Conference PaperPDF Available

THE SYMMETRY ANGLE IDENTIFIES LESS CLINICALLY RELEVANT INTER-LIMB ASYMMETRIES THAN THE SYMMETRY INDEX IN HEALTHY ADULTS

Authors:

Abstract and Figures

There are several methods for calculating inter-limb symmetry, an inter-limb difference ≥15% has been suggested as an indicator of sporting injury risk. The purpose of this study was to compare three common methods for determining symmetry: the Symmetry Index (percentage difference; SI) when referenced to the left limb (SILeft) or the average of both limbs (SIAverage), and the Symmetry Angle (vector difference; SA). 15 recreationally active participants completed a sprint protocol on a non-motorised treadmill. Accelerometers were positioned on both tibias to measure peak resultant acceleration (PRA). The SA identified less clinically relevant PRA inter-limb asymmetries than the SI in healthy adults. Once an appropriate level of asymmetry as measured by the SA is determined, this may help to more correctly identify asymmetry in athletes and patients than the SI.
Content may be subject to copyright.
THE SYMMETRY ANGLE IDENTIFIES LESS CLINICALLY RELEVANT
INTER-LIMB ASYMMETRIES THAN THE SYMMETRY INDEX IN HEALTHY
ADULTS
Daniel J. Glassbrook1, Joel T. Fuller1, Jacqueline A. Alderson2, Jodie A. Wills1
and Tim L. A. Doyle1
Faculty of Medicine and Health Sciences, Macquarie University, Sydney,
Australia1
Faculty of Science, The University of Western Australia, Perth, Australia2
There are several methods for calculating inter-limb symmetry, an inter-limb difference
15% has been suggested as an indicator of sporting injury risk. The purpose of this study
was to compare three common methods for determining symmetry: the Symmetry Index
(percentage difference; SI) when referenced to the left limb (SILeft) or the average of both
limbs (SIAverage), and the Symmetry Angle (vector difference; SA). 15 recreationally active
participants completed a sprint protocol on a non-motorised treadmill. Accelerometers were
positioned on both tibias to measure peak resultant acceleration (PRA). The SA identified
less clinically relevant PRA inter-limb asymmetries than the SI in healthy adults. Once an
appropriate level of asymmetry as measured by the SA is determined, this may help to
more correctly identify asymmetry in athletes and patients than the SI.
KEYWORDS: Accelerometer, non-motorised treadmill, inertial measurement unit.
INTRODUCTION: Measuring biomechanical locomotor asymmetry of the lower-limb is
relevant for athletic and general populations because inter-limb differences of ~15% may be
indicative of injury risk (Knapik, Bauman, Jones, Harris, & Vaughan, 1991; Zifchock, Davis, &
Hamill, 2006). Moreover, lower-limb asymmetry may decrease athletic performance (Bishop,
Turner, & Read, 2018). There are several different methods to calculate locomotor asymmetry
(Bishop, Read, Chavda, & Turner, 2016), two common methods are the Symmetry Index (SI),
and Symmetry Angle (SA). The SI measures the percentage difference between two limbs
relative to the non-injured or dominant limb, or an average of both limbs. However, in healthy
populations there may not be a clear reference limb available and the choice of reference limb
can influence the percentage outcome, as the reference limb may not consistently provide the
smallest or largest value in the equation for each participant measured (Zifchock, Davis,
Higginson, & Royer, 2008). A study by Zifchock et al. (2008) also showed that using an average
of both limbs as reference produces a significantly smaller percentage asymmetry than when
referenced to one limb (left) (Zifchock et al., 2008). The SA was developed as an alternative
to the SI and is calculated by plotting a measure for the right side against a measure for the
left side (Xright,Xleft). A vector line is drawn from this point through the intersection of the x and
y axes, and the angle with respect to the x-axis is calculated. Two identical values will create
a 45° angle, indicating perfect symmetry. The vector angle can be converted to a percentage
and compared to SI results (Zifchock et al., 2008).
Wearable technology such as accelerometers have been suggested as a way to measure
meaningful biomechanical asymmetries in human locomotion (Willy, 2018). They have been
shown to reliably measure impact loading (Crowell & Davis, 2011), and foot-ground collisions
(Lucas-Cuevas, Encarnacion-Martinez, Camacho-Garcia, Llana-Belloch, & Perez-Soriano,
2017). However, it is unclear whether the SI or SA is most suitable to assess these results for
inter-limb asymmetry.
The purpose of this paper was to compare the SI with reference to the left side (SILeft) and the
SI referenced to the average of left and right sides (SIAverage) with the SA in peak resultant
acceleration (PRA) obtained during running.
METHODS: Fifteen recreationally active participants (Male n = 9, 23.9 ± 3.6 yrs, 1.8 ± 0.05 m,
78.3 ± 12.0 kg; Female n = 6, 27.3 ± 6.0 yrs, 1.7 ± 0.05 m, 66.3 ± 10.7 kg) volunteered to
participate in this study. Participants were eligible to participate if, at the time of recruitment,
they were: 1) aged 18-35 years, 2) free of injury, and 3) able to run without restriction. This
study was approved by the Macquarie University Human Research Ethics Committee (ethics
protocol number: 5201700532). Written informed consent was received from each participant
prior to participation.
Participants were required to attend a total of four sessions, separated by a minimum of 24-
hours recovery, over the course of a two-week period. The first three sessions were
familiarisation sessions that allowed participants to become accustomed to running on a non-
motorised treadmill (Force 3, Woodway USA, Inc., Waukesha, WI, USA). All data collection
occurred during the final session and all running was performed on the non-motorised
treadmill. Each session lasted approximately 20 minutes and was identical for both
familiarisation and data collection sessions. Participants performed a standardised warm-up
consisting of dynamic stretches and two minutes of steady-state running at 50-60% of self-
perceived maximal effort. After a 30-60 second standing rest period, participants ran for 60
seconds at 60% of self-perceived maximal effort and then immediately completed a 15 second
sprint at 70% of self-perceived maximal effort. The participant then ran for 60 seconds at 60%
self-perceived maximal effort as an active recovery. This sequence was repeated four more
times, with the sprint efforts of 80%, 90%, 100%, and 100%, respectively. Running at self-
perceived maximal effort is a reliable method of setting running speed on a non-motorised
treadmill (Tofari, McLean, Kemp, & Cormack, 2015). The athlete was tethered to a vertical
strut at the rear of the treadmill using a belt and cable so that they remained in place while
running on the treadmill belt (belt dimensions: 55 cm wide x 173 cm long) (Brown, Brughelli, &
Cross, 2016). Two accelerometers (iMeasureU, Auckland, New Zealand) measuring 40 x 28 x
15 mm and weighing 12 g were used to measure accelerations in three axes (x, y and z) at
500 Hz. The two accelerometers were attached to the distal medial tibial malleolus using velcro
straps in accordance with manufacturer recommendations.
Analysis of step-by-step acceleration data for each accelerometer was performed using
proprietary software (IMU_Step, version 1.0, iMeasureU, Auckland, New Zealand). The
variable of interest was the PRA for each step during the 100% sprint. Only the best quality
100% sprint effort for each participant was used for analysis. Data from each 100% sprint were
visually inspected to qualitatively determine the sprint with highest data quality. For each
participant the selected 100% sprint contained a minimum of 19 steps, and a maximum of 32
steps. Distal tibial PRA provides an indication of lower-limb loading, and increased lower-limb
loading is often suggested to increase injury risk (Crowell & Davis, 2011). Using the mean
absolute PRA value for the 100% sprint, the SILeft (equation 1), SIAverage (equation 2) and SA
(equation 3) were calculated (Zifchock et al., 2008):
   


(1)
     
 
(2)
      

(3)
Consistent with previous research, the left side was chosen for the SI equation with a single
side as a reference value as opposed to the right side (Zifchock et al., 2008).
Paired t-tests were used to assess any systematic bias between SILeft and SA, and between
SIAverage and SA. Pearson’s correlations were calculated to determine the relationship between
SILeft and SA, and between SIAverage and SA. Consistent with prior research, 15% was used as
the threshold for clinically significant asymmetry (Knapik et al., 1991; Zifchock et al., 2006). All
analysis was completed using SPSS software (version 23, IBM, Armonk, NY, USA). The alpha
level was set at 0.05.
RESULTS and DISCUSSION: Symmetry results for all participants, the number of participants
presenting with a clinically significant asymmetry for each equation, and the difference
between the SILeft and SA, and SIAverage and SA symmetry angles are presented in Table 1.
Table 1: Asymmetry results for each method for all participants.
Equation
Asymmetry (%)
No. of participants >15% asymmetry
SILeft
9.7 ± 7.6
5
SIAverage
9.8 ± 7.6
5
SA
3.1 ± 2.4
0
SILeft, Symmetry Index with reference to the left side; SIAverage, Symmetry Index with reference to the
average of left and right sides; SA, Symmetry Angle. Data are presented as mean ± SD.
Significant differences were observed between SILeft and SA (t(14) = 4.927, p < 0.001) and
between SIAverage and SA (t(14) = 4.943, p < 0.001). The SA was positively correlated with
SILeft (r = 0.989) and SIAverage (r = 1.000) (Figure 1).
Figure 1: Relationship between Symmetry Angle and Symmetry Index.
The results of this study support previous research that suggested the SA produces
significantly less asymmetry values than the SI, regardless of whether the left limb or average
of both limbs is used as a reference (Błażkiewicz, Wiszomirska, & Wit, 2014). If only the SI
equations were used, this study would have found that one third (5/15) of the participants in
this study had a clinically significant asymmetry between limbs. However, if the SA was used,
no participants would have been found to have inter-limb asymmetry >15%. Clinical significant
asymmetries were not expected because participants were only eligible to take part in this
study if they were free of injury and able to run without restriction. However, it is possible that
each participant may have presented with a level of inherent inter-limb asymmetry, resulting
from their training or sporting history, limb dominance, or leg-length discrepancies (Perttunen,
Anttila, Södergård, Merikanto, & Komi, 2004; Sadeghi, Allard, & Duhaime, 1997). Indeed, the
mean SI of resultant force metrics during sprinting were 3.9-9.6% in a previous study involving
healthy adults (Korhonen et al., 2010). As a result, it may not be reasonable to expect that
each participant in the present study should be without clinically significant asymmetry
between limbs in PRA.
The results of the SA should be interpreted with some caution because a clinically significant
asymmetry of >15% may not apply to these results (Zifchock et al., 2008). Knapik et al. (1991),
found that inter-limb asymmetries in isokinetic hip and knee strength >15% were a predictor of
injury in female collegiate athletes. Since then, a value of 15% asymmetry has been applied
to a variety of lower-limb measures. The SA was developed more recently (Zifchock et al.,
2008) and it is not known whether a 15% threshold for defining clinically significant asymmetry
is still appropriate. Therefore, when using the SA, the percentage chosen as being
representative of a clinically significant asymmetry may need to be investigated and modified.
Findings from the present study suggest that a threshold of 10% might be more appropriate
because SA values were approximately 6-7% lower than SI values.
CONCLUSION: This study used accelerometers attached to the distal tibia to measure inter-
limb asymmetries in peak acceleration during running. When deriving the percentage
difference between limbs, the result depends on the equation used to calculate the inter-limb
difference. The SA equation identifies less clinically relevant asymmetries than the SI equation
when referenced to the left lower-limb and the average of both lower-limbs. Future research
should validate a clinically relevant threshold of asymmetry using the SA because the 15%
threshold that is commonly used for the SI may not be appropriate.
REFERENCES:
Bishop, C., Read, P., Chavda, S., & Turner, A. (2016). Asymmetries of the Lower Limb: The
Calculation Conundrum in Strength Training and Conditioning. Strength and Conditioning
Journal, 38(6), 27-32.
Bishop, C., Turner, A., & Read, P. (2018). Effects of inter-limb asymmetries on physical and
sports performance: a systematic review. Journal of Sports Sciences, 36(10), 1135-1144.
Błażkiewicz, M., Wiszomirska, I., & Wit, A. (2014). Comparison of four methods of calculating
the symmetry of spatial-temporal parameters of gait. Acta of Bioengineering and
Biomechanics, 16(1), 29--35.
Brown, S., Brughelli, M., & Cross, M. (2016). Profiling sprint mechanics by leg preference and
position in rugby union athletes. International Journal of Sports Medicine, 37(11), 890-897.
Crowell, H. P., & Davis, I. S. (2011). Gait retraining to reduce lower extremity loading in
runners. Clinical Biomechanics, 26(1), 78-83.
Knapik, J. J., Bauman, C. L., Jones, B. H., Harris, J. M., & Vaughan, L. (1991). Preseason
strength and flexibility imbalances associated with athletic injuries in female collegiate
athletes. The American Journal of Sports Medicine, 19(1), 76-81.
Korhonen, M. T., Suominen, H., Viitasalo, J. T., Liikavainio, T., Alen, M., & Mero, A. A. (2010).
Variability and symmetry of force platform variables in maximum-speed running in young
and older athletes. Journal of Applied Biomechanics, 26(3), 357-366.
Lucas-Cuevas, A. G., Encarnacion-Martinez, A., Camacho-Garcia, A., Llana-Belloch, S., &
Perez-Soriano, P. (2017). The location of the tibial accelerometer does influence impact
acceleration parameters during running. Journal of Sports Sciences, 35(17), 1734.
Perttunen, J., Anttila, E., Södergård, J., Merikanto, J., & Komi, P. (2004). Gait asymmetry in
patients with limb length discrepancy. Scandinavian Journal of Medicine & Science in
Sports, 14(1), 49-56.
Sadeghi, H., Allard, P., & Duhaime, M. (1997). Functional gait asymmetry in able-bodied
subjects. Human Movement Science, 16(2), 243-258.
Tofari, P. J., McLean, B. D., Kemp, J., & Cormack, S. (2015). A self-paced intermittent protocol
on a non-motorised treadmill: a reliable alternative to assessing team-sport running
performance. Journal of Sports Science and Medicine, 14(1), 62-68.
Willy, R. W. (2018). Innovations and pitfalls in the use of wearable devices in the prevention
and rehabilitation of running related injuries. Physical Therapy in Sport, 29, 26-33.
Zifchock, R. A., Davis, I., & Hamill, J. (2006). Kinetic asymmetry in female runners with and
without retrospective tibial stress fractures. Journal of Biomechanics, 39(15), 2792-2797.
Zifchock, R. A., Davis, I., Higginson, J., & Royer, T. (2008). The symmetry angle: A novel,
robust method of quantifying asymmetry. Gait and Posture, 27(4), 622-627.
... IR and PCycle were examined as described in Chapters 3 and 4. As suggested by , the FMean/FPeak ratio was calculated for each stroke. Various symmetry index (SI) were calculated to quantify the differences between right and left for kinematic and kinetic variables (Equation 6) (Bishop, Read, Chavda, & Turner, 2016;Fohanno, Nordez, Smith, & Colloud, 2015;Glassbrook, Fuller, Alderson, Wills, & Doyle, 2018). Effective work per stroke (EWS) was calculated for each stroke as well (Equation 7), inspired by its strong positive correlations with boat velocity in other paddle sports (Baudouin & Hawkins, 2002;Kleshnev, 2006). ...
Thesis
Full-text available
Technological innovation has coincided with rapid improvements in performance in kayak sprint. With breakthroughs in materials, sensors, and wireless telemetry, instrumented equipment has unlocked new methods of quantifying and analysing performance. While some research has identified key biomechanical factors associated with performance, there is limited information available from on-water paddling and from new instrumented paddle devices. Thus, this thesis will improve the understanding of kayak performance through the examination of a novel instrumented paddle device. The device, developed by High Performance Sport New Zealand (HPSNZ), is capable of measuring blade force and velocity during normal on-water training conditions and may be adapted for ergometer use. A multi-level paddle protocol was created in partnership with coaches and support staff in order to replicate training and race intensities with maximal specificity. In Chapter 3, a study was performed to examine the validity and comparative reliability of the smart paddle (SP) relative to a popular kayak ergometer (DS). The SP and DS were practically identical in detecting stroke rate (SR) (limits of agreement = 0.02 ± 9.02%; R 2 = 0.98; p < 0.01), but there were detectable differences in pull time (TPull) (limits of agreement = 10.1 ± 18.4%, R2 = 0.78, p < 0.01) and peak force (FPeak) (limits of agreement = 8.8 ± 30.1 N, R 2 = 0.94, p < 0.01). Regardless, cyclical power variables were similar between SP and DS (SP IR and DS power; R 2 = 0.98, SE = 0.045, p < 0.01) across all intensities. Chapter 4 uses the SP to compare kinetic and kinematic variables between on-water and kayak ergometer paddle environments. Large significant differences in TPull (d = 5.9 ± 0.39), air time (TAir) (d = 3.7 ± 0.27), mean force (FMean) (d = 1.06 ± 0.19), peak force (FPeak) (d = 1.92 ± 0.22), Impulse (d = 2.62 ± 0.23), and impulse rate (IR) (d = 2.10 ± 0.21) were found between environments. Kinetic differences expanded at higher intensities, which were visually apparent in statistical parametric mapping (SPM) analyses. Notably, IR was quite similar at maximal intensity (d = 0.28 ± 0.27). In Chapter 5, the previous results and other literature were used to examine correlations between performance variables and boat speed. The strongest correlations and predictive power for kayak velocity (VKayak) were with IR (R 2 = 0.98, SE = 0.31, z = 1.86, p < 0.01) and cycle power (PCycle) (R 2 = 0.95, SE = 0.035, z = 2.08, p < 0.01). Allometric scaling increased the predictive power of most kinetic relationships. Strong correlations were observed between DPS and FPeak (r = 0.69 ± 0.10), FMean (r = 0.68 ± 0.11), peak power (PPeak) (r = 0.72 ± 0.10), and impulse (r = 0.65 ± 0.18). Paddling efficiency (ep) was estimated between 0.65-0.75, on average, for all intensities. These data expand the body of knowledge surrounding kayak sprint biomechanics, suggesting that specific performance variables can predict performance and detect differences between athletes, paddling intensity, and environments. Instrumented paddle devices are a powerful tool with more potential to be explored.
... Asymmetry between limbs in each category of acceleration for AUC and %Time was calculated using the Symmetry Angle Equation (Zifchock et al., 2008). Previous research has suggested that the symmetry angle equation results in significantly smaller asymmetry percentages (6-7% smaller) than traditional asymmetry equations such as the symmetry index (Błazkiewicz, Wiszomirska & Wit, 2014;Glassbrook et al., 2018). The threshold applied for defining clinically significant asymmetry in this study was 10% (Kvist, 2004;Schmitt, Paterno & Hewett, 2012). ...
Article
Full-text available
Background Quantifying lower-limb load and asymmetry during team sport match-play may be important for injury prevention and understanding performance. However, current analysis methods of lower-limb symmetry during match-play employ wearable microtechnology that may not be best suited to the task. A popular microtechnology is global positioning systems (GPS), which are torso worn. The torso location, and the summary workload measures calculated by GPS are not suited to the calculation of lower-limb load. Instead, research grade accelerometers placed directly on the lower-limb may provide better load information than GPS. This study proposes a new technique to quantify external mechanical load, and lower-limb asymmetry during on-field team sport play using inertial measurement units. Methods Four professional rugby league players (Age: 23.4 ± 3.1 years; Height: 1.89 ± 0.05 m; Mass: 107.0 ± 12.9 kg) wore two accelerometers, one attached to each foot by the boot laces, during match simulations. Custom Matlab (R2017b, The Mathworks Inc, Natick, MA) code was used to calculate total time, area under the curve (AUC), and percentage of time (%Time) spent in seven acceleration categories (negative to very high, <0 g to >16 g), as well as minimum and maximum acceleration during match simulations. Lower-limb AUC and %Time asymmetry was calculated using the Symmetry Angle Equation, which does not require normalization to a reference leg. Results The range of accelerations experienced across all participants on the left and right sides were 15.68–17.53 g, and 16.18–17.69 g, respectively. Clinically significant asymmetry in AUC and %Time was observed for all but one participant, and only in negative (<0 g) and very high accelerations (>16 g). Clinically significant AUC differences in very high accelerations ranged from 19.10%–26.71%. Clinically significant %Time differences in negative accelerations ranged from 12.65%–25.14%, and in very high accelerations from 18.59%–25.30%. All participants experienced the most AUC at very low accelerations (2–4 g), and the least AUC at very high accelerations (165.00–194.00 AU vs. 0.32–3.59 AU). The %Time results indicated that all participants spent the majority of match-play (73.82–92.06%) in extremely low (0–2 g) to low (4–6 g) acceleration intensities, and the least %Time in very high accelerations (0.01%–0.05%). Discussion A wearable located on the footwear to measure lower-limb load and asymmetry is feasible to use during rugby league match-play. The location of the sensor on the boot is suited to minimize injury risk occurring from impact to the sensor. This technique is able to quantify external mechanical load and detect inter limb asymmetries during match-play at the source of impact and loading, and is therefore likely to be better than current torso based methods. The results of this study may assist in preparing athletes for match-play, and in preventing injury.
ResearchGate has not been able to resolve any references for this publication.