Participants
Twenty-two runners, 11 females, estimated to fit performance level 4 criteria [32], were recruited from local running clubs via email distribution of digital flyers. Main inclusion criteria were a half marathon personal best of less than 1 h:40 min and 1 h:25 min for female and male runners, respectively. The 15% absolute performance difference accurately reflects the average sex difference in the TOP-100 finishing times of an international half marathon race event (www.twooceansmarathon.org.za) hosted locally. Performance level categorisation of each participant was subsequently verified according to criteria by DePauw and colleagues [32, 33]. Sample size estimation using conventional methods (one-tailed alpha of < .05, β > .80, and large effect size) and G*Power 3 software (G*Power, Version 3.1.9.2, Kiel, Germany) indicated overall 20 participants would be needed. All participants provided prior written informed consent to the procedures used in this study, which were approved by the institutional ethics committee and carried out in accordance with the Declaration of Helsinki.
Study Design
Participants completed in total four visits including three maximal self-paced 20-km treadmill time trials over a simulated profiled course. A 10-km time trial took place after preliminary testing and acted as a submaximal familiarisation time trial (FTT). Before returning for their second visit, participants were instructed to log their training and diet for 48 h prior to the baseline time trial (BTT) and to prepare in a way that resembled their routine before an important race. They were then advised to repeat this ‘mini taper’ for the following experimental time trials. Participants then completed the intervention time trial (ITT) and control time trial (CTT) in a counterbalanced AB/BA crossover design. The ITT was preceded by a standardised drop-jump protocol (for details, see the “Drop-Jump Protocol” section below), while the CTT was preceded by a rest period of equal length. A washout period of 4 weeks took place before the crossover trial to (A) allow for complete recovery of muscle damage and (B) control for menstrual cycle in female runners (two were using hormonal contraception and two were amenorrhoeic), who were instructed to schedule experimental trials 5 to 9 days after the start of menses (early follicular phase).
Procedures
Preliminary Measurements
During the first laboratory attendance, each participant had their age, stature, and body mass recorded. Body fat percentage was estimated using the seven skinfold method [34].
Peak Treadmill Running Speed Test
A peak treadmill running speed (PTRS) test with peak oxygen consumption (VO2peak) was performed on a motor-driven treadmill (Viasys LE500 CE, Hoechberg-Wuerzburg, Germany). The treadmill gradient was set to 1% simulating the energetic costs of outdoor running [35]. After a 10-min self-paced warm-up, participants started the PTRS at 10 and 11 km h−1 for female and male runners, respectively, after which speed increased stepwise by 0.5 km h−1 every minute. PTRS was calculated as the last completed stage added to the product of the speed increment and the completed fraction of the incomplete stage. During the progressive exercise test, expired ventilation volume (VE), oxygen uptake (VO2), and carbon dioxide production (VCO2) were measured with an online breath-by-breath gas analyser and pneumotach (Cosmed Quark b2, Rome, Italy). Calibration took place before each trial according to the manufacturer’s instructions using a 3-l syringe (Hans Rudolph 5530, Inc., Kansas City, MO, USA) and a gas mixture of known composition (5.05% CO2; 15.97% N2). Peak oxygen uptake was determined as the highest recorded VO2 measurement averaged over 30 s. The first ventilatory threshold and respiratory compensation point were determined according to the methods described by Lucia et al. [36]
Maximum Isokinetic Power Output Test
Maximum isokinetic power output of the quadriceps and hamstring muscles of the dominant leg was measured using the Biodex 3 Isokinetic dynamometer (Biodex Medical Systems, Shirley, NY, USA). Participants were familiarised with the procedure before the FTT and BTT. Participants were seated with their arms crossed over their chest, the hip flexed at 90°, and a knee angle of 90° from full leg extension. The lateral condyle of the femur was aligned with the rotational axis of the dynamometer, and the ankle was secured to the lever arm of the dynamometer using a padded Velcro strap. Shoulder and waist straps were used to fixate joint positions during trials. Standardisation was enhanced by performing gravity correction. After a warm-up set of increasing intensity (50, 70, and 90%), participants performed six maximal isokinetic contractions at 120° s−1. Isokinetic measurements were chosen due to the repetitive cyclic nature of running specific muscle contractions, and the above movement velocity was selected as it more closely resembles running specific contraction speeds as well as provides a trade-off between peak and explosive power output usually measured at < 60 s−1 and > 180° s−1, respectively. Power output curves were displayed on a screen in front of the participant, and verbal encouragement was given during maximal contractions.
During the experimental trials, maximum isokinetic power output was assessed twice: (1) after the 10-min self-paced warm-up and (2) within 2 min after completing the drop-jump protocol and control period, respectively.
Running Economy Test
Assessment of running economy took place before each of the three maximal 20-km time trials. After 5 min of running at a constant speed of 10 and 11 km h−1 for female and male runners, respectively, participants ran for 5 min at a speed corresponding to ∆ 1/3 between the first ventilatory threshold and the respiratory compensation point determined during the PTRS. The average values of breath-by-breath VO2 and VCO2 during the final minutes were used to calculate the oxygen cost and energy cost of running. Updated non-protein respiratory quotient equations were used to estimate substrate use (g min−1) during the monitored period [37]. The mean energy content of the metabolised substrates was then calculated according to the methods described by Jeukendrup and Wallis [38], scaled to body mass (BM−1) and expressed per kilometre (km−1) [39].
Time Trial Procedure
Participants performed three maximal self-paced 20-km time trials over a simulated profiled course on the same treadmill. A customised course profile was written using h/p/cosmos para-graphics® software (Version 2.6.14, h/p/cosmos sports and medical GmbH, Nussdorf-Traunstein, Germany) simulating a 20-km long time trial with two uphill sections of 2 km length and a gradient of 7%, starting after 4 and 12 km, respectively. The uphill sections were immediately followed by two downhill sections of the same length and gradient. Participants were assisted with gradient-dependent alterations in running speed, but custom-made modifications of the treadmill enabled participants to self-select running speed at any time with a handheld remote control in 0.1 km h−1 increments. Only course profile, gradient, and distance covered were displayed to participants during each time trial. Participants were instructed to complete the time trial in the shortest possible time, but no verbal encouragement was given during the time trial itself.
A fan was placed 1.5 m in front of the participants, and the fan level was adjusted in accordance with the course profile to simulate speed-related alterations in peripheral cooling. During the uphill sections, fans were set on level 1, during the flat sections on level 2, and during the downhill sections on level 3, respectively, creating average wind speeds of 3.15, 3.80, and 4.25 ms−1. Participants were permitted to consume water ad libitum, and a commercially available carbohydrate drink (Enduren™ Endurance Energy Drink) was provided at 2-km intervals and upon request for an average rate of ≈ 60 g h−1. All experimental trials were conducted under stable climatic conditions (temperature 20.5 ± 0.7 °C; humidity 58.3 ± 5.1%) and commenced at 8 am to control for diurnal variations.
During each running economy test and time trial, speed, gradient, and distance covered were recorded and stored at 1-s intervals using h/p/cosmos para-graphics® software. Heart rate was recorded at 2-s intervals throughout each trial by telemetry (Suunto® T6, Suunto Oy, Vantaa, Finland). Performance and heart rate data were subsequently analysed with TrainingPeaks™ analysis software (WKO edition+, Version 3.0, Lafayette, CO, USA). All continuously captured data were averaged into 2-km bins before statistical analyses were performed.
Drop-Jump Protocol
During the intervention trial only, participants performed muscle-lengthening contraction exercise consisting of 100 drop-jumps from a step of 45 cm height [40] known to induce force losses consistent with mild EIMD and without confounding effects of significant muscle metabolite accumulation (blood lactate concentration [mmol/l]: CTT = 1.5 ± 0.6 vs ITT = 1.4 ± 0.6) and cardiovascular demands (heart rate [bpm]: CTT = 65 ± 10 vs ITT = 100 ± 11) [29]. Participants dropped to an approximate knee angle of 90° before jumping upward as high as possible. Drop-jumps were performed every 20 s for a total time of ≈ 33 min. Vertical displacement of the centre of mass (COM) was calculated by means of a floor-embedded force plate (AMTI, Watertown, MA, USA) and a customised programme written in MATLAB® (R2013a, The Mathworks Inc., Natick, MA, USA) determining vertical take-off velocity of COM [41]:
Equation 1: determining jump-height from vertical take-off velocity
$$ \mathrm{Jump}-\mathrm{height}={\mathrm{TOV}}^2\kern0.5em 2{g}^{-1}, $$
where TOV = vertical velocity of COM at take-off, g = 9.81 ms−2.
Before and after the drop-jump protocol and rest period, perceived muscle discomfort and unpleasantness were assessed by means of 100 mm visual analogue scales (VAS) ranging from ‘no muscle discomfort/unpleasantness at all’ to ‘unbearable muscle discomfort/unpleasantness’. The extent to which participants experienced DOMS in the 84-h recovery period after experimental trials was assessed using a 7-point Likert-type scale [42].
Measurements
Measures of Perceived Fatigability
During the final minutes of the running economy test and at 2-km intervals during each time trial, participants provided ratings on the following four single-item scales presented in random order. All scales were anchored during the PTRS test.
The 15-point (6–20) Borg scale with the indicator terms ‘light’ and ‘strong’ was used to approximate perceived physical strain by phrasing ‘How strong are the physical sensations from your legs, lungs, and body?’ Participants were instructed to include only the subjective perception of physical sensations caused by the task and to focus on location, quality, and intensity of physical sensations before returning an overall perceived physical strain score.
The 15-point (0–14) Borg scale with the indicator terms ‘easy’ and ‘hard’ was used to approximate perceived mental strain by phrasing ‘How difficult is it to run at this pace?’ Participants were instructed to include only perceived task difficulty and mental effort invested into/required to continue with the task before returning a perceived mental strain score.
In line with recent recommendations [43], a detailed three-tiered justification process has been outlined for measurement selection in the assessment of dynamic changes in core affective state during prolonged endurance exercise [44]. Ultimately, the 11-point (− 5 ‘very bad’ to 0 ‘neutral’ to + 5 ‘very good’) Feeling Scale (FS) and 6-point (1 ‘low activation’ to 6 ‘high activation’) Felt Arousal Scale (FAS) were chosen to approximate dynamic changes in valence and felt activation, respectively [45, 46].
The Action Crisis Scale (ACRISS) and short Flow State Scale (FSS) were administered during the 30-min recovery period retrospectively measuring the extent to which participants experienced a shift from an implemental to a deliberative mindset. Both scales were administered for seven sections of the profiled TT course (three flat, two uphill, and two downhill) and rated on 5-point Likert-type scales.
The ACRISS comprises six items: conflict, setbacks, implemental disorientation, rumination, disengagement impulses, and procrastination [47]. The internal consistency estimate of reliability was good with mean Cronbach’s alpha of 0.83 ± 0.08 and 0.89 ± 0.04 during the control and intervention time trial, respectively.
The FSS comprises nine items: challenge-skill balance, action-awareness merging, clear goals, unambiguous feedback, concentration on task at hand, sense of control, loss of self-consciousness, transformation of time, and autotelic experience [48]. The internal consistency estimate of reliability was good with mean Cronbach’s alpha of 0.81 ± 0.06 and 0.81 ± 0.07 during the control and intervention time trial, respectively.Footnote 2
Physiological Measures
Venous blood samples from a superficial antecubital vein were taken at rest, after the drop-jump protocol, after the running economy test, halfway through and at the end of experimental time trials as well as after the 30-min recovery period. Blood samples were placed into four different pre-chilled Vacutainers, respectively containing (A) serum clot activator for the analysis of cortisol, (B) potassium oxalate and sodium fluoride for the analysis of lactate concentrations, and (C) K2-ethylenediaminetetraacetic acid (EDTA) for the analysis of interleukin-6, differentiated white blood cell count, haemoglobin, and haematocrit. Where appropriate, samples were inverted five times, immediately centrifuged at 3000 rpm at 4 °C for 10 min, plasma/serum pipetted off, and kept on ice until stored at − 80 °C for subsequent analysis. Plasma lactate concentrations were determined using glucose oxidase method (YSI 2300 STAT PLUS, Ohio, USA). Serum cortisol was determined using an automated chemiluminescence system (Architect iSR, Abbott Diagnostics, IL, USA) with conventional reagent kits and calibrators. Interleukin-6 was analysed by means of enzyme-linked immunosorbent assay (ELISA) (eBioscience, Bender MedSystems, Vienna, Austria). Differentiated white blood cell count, haematocrit, and haemoglobin were determined by Lancet Laboratories using conventional methods. The degree of haemoconcentration was calculated, and all blood samples were subsequently corrected for plasma volume changes [49].
Measures of Performance Fatigability
Besides the assessment of global time trial performance between experimental conditions, this study aimed to analyse the decline in objective measures of endurance performance over time. The second-half to first-half split time quotient was calculated as a crude assessment of the degree of positive or negative pacing, and repeated measures of 2-km split time intervals were used to locate onset and extent of performance fatigability. For a clearer graphical presentation, dynamics in perceived fatigability were further indicated by the percentage increase in split times during ITT compared to CTT and accordingly assessed via one-way repeated measures ANOVA.
Statistical Analysis
Data were tested for assumptions, normality, equality of variances, equality of covariance matrices, and sphericity where appropriate. Independent samples t tests were used to analyse inter-individual differences in parameters between the sexes, and paired samples t tests were used for intra-individual comparisons in parameters between experimental trials. Between-group effect sizes were calculated using Cohen’s d (trivial < 0.20, small 0.21–0.60, medium 0.61–1.20, large 1.21–2.00, very large 2.01–4.00, and near perfect > 4.00). A non-parametric Mann-Whitney U test was performed when the assumption of equality of variance was violated. Confidence intervals and magnitude-based inferences for meaningful differences between time trial performances were derived from p values in accordance with Hopkins [50]. Two-way repeated measures ANOVAs were used to compare changes in parameters over time between experimental trials. Significant interaction effects were followed up with Bonferroni correction procedure. A Greenhouse-Geisser epsilon adjustment was made when sphericity was violated. Aligned rank transformation procedure (ARTool software 1.5.1, Washington, USA) was used to perform non-parametric factorial analysis on all perceptive data and when assumptions of parametric tests were violated [51]. All ANOVA effect sizes were calculated as partial eta squared (ηp2) and classified as small 0.02–0.13, medium 0.13–0.26, and large > 0.26. All data are presented as mean ± one standard deviation, and an alpha level of < .05 (two-tailed) was used to indicate statistical significance.