Load carriage is defined as locomotion while bearing a mass on the torso supported by shoulder straps and/or a hip belt and is a constituent element in some occupational settings, such as the military or emergency services and also physical activities such as hiking (14). Carrying a backpack remains one of the most convenient and economical ways of transporting an external load (2) particularly in some military contexts where land vehicles may be restricted (15).
Studies have previously investigated the physiological and performance consequences of load carriage; however, protocols typically lack ecological validity for occupational load carriage activities. For example laboratory-based studies have used single trials (normally a time trial over a set distance ∼2.4–20 km), for various durations (up to 30 minutes) and with varying external loads (15–46 kg) (12,15). In a military context, soldiers will often exercise at submaximal levels (e.g., while on patrol) and also at high-intensity levels (e.g., during engagement) (6). Although it is inherently difficult to devise a laboratory-based protocol, which precisely reflects operational requirement as the load, intensity, duration, terrain, and environment vary with each scenario, protocols should more closely reflect the physiological characteristics of activities encountered in military situations such as during training or predeployment assessments. Therefore, the practical question examined here is whether a laboratory-based protocol that incorporates more realistic occupational demands, that is, a combination of constant speed low intensity and self-paced high-intensity exercise can demonstrate sufficiently low between-session variability and be of use to the strength and conditioning coach. Accordingly, this protocol will improve the relevance of future studies that investigate load carriage performance and also providing practitioners with a useful tool to measure load carriage specific fitness, readiness for deployments, and responses to relevant interventions. However, to date, a protocol of this nature has yet to be designed with their reliability yet to be determined.
Therefore the aim of this study was to determine the reliability of a preloaded treadmill time-trial protocol (and also selected physiological parameters) by combining a period of submaximal exercise (preloaded phase), a rest period, and a performance phase (time trial). We hypothesized that the protocol detailed will demonstrate very well between trial reliability and be of great benefit to the strength and conditioning practitioner.
Experimental Approach to the Problem
The importance of load carriage was outlined in an excellent review within this journal (15), discussing at length the implications of load carriage assessment within an occupational context. This review suggested several areas of further study, including the improvement of methodologies used to quantify performance, noting that previously adopted models were confounded on methodical issues (see Introduction). The novel approach outlined here aims to establish a method suitable of quantifying load carriage performance, which could be used by the Strength and Conditioning coach for baseline testing of performance, determining performance improvements after specific interventions, and assessment of the individuals load carriage capability before periods of operational activity or training.
After ethics approval from the host university, 8 healthy, nonsmoking males with experience of load carriage through regular recreational load carriage activities, provided written informed consent to participate in this study. The physical characteristics of the participants are shown in Table 1. All participants were engaged in recreational physical training (strength and endurance) that was constant in the 2 months leading up to the study. A series of additional control measures were adopted before testing: participants did not engage in any strenuous exercise on the day preceding and the day of an exercise trial. Each participant also completed a 24-hour nutrition log, which was replicated for all subsequent trials. All trials were completed at identical times in the day, and participants abstained from alcohol and caffeine in the 24 hours before testing and arrived at the laboratory 2 hour postprandial.
Participants were briefed individually on the experimental design; after this, each completed 3 preliminary trials. The first preliminary trial consisted of a body composition assessment using dual energy x-ray absorptiometry (Lunar iDXA; GE Healthcare, Hertfordshire, United Kingdom) followed by an incremental exercise test on a motorized treadmill (Desmo, Woodway, Germany) for the determination of
. After a 5-minute warm-up at 8 km·h−1 and 1% gradient, the gradient was subsequently increased to 4% and speed increased by 1 km·h−1·min−1 until the limit of volitional tolerance (11). Online breath-by-breath gas analysis (MetaLyser ll, Cortex Biophysik, Birmingham, United Kingdom) was used to determine
defined as the highest 30-second
recorded throughout the test.
During the second preliminary trial, participants were familiarized and fitted with the 25-kg backpack (Web Tex, Bedford, United Kingdom) and completed 20-minute exercise at 0% gradient and 6.5 km·h−1. After 15 minutes of seated recovery, participants then completed a self-paced 2.4 km time-trial. The mass of the load was evenly distributed within the backpack and worn in accordance with manufacturer's guidelines. The backpack incorporated shoulder straps and a waist strap, which were adjusted individually and recorded to the nearest millimeter for subsequent trials. The third and final preliminary trial provided a full familiarization of the experimental trial detailed below.
Participants performed the experimental trial on 2 occasions separated by a minimum of 7 days. Participants walked for 60 minutes, 0% gradient and 6.5 km·h−1 carrying a 25-kg backpack (hereon referred to as LC) (5,13). After 15 minutes of seated recovery, participants then completed a 2.4 km time-trial while bearing the load (LCTT) where the speed of the treadmill was manually adjusted by the individual to complete the distance in the quickest time possible (13). The elapsed time was masked from the participant during all trials. The walking speed and duration, time-trial distance, and absolute mass were selected to more closely reflect realistic occupational requirements in line with previous recommendations for laboratory-based studies (18,19). Throughout LC, physiological parameters were measured at 15-minute intervals, immediately before and after 1.2 and 2.4 km of LCTT. Heart rate (HR) was measured using short-range telemetry (Polar T31, Kempele, Finland), expired pulmonary gases were assessed using Douglas bags (Cranlea and Co, Birmingham, United Kingdom), and blood lactate concentration ([Lac−]B, Accu-Check, Safe T-Pro, Birmingham, United Kingdom) was measured from arterialized-venous fingertip blood samples. Core body temperature was recorded using a tympanic thermometer (IRT 4520; Braun, Nottingham, United Kingdom). Ratings of whole body perceived exertion (RPE) were measured using the Borg scale (8). Perceptions of effort were further separated for leg (RPElegs) and breathing (RPEbreathing) discomfort using a visual analog scale: where 0 = no exertion and 10 = maximal exertion (24). In addition, participants performed a number of respiratory muscle strength and pulmonary function tests. Maximal inspiratory (P Imax) and expiratory pressures (P Emax) were assessed using a hand-held mouth pressure meter (MicroRPM; Micro Medical, Kent, United Kingdom) to provide an index of inspiratory and expiratory muscle strength (13). Pulmonary function was also assessed using a pneumotachograph (MS03; Micro Medical, Buckinghamshire, United Kingdom). All maneuvers were performed pre- and post-LC and post-LCTT (13).
First, mean differences between trials were calculated and paired t-tests were used to determine systematic bias between the experimental trials (apriori α = 0.05) using SPSS for Windows (SPSS, Chicago, IL, USA). Second, the intraclass correlation coefficient (ICC), coefficient of variation (CV), and Cohen's d were calculated as a general indicator of reliability between experimental trials. After this, absolute limits of agreement were calculated using methods detailed previously (7); however, because of the presence of heteroscedasticity in the data (a positive relationship between the absolute measurement differences and their mean) and in line with previous recommendations (1), 95% log ratio limits of agreement (LoA) were calculated for all variables using methods detailed previously (7) providing an average dimensionless measurement bias (e.g., general learning effect) and random measurement error (e.g., level of agreement) using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA). In addition, SE and 95% confidence intervals for the measurement bias and random error components of the limits of agreement were calculated. Finally, the 95% ratio limits of agreement were used to estimate sample sizes required for 5%, 10%, 20%, and 30% effects for a repeated-measures methodological design. Sample size calculations were performed using Microsoft Excel according to published equations (25).
Time-trial performance was 16.71 ± 1.82 minutes in trial 1 and 16.37 ± 1.78 minutes in trial 2 (p > 0.05) with a mean difference of 0.34 ± 0.89 minutes (Figure 1). Velocity was fixed during the preloaded phase at 1.8 m·s−1 but during LCTT increased significantly to 2.42 ± 0.24 m·s−1 in trial 1 and 2.46 ± 0.23 m·s−1 in trial 2 with no difference between trials in the magnitude of this increase (p > 0.05).
Physiological responses to both trials are shown in Table 3. All transient changes in physiological parameters both post-LC and post-LCTT were similar between trials (p > 0.05). Baseline HR was 96 ± 19 b·min−1 in trial 1 and 98 ± 21 b·min−1 in trial 2. Post-LC, HR increased by 34 ± 19 b·min−1 and 39 ± 22 b·min−1 in trials 1 and 2 respectively, which was similar between trials (p > 0.05). Relative to baseline, HR increased post-LCTT by 84 ± 28 b·min−1 and 90 ± 24 b·min−1, respectively (p ≤ 0.05).
increased from 1.49 ± 0.18 L·min−1 to 1.37 ± 0.38 L·min−1 pre-LC by 0.54 ± 0.35 L·min−1 to 0.70 ± 0.42 L·min−1 post-LC and a further 0.27 ± 0.48 L·min−1 to 0.28 ± 0.43 L·min−1 post-LCTT (p ≤ 0.05).
increased from 1.27 ± 0.17 to 1.81 ± 0.35 L·min−1 (absolute increase: 0.54 ± 0.35 L·min−1) in trial 1 and from 1.28 ± 0.23 to 1.98 ± 0.31 L·min−1 (absolute increase: 0.70 ± 0.42 L·min−1) in trial 2 post-LC (p > 0.05). Post-LCTT
increased a further 0.27 ± 0.48 L·min−1 to 0.57 ± 0.55 L·min−1 post-LCTT (p > 0.05). Baseline
was 35.65 ± 4.36 L·min−1 in trial 1 and 35.80 ± 6.09 L·min−1 in trial 2. Post-LC,
increased to 49.54 ± 7.42 L·min−1 and 52.94 ± 7.77 L·min−1 in trials 1 and 2, respectively (p ≤ 0.05) and post-LCTT increased to 60.52 ± 8.93 L·min−1 and 64.08 ± 18.17 L·min−1 in trials 1 and 2, respectively (p ≤ 0.05).
Relative to baseline, P Imax was reduced by 13 ± 11 cm H2O (11%) and 18 ± 11 cm H2O (15%) post-LC (p ≤ 0.05) and reduced 20 ± 8 cm H2O (15%) and 21 ± 9 cm H2O (17%) post-LCTT (Baseline to post-LCTT, p ≤ 0.05) in trials 1 and 2, respectively. Relative to baseline, P Emax was reduced 14 ± 11 cm H2O and 1 ± 10 cm H2O in trials 1 and 2, respectively; with further reductions of 1 ± 9 cm H2O (1%) and 3 ± 11 cm H2O (3%) in trials 1 and 2, respectively post-LCTT (p ≤ 0.05). Baseline pulmonary was within normal limits (17). Relative to baseline; FVC was reduced by 0.35 ± 0.24 L in trial 1 and 0.42 ± 0.26 L in trial 2 (8 and 12%, respectively) post-LC and 0.31 ± 027 L in trial 1 and 0.46 ± 0.31 L in trial 2 (10 and 11%, respectively) post-LCTT. FEV1 was reduced by 0.35 ± 0.24 L in trial 1 and 0.31 ± 0.20 L in trial 2 (9 and 8%, respectively) post-LC; further similar differences were observed post-LCTT, trial 1: 0.31 ± 0.27 L, trial 2: 0.46 ± 0.31 L (8 and 11%, respectively). Increases in FEV1/FVC were observed relative to baseline during both trials and to a similar magnitude.
The test-retest and mean difference data for time-trial performance and physiological measurements for LC and LCTT are shown in Table 2, and absolute 95% limits of agreement for the LCTT experimental trials 1 and 2 are shown in Figure 1. Correlation coefficients of the absolute differences vs. their mean in some variables between trials demonstrated heteroscedasticity; therefore, log ratio limits of agreement were calculated for all data sets in addition to ICC, CV, and Cohen's d. Tables 3 and 4 show the ratio limits of agreement for time-trial performance and respiratory function tests post-LC and post-LCTT, respectively and Tables 5 and 6 show the ratio limits of agreement for physiological and perceptual responses post-LC and post-LCTT, respectively. The estimated sample sizes are displayed in Table 7. Overall, for an alpha level of 0.05 and 10% effect, mean sample sizes for all variables were 20 and 22 for LC and LCTT, respectively.
We are the first to present a reliable load carriage protocol that more closely reflects the intensity and physiological demands of some occupational activities (18). We also demonstrate that a range of physiological and perceptual responses can be assessed between trials with acceptable between-session variability (1). We calculated ratio LoA between each trial and the ICC, CV, and Cohen's d, for which combination of analyses provides empirical evidence supporting the inclusion of this performance assessment protocol within future studies and within future practice (1).
Time-trial performance was similar to previous study using an identical protocol (13) and also similar to British Infantry recruits carrying the same mass in a backpack over the same distance, however, without a 60-minute prior constant-intensity bout (9,10). In this study, mean difference between efforts was 0.34 ± 0.89 minutes demonstrating excellent agreement (measurement bias: 1.02; random error ×/÷ 1.11). This is supported by narrow confidence intervals as depicted in Table 3 and Figure 1. The ICC (0.85) and CV (10.5%) values demonstrated good reliability, and Cohen's d (0.36) suggests that differences between trial variance was moderate and inconsequential demonstrating that this protocol is suitable for the Strength and Conditioning coach to use to assess load carriage performance.
Between-session variability of physiological and perceptual parameters post-LC and LCTT are presented in Tables 5 and 6, illustrating good agreement and small error ratios. The variation between trials is greater during LCTT than LC, most likely explained by the nature of the time-trial, which was self-paced and hence sensitive to changes in the adopted pace. However, the very good agreement should give confidence to the strength and conditioning practitioner that the physiological parameters presented herein can be reliably performed between trials and used to quantify the physiological responses to steady-state and time-trial load carriage activity. Data represented here show similar temporal changes as existing literature investigating the physiological responses to treadmill marching at 6.5 km·h−1 with 25 kg followed by a 2.4-km time trial (4,13), which are also key determinants of load carriage performance (9). Accordingly, the protocol presented presents a useful mode of exercise, which reflects the physiological demands of some operational tasks.
Baseline and changes in P Imax and P Emax were similar (p > 0.05) between trials demonstrating excellent agreement (Tables 3 and 4). There was low bias post-LC (P Imax 1.08, P Emax 0.91) and post-LCTT (P Imax1.09, P Emax 0.91) and small random error ratios post-LC (×/÷ P Imax 1.11, P Emax 1.24) and post-LCTT (×/÷ P Imax 1.30, P Emax 1.37) (Figures 4 and 5). The agreement ratio's for P Emax are larger than anticipated; however, data in Tables 3 and 4 demonstrate narrow 95% confidence intervals for both the upper and lower bound limits. It is likely that greater variability was witnessed in the post-LCTT measures because of the individual differences in recovery rate post-LC, adopted running speeds, and time taken to complete the time trial because reductions in respiratory muscle pressures are inversely related to exercise intensity. However, excellent ICC values post-LC (P Imax 0.93, P Emax 0.86) and post-LCTT (P Imax 0.86, P Emax 0.81), where >0.8 demonstrates a very good level of reliability (1) qualify their use in this protocol. We observed reductions in respiratory muscle pressure generation, which is illustrative of respiratory muscle fatigue, which are similar to reductions demonstrated recently from our laboratory (13). The implications of this reduction are far reaching in an occupational context because respiratory muscle fatigue may exacerbate limb muscle fatigue and impair performance through a sympathetically mediated reflex reduction in limb blood flow (13).
Reductions in pulmonary function under load carriage conditions have been previously observed (16). Thoracic restriction modifies the normal breathing mechanics of the exercise hyperpnoea response and consequently heightens the work of breathing (22). Measurements of pulmonary function were similar between trials at baseline, post-LC, and post-LCTT (p > 0.05; Table 2) and demonstrated excellent agreement with narrow confidence intervals and trial bias (Tables 3 and 4). Interestingly, however, agreement of peak expiratory flow (PEF) was poor, probably because of the effort dependence of this parameter during the initial high flow low volume phase of the maneuver. Large variations in PEF occur with only slight changes in the inspiratory muscle recruitment pattern during inspiration by the participants before expiration (20). If inspiration from functional residual capacity to total lung capacity (TLC) occurs rapidly and is subsequently followed by a forceful exhalation, a greater PEF will result compared with a controlled inspiration and longer time spent at TLC because of the elastic recoil properties of the thorax (23). Neither timing nor inspiration strategy was controlled here; and similar to existing research, we offer this as a potential explanation for the heightened variability in PEF postexercise (20). On this basis, studies should control inspiratory flow rate and the time spent at TLC during initial phase of the maneuver when conducting this measurement within a load carriage setting.
We present a novel protocol that can assess load carriage performance through a preloaded time-trial and a host of physiological markers that are relevant to the analysis of load carriage performance. This protocol can be adopted in settings where it is necessary to quantify load carriage performance, before and after a training intervention where load carriage performance is the dependant variable and before an operational deployment, where load carriage is considered to be a critical role-related task. Within physical selection tests, armed forces organizations do not currently use a loaded time trial, rather, they use unloaded running (typically 2.4–3.2 km distance). However, this is known to be a very poor predictor of occupational (i.e., load carriage) performance (3). The omission of a loaded time trial is likely to mitigate associated injury, although our findings suggests that including this test would allow the strength and conditioning practitioner to reliably capture changes in military specific fitness. It is impossible to devise a laboratory-based protocol that precisely reflects operational requirements because the load, intensity, duration, terrain, and environment vary with each deployment or training scenario. However, the protocol we present here is similar to current British Army assessments used in predeployment training and fitness assessments. For example, the 2.4-km loaded time trial has been used to determine the physical and physiological responses to acute changes in British Army infantry training programs (9,10). In addition, we present a protocol that is similar to the Infantry Basic Combat Fitness Test where soldiers complete an 8-mile course, carrying a load or 25 kg at 15 minutes per mile (6.4 km·h−1) and the Advanced Combat Fitness Test 1, whereby after an 800-m warm-up at 15 minutes per mile, participants perform a 2.4-km time trial carrying a 20-kg load (21). Thus, our protocol provides the strength and conditioning coach with a useful tool in assessing responses to training interventions and readiness for operations. In addition, as shown by the sample size calculations (Table 7), controlled trials can be performed effectively typically with a sample of 20 and 22 for the LC and LCTT trials, respectively during a repeated-measures design offering a large effect size (≥10%).
The results from this study suggest that the variability in parameters during LC and LCTT are negligible, and the load carriage protocol presented provides a reliable measure for assessing the performance and physiological effects of bearing an external load on the thorax, while also containing improved ecological validity on submaximal and high-intensity load carriage activities compared with previous measures. To our knowledge, this is the first study that has attempted to address the reliability of preloaded time-trial performance and has important implications for future research design and practice by strength and conditioning professionals.
1. Atkinson G, Nevill AM. Statistical methods for assessing measurement error (reliability
) in variables relevant to sports medicine. Sports Med 26: 217–238, 1998.
2. Bastien GJ, Willems PA, Schepens B, Heglund NC. Effect of load and speed on the energetic cost of human walking. Eur J Appl Physiol 94: 76–83, 2005.
3. Bilzon JLJ, Allsopp AJ, Tipton MJ. Assessment of physical fitness for occupations encompassing load-carriage tasks. Occup Med 51: 357–361, 2001.
4. Blacker SD, Fallowfield JL, Bilzon JL, Willems ME. Physiological responses to load carriage during level and downhill treadmill walking. Med Sport 13: 116–124, 2009.
5. Blacker SD, Williams NC, Fallowfield JL, Bilzon JL, Willems ME. Carbohydrate vs protein supplementation for recovery of neuromuscular function following prolonged load carriage. J Int Soc Sports Nutr 7: 1–11, 2010.
6. Blacker SD, Williams NC, Fallowfield JL, Willems ME. The effect of a carbohydrate beverage on the physiological responses during prolonged load carriage. Eur J Appl Physiol 111: 1901–1908, 2011.
7. Bland M, Altman D. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 327: 307–310, 1986.
8. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 14: 377–381, 1982.
9. Brown PE, Fallowfield JL, Blacker SD, Izard RM, Wilkinson DM, Bilzon JL. Does loaded marching economy explain some of the variance in loaded marching performance? Med Sci Sports Exerc 39: 154, 2007.
10. Brown PEH, Fallowfield JL, Blacker SD, Wilkinson DM, Izard RM, Bilzon JLJ. Physical, and Physiological responses to a modified british army infantry physical training programme: 1118. Med Sci Sports Exerc 42: 166, 2010.
11. Brown PI, Hughes MG, Tong RJ. The effect of warm-up on high-intensity, intermittent running using nonmotorized treadmill ergometry. J Strength Cond Res 22: 801–808, 2008.
12. Brown PI, McConnell AK. Respiratory-related limitations in physically demanding occupations. Aviat Space Environ Med 83: 424–430, 2012.
13. Faghy MA, Brown PI. Thoracic load carriage-induced respiratory muscle fatigue. Eur J Appl Physiol 114:1085–1093, 2014.
14. Knapik J, Harman E, Reynolds K. Load carriage using packs: A review of physiological, biomechanical and medical aspects. Appl Ergon 27: 207–216, 1996.
15. Knapik J, Harman EA, Steelman RA, Graham BS. A systematic review of the effects of physical training on load carriage performance. J Strength Cond Res 26: 585–597, 2012.
16. Muza SR, Latzka WA, Epstein Y, Pandolf KB. Load carriage induced alterations of pulmonary function. Int J Ind Ergon 3: 221–227, 1989.
17. Quanjer PH, Stanojevic S, Cole TJ, Baur X, Hall GL, Culver BH, Enright PL, Hankinson JL, Ip MS, Zheng J, Stocks J; ERS Global Lung Function Initiative. Multi-ethnic reference values for spirometry for the 3-95-yr age range: The global lung function 2012 equations. Eur Respir J 40: 1324–1343, 2012.
18. Rayson M, Holliman D, Belyavin A. Development of physical selection procedures for the British Army. Phase 2: Relationship between physical performance tests and criterion tasks. Ergonomics 43: 73–105, 2000.
19. Richmond VL, Rayson MP, Wilkinson DM, Carter JM, Blacker SD. Physical demands of firefighter search and rescue in ambient environmental conditions. Ergonomics 51: 1023–1031, 2008.
20. Romer LM, McConnell AK. Inter-test reliability
for non-invasive measures of respiratory muscle function in healthy humans. Eur J Appl Physiol 91: 167–176, 2004.
21. The British Army email@example.com. The British Army—Homepage [Internet]. 2014. Available at: http://www.army.mod.uk/
. Accessed on October 21, 2013.
22. Tomczak SE, Guenette JA, Reid WD, McKenzie DC, Sheel AW. Diaphragm fatigue after submaximal exercise with chest wall restriction. Med Sci Sports Exerc 43: 416–424, 2011.
23. Tzelepis GE, Zakynthinos S, Vassilakopoulos T, Geroulanos S, Roussos C. Inspiratory maneuver effects on peak expiratory flow: Role of lung elastic recoil and expiratory pressure. Am J Respir Crit Care Med 156: 1399–1404, 1997.
24. Verges S, Lenherr O, Haner AC, Schulz C, Spengler CM. Increased fatigue resistance of respiratory muscles during exercise after respiratory muscle endurance training. Am J Physiol-regul Integr Comp Physiol 292: R1246–R1253, 2007.
25. Zar JH. Biostatistical Analysis. Edinburgh Gate, Harlow, UK: Prentice-Hall/Pearson, 2010.