Abstract
BACKGROUND:
Modifying the horizontal position of the load in a backpack will change the size of the external torque it creates on the wearer but the effect on postural stability is unclear.
OBJECTIVE:
To determine if changing the horizontal position of a fixed backpack load affects postural stability in young adults.
METHODS:
A backpack was attached to a steel frame with a bar protruding posteriorly. A fixed load (5% body mass) was placed at three distances along the bar – 0 m, 0.20 m, and 0.40 m. Centre of pressure (CoP) derived measurements were recorded from a force platform sampling at 100 Hz. For each condition participants performed three 90s narrow stance trials with their eyes closed whilst standing on a firm surface. A comparison was made across unloaded (no backpack) and loaded conditions.
RESULTS:
There was an immediate decrease in postural stability when a loaded backpack was worn. Only two of the CoP derived measures (Total Excursion - TEx, and Mean Velocity Total Excursion - MVel TEx) differed between the loaded at 0.20 m and loaded at 0 m conditions. All CoP derived measures differed between the loaded at 0.40 m and loaded at 0 m conditions. Furthermore, three of the CoP derived measures (Anterior/Posterior Root Mean Square - A/P RMSq, TEx, and MVel TEx) differed between the loaded at 0.40 m and loaded at 0.20 m conditions.
CONCLUSION:
The distribution of a load within a backpack must be carefully considered. The findings for the 0.40 m condition are important for the use and design of large backpacks used by multi-day hikers, travellers, and the military.
Introduction
People use backpacks to carry loads in certain occupations (e.g., emergency services and the military), when attending educational institutions (school and university), and during outdoor recreational activities (e.g., hiking and camping). When wearing a backpack, the centre of mass (CoM) of the system (body + backpack) is displaced vertically upwards and posteriorly. The vertical translation in the CoM may decrease stability because, according to the inverted pendulum model, stability is inversely related to the height of the CoM above the base of support [1]. Meanwhile the posterior translation of the CoM may cause an individual to exhibit anterior trunk lean to ensure that the CoM of the system stays within the base of support and a stable position can be maintained [2, 3].
Postural stability is maintained through the integration of visual, vestibular, and somatosensory input [3, 4]. This complex coordinated process can be assessed using centre of pressure (CoP) derived measurements from a force platform while a person is positioned in quiet standing [4]. Research has generally shown that a backpack loaded at approximately 15–30% of body mass will reduce the postural stability of the wearer compared to when standing without a backpack [2, 6]. Furthermore, as the magnitude of the load in the backpack increases there tends to be a progressive reduction in postural stability [7, 8]. Collectively, this would seem to indicate that wearing a loaded backpack can increase the potential for a fall and subsequent injury [5, 8].
What is less clear is how altering the position of the load within a backpack affects postural stability. There is some evidence that changing the vertical position of the load within a backpack does not affect postural stability [6]. Most backpacks, however, have internal compartments and/or external pockets that allow items to be separated from each other. This often causes some of the load to be shifted posteriorly and away from the wearer’s CoM. It has been hypothesised that adjusting the horizontal position of a load would adversely affect postural stability [2]. However, it is currently unclear whether this is the case or not [2, 7]. Therefore, the aim of this study was to determine if changes in the horizontal position of a fixed load when wearing a backpack will affect postural stability in a large sample of young adults.
Methods
Participants
This study was part of a larger project investigating sagittal plane cervical spine alignment and postural stability. Participants were primarily obtained from a university student population but offers were also extended to their families and friends. This resulted in 182 respondents. A comprehensive series of clinical and physical assessments were used to assess each person’s eligibility [9]. Thirty-two people failed at least one aspect of the screening process resulting in a final sample of 150 asymptomatic participants whose age, height and mass can be seen in Table 1. Ethics approval was granted from an institutional Research Ethics Committee (S/14/607).
Participant characteristics displayed as means (1 SD)
Participant characteristics displayed as means (1 SD)
n = Number of Participants.
Four conditions were established after consideration was given to the many different scenarios in which a backpack may be used: unloaded (no backpack), loaded at 0 m (i.e., at the base of the protruding bar), loaded at 0.20 m, and loaded at 0.40 m [10]. The loaded at 0 m and loaded at 0.20 m conditions were devised to represent the depths of typical backpacks. The loaded at 0.40 m condition was added firstly because it represents the depths of large backpacks used by multi-day hikers, travellers, and the military, and secondly to test the adaptive mechanisms of the participants. The unloaded condition was always completed first. The order of the loaded conditions was then randomized to address possible learning and fatigue effects.
To accommodate the loaded conditions a customised backpack was designed for this study (see Fig. 1). This involved attaching a steel frame with a bar extending posteriorly at approximately the height of the lumbosacral junction to be attached to a commercially available backpack (Kathmandu, Entrada Pack). This allowed a mass to be fixed at the desired distances along the bar. The mass of the backpack and frame was 6.1 kg. Pilot testing indicated that an additional load of 5% of the participant’s body mass would ensure the overall mass of the loaded backpack would pose a physical challenge without being a limiting factor that caused excessive muscular fatigue and prevented participants from completing any of the trials. The mean total mass of the backpack was 9.6 kg±0.7. Adjustments were made to the shoulder and waist straps so that the backpack was positioned comfortably but securely against the posterior trunk of each participant.

The backpack with the three loaded positions - left image is 0 m (i.e., at the base of the bar), middle image is loaded at 0.20 m, and right image is loaded at 0.40 m.
Ground reaction force data from a 400×600 mm force platform (Bertec Corporation, Columbus, USA) were recorded by a Qualisys motion capture system (Qualisys AB, Gothenburg, Sweden) at a sampling rate of 100 Hz [4, 11]. All six ‘best practice’ procedures outlined by Ruhe, Fejer [4] were followed to improve the reliability of CoP data. Participants had their eyes closed and stood barefoot on a firm surface with their great toes and heels touching (narrow stance position) so their feet were parallel [11–13]. A standardized set of instructions were provided to all participants. A trial was terminated and repeated if a participant: (1) opened their eyes during the trial, (2) lifted an arm or arms higher than 45° in any direction, (3) stepped off the platform, (4) stumbled or fell out of position, (5) lifted the forefoot or heel, and (6) was unable to correct the test position in 5s [14]. Three 90s trials were performed for each of the four conditions (unloaded, loaded at 0 m, loaded at 0.20 m and loaded at 0.40 m). A rest period of 60–90s was allowed between trials [11, 14].
Data analysis
Standard biomechanical analysis software (Visual3D, C-Motion, Inc. Maryland, USA) was used to process the data. Data were filtered using a fourth order zero lag Butterworth filter with a cut off frequency of 10 Hz [4, 13]. Eight standard CoP parameters were calculated for this study: Anterior/Posterior Range (A/P Range), Medial/Lateral Range (M/L Range), Anterior/Posterior Root Mean Square (A/P RMSq), Medial/Lateral Root Mean Square (M/L RMSq), Root Mean Square Distance (RMSq Dist), Total Excursion (TEx), Mean Velocity Total Excursion (MVel TEx), and 95% Confidence Circle (95% CC) as previously described [9].
Descriptive statistics for the postural stability data were calculated. The normality of the data was checked using standard procedures. A within-subjects repeated measures multivariate analysis of variance test was performed to determine changes in postural stability between the conditions. If significant findings were produced, then post hoc repeated-measures analysis of variance tests and pairwise comparisons (with Bonferroni adjustment) were performed. If there was a violation in the assumption of sphericity (Mauchly’s test) the analysis was corrected using either the Greenhouse-Geisser epsilon (if Greenhouse-Geisser epsilon <0.75) or the Huynh-Feldt epsilon (if Greenhouse-Geisser epsilon >0.75) [15]. Given the relatively large sample size, an alpha level of 0.01 rather than 0.05 was used to determine statistical significance to minimise the potential for a Type-1 error and provide more confidence in the statistical findings. Partial eta squared (
Results
The mean±SD for each of the postural stability measures across the unloaded and loaded conditions can be seen in Table 2. Using Pillai’s trace, results of the within-subjects repeated measures multivariate analysis of variance test revealed a statistically significant and large effect of the horizontal position of the load on CoP postural sway parameters (V = 0.73, F (22,128) = 15.3, p < 0.001,
The mean±SD for each postural stability measure across unloaded and loaded conditions
The mean±SD for each postural stability measure across unloaded and loaded conditions
*indicates significantly different from the unloaded condition, ∧indicates significantly different from the loaded @ 0 m condition, +indicates significantly different from the loaded @ 0.20 m condition.
Pairwise comparisons further revealed that: A/P Range significantly increased from unloaded to loaded at 0 m, 0.20 m, and 0.40 m (p < 0.001), and from loaded at 0 m to loaded at 0.40 m (p < 0.001); M/L Range significantly increased from unloaded to loaded at 0 m, 0.20 m, and 0.40 m (p < 0.001), and from loaded at 0 m to loaded at 0.40 m (p < 0.001); A/P RMSq significantly decreased from unloaded to loaded at 0 m (p = 0.004), significantly increased from loaded at 0 m to loaded at 0.40 m (p < 0.001), and significantly increased from loaded at 0.20 m to loaded at 0.40 m (p = 0.009); M/L RMSq significantly increased from unloaded to loaded at 0 m, 0.20 m, and 0.40 m (p < 0.001), and from loaded at 0 m to loaded at 0.40 m (p = 0.006); RMSq Dist significantly increased from unloaded to loaded at 0 m, 0.20 m, and 0.40 m (p < 0.001), and from loaded at 0 m to loaded at 0.40 m (p < 0.001); TEx significantly increased from unloaded to loaded at 0 m, 0.20 m, and 0.40 m (p < 0.001), from loaded at 0 m to loaded at 0.20 m (p = 0.001) and 0.40 m (p < 0.001), and from loaded at 0.20 m to loaded at 0.40 m (p < 0.001); MVel TEx significantly increased from unloaded to loaded at 0 m, 0.20 m, and 0.40 m (p < 0.001), from loaded at 0 m to loaded at 0.20 m (p = 0.001) and 0.40 m (p < 0.001), and from loaded at 0.20 m to loaded at 0.40 m (p < 0.001); and 95% CC significantly increased from unloaded to loaded at 0 m, 0.20 m, and 0.40 m (p < 0.001), and from loaded at 0 m to loaded at 0.40 m (p = 0.001). See Table 2.
People use backpacks to carry items for occupation and recreation. There has been a lot of research on how increasing backpack weight and/or altering the vertical position of the weight affects the wearer [17, 18]. Another important consideration is the way in which individual items are positioned within the backpack. Two backpacks may weigh the same amount but differences in the horizontal position of the weight within each backpack may change the external torque created by the backpack and therefore the response of the wearer. Therefore, the aim of this study was to determine if modifying the horizontal position of a fixed backpack load affects postural stability in young adults. This information could be used to inform recommendations on the use and design of backpacks across a number of different settings.
There was a significant difference between the unloaded condition and the loaded at 0 m condition for all the CoP derived measures of postural stability. All but one of these changes indicates an immediate decrease in postural stability when a loaded backpack is worn. This finding is generally consistent with most previous studies [2, 6]. From a mechanical perspective, upright stance is often modelled as an inverted pendulum, whereby the addition of a weighted backpack (0 m condition) translates the CoM of the system (body + backpack) vertically upwards and posteriorly and causes this system (body + backpack) to become more unstable [5]. It is interesting to note that wearing a backpack loaded at 0 m already appeared to challenge mechanisms responsible for controlling posture in both the anterior-posterior (primarily ankle dorsi and plantar flexors) and medial-lateral (primarily hip abductors and adductors - load-unload strategy) directions [19]. The narrow stance position adopted in this study to standardise the feet position of the participants created a small base of support that was narrower in the medial-lateral direction than in the anterior-posterior direction and this may have contributed to this finding.
Only two of the CoP derived measures of postural stability (TEx and MVeL TEx) differed between the loaded at 0.20 m condition and the loaded at 0 m condition. Interestingly none of the sagittal plane specific measures (A/P range, A/P RMSq) differed between these two conditions even though the load was centrally located and only adjusted in the posterior direction. The loaded at 0.20 m condition encompasses the depths of typical backpacks so this finding would seem to indicate that small additional loads (up to 5% of body mass) can be placed in external pockets of a backpack without adversely affecting most measures of postural stability. It may be important, however, that backpack wearers undergo the appropriate training to gradually become accustomed to the load sizes and placements that will be experienced in occupational or recreational activities [20].
All of the CoP derived measures of postural stability differed between the loaded at 0.40 m condition and the loaded at 0 m condition. Furthermore, three of the CoP derived measures of postural stability (A/P RMSq, TEx and MVeL TEx) differed between the loaded at 0.40 m condition and the loaded at 0.20 m condition. Collectively, these findings indicate that once the load is shifted more posteriorly to the 0.40 m position, participants are likely to experience significant reductions in postural stability. The loaded at 0.40 m condition further increases the moment arm for the external load and therefore the external torque created by the external load. Participants often exhibited anterior trunk lean (see Fig. 1) to help keep the CoM of the system (body + backpack) within the base of support and make sure the system’s rotational equilibrium was maintained. It may be postulated that the angular momentum created by the large external torque produced a delay in the capacity of feedback and control mechanisms to adjust and maintain upright posture, which may have increased postural sway away [7]. To avoid a loss of balance, corrective adjustments would have been required to keep the system (body + backpack) in rotational equilibrium, and this was seen through changes in the CoP derived measures of postural stability. Our study participants appeared to tolerate the loaded at 0.40 m condition for the duration of the test. However, this may not be the case for real world activities where people may be required to carry backpacks loaded in this way for extended periods of time. This could increase the potential for a fall and subsequent injury [5]. These efforts to maintain rotational equilibrium may have other implications for the musculoskeletal system as previous research by Daffin, Stuelcken [10] has shown that the loaded at 0.40 m condition adversely affected craniovertebral alignment in the sagittal plane to the extent that some people may be susceptible to neck pain and headaches.
The loaded at 0.40 m condition was used to test the adaptive mechanisms of the participants. It represents the depths of large traditional backpacks used by and recommended for multi-day hikers, travellers, and the military [21] but results from this study indicate that backpack wearers should be mindful of how items are positioned within the compartments and pockets and of the backpack and, if possible, ensure that heavier items are placed as close as possible to the trunk. Alternatively, large backpacks with compartments or pockets where load could be placed further than 0.20 m from the trunk should be avoided and potentially replaced by double packs (load positioned both in front and behind the wearer) of various designs [22, 23].
There are some limitations that need to be considered. Firstly, the study used young asymptomatic adults and findings may not be able to be generalised to different age groups. Secondly, the load was positioned at the height of the lumbosacral junction and results may have differed if the load was positioned more superiorly. The load was positioned at the height of the lumbosacral junction because it was the best option for minimizing the potential discomfort during the testing session. Thirdly, whilst it was ensured that the backpack was positioned comfortably but securely against the posterior trunk, it is unclear how the load was shared between the shoulder and waist straps for each participant. Finally, a fixed load was maintained throughout the different loaded conditions for each participant, and it is unclear how postural stability may have changed if loads of different magnitudes had been used.
Conclusion
To our knowledge, this is the first study to investigate whether modifying the horizontal position of a fixed backpack load affects postural stability in young adults. There was an immediate decrease in postural stability when a loaded backpack was worn. Only two of the CoP derived measures (TEx and MVeL TEx) differed between the loaded at 0.20 m and loaded at 0 m conditions. This condition encompasses the depths of typical backpacks so this finding would seem to indicate that small additional loads (up to 5% of body mass) can be placed in external pockets of a backpack without adversely affecting most measures of postural stability. All CoP derived measures differed between the loaded at 0.40 m and loaded at 0 m conditions. Furthermore, three of the CoP derived measures (A/P RMSq, TEx and MVeL TEx) differed between the loaded at 0.40 m and loaded at 0.20 m conditions. Collectively, these findings indicate that once the load is shifted posteriorly to the 0.40 m position, participants experienced significant reductions in postural stability. Therefore, people who use large backpacks (multi-day hikers, travellers, and the military) should be mindful of how items are positioned within compartments and pockets of the backpack and, if possible, ensure that heavier items are placed as close as possible to the trunk. This information could be used to inform recommendations on the use and design of backpacks across different settings.
Footnotes
Acknowledgments
The authors would like to acknowledge Dr. Mark Sayers for the guidance and mentoring he provided Lee Daffin throughout his candidature.
Conflict of interest
None to report.
