ACSM Cycle Ergometry Equation Validity in Trained Cyclists

Introduction

Several methods have been developed to estimate oxygen consumption (VO2) during exercise. The American College of Sports Medicine (ACSM) developed equations to predict the energy cost of various activities, including walking, running, and arm and cycle ergometry. The ACSM cycle ergometry equation uses pedal frequency (rpm), distance of flywheel travel (meters), applied resistance to the flywheel (kp), and an estimation of resting metabolism to predict oxygen cost during submaximal cycle exercise between 50 and 200 watts (W). This equation appears as:

VO2 = (kg·m·min⁻¹ × 2 ml·kg·m⁻¹) + (3.5 kg·m·min⁻¹ × M)

where VO2 is in ml·min⁻¹ and M is the subject's body mass in kg (Franklin, 2000).

Methods

The variability in direct VO2 measures has been shown to have a standard error of the estimate of up to 7%; the variability when using the prediction model is even greater (Stanforth et al., 1999). Recent studies have focused on determining the validity of the ACSM prediction equation for cycle ergometry (Lang, Latin, Berg, & Mellion, 1992; Stanforth et al., 1999). In general, these studies have shown the ACSM prediction equation to underestimate the actual VO2 by 0% to 16% at power outputs from 30 to 150 W.

The primary aim of this study was to examine the validity of the ACSM prediction equation for estimating oxygen consumption during submaximal cycling in aerobically trained males. The alternate hypothesis was that the estimated vs. actual VO2 during submaximal cycle ergometry would be similar from 50–200 W. A secondary focus was to analyze and compare the accuracy of the equation between two distinct groups of subjects: trained male cyclists vs. aerobically trained male non-cyclists.

This study used a prospective, nonrandomized, noncontrolled study design. A convenience sample of 14 apparently healthy males (26.4 ± 1.2 yr, 179.1 ± 1.8 cm, 79.5 ± 3.4 kg, BMI 24.8 ± 0.9) volunteered for the study. Inclusion criteria included: (a) all "no" answers on the PAR-Q questionnaire, (b) no contraindications to exercise as indicated by responses on the health history questionnaire, (c) informed consent given, and (d) BMI of less than 32 kg·m⁻². Seven subjects were trained cyclists, defined as those who had cycled at least three times per week for at least 30 minutes per training session for a minimum of six months prior to testing. Seven subjects were aerobically trained non-cyclists who engaged in aerobic activity at least three times per week for at least 30 minutes per session during the previous six months; however, their cycling did not exceed two sessions per week or 15 minutes per session. No subject in this group had ever cycled more than 50 miles per week for four or more consecutive weeks.

Subjects were instructed to avoid food, alcohol, and tobacco for three hours prior to testing and to avoid strenuous physical activity for the 24 hours prior to the test. Each subject's height and weight were recorded. Three-site skinfold tests to estimate body density were performed with a Lange skinfold caliper (Beta Technology, Inc., Santa Cruz, CA). Body fat percentage was estimated using the Siri equation. A cycle ergometer (Monark, Vansbro, Sweden) was used to conduct each test. Seat height was adjusted to allow for 5–10 degrees of knee flexion for each subject.

Results

Subjects pedaled at 65 rpm against no resistance for one minute to warm up. Resistance was then increased by 0.75 kg·m·min⁻¹ per stage to yield workloads of 50, 100, 150, and 200 W. Each stage lasted three minutes, provided steady state had been achieved. To verify steady state, the absolute difference in heart rate between the second and third minute of each stage could not exceed five beats per minute. When this criterion was not met, the stage continued at the same workload until the heart rate response satisfied the criterion. Gas exchange variables were analyzed using a SensorMedics gas exchange system (Yorba Linda, CA), calibrated prior to each test. Heart rate was monitored with a 3-lead single-channel electrocardiogram.

Statistical analyses were performed with SPSS 11.5 (SPSS, Inc., Chicago, IL). Means and total error were calculated to show the accuracy of predicted oxygen cost values. Pearson correlation coefficients were used to quantify the relationship between predicted and actual oxygen consumption. A paired samples t-test was used to compare predicted and actual test-group means. An independent t-test compared VO2 differences between cyclists and non-cyclists. Significance was set at α = 0.05 for all tests. All values are reported as mean ± SE.