Many athletes incur lower extremity injuries that necessitate a reduction in training volume. Research has shown that 6 wk of cardiovascular detraining may result in a 14% to 16% decrease in V˙O2max (4). After such injuries, the clinician's objective is to prevent deconditioning of the injured athlete while properly rehabilitating the injury to provide the athlete the best possible outcome. During the rehabilitation process, the early introduction of movement and restoration of normal function is of paramount importance (7). Water running has been used in the rehabilitation process in an effort to minimize the deleterious effects of detraining on cardiovascular fitness. The properties associated with water, such as buoyancy and viscosity, make it an appealing exercise medium for injured populations (6). Previous research has investigated the physiological responses of running in the deep or shallow end of a pool, but little information has been obtained regarding the physiological responses obtained during water treadmill running (WTR).
Studies indicate that V˙O2max and maximal HR (HRmax) are significantly lower in deep-water running (DWR) when compared with land treadmill running (LTR). Investigators had subjects in these studies mimic land running while DWR. They monitored HR and step cadence while having the subjects subjectively increase the difficulty level and by adding weight to the subject (3,5,11,12). Shallow-water running (SWR) resulted in lower V˙O2max and HRmax values than LTR (12). Stride frequencies (SF) during DWR and SWR are significantly lower than during LTR (5,12).
Research has shown that maximal effort SWR elicits greater V˙O2 than DWR and that the V˙O2, HR, and SF associated with SWR are more closely related to LTR than DWR (12). One explanation for this is that during SWR, the subject is able to mimic the normal land running gait. It is reasonable to expect that for the same reason, WTR would elicit an exercise response similar to that of LTR. The physiological responses to WTR may vary depending on whether the subject is wearing a water shoe. Research addressing WTR (with and without water shoes) would provide beneficial information with respect to the prescription of water training for healthy and injured populations.
The purpose of this study was to establish WTR parameters needed to obtain cardiorespiratory responses comparable to LTR. To this end, we analyzed the differences in HR, treadmill speed, SF, and gender between LTR and WTR when wearing a water shoe and when not wearing a water shoe at four exercise intensities representing 50%, 60%, 70%, and 80% of land V˙O2max.
Participant characteristics are listed in Table 1. Eighteen trained male (n = 9) and female (n = 9) runners between 18 and 30 yr were subjects for this study. Subjects had been engaged in a running program consisting of at least three 30-min training sessions per week for at least 2 months before the commencement of the study. All subjects signed an approved institutional review board consent form before participating in this study. Subjects also completed a preexercise testing questionnaire to determine history of injuries, neurological impairments, and risk of untoward cardiovascular, pulmonary, or metabolic events during exercise. All subjects were classified as "low risk" according to the American College of Sports Medicine risk stratification (15).
A repeated-measures experimental design across three running conditions and four exercise intensities and gender guided this study. The primary variables of interest were V˙O2, HR, treadmill speed, and SF.
All subjects completed a maximal graded exercise test (GXT) and a submaximal exercise test on a land-based treadmill (Model TMX425C; Full Vision, Inc., Newton, KS) and two submaximal exercise tests on a motor-driven treadmill in a hydrotherapy pool (Model 500; HydroWorx, Middletown, PA). Before the GXT and each of the submaximal exercise tests, subjects were fitted with a chest-strap HR monitor (Polar Electro Oy, Hong Kong) to measure HR during exercise. Metabolic responses to exercise were measured using a TrueMax 2400 metabolic cart (Consentious Technologies, Sandy, UT). Before testing each subject, the flow meter was calibrated using a 3-L syringe at five different flow rates. The oxygen (O2) and carbon dioxide (CO2) analyzers were calibrated using room air and a medical grade calibration gas of known concentrations. The metabolic cart was programmed to display and print metabolic and ventilatory data every 15 s.
To determine appropriate exercise intensities for the three running conditions, all subjects completed the GXT first. Subjects began the GXT by walking at a self-selected brisk walking pace at level grade for 3 min. This was followed by jogging at a self-selected, submaximal pace between 1.92 and 3.35 m·s−1 at level grade for 3 min. The treadmill speed remained constant throughout the remaining stages of the exercise test, as the grade was increased 1.5% each additional minute until the subject voluntarily terminated the test because of fatigue despite verbal encouragement. The subjects' efforts were considered maximal if physical signs suggestive of exhaustion were apparent and at least two of the following three criteria were met: (a) maximal RER > 1.10, (b) maximal HR (HRmax) no less than 15 beats below age-predicted maximal HR, and (c) leveling off of V˙O2 despite an increase in workload (6,8,13,14). V˙O2max was defined as the highest 30-s average V˙O2 value. HRmax was defined as the highest single HR value recorded during the GXT.
After the GXT, subjects completed three submaximal exercise tests under the following three running conditions: 1) LTR, 2) WTR while wearing water shoes (WTR-S), and 3) WTR when not wearing water shoes (WTR-NS). At least 48 h after the GXT, all subjects completed the LTR submaximal exercise test first to determine appropriate water running speeds for the two subsequent submaximal exercise tests in the water. The order of the water running conditions was randomized, and all exercise tests were completed within 2 wk.
Each submaximal exercise test included running at 50%, 60%, 70%, and 80% of the V˙O2max determined from the maximal GXT. Each submaximal exercise test began with a 3-min warm-up at a self-selected walking speed. The speed of the treadmill was then gradually increased until the subjects reached a randomly selected intensity of exercise (50%, 60%, 70%, or 80% of V˙O2max). Subjects jogged at this intensity for at least 5 min, and steady-state V˙O2, HR, and SF were recorded. Stride frequency (strides per minute) was visually counted. A stride was considered to be a complete gait cycle of the running motion (right foot heal strike to right foot heal strike). After a 5-min rest period, subjects repeated this process until they had run at all four intensities of exercise.
During WTR-S, subjects were fitted with a water running shoe (2) (AQx Sports, Inc., McMinnville, OR) (Fig. 1). During the WTR-NS and WTR-S exercise tests, the water level was set at chest height. Water temperature was maintained at 32.2° ± 2°C. Subjects were instructed to jog looking straight ahead with their hands free of the hand rail. To prevent sculling motions of the arms, subjects were instructed to keep their arms at their sides with their hands in a loose-fisted position and their elbows flexed at 90° (16). Because excess hang time increases the time between foot contacts and reduces effort, subjects were instructed to jog with as normal a gait as possible to minimize hang time after rear foot push off. Several cues were provided to assist the runners in their efforts to minimize hang time. The water treadmill has a rail along the front; subjects were instructed to watch their vertical movement in relationship to the rail through live videos from frontal and sagittal cameras. In addition, the primary investigator monitored each trial. If a subject began to demonstrate excessive vertical movement, he/she was immediately advised to correct his/her gait. Although hang time was not reduced to that of LTR, excessive hang time was successfully minimized.
All data were analyzed with SAS statistical software (SAS Institute, Inc., Cary, NC). All tests were Bonferroni-corrected for multiple tests to maintain a familywise critical α value of P = 0.05. Eighteen subjects participated in this study, each of whom performed a submaximal exercise test under all three running conditions (i.e., LTR, WTR-S, and WTR-NS). Steady-state data from all four intensities of exercise (i.e., 50%, 60%, 70%, and 80% of V˙O2max) were obtained for all three running conditions. Thus, each subject (n = 18) contributed 4 data points to each running condition, and each running condition was represented by 72 data points.
Statistical analyses were conducted to determine differences in the HR and V˙O2 responses by gender and running condition. Analyses were also conducted to determine differences in stride frequency and V˙O2 as a function of treadmill speed among the three running conditions. Because slopes and intercepts were used in the data analysis, it was appropriate to center the data; so, intercepts were calculated within the range of data observed rather than when the independent variable was 0. The variables were centered at the average response for each variable. To center the variable, the average response for that variable was subtracted from the individual response. For example, when analyzing the V˙O2 response as a function of HR, the HR data were centered at HR of 150 bpm. When analyzing the V˙O2 response as a function of %HRmax, the HR data were centered at 75% of HRmax. When analyzing the differences in stride frequency as a function of treadmill speed, treadmill speed was centered at 160.9 m·min−1.
Linear mixed models (Proc Mixed in SAS) were used in all analyses so that within-subject covariances could be appropriately accounted for. Because there were multiple measures per subject, failure to account for covariances within subjects would lead to underestimated standard errors for the model terms. All data were analyzed as linear growth curves. That is, the V˙O2 response was represented as a function of HR, %V˙O2max, and %HRmax; stride frequency was represented as a function of the treadmill speed; and V˙O2 was represented as a function of treadmill speed. In all cases, a linear growth curve was appropriate for the range of values tested (Figs. 2-7). Because we centered the data, tests on intercepts were conducted at the center of the independent variable values and thus are appropriate even when the slopes are not parallel.
Personal characteristics of the 18 participants are shown in Table 1. All 18 participants demonstrated maximal efforts during the maximal GXT as demonstrated by RER values greater than 1.10 and HRmax values ranging from 91.8% to 101% of age-predicted HRmax (mean ± SD = 97.6% ± 2.46%).
The absolute HR-V˙O2 relationship in each of the three running conditions was averaged across gender (Fig. 2). The growth curve analysis revealed that the intercept of the HR-V˙O2 relationship at an HR of 150 bpm was significantly less (P < 0.05) when running on a land treadmill (34.66 mL·kg−1·min−1) compared with a water treadmill. The intercepts of the HR-V˙O2 relationship during WTR-S (37.51 mL·kg−1·min−1) and WTR-NS (37.21 mL·kg−1·min−1) were nearly identical (P = 0.4115). The slopes of the HR-V˙O2 relationships during LTR (0.3953), WTR-S (0.4004), and WTR-NS (0.3938) were not significantly different (P = 0.72-0.94).
Figure 3 illustrates the %HRmax-%V˙O2max relationship in each of the three running conditions averaged across gender. The growth curve analysis revealed that the intercept of the %HRmax-%V˙O2max relationship at an HR of 75% of HRmax was significantly less (P < 0.05) when running on a land treadmill (58.82% of V˙O2max) compared with a water treadmill. The intercepts of the %HRmax-%V˙O2max relationship during WTR-S (63.71% of V˙O2max) and WTR-NS (63.17% of V˙O2max) were not significant (P = 0.3914). The slopes of the %HRmax-%V˙O2max relationships during LTR (1.3277), WTR-S (1.3829), and WTR-NS (1.3749) were not significant (P > 0.05).
There were significant differences in SF between running conditions. Figure 4 illustrates the differences in SF at various treadmill speeds during LTR, WTR-S, and WTR-NS. At a treadmill speed of 160.9 m·min−1, stride frequency during LTR was 23.6 strides per minute greater (P < 0.05) than the SF during WTR-S and 21.8 strides per minute greater than SF during WTR-NS. There were no significant differences (P > 0.05) in the slopes of the SF data during LTR, WTR-S, and WTR-NS conditions. The relationship between V˙O2 and various water treadmill speeds during WTR-S and WTR-NS conditions is plotted in Figure 5. At a given treadmill speed, V˙O2 was, on average, 4.12 mL·kg−1·min−1 significantly higher (P < 0.05) during WTR-S compared with WTR-NS condition. There were no differences in the slopes between WTR-S and WTR-NS (P = 0.8498).
Figure 6 illustrates the absolute HR-V˙O2 relationship in male and female participants averaged over the running conditions for all exercise intensities. The growth curve analysis revealed a significant gender effect (P < 0.05) in intercepts of the growth curves. At an HR of 150 bpm, the V˙O2 in males (40.52 mL·kg−1·min−1) was 8.12 mL·kg−1·min−1 higher than that of their female (32.40 mL·kg−1·min−1) counterparts. The slope of the HR-V˙O2 relationship was significantly greater (P < 0.05) in males (0.4453) than in females (0.3477).
The relative %HRmax-%V˙O2max relationship in male and female participants was averaged across the three running conditions (Fig. 7). The growth curve analysis revealed that, at an HR of 75% of HRmax, the %V˙O2max in males (63.44%) was not significantly different (P = 0.2335) from that in females (60.36%). The slope of the %HRmax-V˙O2max relationship was not significantly different (P = 0.2827) in males (1.4022) compared with that in females (1.3214).
We found that there were no gender differences (P = 0.5287) in SF at the various treadmill speeds used to elicit the desired intensity of exercise.
The findings of this study support the use of WTR as an effective training alternative to LTR. Although WTR can be used by any athlete, it is particularly efficacious for injured athletes who would benefit from an exercise modality that can produce an overload significant enough to maintain cardiorespiratory fitness, where full weight bearing is contraindicated (4,12). To achieve metabolic oxygen demands equivalent to intensities from 50% to 80% of V˙O2max on the land, WTR parameters have to be changed from those used on LTR.
This study recorded the cardiovascular and metabolic responses of trained runners at 50%, 60%, 70%, and 80% of their land V˙O2max. All subjects were able to exercise on the water treadmill at intensities equivalent to 80% of their V˙O2max and at intensities equivalent to 55% to 94% of their land HRmax. These intensities fall within the recommendations of the American College of Sports Medicine to improve or maintain cardiorespiratory fitness (1).
We found that the most practical way to induce similar metabolic demands of running on a water treadmill to that of running on a land treadmill is by monitoring HR. The results of this study indicate that individuals can select a treadmill speed during WTR that elicits an HR of approximately 7 bpm less than their LTR to obtain an equivalent cardiorespiratory overload (Fig. 2). Svedenhag and Seger (11) reported that HR during running in water with a buoyancy jacket was 8-11 bpm lower for a given V˙O2 than during LTR. In contrast, Pohl and McNaughton (10) reported that V˙O2 and HR were higher while walking on a water treadmill when compared with land walking. One explanation for the discrepancy between our findings and those of Pohl and McNaughton (10) is that they had subjects walk on a water treadmill with the water level at mid thigh. A lower water level results in less buoyancy. They also reported that when water levels were increased to waist depth, the V˙O2 and HR decreased because of the increase in buoyancy, which supports the findings of this and other research (11). Buoyancy is thought to reduce the metabolic demands of exercising in water. Buoyancy decreases the weight-bearing component of exercise by increasing the subject's hang time while running, thereby decreasing physical effort and HR.
The viscous properties of water also play a role in water running. Water is at least 800 times more viscous than air (9), creating a greater resistance and longer muscle contraction time. During WTR, subjects took 22 fewer strides per minute than running at the same treadmill speed on land. Our data concur with those previously reported (5,12,16). Subjects in our study had increased resistance, because of water viscosity, with minimal hang time, thus maintaining a similar metabolic demand (V˙O2) as on land. The lower HR at any given V˙O2 during WTR compared with LTR may be because of the additional hydrostatic pressure associated with WTR. Increased hydrostatic pressure while running in the water is thought to contribute to a central shift in blood volume, which facilitates a greater venous return and a greater stroke volume, thereby decreasing HR while running in the water (5). Therefore, the water treadmill could be beneficial to injured and healthy individuals who are seeking an exercise alternative that is metabolically comparable to land-based running while minimizing impact forces.
The results of this study apply to males and females alike. Gender differences existed in the absolute HR-V˙O2 relationship (Fig. 6) but not in the relative %HRmax-%V˙O2max relationship (Fig. 7) among the three running conditions. The gender differences observed in the absolute HR-V˙O2 relationship (Fig. 6) among the three running conditions may be because male subjects had higher V˙O2max values than their female counterparts (Table 1). The V˙O2max in men (59.6 ± 4.6 mL·kg−1·min−1) was, on average, 8.7 mL·kg−1·min−1 higher than that in females (50.9 ± 2.9 mL·kg−1·min−1). Compared with males, at any given V˙O2, females had a higher HR. This can be explained by the fact that any given V˙O2 represented a higher percentage of the females' V˙O2max. When the HR-V˙O2 relationship was compared relative to maximal values (Fig. 7), there were no gender differences. Therefore, exercise intensity on the basis of HR for WTR should be based on relative terms (i.e., %HRmax or %V˙O2max).
The use of the water running shoe while running on a water treadmill provides additional training options. The AQinc shoe used in this study creates greater resistance during water running because of the small cups placed on the sole of the shoe (Fig. 1). The effect that water running shoes have on the relationship between V˙O2 and treadmill speed is plotted in Figure 5 The results indicate that, at any given treadmill speed, V˙O2 was, on average, significantly increased (4.12 mL·kg−1·min−1, P < 0.05) during WTR with the water running shoes than without the shoes. Practically, this related to subjects being able to run approximately 16 m·min−1 slower when wearing the water running shoe compared to when running without the shoe to get the same cardiovascular response. Likewise, at any given water treadmill speed, wearing the water running shoe elicited a higher V˙O2 response than when not wearing the water running shoe. Athletes should recognize that wearing or not wearing a water running shoe during WTR will affect their cardiorespiratory responses to WTR and should therefore adjust treadmill speed accordingly to achieve the desired HR response. In the event that athletes cannot achieve the desired exercise intensity because of limitations in the water treadmill maximum speed (12.07 km·h−1), for instance, the workload can be increased with the addition of the use of water running shoes. In addition, the water running shoe provides padding between the subjects' feet and the treadmill surface, which tends to make training on the water treadmill more comfortable.
In conclusion, the water treadmill provides athletes an alternative method of training to maintain cardiovascular fitness without the weight bearing demands of land running. By monitoring HR, both males and females can use WTR to induce a similar overload as LTR. Subjects should select water treadmill speeds that elicit an HR response that is 7 bpm less than their typical training HR during land-based running. Wearing the AQinc water running shoe increases the metabolic demand (i.e., V˙O2) at any given water treadmill speed. Further research reporting the usefulness of WTR during rehabilitation of lower extremity injuries would supplement the findings of this research. A biomechanical analysis of WTR compared with LTR could also be a useful information for people who are interested in WTR for cross-training.
Partial funding for this research was received from the Mary Lou Fulton Chair and AQx Sports, Inc.
The results of the present study do not constitute endorsement by American College of Sports Medicine.
3. Brown S, Chitwood L, Beason K, McLemore D. Deep water running
physiologic responses: gender differences at treadmill-matched walking/running
cadences. J Strength Cond Res
4. Eyestone E, Fellingham G, George J, Fisher G. Effect of water running
and cycling on maximum oxygen consumption and 2-mile run performance. Am J Sports Med
5. Frangolias D, Rhodes E. Maximal and ventilatory threshold responses to treadmill and water immersion running
. Med Sci Sports Exerc
6. George JD. Alternative approach to maximal exercise testing and V˙O2max
prediction in college students. Res Q Exerc Sport
7. Hall J, Grant J, Blake D, Taylor G, Garbutt G. Cardiorespiratory responses to aquatic treadmill walking in patients with rheumatoid arthritis. Physiother Res Int
8. Larsen GE, George JD, Alexander JL, Fellingham GW, Aldana SG, Parcell AC. Prediction of maximum oxygen consumption from walking, jogging, or running
. Res Q Exerc Sport
9. Miyoshi T, Shirota T, Yamamoto S, Nakazawa K, Akai M. Functional roles of lower-limb joint moments while walking in water. Clin Biomech (Bristol, Avon)
10. Pohl M, McNaughton L. The physiological responses to running
and walking in water at different depths. Res Sports Med
11. Svedenhag J, Seger J. Running
on land and in water: comparative exercise physiology. Med Sci Sports Exerc
12. Town G, Bradley S. Maximal metabolic responses of deep and shallow water running
in trained runners. Med Sci Sports Exerc
13. Vehrs PR, Fellingham GW. Heart rate and V˙O2
responses to cycle ergometry in white and African American men. Meas Phys Educ Exerc Sci.
14. Vehrs PR, George JD, Fellingham GW. Prediction of V˙O2max
before, during, and after 16 weeks of endurance training. Res Q Exerc Sport
15. Whaley MH, ed. ACSM's Guidelines for Exercise Testing and Prescription
. 7th ed. Philadelphia (PA): Lippincott Williams & Wilkins; 2006. p. 22-7.
16. Wilder R, Brennan D, Schotte D. A standard measure for exercise prescription for aqua running
. Am J Sports Med
Keywords:©2010The American College of Sports Medicine
CARDIOVASCULAR TRAINING; REHABILITATION; LOWER EXTREMITY; RUNNING