position. This difference diminishes once the subject starts exercise, when more muscles are active. HR is higher during upright LBPP than supine LBNP while simulating lunar gravity. This reflects the reduced preload of the heart; as venous return is attenuated when the subject is positioned upright instead of supine. This effect is also diminished when subjects are walking. We explain this finding by increased venous return from the leg muscle pump toward the heart during walking (4). MAP during rest and walking is unaffected by the simulation method used. These results are in line with previous measurements taken at heart level, as described by Watenpaugh and Hargens (27). The differences between supine LBNP and upright LBPP that we found during rest, however, largely vanished during walking. We surmise that the applied pressures are better tolerated by the subjects when they are exercising. Supine LBNP is known to cause syncope if applied to resting subjects at high negative pressures. The negative atmospheric pressure decreases venous return to the heart and thereby ultimately reduces HR, vascular tone, and blood pressure, which, consequently, leads to a vasovagal syncope, also called “cardiac depressor reflex” (10). Exercise, however, attenuates these untoward effects and increases the similarities between supine LBNP and upright LBPP.

Walking Biomechanics During Supine LBNP and Upright LBPP

Although we did not find significant differences in gait between the simulations, we only investigated four parameters (cadence, normalized stride length, duty factor, and nVGRF) that represent important characteristics for a global evaluation of gait. Boda and coworkers (2), however, performed more detailed analyses of gait within supine LBNP and compared it to upright ambulation, but only at 1 G and without comparison to upright LBPP. They find no major differences in gait, but detected minor differences in knee flexion, hip flexion, and maximum rise of the foot during the swing phase of gait. They attribute these findings to the waist seal and the suspension system, which is used to support the subject's back and extremities against Earth's gravity while in supine LBNP. Moreover, Boda and coworkers (2) also remark that exercise within supine LBNP can only simulate inertial accelerations of gait but cannot mimic the static accelerations that gravity imposes on the vestibular system of the human body. The lack of these afferent impulses to the vestibular system might also contribute to alterations in gait. The above-mentioned differences of gait between supine LBNP and upright gait in 1 G were also described as trends by Cutuk and associates (8) for upright LBPP. Although not statistically significant, these authors report a lowered dynamic knee range of motion due to the waist seal, which limits upward displacement during ambulation.

Operational Considerations

While we find no difference in simulating exercise in lunar and Martian gravity with supine LBNP or upright LBPP, there are some operational considerations one should regard when deciding for or against a certain simulation method. When using simulations of partial gravity in the course of bed-rest studies, for example countermeasure development, supine LBNP may be preferred. There subjects can remain in their supine posture and are not exposed to the effects of upright posture on various physiological systems. In view of future space missions to the moon or Mars, the question arises: how to simulate activity in these partial gravity environments while wearing an astronaut's suit for EVA. Our laboratory already performed pilot studies and found that the weight bearing of an EVA suit is better simulated in upright LBPP than in supine LBNP. While in upright posture it is possible to distribute the weight equally over the subject's body, the supine posture in LBNP only allows the additional weight bearing over the upper body by attaching shoulder straps attached to the waist seal (5). This latter condition is uncomfortable for the subject and not a good EVA suit simulation, as it only applies the weight of the simulated EVA suit over the subject's shoulders. Further studies on EVA suit simulations are, of course, indicated. Furthermore, it should also be considered that supine LBNP can cause discomforts for the subject at increasing negative pressure levels. Our results indicate that, during the Martian gravity simulation, supine LBNP was perceived to be less comfortable than upright LBPP. Verbal feedback from our subject mainly referred to mild abdominal discomforts due to the suction. This side effect is commonly seen with LBNP and, for example, described by Khan and coworkers (14).

Effects of Gravity and Activity on Cardiovascular Parameters

When comparing the two methods for simulating moon and Mars, raising gravitational load and activity increase most cardiovascular parameters. Elevated HR and V̇o₂ during walking at Martian gravity are a result of increased muscle activity and oxygen demand. Our results show that this effect is sex related, as men have a larger muscle mass that is employed during walking. Additionally, increased gravitational loads elevate the hydrostatic pressure gradient along the vertical body axis and reduce venous return to the heart (27). As a response to maintain cardiac output, HR rises (27). The heightened activity of the heart muscle causes a further increase in oxygen demand, leading to a higher V̇o₂. The increases in MAP as a result of walking at higher gravitational loads, although statistically significant, are trends only. The maximum differences in MAP we detected were 5 mmHg and lie within the limits of bias (<5 mmHg) and precision (<8 mmHg) for the Finometer device (28).

Limitations

Ruckstuhl and coworkers investigated HR and V̇o₂ during exercise under different body unloading conditions and treadmill speeds (23). In their study, unloading was achieved by applying LBPP. The specific LBPP level they applied unloaded each subject to a certain fractional body weight and was not specifically prescribed to alter the hydrostatic pressure gradients along the subject's vertical axis. Consequently, they state as a limitation that the cardiovascular system is not well simulated to match the reduced gravity conditions (23). Individual pressure prescriptions in supine LBNP are based on the same principle as in upright LBPP. Our results, however, show good agreement between supine LBNP and upright LBPP for responses of the cardiovascular system for both lunar and Martian gravity conditions. To achieve these gravity conditions, different absolute pressures are employed to load or unload the subject to their fractional body weight on moon or Mars. Interestingly, the resulting HR, MAP, and V̇o₂ are the same. Therefore, we suggest that pressure levels needed to achieve fractional body weights might be suited to produce cardiovascular responses that match cardiovascular responses in reduced gravity quite well.

We assumed that each parameter tested is a linear function of partial gravity over the interval (0 ≤ G ≤ 1) to predict cardiovascular and biomechanical parameters in lunar and Martian gravities. This assumption is based on pilot studies performed in our laboratory that proved this linear relation exists for HR (R² = 0.99) and V̇o₂ (R² = 0.96). Moreover, our study also detects a linear correlation between HR, V̇o₂, nVGRF, and simulated gravity within the interval of 0 ≤ G ≤ 1 (Fig. 5). Other investigators find this linear relation existing for GRFs and other musculoskeletal parameters as well (9, 16). Although only observed during simulated fractional gravity, we speculate that this linear relation also exists under real fractional gravity conditions (moon or Mars). However, with presently available data, we cannot exclude the possibility that the extremes (1 G and 0 G) have an undue influence on the linear regression. This is a possible limitation of our study.

The assessment of comfort was performed by using a visual analog scale, which was originally designed to assess foot comfort. Before the experiment, subjects received clear instructions that they should consider their overall comfort while performing the experimental steps, with no particular focus on the waist seal, loading of joints, or pressure on their feet. Moreover, due to high signal noise, force signals had to be low-pass filtered at 20 Hz. We consider this limitation as minor, because peak GRF may be only minimally affected by the low cutoff frequency, as we just investigated slow walking.

Conclusions

Our study documents some significant differences between supine LBNP and upright LBPP in terms of simulating lunar and Martian gravities on Earth. During rest, HR and V̇o₂ were lower in supine LBNP compared with upright LBPP. These effects, however, largely vanished during exercise. Cardiovascular and biomechanical parameters, as well as Borg rating, are not influenced by the simulation device during walking. There is, however, a slight tendency for lower comfort during supine LBNP compared with upright LBPP, especially at higher gravitational loads. Nevertheless, we conclude that both simulations are suited to simulate the cardiovascular and biomechanical conditions during activity in lunar and Martian gravities. Furthermore, our results demonstrate very clearly that activity and gravity are important drivers of the cardiovascular system and biomechanics of walking. HR, V̇o₂, and nVGRF significantly increase with activity and rising gravity levels. Overall, our results contribute to a better understanding of the Earth-based simulations of partial gravity and serve as guidance for further studies involving Earth-based simulation of reduced gravity. The decision to use either supine LBNP or upright LBPP to simulate partial gravities on Earth probably depends on the particular application. Moreover, further studies investigating the effect of lower body pressure application on the hydrostatic blood pressure gradient are needed.

GRANTS

These studies were partially supported by National Aeronautics and Space Administration Grant NMX10AM18G and by the UCSD General Clinical Research Center with funding provided by National Institutes of Health Division of Research Resources Grant M01RR-000827. T. Schlabs received a research scholarship grant from the German Academic Exchange Service (DAAD).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: T.S. and A.R.H. conception and design of research; T.S. and A.R.-V. performed experiments; T.S., A.R.-V., H.R., and A.R.H. analyzed data; T.S., A.R.-V., H.R., A.C.S., and A.R.H. interpreted results of experiments; T.S., H.R., and A.C.S. prepared figures; T.S. and A.R.H. drafted manuscript; T.S., A.R.-V., H.R., A.C.S., and A.R.H. edited and revised manuscript; T.S., A.R.-V., H.R., A.C.S., and A.R.H. approved final version of manuscript.