The Wingate anaerobic power test (WAnT) is a widely used protocol for assessing anaerobic power characteristics that has been shown to be valid and reliable (2,4,8). The 30-second cycle ergometer protocol requires a maximal effort by the subject against a resistance based on body weight (traditionally 0.075 kp·kg−1 body weight). Results of this test provide data including peak, mean, and minimum power output and percentage of fatigue over the 30-second period. Though widely used, the WAnT often elicits mild to severe physical discomfort in subjects after completion of the test; Marquardt et al. (9) reported decreased physical discomfort in 93% of their subjects following a 20-second protocol compared to a traditional 30-second WAnT. Common complaints include nausea and vomiting (6) and slight to severe headaches, localized muscle fatigue, and dizziness or syncope (6), which may affect testing validity or test-retest reliability.
For this reason, a shortened WAnT protocol that provides valid results when compared to data collected using the established procedure while minimizing negative side effects would be valuable to researchers and clinicians for testing anaerobic power. Smith and Hill (10) concluded that peak power was attained during the first 15 seconds of a WAnT protocol. Therefore, an abbreviated protocol of at least 15 seconds in duration would still provide adequate time to assess peak power output. Previous studies have attempted to establish the validity of a shortened WAnT protocol (3,9). Marquardt et al. (9), using a small sample (n = 14) of male subjects, found that maximum power output was not significantly different between 20- and 30-second testing procedures and that 30-second minimum power output could be accurately predicted from 15- and 20-second values of a 20-second protocol. Perhaps the most noteworthy outcome of this study was that 13 of 14 subjects reported decreased physical discomfort from the 20-second test. However, their study did not establish the validity of a shortened WAnT procedure in predicting mean power output or percentage of fatigue. Bar-Or et al. (3) found that a 15-second abbreviated WAnT was a safe and reliable measure of peak and mean power (over 15 seconds) for men and women of advanced age.
It has yet to be determined whether mean power values can be accurately estimated from an abbreviated WAnT protocol. Because mean power values from a shortened WAnT protocol cannot be directly compared to data from traditional 30-second tests, an abbreviated protocol must be able to accurately predict 30-second values for mean power to allow comparisons with previous studies. Therefore, the purpose of this study was to determine whether mean power and other commonly reported data for a traditional 30-second WAnT can be accurately predicted from an abbreviated 20-second WAnT protocol.
Experimental Approach to the Problem
Because of its well-established validity and reliability, the WAnT is widely used for assessing anaerobic power (2,4,8). This procedure is relatively simple to complete, requiring only a limited time commitment by the subject and basic equipment considerations. Additionally, normative data for a traditional 30-second WAnT protocol are available for a variety of populations (6). However, a variety of unwanted side effects have been associated with the completion of the WAnT (6). Marquardt et al. (9) reported a significant decrease in symptomatology following a 20-second test relative to completion of a traditional 30-second WAnT. However, the disadvantage of an abbreviated protocol is the inability to compare collected data with established norms. Therefore, data from a traditional 30-second WAnT was used as the basis for examining anaerobic power production in order to establish the predictability of 30-second WAnT values from an abbreviated protocol.
Subjects were 96 healthy female college students between 18 and 35 years of age (mean = 23.2 ± 3.5 years) recruited from university health and physical education classes and were believed to represent a heterogeneous sample ranging from untrained to trained. No attempt was made to quantify the training status of participants. Subject data are summarized in Table 1. Subjects were within the average range for females for body mass index (21.84 ± 2.4) and body fat percentage (22.46 ± 6.36%) (1). Before participation, each subject completed an institutional human subjects committee-approved written informed consent form as well as a Physical Activity Readiness Questionnaire to screen for contraindications to participation in the study. Before participation in the test protocol, subjects were familiarized with the testing protocol including instruction as to the importance of giving a maximal effort without pacing, throughout the 30-second Wingate test. All subjects completed a shortened (15 seconds) familiarization trial. Subjects who felt they could not provide a maximal effort throughout the testing protocol were not included in the testing group and removed from the study.
Height was collected for each subject using a stadiometer and weight was determined using a Detecto Certifier Scale. Body mass index (weight (kg) ÷ height (m2)) was calculated for each subject. Body composition was estimated using the 7-site skinfold method described by Jackson et al. (7).
All data were collected on a mechanically braked Monarch 818e cycle ergometer. This ergometer is specifically designed for WAnT testing, allowing tension to be applied to the flywheel using the basket technique. The ergometer is equipped with a commercially available photoelectric data collection system (Sports Medicine Industries, Inc.) that measures flywheel revolutions by bouncing light off of a set of reflective markers on the flywheel. Each subject completed a 5-minute low-intensity warm-up (moderate pedal rate with no resistance) including a 4- to 5-second sprint interval at the end of each minute of warm-up. After a recovery interval of 2-5 minutes, each subject completed a standard 30-second Wingate test (6) using a resistance of 0.075 kp·kg−1 body weight. Data representing both the 20- and 30-second WAnT protocol were collected from a single 30-second period using Sports Medicine Industries, Inc. Power 2000 program version 1.02. This software reports power output second by second throughout the test period. Peak and minimum power values were determined as the average of the highest and lowest 5-second interval, respectively. Mean power was determined based on average values over the entire 30-second period. Data for 10, 15, 20, and 30 seconds were determined as the single-point power values for each designated time. All subjects were instructed to complete an active cool-down of at least 5 minutes of low-intensity pedaling. Subjects who experienced dizziness, headache, or nausea were monitored at the testing facility until fully recovered. A limitation of the present study was that blood lactate values and acid-base responses were not determined.
Data were analyzed using the SAS version 9.01 stepwise regression procedure. Subjects were randomly divided into 2 groups. A group of 60 subjects was used to develop regression equations to predict mean power, minimal power, 30-second power values, and percentage of fatigue from data collected during the first 20 seconds of the test. Validity of the equations was tested by applying the equations to the remaining 36 subjects. No significant differences were revealed between predicted and actual scores using 1-way analysis of variance (ANOVA). Therefore, the subject groups were combined, and final regression equations were developed using all 96 subjects. Two separate sets of final regression equations were developed, one including data up to 20 seconds and one including data only up to 15 seconds, to simulate abbreviated WAnT protocols of 15 and 20 seconds in length. Bland-Altman plots were developed to compare goodness of fit (5). The α level was set at p ≤ 0.05.
Summary results for the WAnT protocol are shown in Table 2. Average peak power values (473.89 ± 78.73 W) were slightly below those reported by Inbar et al. (6) for females (average range, 482-524 W; N = 18) within the 18-25 age group. The average mean power (370.19 ± 60.92 W) was slightly above averages reported by Inbar et al. (6) (317-252 W; N = 18). The average minimum power was 278.8 ± 56.83 W, while the average fatigue percentage was 40.78 ± 9.95%. Results of the equation validity testing procedure, described previously, showed no significant differences between predicted and actual values for any of the WAnT test parameters using 1-way ANOVA.
The R2 values were considerably different between regression equations developed using only the first 15 seconds of data and those established from inclusion of 20 seconds of WAnT data. The inclusion of 20-second data resulted in a 2.7% increase in R2 value for mean power from 0.9654 to 0.9920, an 11% increase in R2 value for minimum power from 0.7963 to 0.9064, a 13.7% increase in R2 value for 30-second power from 0.7033 to 0.8407, and a 28.9% increase in R2 value for fatigue from 0.5482 to 0.8377. Comparison of R2 values for simulated 15- and 20-second protocols is shown in Table 3.
Data describing 20-second prediction equations are shown in Table 4. The equation for mean power accounted for the greatest amount of variance (R2 = 0.9920); maximum power and 10-, 15-, and 20-second power values were significant predictors in the most parsimonious model. The average actual mean power was 370.19 W, while the predicted average mean power was 370.17 W. An R2 = 0.9064 was attained when predicting minimum power; significant predictors included 15- and 20-second power values. An R2 = 0.84 was attained when predicting both percentage of fatigue and 30-second power; significant predictors included 15-second, 20-second, and peak power and 15- and 20-second power values, respectively. Retrospective power analysis produced statistical power values of ≥0.86 for all predictors of each dependent variable with the exception of percentage of fatigue. The observed power of the 15-second value within the regression procedure predicting percentage of fatigue was 0.669. The remaining source variables (20-second and maximum power) produced statistical power values of 0.999 within the regression procedure predicting percentage of fatigue.
Bland-Altman plots demonstrating goodness of fit for mean power, minimum power, 30-second power, and percentage of fatigue are shown in Figures 1-4. Delta values (predicted minus actual) for no more than 6 subjects fell outside of the ±2 SD range for any of the WAnT variables analyzed. Despite the seemingly large size of the 2-SD range, this range represented only a small percentage of the overall range of absolute values for each parameter. For example, the 2-SD value of 10.9 W for mean power represented only 3.2% of the overall range (342 W). The 2-SD value of 7.8 W represented 15.2% of the overall range (51.2 W) for fatigue percentage. The other two WAnT parameters measured (minimum power and 30-second value) fell between WAnT mean and fatigue. The 2-SD value of 34.66 W represented 10.9% of the overall range (320 W) for minimum power, while the 2-SD value of 44.04 W represented 13.9% of the overall range (316 W) for fatigue.
The results of this study suggest that, because of the near-linear decrease in power over the 30-second period, the parameters measured during a 30-second WAnT protocol can be accurately predicted for college-age women based on data collected during an abbreviated (20 seconds) WAnT procedure (Figure 5). Perhaps the most used parameters measured in WAnT testing for comparison to normative data are peak power and mean power. The results of the present study show that these values can be accurately assessed from a 20-second protocol, with peak power measured directly and mean power effectively estimated. Therefore, investigators concerned with the negative side effects associated with the traditional 30-second protocol may substitute a 20-second protocol without sacrificing the ability to accurately assess mean power.
The findings of the present study agreed with those of a previous study by Marquardt et al. (9) who found that minimum power could be effectively predicted from a 20-second WAnT protocol. Unlike their study, the present study did not measure differences in blood lactate level (BLa) attained during a 30- vs. a 20-second WAnT protocol. However, though Marquardt et al. (9) found a significant difference in BLa between the 20- and 30-second tests, this difference is likely of little importance in most test settings since parameter results of the WAnT alone may be used to assess anaerobic power without the required invasiveness of blood collection.
The ability to predict percentage of fatigue and 30-second values from a 20-second test was lower than for mean power but still relatively high (R2 = 0.84 for both). However, the value of these 2 parameters in assessing anaerobic power via the WAnT protocol is questionable. Because 30-second value is a single-point measure, minimum power, measured as the average of the lowest 5-second power values, is more representative of the actual performance. Similarly, the meaning of percentage of fatigue as measured during the WAnT protocol and its relationship to other factors associated with anaerobic capacity are unclear (6). Therefore, the lower predictability for fatigue rate and 30-second power values from a 20-second test does not appear to limit the usefulness of an abbreviated WAnT protocol.
If peak power were the only variable of interest, the WAnT protocol could possibly be shortened to 10 seconds. However, examination of the data from the present study revealed that estimations of mean power, minimum power, 30-second power, and fatigue, as determined in a traditional 30-second WAnT, are significantly improved through the inclusion of 20 seconds of data over estimations using only 15 seconds of data. Though the development of a WAnT protocol that is even shorter than the 20-second test considered in this study may seem more beneficial for decreasing the negative side effects associated with the traditional 30-second WAnT protocol, the regression equations developed using only 15 seconds of data were less effective in predicting 30-second WAnT values for each of the parameters assessed (Table 3). Additionally, because Marquardt et al. (9) found that a 20-second protocol was sufficient to decrease side effects in 93% of participants, it does not seem that the possible additional benefits from a further decrease in test length would outweigh the decreased predictability of a 15-second test.
A possible limitation of the current study is the assumption that a power curve generated during a 20-second protocol will be nearly identical to the power curve generated during the first 20 seconds of a traditional 30-second WAnT. However, a maximal effort without pacing on the part subjects is an assumption that is inherent to a traditional 30-second WAnT protocol. Attempts to conserve energy and submaximal efforts result in invalid test results (6) and nonlinear force curve declines. Therefore, if subjects complied with these instructions in both a standard and abbreviated test, power curves should be nearly identical. Evidence of subject compliance in the current study is demonstrated by the near-linear power decline seen in Figure 5.
It is worth noting that during the stepwise regression procedure, the percentage of body fat (%BF) reached significance for entry into the overall model for minimum power and 30-second power values. However, entry of %BF into the model only increased the R2 by 0.52% and 1.05% for minimum and 30-second power, respectively. Because %BF is an external variable, not measured directly from the WAnT procedure, and because of the limited increase in predictability from its entry into the stepwise procedure, %BF was not included in the final regression analysis
The WAnT is a widely used procedure that often produces unwanted physical side effects that may affect athlete/client motivation for testing as well as testing validity and reliability (4). Based on the individuals' previous testing experiences, coaches and practitioners may find it difficult to convince athletes or clients to participate or to obtain accurate and meaningful results if participants do not give maximal effort for fear of these side effects. However, because an abbreviated test may reduce the incidence of these side effects (9), a shortened protocol capable of effectively assessing the parameters measured in a traditional 30-second test would be highly valuable for practitioners wishing to use the WAnT in assessing anaerobic power. The results of this study showed that mean power, minimum power, percentage of fatigue, and 30-second power can all be accurately predicted from an abbreviated (20 seconds) WAnT protocol. Therefore, in conclusion, practitioners may use a 20-second WAnT with college-age females in an attempt to decrease negative side effects without sacrificing the ability to accurately assess maximum power and mean power, the 2 WAnT parameters most often used for comparison to normative data.
1. American College of Sports Medicine, Roitman, JL and Herridge, M, ACSM's Resource Manual for Guidelines for Exercise Testing and Prescription
(2nd ed.) Philadelphia: Lippincott Williams & Wilkins; 1993. p. 235.
2. Bar-Or, O. The Wingate anaerobic test. An update on methodology, reliability and validity. Sports Med
4: 381-394, 1987.
3. Bar-Or, O, Berman, L, and Salsberg, A. An abbreviated Wingate anaerobic test for women and men of advanced age. Med Sci Sports Exerc
24(Suppl.): S22, 1992.
4. Bar-Or, O, Doktan, R, and Inbar, O. A 30-sec all-out ergometric test: its reliability and validity for anaerobic capacity. Isr J Med Sci
13: 326-327, 1977.
5. Bland, J and Altman, D. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
1: 307-310, 1986.
6. Inbar, O, Bar-Or, O, and Skinner, J. The Wingate Anaerobic Test
. Champaign, Ill: Human Kinetics; 1996. pp. 12, 16, 23, 87-88.
7. Jackson, AS, Pollock, ML, and Ward, A. Generalized equations for predicting body density of women. Med Sci Sports Exerc
12: 175-181, 1980.
8. Jacobs, PL, Mahoney, ET, and Johnson, B. Reliability of arm Wingate Anaerobic Testing in persons with complete paraplegia. J Spinal Cord Med
26: 141-144, 2003.
9. Marquardt, J, Bacharach, D, and Kelly, J. Predicting 30 second minimum power from a 20 second Wingate test. Med Sci Sports Exerc
10. Smith, JC and Hill, DW. Contribution of energy systems during a Wingate power test. Br J Sports Med
25: 196-199, 1991.
Keywords:© 2008 National Strength and Conditioning Association
Wingate anaerobic power test; shortened Wingate; Wingate prediction; Wingate side effects; anaerobic power testing