Altering Work to Rest Ratios Differentially Influences Fatigue Indices During Repeated Sprint Ability Testing

Abstract

La Monica, MB, Fukuda, DH, Beyer, KS, Hoffman, MW, Miramonti, AA, Riffe, JJ, Baker, KM, Fragala, MS, Hoffman, JR, and Stout, JR. Altering work to rest ratios differentially influences fatigue indices during repeated sprint ability testing. J Strength Cond Res 30(2): 400–406, 2016—This study examined the influence of recovery time on fatigue indices, performance (total work [TW], peak power [PP], and mean power [MP]), and oxygen consumption during repeated sprint ability (RSA) on a cycle ergometer. Eight recreationally-trained men performed 3 RSA protocols consisting of 10 × 6 s sprints with 12 s, 18 s, and 24 s rest intervals between each sprint. Fatigue indices were determined as percent decrement (%Dec) and rate of decline using either a log transform method or standard slope approach for TW, PP, and MP during respective RSA protocols. The maximal V[Combining Dot Above]O₂ value in response to given sprint intervals and the minimal V[Combining Dot Above]O₂ value in response to given rest periods (V[Combining Dot Above]O₂work and V[Combining Dot Above]O₂rest, respectively) were recorded. A repeated measures analysis of variance was used to analyze all variables. Average V[Combining Dot Above]O₂work was not different among rest interval trials. Average V[Combining Dot Above]O₂rest with 12 s rest was greater than 18 s and 24 s (2.16 ± 0.17 L·min⁻¹, 1.91 ± 0.18 L·min⁻¹, 1.72 ± 0.15 L·min⁻¹, respectively), while 18 s was greater than 24 s. Average TW and MP were greater with 24 s rest than 12 s (4,604.44 ± 915.98 J vs. 4,305.46 ± 727.17 J, respectively), with no differences between RSA protocols for PP. No differences in %Dec were observed. Both methods of calculating rates of decline per sprint for PP and TW were greater during 12 s than 18 s or 24 s. Since changes were only noted between the 12 s and 24 s protocols, a 6 s differential in rest intervals may not be enough to elicit alterations in TW, PP, MP, or %Dec in RSA performance. Rate of decline may be a more sensitive measure of fatigue than %Dec.

Introduction

Interval training has become a popular method incorporated into many conditioning programs for individuals and team sport athletes from recreational to elite levels (^12,19,26,38). Individuals who can maintain near maximal sprints for a longer duration have a greater resistance to fatigue which allows them to perform at higher levels further into a competition or training session (^12,15,19,27). In an effort to quantify potential performance, and training response, exercise scientists have developed protocols to measure repeated sprint ability (RSA), which is defined as the athlete's ability to recover and maintain maximal effort during successive sprints (¹³). The appropriate RSA protocol should be representative of the work to rest ratio similar to that required in the competitive environment (³⁸). For example, a range of work to rest ratios (1:1–1:5) have been recommended for preseason soccer training and basketball (^1,19,33). Depending on the number of sprints needed, the rest period will need to be sufficient enough to resynthesize phosphocreatine (PCr), remove metabolic waste, and oxidize lactate (^3,15,38).

RSA protocols have been investigated with numerous work to rest ratios during cycle ergometry (^3,4,11,15,23). According to an investigation by Morin et al. (²⁵), the most common cycling protocol reported in the existing literature is 6 seconds of maximal sprints with 24 seconds active, passive, or static rest. Furthermore, high intensity efforts with less than 30 seconds of rest can lower adenosine triphosphate (ATP) concentrations, inhibit the rate of PCr resynthesis, and increase fatigue, thereby decreasing subsequent sprint performance (³⁷). Longer rest intervals have elicited greater peak power (PP), mean power (MP), and lower V[Combining Dot Above]O₂ at rest (¹⁵) potentially due to enhanced PCr contribution to ATP resynthesis. Although recovery duration has been shown to significantly affect acute V[Combining Dot Above]O₂ response, muscle oxygen demands may supersede pulmonary V[Combining Dot Above]O₂ (^2,6). Nonetheless, oxygen during rest periods play an integral part in PCr resynthesis, subsequent performance, and has been linked to percent decrement (%Dec) (^5,10).

Previous studies have used fatigue indices to quantify the amount of fatigue a participant endures (^{4,12,14,17,20,38}); yet the methods of calculating fatigue indices in protocols greater than 10 sprints may fail to observe decrements because of a plateau in performance variables (¹⁴). Percent decrement is a relative function of total and ideal sprint performance and is the most widely accepted method of estimating fatigue (^{8,14,20,24,38}); however its associated variability has been questioned when measuring RSA performance. Buchheit et al. (⁷) measured poor reliability (coefficient of variation [CV] >20%) in %Dec with 6 shuttle runs lasting 5 seconds and resting 25 seconds in between. Likewise, Hughes et al. (²⁰) found poor reliability (CV∼30%) in %Dec in 6 six-second sprints with 30 seconds of rest on a nonmotorized treadmill. The high variance of %Dec may inhibit significant relationships in calculated fatigue rates between RSA tests (^12,29,30). Because the literature has not appeared to establish a gold-standard (^28,30), the need for an index that can distinguish between work to rest ratios may be warranted.

Glaister et al. (¹⁴) proposed an alternative rate of decline formula that used the back transformation of the regression line for log transformed MP values from RSA testing on a cycle ergometer. As opposed to the traditional %Dec formula, that uses values from either the first sprint or the sprint with the best performance, the alternative rate of decline formula considers power output from each sprint and a linear relationship between sprint number and performance data (¹⁴). Similar intraclass correlation coefficients (ICCs) have been shown between the 2 measures and between different rest intervals (¹⁴). Interestingly, a follow-up investigation (¹⁷) demonstrated poor reliability in running-based RSA protocols.

The aim of this study was to distinguish between fatigue indices, oxygen consumption, and sprint performance within closely bound work to rest ratios, similar to those required in a competitive environment during RSA testing on a cycle ergometer. Our hypothesis is that larger work to rest ratios of 1:2, 1:3, and 1:4 will elicit greater fatigue rates and oxygen consumption accompanied by decreased performance (PP, MP, and total work [TW]).

Methods

Experimental Approach to the Problem

A within-subject, repeated measures study design was used to compare physiological response (V[Combining Dot Above]O₂) and performance (PP–the highest power output achieved during a 6-second bout, TW–the accumulated work during a 6-second bout, and MP–the average power output maintained throughout a 6-second bout), during RSA testing. Percent decrement in PP, MP, and TW over the RSA was calculated as previously described (^12,15) using the following formula (F1):

Secondly, the rate of decline was calculated as the back transformation of the slope of the line of best fit for log-transformed PP, MP, or TW values over 10 sprints (^14,17,32) using the following formula (F2):

where

Lastly, the rate of decline was also calculated as the slope coefficient from the regression line formed by the relationship between the number of sprints completed and the relevant RSA performance variables, including PP (W·sprint⁻¹), MP (W·sprint⁻¹), and TW (J·sprint⁻¹) (F3).

Each participant visited the Human Performance Laboratory on 5 occasions. The first visit was for initial screening during which anthropometric values were collected and paperwork was completed; the second visit involved a familiarization session during which each participant performed a series of familiarization sprints. The final 3 visits were the actual testing sessions. During the testing visits, participants completed 3 different RSA protocols to compare differences in oxygen consumption and performance variables among varying work to rest ratios. Participants completed the entire study over 3 weeks.

Subjects

Eight healthy, recreationally-trained men (23.5 ± 4.0 years; 174.3 ± 6.1 cm; 75.1 ± 10.9 kg) between the ages of 19 and 30 volunteered for this study. It was required that these individuals exercise at least 2 days per week. Before enrolling in the study, all participants completed a confidential medical and activity questionnaire and a physical activity readiness questionnaire to determine if they had any physical limitations or chronic illnesses that would affect their performance. Throughout the study, participants were asked to maintain their normal dietary and nutritional intakes around each scheduled testing day. Participants were also told to refrain from vigorous exercise within 48 hours before each testing day which would result in residual soreness or fatigue. The study was approved by the university institutional review board and all participants provided written informed consent before beginning the study.

Procedures

Familiarization

Participants who met the preliminary study criteria returned to the Human Performance Laboratory and were familiarized with the experimental procedures. The participants performed approximately 10 six-second maximal sprints against a load, equivalent to 0.75 N·kg⁻¹ of body mass interspersed by 24 seconds of unloaded cycling.

Gas Exchange Analysis

All gas exchange data were collected using open circuit spirometry (True One Metabolic Cart, Parvo Medics, Inc., Sandy, UT, USA). Twenty minutes before each RSA protocol, the unit was calibrated with room air and gases of known concentration. Flowmeter calibration was also performed before exercise to determine the accuracy of flow volume while collecting data. Participants wore a head unit and mouth piece that stabilized a one-way valve around their mouth. Breath-by-breath oxygen and carbon dioxide values were analyzed through a sampling line after the gases pass through a heated pneumotach and mixing chamber.

Repeated Sprint Ability Testing

The participants performed a 4 minute warm-up at a self-selected intensity interspersed with 4 submaximal sprints lasting 4 to 6 seconds on a cycle ergometer. After the warm-up, participants were asked to perform 3 maximal 6-second sprints interspersed with 24 seconds of rest to establish PP output. The average of the 3 maximal sprints was used to establish PP output (¹¹) and the first sprint was verified, at or greater than 85% PP output, to ensure participants were giving near maximal effort throughout the trials. The warm-up and peak power establishment was followed by 5 minutes of passive rest to ensure adequate recovery (¹). The participants then performed 10 six-second sprints against a load, equivalent to 0.75 N·kg⁻¹ of body mass interspersed by 24 seconds of unloaded cycling. The participants were asked to place their legs in a standardized starting position with their feet firmly strapped into the metal pedals. Before each sprint, the dominant foot was set in a standardized position as previously described (³).

This protocol was repeated during 2 additional testing sessions (visit 4 and visit 5) with a minimum of 48 hours of recovery between each visit. The rest intervals during successive repeated sprint protocols were 18 seconds and 12 seconds, respectively. Gas exchange analysis (as described above) was conducted throughout the test to determine oxygen uptake (V[Combining Dot Above]O₂). During breath-by-breath V[Combining Dot Above]O₂ analysis, the maximal V[Combining Dot Above]O₂ value attained in response to a given sprint interval was deemed V[Combining Dot Above]O₂work and the lowest V[Combining Dot Above]O₂ value attained in response to a given rest period was deemed V[Combining Dot Above]O₂rest.

Statistical Analyses

Breath-by-breath gas exchange analysis data was converted to text files and further analyzed using a custom program within the LabVIEW software (National Instruments, Austin, TX, USA). All RSA V[Combining Dot Above]O₂ data were fit with a cubic spline interpolation function and plotted over time for each trial and analyzed for each high and low value throughout the trial. To smooth out the V[Combining Dot Above]O₂ data, a low pass Butterworth digital filter with a cut-off frequency of 0.06 Hz was used as suggested by Robergs, et al. (³⁵). V[Combining Dot Above]O₂work and V[Combining Dot Above]O₂rest were also collated as averages for each RSA trial to further compare physiological response between the work to rest ratios. Statistically significant changes and differences between trials were determined using SPSS (version 21.0; SPSS, Inc., Chicago, IL, USA). Data were analyzed for normality via the Kolmogorov–Smirnov test and parametric statistics were conducted. Repeated measures analysis of variance was used to analyze the changes in all variables. In the event of statistical significance, least significant difference (LSD) tests were used for post hoc analyses. Results were considered significant at an alpha level of p ≤ 0.05.

Results

The physiological response and performance values are listed in Table 1.

Table 1: Comparison of physiological and performance variables.*

Physiological Response

When analyzing V[Combining Dot Above]O₂work during RSA trials there were no significant differences among trials (F = 0.755, p = 0.488, η² = 0.097). However, V[Combining Dot Above]O₂rest was significantly different between trials (F = 21.084, p < 0.001, η² = 0.751) with 12 seconds being significantly greater than 18 seconds (p = 0.001) and 24 seconds (p = 0.001), and 18 s being significantly greater than 24 seconds (p = 0.045). The average oxygen response (V[Combining Dot Above]O₂work and V[Combining Dot Above]O₂rest) for each sprint across the RSAs are represented in Figure 1.

Figure 1: Breath-by-breath V[Combining Dot Above]O₂ response in relation to each sprint. W = V[Combining Dot Above]O₂work (maximal V[Combining Dot Above]O₂ value attained in response to a given sprint interval). R = V[Combining Dot Above]O₂rest (lowest V[Combining Dot Above]O₂ value attained in response to a given rest period). The rest bouts (24 s, 18 s, and 12 s) illustrate each of the 3 RSA protocols all with 6 s maximal sprints as the work bout. V[Combining Dot Above]O₂ = oxygen response.

Performance Response

When analyzing average TW throughout the RSA trials, there were significant differences between trials (F = 4.867, p = 0.025, η² = 0.410) with 12 seconds significantly less than 24 seconds (p = 0.012) (Table 1). For average MP, there was a significant difference between trials (F = 5.412, p = 0.018, η² = 0.436) with 12 seconds significant less than 24 seconds (p = 0.013). However, for average PP, there were no significant differences among trials (F = 0.146, p = 0.866, η² = 0.020) (Table 1). The average MP and TW for each sprint across the RSAs are represented in Figures 2 and 3.

Figure 2: Average mean power data for the 3 RSA protocols. The rest bouts (24 s, 18 s, and 12 s) illustrate each of the 3 RSA protocols all with 6 s maximal sprints as the work bout.

Figure 3: Average total work data for the 3 RSA protocols. The rest bouts (24 s, 18 s, and 12 s) illustrate each of the 3 RSA protocols all with 6 s maximal sprints as the work bout.

Fatigue Indices

Fatigue indices are listed in Table 2. When analyzing the %Dec during the RSA trials, there were no significant differences among trials in PP (F = 17.608, p = 0.071, η² = 0.315), TW (F = 29.266, p = 0.141, η² = 0.244), or MP (F = 0.760, p = 0.486, η² = 0.098). When analyzing the rate of decline (F2), there was significant difference between trials for PP (F = 4.045, p = 0.041, η² = 0.366) with 12 seconds significantly greater than 24 seconds (p = 0.019), and 12 seconds significantly greater than 18 seconds (p = 0.028) and 24 seconds (p < 0.001) for TW (F = 6.465, p = 0.031, η² = 0.480). When analyzing the rate of decline (F3) during RSA trials, there were significant differences between trials for PP (F = 3.921, p = 0.044, η² = 0.359) with 12 seconds significantly greater than 24 seconds (p = 0.034) and for TW (F = 6.067, p = 0.035, η² = 0.464), with 12 seconds significantly greater than 18 seconds (p = 0.034) and 24 seconds (p < 0.001). However, there were no significant differences among trials for MP (F = 4.624, p = 0.058, η² = 0.398).

Table 2: Comparison of various fatigue calculations among rest intervals.*

Discussion

This investigation examined the influence of recovery on oxygen consumption, and performance measures utilizing different work to rest ratios during RSA testing. Differences, over 10 sprints, were observed in average V[Combining Dot Above]O₂rest between trials; however, no differences were seen in average V[Combining Dot Above]O₂work. Furthermore, performance differences were significant among RSA trials in average TW and MP. Although no differences were observed in %Dec for TW, MP, or PP, differences were seen in both the log-transformed and standard slope methods of calculating the rate of decline for PP and TW.

In this investigation, V[Combining Dot Above]O₂rest was significantly greater in the 12-second trial as compared with the 18-second and 24-second trials (∼12% and ∼20% greater, respectively) and the 18-second trial was significantly greater than the 24-second trial (∼10% greater). These results were similar to those reported by Glaister et al. (¹⁵) who reported a ∼16% difference in mean V[Combining Dot Above]O₂ during recovery with the shorter rest protocol (10 seconds) being greater than the longer rest protocol (30 seconds), but over 20 sprints. Likewise, Balsom et al. (²) observed higher V[Combining Dot Above]O₂ at the end of shorter rest periods (60 seconds) as compared with longer rest periods (120 seconds) in 15 × 40 m sprints. V[Combining Dot Above]O₂ off kinetics may help explain why the oxygen consumption during the rest periods was significantly different between RSA protocols. Because maximal V[Combining Dot Above]O₂ is not easily adapted, especially in short term (¹), the ability to return to pulmonary homeostasis can severely be affected by the amount of rest because of characteristics of V[Combining Dot Above]O₂ off kinetics and subsequent PCr resynthesis (³⁴). This explanation is further supported by the relationship between V[Combining Dot Above]O₂ kinetics and RSA performance. Dupont et al. (¹⁰) found a significant positive correlation with the time constant associated with the fast component of V[Combining Dot Above]O₂ off kinetics following severe intensity exercise and %Dec of power over the course of repeated sprints. Although pulmonary V[Combining Dot Above]O₂ in relation to each sprint bout did not differ between protocols, oxygen demand from the active muscles was likely high and there may have been a greater reliance on oxidative metabolism to restore PCr (^5,31).

The manipulation of rest intervals also had a significant effect on TW and MP, that were significantly lower in the 12 seconds compared with 24 seconds protocol. The elevated V[Combining Dot Above]O₂rest observed during the shorter (e.g., 12 seconds) rest intervals indicated that insufficient recovery was likely related to insufficient PCr resynthesis over the course of 10 sprints which may explain the lower TW and MP compared with the 24 seconds rest protocol. Dawson et al. (⁹) reported that PCr concentrations after a single 6-second sprint were 55% of pre-exercise values and 27% of pre-exercise values after 5 six-second sprints on a cycle ergometer indicating that PCr levels are reduced for ATP resynthesis during recovery (^9,16,22,38). Glaister et al. (¹⁵) reported similar findings among rest periods of 10 seconds and 30 seconds. In addition, an accumulation of metabolic by-products such as inorganic phosphate (Pi), which can affect calcium release, may also play a significant role in fatigue (^15,22). Since removal of Pi is an oxygen dependent process, it appears that the 24 seconds rest interval allowed for the greatest reduction in oxygen consumption and enhanced Pi removal compared with the shorter rest intervals. However, the rate of PCr resynthesis and removal of Pi likely did not affect PP among the rest intervals; potentially due to the initial phase of PCr resynthesis being unaffected by the metabolic environment (^22,36,39). This is supported by Harris et al. (¹⁸) who reported a time of 21–22 seconds to replenish half of PCr stores in the initial fast phase of PCr resynthesis after dynamic movement. Thus, the used work to rest ratios may not have been high enough to stimulate changes in PP during the 10 sprints. Therefore, the difference observed among rest intervals may have been a result of the rate of PCr resynthesis (¹⁵).

Despite differences in performance variables, there were no significant differences in %Dec among the RSA protocols in the current investigation. Negligible fatigue rates may stem from the number of subsequent sprints and the duration or intensity of the sprint and rest bouts (^4,30). Another possibility could be that the participants experienced a training effect and became more economical as a 20% decrease in muscle oxygen utilization has been shown in as little as 2 high-intensity interval training sessions (²¹). Many RSA studies have been conducted using recreationally trained individuals which may lead to a lack of effort given (^14,17,20,24). While %Dec is commonly evaluated and reported within the RSA literature, and it has been suggested as the most appropriate fatigue rate measure (¹⁴), some evidence suggests lack of validity and reliability (^4,20,29,30). The rates of decline produced using F2 and F3 demonstrated the ability to discriminate between differing work to rest ratios whereas %Dec did not. Both F2 and F3 took into account performance for each of the 10 sprints, however, the assumption that the performance variables over 10 sprints are linear may cause potential issues (¹⁴). Because there is inconclusive evidence as to which fatigue rate is most appropriate and there is not a gold-standard (^14,17), the rate of decline using either F2 or F3 should be considered viable options. Given the similar discriminative ability between these formulae, F3, with a less involved calculation procedure, may provide a more straightforward and practical method of calculating fatigue during RSA testing.

Follow-up studies should consider the use of blood lactate concentrations during RSA trials which may provide greater insight on performance variables and fatigue rates. In addition, greater than 6 seconds between RSA protocols should be used to elicit substantial differences in performance variables over 10 sprints of 6-second maximal work bouts. Because RSA testing was conducted on a cycle ergometer, and most sports emphasize running as their primary means, future studies should investigate RSA with varying rest intervals on a treadmill or other sport-specific modes.

Practical Applications

In summary, this study indicates that V[Combining Dot Above]O₂ response during rest periods remains higher in RSA protocols with larger work to rest ratios despite the absence of differences in %Dec. Researchers should be aware that despite small changes in rest periods, elevated V[Combining Dot Above]O₂ during rest periods may have profound effects on prolonged duration sprints and possibly time to exhaustion. Because of the uncertainties in %Dec and the results of this study, rate of decline may be a more sensitive measure of fatigue because extremely high performance values may affect the %Dec calculation. In particular, the slope of the regression line formed by the number of sprints completed and PP, MP, or TW may provide a similar method of calculating rate of decline compared with the back transformation of the slope of the line of best fit for log-transformed PP, MP, or TW. Future investigations of RSA should measure V[Combining Dot Above]O₂max for baseline measurements and use testing protocols with varying work to rest ratios. These protocols should implement differences in the rest intervals of greater than 6 seconds to elicit potential differences in performance. Furthermore, studies should examine varying work to rest ratios on a treadmill to examine performance differences with different modes of exercise.

Acknowledgments

The authors thank the group of participants for their dedication and the staff of the Institute of Exercise Physiology and Wellness for their help with data collection and analysis.