All running trials were conducted under stable laboratory conditions (ambient conditions: 21 ± 3 °C, <60% relative humidity) on a motorized treadmill (Woodway, Waukesha, WI, USA) with the gradient set a 1% [12
]. The HR response data was collected using an HR monitor (Polar A300, China) and the oxygen consumption data was measured using a carbon dioxide and oxygen analyzer (Metalyzer Cortex, Biophysik GmbH, Leipzig, Germany), which was calibrated before each testing session according to the manufacturer’s specifications. All of the capillary samples were drawn from the preferred finger of the runner and lactate (LA) accumulation was measured using a blood LA analyzer (LA Pro 2, Shiga, Japan). Any subjective data was measured by way of the rate of perceived exertion (RPE) using a modified BORG 10-point scale [13
]. For the WR conditions, subjects were required to wear a pair of compression shorts with associated loads (LilaTM
, Wilayah Persekutuan Kuala Lumpur, Malaysia). The WR was added via 100 or 200 g increments and the total load for each trial was rounded to the nearest 100 g. The loading schemes began with each WR load placed alternatively from the anterior to the posterior, in a stacked balance of distal to proximal (see Figure 1
and Figure 2
). A spectrum of WR loading was used (0%, 1%, 2%, 3%, 4%, and 5% BM) which was comparable to previous WR studies [11
], to examine changes between the loading magnitudes.
For each participant, the study was conducted over a maximum of 15 days. This included one familiarization session and three testing sessions (see Figure 3
) under laboratory conditions. The purpose of the familiarization session was to allow each runner to become accustomed to treadmill running while wearing both the compression shorts and all the metabolic measuring equipment. The participants completed a self-paced run for 20 min followed by a 10 min recovery. During this recovery time, the graded exercise test (GXT) protocol was discussed and an HR monitor and gas mask fitted. The participants then completed a further 10 min run—including a sufficient amount of the GXT incremental protocol, to feel comfortable with the procedures—and no data was collected. The participants were instructed to refrain from training on the day of testing session one and to avoid any strenuous training sessions 24 h prior.
Testing session one occurred within seven days of completing the familiarization session. The purpose was to generate a VO2
response profile to graded exercise to establish VT1
max. The participants completed a self-paced 20 min warm-up on a treadmill and were given a recovery period of 10 min prior to the commencement of the GXT. The starting speed was maintained for 1 min, followed by an increase of 0.5 km·h−1
every 30 s until the point of voluntary exhaustion [18
]. The starting speed was adjusted on an individual basis, to ensure volitional exhaustion at between 8 and 12 min. VO2
was tracked continuously at a sampling rate of 0.1 Hz, and the HR and RPE were recorded at each speed increment, with LA being measured immediately after completion of the test. The maximum oxygen consumption was, on average, over 30 s and was considered to be achieved if any one of the following criteria were met: a plateau in VO2
was reached, despite an increase in workload; a respiratory exchange ratio (RER) >1.15 was observed; an HR within five beats of the age predicted maximum (220-AGE) was reached; or a peak exercise blood LA concentration >8 mmol/L was achieved [19
]. The testing sessions two and three included all submaximal running trials, to measure metabolic and subjective responses while un-loaded and loaded. Testing session two occurred within 2–5 days of testing session one and testing session three occurred within 2–3 days of testing session two, to ensure no fatigue between all three sessions (Barnett, 2006). Testing session two included three WR loads, testing session three included the final three WR loads and load order was randomized. At the start of both testing sessions two and three, an 8 min warm up—set at a running speed equivalent to VT1
—was completed, followed by a 10 min recovery. VT1
was chosen as this is close to the typical training intensity in endurance sports, in line with the polarized model of training. The polarized training model has been shown to be common practice among elite endurance runners, for whom long, slow distance training at lower intensities (<VT2
) makes up 75% of an individual’s training volume, with shorter, higher intensity bouts of effort (>VT2
) making up the remainder of the training program [13
]. Each submaximal running trial lasted 8 min, with 10 min seated recovery between each subsequent trial. The oxygen consumption and HR were tracked for 2 min prior to each trial starting, for the 8 min of each trial (with the final 2 min used for analysis) and for 2 min post trial. The rate of perceived exertion and LA was recorded immediately post completion of each 8 min trial.
Descriptive statistics—including the means and standard deviations—were calculated for each measure. The statistical aim of this study was to make an inference about the impact on metabolic stress of submaximal running with WR, which requires a determination of the magnitude of an outcome. The traditional sample size estimation and hypothesis testing approach was not appropriate for this study design [20
]. Accordingly, inferential statistics were used to examine the qualitative meaning of the observed changes in the metabolic cost (VO2
, HR, LA) and perception (RPE) of submaximal running, with loaded compared to unloaded examples. The collected data was presented as the mean value for each, with the reported effect size (ES) and percent differences at a 90% confidence interval (CI). The smallest worthwhile change was used to determine if any observed changes were considered trivial, possible or likely, including the magnitude of each change, calculated as a change in score standardized to 0.2 of the between–subject SD from the unloaded condition [21
]. The qualitative probabilities were defined by the scale <0.5% most likely trivial increase, <5% very likely trivial increase, <25% likely trivial increase, 25–75% possible small increase, >75% likely moderate increase, >95% very likely large increase, >99.5% most likely very large increase and the outcome was deemed unclear where the 5% and 90% CI of the mean change overlapped with both the positive and negative outcomes [20
]. To help quantify the metabolic cost of WR based on relative exercise intensity and duration, HRs were used to extrapolate a training load score (TLS) for each load [22
] for 10 min of running. To understand the relationship between metabolic variables (VO2
, HR and TLS) and load, a scatterplot was created in excel to establish a linear equation and R2
value for each variable. The formula used for calculating the Training Load Score (Training Stress Score (TSS) [22
TLS = (sec × HR × IF)/(VT2 × 3600) × 100
IF (intensity factor) = HR/VT2
Key: TLS: Training load score, HR: Heart rate (average heart rate during exercise), IF: Intensity factor, VT2: Second ventilatory threshold (the point at which LA accumulation exceeds clearance).