Efficiency of Walking and Stepping: Relationship to Body Fatness
The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Abstract
Objective: To determine energetic efficiency of walking and stepping in a heterogeneous normal adult population and its association with body fatness and to assess within- and between-individual variations.
Research Methods and Procedures: Using a combination of a whole-room indirect calorimeter and a large precision force platform, we simultaneously measured minute-by-minute energy expenditure (EE) and mechanical work during walking and stepping in healthy adult men (n = 60) and women (n = 85). Efficiency was calculated as a ratio (percentage) of mechanical work and EE of activity. Efficiency of walking and stepping performed at various intensities was compared for reproducibility within the same day (morning and afternoon) and correlated with a subject's characteristics.
Results: The efficiency of walking was negatively correlated with body fatness in both men and women at 0.9 to 1.2 m/s but positively correlated with body fatness in men and not correlated in women at the slowest speed tested (0.6 m/s). Efficiency of walking and stepping of various intensities was reproducible during the same day. Compared at similar EE levels, walking was more efficient than stepping (26% to 27% vs. 18% to 22%, p < 0.01). Women were significantly (p < 0.01) more efficient than men during stepping. Age, sex, body mass, fat-free mass, fitness (maximal oxygen uptake), height, and speed variations contributed to the between-subject differences in efficiency.
Discussion: Obese individuals were less efficient than lean individuals during normal-speed walking. Significant interindividual variations in efficiency of walking and stepping may be attributed to habituation and physical characteristics such as age, sex, and fitness level.
Introduction
The increased incidence of obesity has become a major public health problem in the U.S. and throughout the world (1,2). The most important factor for weight gain in most individuals without low resting energy expenditure (EE)1 (3) is the long-term imbalance between energy intake and energy expenditure of physical activity (EEACT). Low EEACT could be a result of a reduced physical activity level or increased energetic efficiency. The amount of physical activity plays an important role in the prevention of overweight/obesity and maintaining weight loss (4). However, the role of efficiency has not been extensively studied.
EEACT is determined by the ratio of mechanical work (MW) exerted by body movements during physical activities and the energetic efficiency of these activities. A decrease in the amount of MW (i.e., inactivity) or increase in efficiency will reduce the amount of EEACT. Efficiency of physical activity can be calculated from simultaneously measuring EEACT and MW under well-controlled laboratory conditions. Previous studies on the efficiency of physical activity were conducted mainly in small and homogeneous populations, such as athletes (5,6) or amputees (7,8), and were limited to short measurement intervals of a particular type of human movement (9). For general populations, an interesting question highly relevant to the obesity epidemic remains unanswered: Are obese people more efficient than lean individuals in performing common physical activities, such as walking and step climbing?
The purpose of this cross-sectional study conducted in a heterogeneous group of healthy adults was to assess efficiency of walking and stepping at close-to-normal conditions by simultaneously measuring minute-by-minute EE and MW using a combination of a state-of-the-art whole-room indirect calorimeter and a floor force platform system. The characteristics of efficiency, such as reproducibility between separate morning and afternoon exercise sessions, and within-subject variations with changes in activity types and intensities were measured. Our primary hypothesis was that the efficiency of common physical activity in a heterogeneous group of healthy normal adults could be associated with the subject's body fatness. In addition, we also further explored the within- and between-individual differences in efficiency of walking and stepping at various speeds in our study.
Research Methods and Procedures
Subjects
One hundred forty-five healthy adult volunteers (60 men and 85 women) were recruited from the Nashville area using posters, advertisements in the university periodicals, and word of mouth. All participants signed an informed consent form approved by Vanderbilt's Institutional Review Board. They were generally healthy nonsmokers, with no evidence of past or present metabolic disorders, including thyroid disorders or diabetes, weight stable (<2 kg of change in the last 6 months), not using medications or supplements known to change EE, and able to engage in normal walking and stepping. Women participants were studied during the follicular phase (within the first 10 days after menstruation) of their self-reported menstrual cycle to minimize the effects of hormonal changes on EE (10). All participants were encouraged to maintain their normal pattern of physical activity and dietary intake for the week before the study.
Experimental Procedures
The study consisted of an ∼24-hour stay in a room calorimeter during which the participants were asked to perform regular personal care activities and to engage in a standardized physical activity protocol consisting of walking diagonally across the room (3 m) and stepping up and down on a step platform (15.2-cm height). The written description of the protocol given to subjects included a list of segments, sequence, speed or pace, and duration. Walking and stepping were close in intensity and energy cost to activities participants would perform outside the chamber. Specifically, it consisted of a 10-minute practice period, three 10-minute walking periods (denoted W1, W2, and W3, with average speeds of 0.6, 0.9, and 1.2 m/s, respectively), and four 10-minute stepping periods (denoted S1, S2, S3, and S4, with average speeds of 0.2, 0.3, 0.4, and 0.5 steps/s, respectively). Self-termination of the activity was allowed only if the subject felt that the intensity was too high to sustain. A 10-minute rest period was scheduled between each 10-minute walking and stepping period and a 20-minute rest between walking and stepping segments. During the walking and stepping segments, subjects followed the appropriate exercise cadence set by a metronome, such that all subjects would exercise at closely comparable speed and pace. The exercise protocol was performed twice, in the morning (9:30 AM to 12:30 pm) and afternoon (2:30 to 5:30 pm). To minimize the thermic effect of food, exercise sessions were scheduled between meals (11). Meals, designed to achieve approximate energy balance, were provided at 7:30 am, 1:00 pm, and 6:00 pm. After an overnight sleep and fast, resting EE was measured during a 30-minute period (6:15 to 6:45 am) while the subject was awake but resting quietly in bed.
Activity-Energy Measurement System (AEMS)
The AEMS is housed in the General Clinical Research Center at Vanderbilt University Medical Center. It combines a whole-room indirect calorimeter unit with a large precision force platform for the measurements of EE and MW, respectively. The force platform, measuring 2.5 × 2.5 m, covers the entire living area inside the room calorimeter and is supported by multiple transducers. During a subject's stay, the platform allows computer-aided measurement (60 times/s) of body position, displacement, and mechanical forces with an accuracy of >97%. The AEMS allowed simultaneous measurement of EE and MW as described previously (12,13).
Measurement of MW
Body motion in terms of speed was defined as the distance a subject travels over time in the room calorimeter. MW was estimated by calculating force, acceleration in x, y, and z axes (time derivative from speed), and body mass, as described previously (12). Measurements included both horizontal and vertical MW components, which provided an accurate measure of the distance and speed in which a subject traveled inside the room.
Measurement of EE
EE was derived from the rates of oxygen consumption and carbon dioxide production using measured concentrations of O2 and CO2 in the air inside the room calorimeter and multiplying the flow rate of the purged air. This allowed variables such as EE to be calculated on a minute-by-minute basis (13), with the accuracy of >99% for a 24-hour study period and >90% within a minute. Such methodology allowed precise measurements of EEACT during short bouts of exercise. EEACT was defined as the increase in EE during walking and stepping above the resting EE, which was measured during a 30-minute waking period after an overnight fast. The force platform monitored body motion during the resting EE measurements. Detected movements were removed from resting EE calculations as described previously (14). In each exercise bout, the rapid increase/decrease in MW was marked as the starting and ending time of each exercise bout. EEACT was then integrated as the area under the EE curve between this period, whereas the resting EE was subtracted before the integration. EEACT was also established in terms of its relative intensity [EEACT/resting EE, or metabolic equivalents (METs)] during each exercise. If the duration of the exercise bout was <5 minutes, it was marked as an incomplete trial. These processes were performed using automated computer programs to minimize tester error. The respiratory quotient (RQ) was calculated as the ratio of the carbon-dioxide production and oxygen consumption in each minute.
Exercise Efficiency
Exercise efficiency was defined as the MW performed by a subject during each walking and stepping bout divided by the associated EEACT (efficiency = MW/EEACT) (15). Figure 1 demonstrates an example of the MW and EEACT output and the determination of the efficiency parameter.
A sample output of simultaneous measurements of EE and MW from the whole-room indirect calorimeter/force platform system for a 24-hour study period and efficiency derivation of an exercise bout.
Anthropometric and Body Composition Measurements
Before entering the AEMS in the morning of each study visit, body mass was measured to the nearest 0.05 kg using a digital scale. Height was measured to the nearest 0.5 cm using a stadiometer. Fat mass (FM) and fat-free mass (FFM) were determined by hydrodensitometry (underwater weighing) within a week of the study visit. The subjects were weighed underwater, and their residual lung volume was measured using the nitrogen dilution technique while subjects were submerged in water to chest level (16). Body fat percentage (% fat) was calculated from body density using Schutte's equation (17) for African Americans and Siri's equation (18) for all others, whereas FM and FFM were calculated from the subject's body mass.
Maximal Fitness Testing (Maximal Oxygen Uptake)
Also within a week of the study visit, physical fitness was assessed by measuring the maximal oxygen consumption during the standard Bruce treadmill protocol (19). The maximal oxygen uptake score is expressed in milliliters per kilogram of body weight per minute.
Statistical Analysis
Calculations of EEACT, MW, and efficiency were performed using automated programs written using the MATLAB software package (version 12, MathWorks Inc., Natick, MA). Statistical analyses were performed using the SPSS/PC statistical program (version 11.0 for Windows; SPSS, Inc., Chicago, IL).
All descriptive statistics were presented as mean ± SD. Differences between men and women in demographic and measured parameters were compared using independent-sample Student's t tests. Morning and afternoon comparisons were made in each walking and stepping bout using paired Student's t tests. Pearson's correlation coefficients (r) were computed to examine the morning and afternoon efficiencies to further validate reproducibility. For men and women separately, repeated measures ANOVA and contrasts were conducted to confirm that parameters of exercise intensity varied from one level of exercise to the next and to examine gender differences in intensity across exercise levels. Similar analyses were conducted to examine patterns of exercise efficiency across levels of exercise intensity. ANOVA (Tukey's test) was used to compare differences among efficiencies during different intensities of walking and stepping, separately. To compare differences in efficiency between walking and stepping, we performed paired Student's t test on efficiency for two bout pairs (W2 and S1 and W3 and S3), which had similar intensities of EEACT (expressed in METs).
Between-subject differences in efficiency were explored among different bouts and subject characteristics using Pearson's correlation coefficients. Particularly, the relationships between efficiency and body composition were demonstrated with scatter plots and linear regression analysis. Multiple step-wise regression analyses were then conducted to ascertain how subject characteristics influenced efficiency across levels of intensity for men and women separately. Independent variables included in the model were: age, weight, % fat, FFM, height, and maximal oxygen uptake. Furthermore, considering that between-subject variances in speed could influence efficiency, the difference between actual measured and designed speeds for walking W1 to 3 and between measured and group average speeds for stepping S1 to 4 for each individual was entered in the regression model as a covariate. All statistical analyses used two-sided p < 0.05 for significance.
Results
Of the 145 subjects participating in this study, 139 completed all exercise bouts within each exercise protocol.
Subject Characteristics
Table 1 presents descriptive data for the study participants. Compared with women, men were heavier, taller, had lower % fat and FM, and higher FFM.
| Women (n = 85) | Men (n = 60) | |
|---|---|---|
| Body mass (kg)* | 76.1 ± 23.3 (46.0 to 154.4) | 88.9 ± 20.6 (55.5 to 143.5) |
| Height (cm)* | 163.5 ± 6.1 (150.5 to 184.0) | 178.1 ± 6.9 (162.0 to 191.0) |
| Age (years) | 36.5 ± 9.6 (20.0 to 62.0) | 35.6 ± 10.7 (18.0 to 57.0) |
| BMI (kg/m2) | 28.3 ± 8.0 (17.8 to 53.4) | 28.0 ± 6.4 (18.8 to 45.5) |
| Body fat (%)* | 34.1 ± 10.5 (5.4 to 58.9) | 23.5 ± 9.9 (7.9 to 44.9) |
| FFM (kg)* | 48.1 ± 8.5 (35.3 to 83.4) | 66.4 ± 8.7 (50.2 to 90.5) |
| Maximal oxygen uptake (mL/kg per minute)* | 28.8 ± 7.7 (15.1 to 47.8) | 40.2 ± 10.9 (20.2 to 62.0) |
- Values are mean ± SD and (ranges).
- * Significantly different between men and women (p < 0.05).
Intensity of Walking and Stepping
The intensity parameters of walking and stepping are represented by speed, RQ, and EEACT (in METs) (Figures 2 A to C, respectively). The differences between the morning and afternoon sessions were nonsignificant in terms of speed (<5.1%, p > 0.64), RQ (<3.4%, p > 0.84), or EEACT (<7.1%, p > 0.45) for each walking or stepping bout.
Mean and SD of speed (A), RQ (B), and EEACT (C) (EEACT in METs, EEACT/resting EE) during each exercise bout. (W1, W2, and W3) Walking at 0.6, 0.9, and 1.2 m/s. (S1, S2, S3, and S4) Stepping at 0.2, 0.3, 0.4, and 0.5 steps/s. (Solid bars) Men. (Open bars) Women.
Efficiency of Walking and Stepping
Within-Subject Differences
There was no significant difference (all paired Student's t tests p > 0.4) in efficiency of walking and stepping (W1 to 3 and S1 to 4) performed in the morning or afternoon. Furthermore, significant Pearson's correlation coefficients between the morning and afternoon bouts of walking and stepping confirmed that the efficiency of walking and stepping was reproducible within the same day (all r > 0.68, p < 0.01). Thus, for the further analyses, walking and stepping efficiency was calculated as the average efficiency of morning and afternoon bouts.
Efficiency during walking and stepping was significantly different among intensities (p < 0.01). Efficiencies of W2 and W3 were similar (p = 0.50) but were significantly (p < 0.01) higher than W1. With the increasing stepping rate, the general trend of efficiency increased then decreased (Figure 3), whereas efficiency of S1 was significantly (p < 0.05) less than S2 to 4. The effect of intensity, measured by the speed of walking or the pace of stepping, contributed significantly to the changes in efficiency (all p < 0.01). This effect was also significant after adjusting for sex (p = 0.01). When compared at similar EEACT (p > 0.05), efficiencies of W2 (25.8 ± 4.3%) and W3 (26.4 ± 4.5%) were significantly (p < 0.01) higher than S1 (18.2 ± 3.8%) and S3 (22.4 ± 4.2%), respectively (Figures 2C and 3).
Mean and SD of efficiency (MW/EEACT) during each exercise bout. (W1, W2, and W3) Walking at 0.6, 0.9 and 1.2 m/s. (S1, S2, S3, and S4) Stepping at 0.2, 0.3, 0.4, and 0.5 steps/s. (Solid bars) Men. (Open bars) Women. (*) Different between men and women, p < 0.05. (†) Difference between W1 and W2 to 3.
Between-Subject Differences
Men were more efficient than women (19.0 ± 5.3% vs. 17.1 ± 3.6%, p < 0.01) during W1 but were less efficient during S1 (17.2 ± 3.6% vs. 18.9 ± 4.0%, p = 0.03) and S3 (21.4 ± 4.0% vs. 23.3 ± 4.4%, p = 0.03), and the difference nearly reached significance for S2 (20.0 ± 3.6% vs. 21.7 ± 4.3%, p = 0.06). After adjusting for individual speed variations, women were more efficient during the overall stepping exercise than men (p < 0.01), but efficiency of walking was not different between sexes (p = 0.96).
The efficiencies were positively and significantly correlated among the exercise bouts (r = 0.23 to 0.83, p < 0.01), except those between W1 and S1 to 3 (p = 0.07, 0.12, and 0.12, respectively). The result of the correlation analysis among efficiencies and subject characteristics is shown in Table 2.
| Activity type | Body fat (%) | Age (years) | Body mass (kg) | Height (cm) | FFM (kg) | Vo2 max‡ [mL/(kg · min)] |
|---|---|---|---|---|---|---|
| W1 | −0.024 | −0.088 | 0.138 | 0.110 | 0.200 | 0.040 |
| W2 | −0.205* | −0.273† | −0.093 | 0.031 | 0.071 | 0.166* |
| W3 | −0.218† | −0.201* | −0.160 | 0.003 | −0.001 | 0.182* |
| S1 | 0.020 | −0.069 | −0.075 | −0.120 | −0.112 | −0.065 |
| S2 | 0.054 | −0.027 | −0.070 | −0.135 | −0.114 | −0.110 |
| S3 | 0.009 | −0.109 | −0.113 | −0.164* | −0.132 | 0.014 |
| S4 | −0.016 | −0.072 | −0.099 | −0.134 | −0.117 | 0.040 |
- * Significant at p < 0.05.
- † Significant at p < 0.01.
- ‡ Vo2 max, maximal oxygen uptake.
We found that significant (p < 0.04) and negative associations existed between body fatness (% fat) and walking efficiency during W2 and W3 for both men and women (Figure 4), contrasted by a positive association in men during W1. However, the associations were weak (slope not significantly different from 0, p > 0.05) for all stepping in men and women.
Relationship between body fatness (% fat) and efficiency of walking at various intensities: W1, W2, and W3, walking at 0.6, 0.9, and 1.2 m/s. For W2 to 3 in women and W1 to 3 in men, all linear regression slopes are statistically different from 0, p < 0.05.
Multiple linear regression analyses performed for each exercise bout generally indicated that the efficiency of walking in men and stepping in women significantly (p < 0.05) decreased with body mass. Efficiency during W1 and W2 decreased (p < 0.01) with age in women but not in men. FFM was positively (p < 0.05) correlated with efficiencies during W2 to 3 in men and during S1 to 3 in women.
Discussion
In this study, we assessed the efficiency of walking and stepping of various intensities in a heterogeneous group of healthy normal adults using simultaneous measurements of EEACT and MW. The major finding of this study is that the efficiency of walking at normal speed was negatively correlated with body fatness in both men and women. This finding does not support the notion that obesity is associated with increased efficiency of common physical activities, such as walking and stepping at normal speeds. We also found that efficiency of these activities was reproducible within the same day for each individual. Walking was significantly more efficient than stepping when the EEACT was comparable. Women also appeared to have higher stepping efficiency than men.
Efficiency of walking (20,21), running (5,6,20), cycling (22), and rowing (23) measured using various techniques have been reported. Most studies have shown within-individual variations of exercise efficiency changing with various types and intensities of the workload (9,22,24). These results are in general agreement with our findings in the current study. Using treadmills or cycle ergometers, some studies have detected changes in efficiency of physical activity with changes in body mass or body composition (25,26). For example, Dempsey et al. (25) reported in a cross-sectional study that as the workload increased, oxygen consumption increased more rapidly in obese than in lean subjects, leading to the decreased efficiency in obese subjects. Although the activity type and measurement methodology are different from our study, results are similar. Between-subject variations in efficiency may also be related to history of weight loss. Geissler et al. (27) reported that adults who lost weight had a lower EE during exercise (greater efficiency) than weight-matched controls with no history of obesity. Recently, weight change was found to be a robust and negative predictor of work efficiency by cycle ergometry in an interventional trial (28). However, other studies have not found that efficiency in humans is related to body mass or body fatness (29,30,31,32) after weight changes. Our subjects were all self-reportedly weight stable for the past 6 months or longer; thus, we were unable to explore such contribution.
In the present study, we observed large within-subject variations in efficiency of walking and stepping at various intensities. As illustrated in Figure 3, efficiency increased nonlinearly with increased walking and stepping intensity. These results are similar to the finding of Gaesser et al. (22) using a cycle ergometer. The nonlinear relationship between efficiency and work rate could be explained by the nonzero intercept of the regression of EEACT on MW (15,22). Such observations suggest that there is an optimal efficiency at particular activity intensity within a certain type of activity or exercise, which can be at least partially explained by energy optimizing mechanisms (20,21).
Efficiency of certain activities of certain types and intensities familiar to an individual could be affected or improved by habituation. Habituation could be viewed as a form of negative conditioning caused by long-term participation of a certain movement, thus leading to a decreased anxiety level (33), increased tolerance (5), and neuromuscular adaptations in muscle cells (34). We have previously demonstrated that efficiency during step aerobic exercise is higher in aerobic instructors compared with untrained controls (35).
Results from this study also demonstrated that even when the EEACT was comparable (W2-S1 and W2-S3), walking was a more efficient activity than stepping. There may be several factors contributing to this difference. In contrast to walking, stepping mostly consists of vertical movements. These movements require a constant exchange of kinetic and potential energy, which could result in a lower efficiency. Moreover, most individuals are likely to be more familiar with walking than with stepping up and down on a stepper.
Between-individual differences in efficiency in this population with a wide range of physical and physiological characteristics were the most important features of this study. For example, we found that women were more efficient than men in stepping. A plausible explanation of this difference could be a lower body mass and shorter height of women than men because both variables were negatively associated with stepping efficiency.
This study did not show an increase in efficiency of physical activity with the increase in body fatness. To the contrary, efficiency during level walking at normal speed (0.9 to 1.2 m/s or 2.0 to 2.7 mile/h, W2 and W3) decreased with increased body fatness. Our data also suggest that adult men and women with lower body fatness are likely to walk and step more efficiently than individuals with higher body fatness. This observation may be explained, in part, by the concept of habituation assuming that individuals with lower body fatness and higher FFM engage in such activities on a more regular basis than more obese individuals. The fact that during W1 (0.6 m/s or 1.4 mile/h), an abnormally slow walking pace, the correlation between body fatness and efficiency was positive in men might also be viewed as evidence of habituation. The practical implication of these findings should be viewed beyond the scope of efficiency itself because the general population who engages in physical activity has limited concern about efficiency or competitive performance. Many individuals exercise to lose weight by increasing their EE. Therefore, physical activity designed to maximize EE should involve different types and intensities of movement to minimize the effect of increased efficiency due to habituation.
This study has some limitations. First, walking across the force platform (3 m) and stepping up and down required repeated stopping and turning (walking), and this may not represent the normal patterns in the free living. Compared with continuous movements, we anticipate that such stopping and changing directions would reduce the efficiency due to changing momentum. It is unknown whether this effect would differ among subjects with different body fatness. Use of exercise equipment such as a treadmill or step master, however, would not allow us to measure MW and calculate efficiency. The negative association between the efficiency of walking and body fatness should not be considered as the association between EE and body fatness because the energy costs of physical activity are also dependent on the person's body weight and other parameters. The observed association between age and efficiency of walking in women warrants further investigation. Finally, such cross-sectional associations, although statistically significant (p < 0.05), may have limited clinical significance in terms of explaining the variance (r2 < 0.1) in efficiency of physical activity.
Our study has several strengths. The availability of several subject characteristics allowed us to evaluate associations of these parameters with the efficiency of physical activities commonly performed in everyday life. The study includes a relatively large (n = 145) number of men and women observed under strictly controlled laboratory conditions. To our knowledge, this is the first study to assess efficiency of walking and stepping in a heterogeneous group of adults with a wide spectrum of body fatness, which is an important step in exploring the relationship between energetic efficiency of physical activity and body weight regulation. The results could help in designing further studies using similar or different techniques intended to assess efficiency of physical activities in various populations.
In summary, we determined efficiency of walking and stepping by simultaneous measurements of EE and MW in a group of 145 heterogeneous healthy adults. We found large within-subjects variation in efficiency of walking and stepping at different intensities. Our null hypothesis that obese men and women walk and step more efficiently than leaner adults was not supported by the results of this study. Rather, we observed a decrease in efficiency of walking at normal speed with the increase in body fatness in both men and women.
Acknowledgment
This research has been supported by NIH Grants RR00095, DK26657, DK02973, and DK46084. The authors thank J.L. Darud, C.N. Sun, and L. Baker for their help in conducting the experiments and data processing and H. Gwirtsman and K. Virts for their editorial review. We also thank personnel of the General Clinical Research Center at Vanderbilt University Medical Center for their professional assistance.
