Participants
Twenty-three women (22.3 ± 2.6 years of age, 168.5 ± 5.0 cm height, 66.3 ± 6.5 kg body mass) who were professionally fitted [22] to wear a D+ cup bra size (average cup size DD, range D to E; average band size 10, range 10 to 16) were recruited as representative of women with large breasts. The cohort was deliberately homogenous in age range, mass and physical activity levels (all reported to exercise in a sports bra approximately 5 h per week) [23] to eliminate differences due to these variables and to ensure that sports bras were commonly used by the participants. Exclusion criteria included previous breast surgery, currently experiencing menopause, being pregnant or breastfeeding, or suffering from any musculoskeletal disorder or pain that prevented treadmill running. Based on a power analysis using G*power 3.1.3 and that a difference of ±2 on a visual analogue scale (VAS; rated 0 to 10) measuring discomfort was deemed significant [19], it was estimated that a minimum sample size of 22 participants was required to achieve statistical power of at least 80% (with a significance level of p < 0.05). To account for potential participant drop out, 23 participants were recruited for the present study. Participants were not tested if they were experiencing any breast tenderness associated with their menstrual cycle. Each participant completed a short questionnaire about their current sports bra usage and provided written informed consent before participating in the study. The University of Wollongong Human Research Ethics Committee (HE12/118) approved recruitment and testing procedures, and all testing was conducted in accordance with the National Statement on Ethical Conduct in Human Research [24].
Experimental design
A within-subject design was used, where participants ran in a standardised manner and speed (average speed: 9.1 ± 0.3 kph) on a treadmill (PowerJog, GX-100; Expert Fitness UK, Glamorgan, UK) while wearing the same style of encapsulation sports bra (New Legend Underwire sports bra, Berlei, Wentworthville, NSW, Australia), which provided a high level of support and is recommended for women with large breasts to wear during high-impact physical activity [22]. Six randomly [25] allocated bra strap conditions were trialled, including two bra strap orientations (vertical and cross-back) and three different bra strap designs (standard width, wide and gel). All of the straps were made and sewn by the primary investigator [CEC] to ensure a standardised bra strap structure across the three strap designs and a standardised length of the two strap orientations per participant, which were longer in the cross-back orientation compared to the vertical orientation. Professional bra fitting criteria [5,22] were used to ensure the bras fitted the participants correctly and that each strap was the correct length for each participant, in each strap condition (strap length was adjusted once the gel pad was added), as both incorrect bra fit and insufficient strap length could bias both bra strap discomfort and pressure measurements. Following adequate familiarisation, the participants ran in each bra strap condition for 3 min, with data collected while the participants stood motionless (static condition) prior to running and then between the first and third minute of running (dynamic condition). A three-minute duration was chosen as the baseline duration for each condition to minimise participant burden and to facilitate comparison of the strap conditions by limiting the time between them, although it is acknowledged that women commonly exercise and wear sports bras for much longer durations. All the running trials were closely supervised to standardise both the mode of running and the speed, and at least 5 min of rest was allowed between conditions. Participants wore their own running shorts and shoes, which were checked to ensure that they were appropriate for the running task.
Experimental bra strap conditions
The standard width and wide strap designs were made of the same material, consistent with commercially available encapsulation sports bra straps (see Figure 1). The gel strap design consisted of the standard width strap, with the addition of a 2.5-cm-wide gel pad (Dermis Plus Polymer gel; MacMed Health Care, Mudgeeraba, Queensland, Australia) placed under the bra strap (see Figure 2). Each strap was secured to the test bra in the same manner, anteriorly with bikini hooks (15 mm × 2 mm white plastic hooks; Birch Haberdashery & Craft, Heidelberg, Victoria, Australia) and posteriorly with hook and loop tape (Birch Haberdashery & Craft, Heidelberg, Victoria, Australia), staples and strapping tape. Each participant was provided with a new test bra, two new sets of bra straps (standard width (2.5 cm) and wide (4.5 cm)) and two unused gel strips for hygiene purposes and to eliminate potential effects of wear or washing on the bra straps.
Bra strap discomfort and strap preference
Bra strap discomfort was measured immediately after the running trials using a VAS (rated 0 to 10), whereby 0 represented ‘no discomfort’ and 10 represented ‘worst possible discomfort’. The perceived reason for any bra strap discomfort was also reported, and at the end of the test session, participants selected their most and least preferred bra strap orientation and design.
Bra strap pressure
Strap pressure exerted on the shoulders (bra strap-shoulder interface) was measured while the participants stood stationary and upright prior to running (static condition) and then while they ran on the treadmill (dynamic condition) during each strap condition. Pressure was measured using a custom-designed 10 mm2 calibrated pressure sensor (0.5 to 24 kPa range, 50 Hz, S2011, Novel GmbH, Munich, Germany) placed under the right bra strap at the crest of each participant’s shoulder, where the bra strap traversed the shoulder and exerted a downward force on the sensor. The sensors were secured with micropore surgical tape (3M™ Australia, Sydney, NSW, Australia) and zeroed prior to the 10-s of static pressure data collection, the six 10-s samples during the steady state treadmill running and the 10-s static pressure recording once the running was completed. Pliance-x Expert Online software (Version 10.3, Novel GmbH, Munich, Germany) was used to calculate the average static pressure (kPa), and the average dynamic peak pressure (kPa), which was taken as the average of the six 10-s periods per bra strap condition.
Vertical breast displacement
Vertical breast displacement (VBD; cm) relative to the torso was measured using an Optotrak Certus® motion capture system (200 Hz, Northern Digital, Ontario, Canada) during dynamic treadmill running to determine whether breast motion was consistent among the different strap conditions. Two infrared-emitting diodes (2 mm diameter) were placed on each nipple using double-sided toupee tape (Creative Hair Products, Melbourne, Victoria, Australia), which was placed over micropore surgical tape (3M™ Australia, NSW). A third diode was placed on the sternal notch as a reference point to characterise trunk motion in the vertical plane. Three-dimensional motion of the three markers was recorded during each running trial for six 10-s periods using First Principles software (Version 1.2.2, Northern Digital Inc., Ontario, Canada). The average VBD (minimum from maximum during dynamic treadmill running) relative to the trunk was calculated from a representative 8-s epoch (equivalent to 15 to 20 consecutive breast cycles) for each of the six 10-s data recordings per condition.
Statistical analysis
Frequencies of the questionnaire responses and the bra strap preference data for each bra strap condition were calculated. After confirming the data were normally distributed, means and standard deviations were calculated for, these data, as well as, bra strap pressure and VBD data for each strap condition. A two-way ANOVA design with two within factors (strap orientation and strap design) was then used to determine whether there were any significant main effects or interactions of strap orientation (vertical, cross-back) or strap design (standard width, wide, gel) on the outcome variables, with Tukey post hoc analyses used to determine where any significant difference lay. All statistical procedures were conducted using the Statistical Package for the Social Sciences (Version 15.0; SPSS Inc., Chicago, IL, USA).