Accepted Manuscript

Title: Effects of external pelvic compression on isokinetic

strength of the thigh muscles in sportsmen with and without
hamstring injuries

Author: Ashokan Arumugam Stephan Milosavljevic

Stephanie Woodley Gisela Sole

PII: S1440-2440(14)00092-9
DOI: http://dx.doi.org/doi:10.1016/j.jsams.2014.05.009
Reference: JSAMS 1036

To appear in: Journal of Science and Medicine in Sport

Received date: 30-12-2013

Revised date: 7-5-2014
Accepted date: 17-5-2014

Please cite this article as: Arumugam A, Milosavljevic S, Woodley S, Sole G, Effects
of external pelvic compression on isokinetic strength of the thigh muscles in sportsmen
with and without hamstring injuries, Journal of Science and Medicine in Sport (2014),
http://dx.doi.org/10.1016/j.jsams.2014.05.009

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 1  Effects of external pelvic compression on isokinetic strength of the thigh muscles in
2  sportsmen with and without hamstring injuries
3  Authors:
4  1. Ashokan Arumugama
5  2. Stephan Milosavljevicb
3. Stephanie Woodleyc

t
6 

ip
7  4. Gisela Solea
8  Affiliations:

cr
a
9  School of Physiotherapy, University of Otago, New Zealand
b
10  School of Physiotherapy, University of Saskatchewan, Canada

us
c
11  Department of Anatomy, University of Otago, New Zealand
12 

an
13  Correspondence:
Name Ashokan Arumugam
Department School of Physiotherapy
Institution University of Otago
M
Country New Zealand
Tel 64 3 479 9619
Mob 64 21 0238 8003
Fax 64 3 479 8367
d

Email ashokan.arumugam@otago.ac.nz; ashokanpt@gmail.com
te

14 
15 
16  Word Count: 251 (Abstract)
p

17  3322 (Main body of the text)

18 
ce

19  References: 35
20 
21 
The University of Otago Human Ethics Committee approved this
Ac

22  Ethics approval:

23  study. Written informed consent was obtained from all participants
24  before data collection began.
25 
26  Source(s) of support: An internal grant from the Mark Steptoe Memorial Trust of School of
27  Physiotherapy, University of Otago; No external funding was received.
28 
29  Competing interests: None.

30   

31 

1 
 
Page 1 of 23
31  Abstract

32  Objective: To investigate whether application of a pelvic compression belt (PCB) affects isokinetic

33  strength of the thigh muscles in sportsmen with and without hamstring injuries.

34  Design: Randomised crossover, cross-sectional.

t
ip
35  Methods: Twenty sportsmen (age 22.0 ± 1.5 years) with hamstring injuries (hamstring-injured group)

36  and 29 (age 23.5 ± 1.5 years) without hamstring injuries (control group) underwent isokinetic testing

cr
37  of the thigh muscles. Testing included five reciprocal concentric quadriceps and hamstring

contractions, and five eccentric hamstring contractions at an angular velocity of 60°/s, with and

us
38 

39  without a PCB in randomised order. The outcome measures were average torque normalised to

bodyweight for terminal range eccentric hamstring contractions and peak torque normalised to

an
40 

41  bodyweight for concentric quadriceps, concentric hamstring and eccentric hamstring contractions.

42  Results: There was a significant increase in normalised average torque of eccentric hamstring
M
43  contractions in the terminal range for both groups (p ≤ 0.044) and normalised peak torque of eccentric

44  hamstring contractions for injured hamstrings (p = 0.025) while wearing the PCB. No significant
d

45  changes were found for other torque variables. Injured hamstrings were weaker than the contralateral
te

46  uninjured hamstrings during terminal range eccentric hamstring (p = 0.040), and concentric hamstring

(p = 0.020) contractions recorded without the PCB. However, no between-group differences were
p

47 

48  found for any of the investigated variables.

ce

49  Conclusion: Wearing the PCB appears to have a facilitatory effect on terminal range eccentric

50  hamstring strength in sportsmen with and without hamstring injuries. Future investigations should
Ac

51  ascertain whether there is a role for using a PCB for rehabilitation of hamstring injuries.

52  Key words - Athletic injury, Muscle strength dynamometer, Orthotic devices

2 
 
Page 2 of 23
53 

54  1. Introduction

55  Hamstring injury is reported to most commonly occur in either the terminal stance1 or swing phases1,2

t
56  of sprinting, associated with eccentric loading and lengthening of this bi-articular muscle group.

ip
57  Injured hamstrings also exhibit decreased torque and electromyographic (EMG) activity in the

cr
58  terminal range of eccentric isokinetic contractions.3 Assessment and rehabilitation of hamstring

59  injuries include multi-factorial strategies including examination of hamstring neuromotor control and

us
60  strength. Recent literature has also emphasised examination of lumbopelvic spine biomechanics and

61  motor control as potential factors contributing to hamstring injury.4 Moreover, an increase in

an
62  isokinetic concentric peak torque of injured hamstrings following manipulation of the sacroiliac joint

63  (SIJ) has been reported.5 Anatomically, the proximal tendon of the biceps femoris (long head) is
M
64  continuous in part with the sacrotuberous ligament.6 Thus, there appears to be a functional

65  relationship between the hamstring muscles and the lumbopelvic spine.
d
te

66  The use of a pelvic compression belt (PCB) is found to directly influence stability and mobility of the

67  SIJ,7 and also claimed to indirectly influence function of the hamstrings.8 While application of a PCB
p

68  appears to affect hamstring neuromotor control and strength,7,8 these relationships need to be explored
ce

69  further. Weakness of injured hamstrings has been hypothetically linked to injury recurrence9 and, if

70  so, there may be some merit in examining the effects of external pelvic compression on hamstring
Ac

71  neuromotor control as part of a multi-modal intervention plan.

72  According to a recent systematic review,7 studies have investigated the effects of external pelvic

73  compression on the isometric strength of muscles during a task (for example during lifting, and the

74  active straight leg raise) or certain muscle groups (low back, hip adductor) in individuals with or

75  without lumbopelvic dysfunction. There is some evidence that interventions (manipulation) directed

76  at the SIJ/lumbopelvic joints can affect thigh muscle strength5,10 supporting an argument for a

3 
 
Page 3 of 23
 77  neuromotor link between the pelvic and thigh regions. These putative structural and neuroreflexive

78  links between the hamstrings, pelvis and lumbar spine provide a research focus to determine whether

79  application of a pelvic compression belt (PCB) can alter the isokinetic strength of the thigh muscles in

80  sportsmen with and without hamstring injuries.

t
ip
81  2. Methods

cr
82  A randomised cross-over experimental design was used and the study was conducted in the Mark

83  Steptoe Laboratory of the School of Physiotherapy at the University of Otago. Ethical approval was

us
84  granted by the University of Otago Human Ethics Committee (Reference no. 11/115) and written

85  informed consent was obtained from all participants.

86 
an
Participants aged between 18 to 35 years were recruited by email, word of mouth, and adverts
M
87  displayed around the University. Sportsmen were included in the hamstring-injured group based on

88  self-report of injury3,11 if they had experienced an immediate onset of pain in the posterior aspect of
d

89  the thigh while playing sport12 within the previous 12 months, but not less than a month; the injury
te

90  necessitated intervention from a health professional and prevented participation in at least one match

91  or competition,13 or at least one week of usual sports training,14 within the previous 12 months.
p

92  Unilateral or bilateral, first-time or recurrent hamstring injuries were eligible for inclusion. Sportsmen
ce

93  without any previously diagnosed hamstring injury were recruited for the control group. A known

94  history of trauma/dysfunction in the lower limb (other than hamstring injury) or lumbopelvic region
Ac

95  within the previous six months that required intervention by a health professional excluded potential

96  participants from both groups. Further, those with any evidence of abnormal signs and symptoms

97  (other than those related to hamstring injuries) during clinical examination of the lumbopelvic region

98  and/or the lower limb were excluded. The ability of sportsmen to recall the history of injury within the

99  previous 12 months has been reported to be valid11 and, therefore, participants were recruited based

100  on their self-declaration of history of hamstring injury.

4 
 
Page 4 of 23
101 

102  Before isokinetic testing, anthropometric measurements (height, body mass and 4-point skin fold

103  measures) were recorded. To estimate body fat percentage, skin fold measurements were taken using

104  calipers (Slim Guide® caliper, Creative Health Products, MI) for the triceps, infrascapular, suprailiac

t
ip
105  and mid-thigh regions using standard guidelines.15 The sit-and-reach test was used to assess bilateral

106  hamstring flexibility.16

cr
us
107  Isokinetic tests were performed under two conditions, with and without the PCB. Data were collected

108  from both sides for the hamstring-injured participants and only one side (left or right) for the control

an
109  participants. The leg to be tested and the order of test conditions (PCB vs. no PCB) were randomized

110  using computer generated numbers.

M
111  The PCB (SI-brace neoprene-ADL-anatomisch; Rafys, The Netherlands) was applied just below the

112  anterior superior iliac spines (Fig. 1),7,17 and tightened maximally by the primary investigator (AA)
d

113  without any discomfort to participants. The amount of PCB tension achieved during isokinetic tests
te

114  was recorded using a load cell in a separate study on 10 healthy men. The mean PCB tension was
p

115  found to be 63.43 (± 9.90) N for reciprocal concentric quadriceps (ConQ) and concentric hamstrings
ce

116  (ConH) contractions, and 49.78 (± 5.70) N for eccentric hamstring (EccH) contractions. Participants

117  walked around the room between the conditions for at least 5 min18 to provide an adequate wash-out
Ac

118  effect.

119  A warm-up of 5 min of static cycling (60 rpm) was undertaken prior to testing. Participants were then

120  seated on a BiodexTM system 3 pro isokinetic dynamometer (Biodex Medical systems, NY) with a

121  trunk-hip angle of 100° (Supplementary Fig. 1). The mechanical axis of rotation of the dynamometer

122  was aligned with the lateral femoral epicondyle, and the shin pad was placed about 2 cm above the

123  medial malleolus. The effect of gravity on the leg was adjusted using the Biodex software after

124  placing the knee between 25° to 30° of extension. Participants were familiarised with the

5 
 
Page 5 of 23
125  dynamometer and a warm-up of the thigh muscles included a minimum of 10 sub-maximal

126  contractions followed by two maximal concentric and eccentric contractions at 60°/s.3 Five reciprocal

127  ConQ and ConH maximal contractions were then performed at an angular velocity of 60°/s followed

128  by five EccH contractions at 60°/s. The torque and velocity data were recorded at 200 Hz with the

129  Biodex software (version 3.30), within a range of motion of 90° during each contraction: 0° of

t
ip
130  extension (starting position) to 90° of flexion (end position) for ConH, and 90° of flexion to 0° of

131  extension for ConQ and EccH contractions. A rest period of 2 min was allowed between concentric

cr
132  and eccentric trials to minimise fatigue.19

us
133  Outcome measures included (gravity-corrected) peak torque (PT) normalised to bodyweight for

an
134  ConQ, ConH and EccH contractions, average torque normalised to bodyweight for the terminal range

135  of EccH contractions, and the functional torque ratio (PT EccH:PT ConQ). Further, the torque data of
M
136  EccH contraction were analysed from 85˚ to 5˚ knee flexion using a 50 ms epochs approach; the

137  initial and terminal 5˚ were omitted because they are essentially non-isokinetic. The outer range (≈

138  25˚ to 5˚ of knee extension) corresponded to the last six 50 ms epochs.3 The average torque of the
d

139  terminal movement quartile from five repetitions was normalized to bodyweight to allow comparison
te

140  between the test conditions (PCB vs. no PCB). The obtained value was multiplied by 100 to ensure
p

141  consistency with the results of the Biodex software.20 As the knee joint angle can vary from 20˚ to 33˚
ce

142  compared to dynamometer lever angle,3,21 the range of motion was calculated based on time instead of

143  dynamometer lever angle.

Ac

144  The effects of the PCB on dependent variables were investigated using paired t tests; the between-

145  group comparison of magnitude of change induced by the PCB was done using independent t tests for

146  any variable showing significant difference between the test conditions among groups. In addition,

147  independent t tests were used to explore between-group differences, and paired t tests to compare

148  within-group differences for the trials without the PCB. All statistical analyses were performed using

149  the IBM-SPSS software (Version 20, IBM, NY).

6 
 
Page 6 of 23
150  To compute effect size (d) for estimating the magnitude of change induced by the PCB on the

151  dependent variables, the formula devised by Cohen22 was used, assuming that the SDs of the test

152  conditions were not different. The following index was used to interpret effect sizes: small (0.20 ≤ d ≤

153  0.50), medium (0.50 ≤ d ≤ 0.80) and large (≥ 0.80).22

t
ip
154  3. Results

cr
155  Twenty sportsmen with hamstring injuries and 29 healthy sportsmen were included in the hamstring-

156  injured group and control group respectively. Demographic and anthropometric data and sporting

us
157  activities of all participants are presented in Table 1 and history relevant to those in the hamstring-

158  injured group and strength differences between limbs in unilateral hamstring-injured participants are

an
159  summarised in Table 2. Three participants had bilateral hamstring injuries. Among them, the onset of

160  injury for one limb of one participant did not occur within the required 12 month timeframe and,
M
161  therefore, this limb was excluded from statistical analyses. There was no significant difference

162  between the two groups for body fat (%) (p = 0.835); however, significant differences were found for
d

163  height, weight and BMI (p < 0.050). Therefore, instead of absolute values, normalisation using
te

164  bodyweight was used to reduce inter-subject variability within- and between-groups.
p

165  All participants in the hamstring-injured group had returned either partially or fully to sports training
ce

166  prior to data collection. Six reported continued discomfort during moderate exertion or sports

167  activities. Six others reported minor discomfort and/or soreness of the injured hamstrings during or
Ac

168  after strenuous/sporting activities. The remaining four participants described that they had returned to

169  a level of pre-injury training. Four participants were still undergoing rehabilitative exercises

170  prescribed by a physiotherapist.

171  There was a significant increase in the normalised average torque for EccH contractions in the

172  terminal range for participants with (p = 0.003) and without hamstring injury (p = 0.044), amounting

173  to 18.07 Nm/Kg (10%) and 10.08 Nm/Kg (5%), respectively, while wearing the PCB (d ≤ 0.33, Table

174  3). The magnitude of increase with the PCB was not significantly different between the groups (p =

7 
 
Page 7 of 23
175  0.275). In addition, with application of the PCB there was a significant increase in normalised PT

176  value of EccH contractions for the injured side by 11.44 Nm/Kg (5%), but not for other contractions,

177  in the hamstring-injured (p = 0.025, d = 0.22) but not the control group (p = 0.313, Table 3). There

178  was no significant difference between the test conditions for normalised PT values of ConQ and

179  ConH contractions, and the functional torque ratio (EccH:ConQ) for participants in both groups.

t
ip
180  Additional findings from the study are shown in a Supplementary Table. Normalised torque values

cr
181  and the functional toque ratio did not show statistically significant between-group differences.

us
182  However, the injured side was significantly weaker than the contralateral uninjured side by 18.63

183  Nm/kg (10%) during the terminal range of EccH contractions (p = 0.040, d = 0.34) and by 10.12

an
184  Nm/kg (7%) (p = 0.020, d = 0.27) during ConH contractions (Table 2). No statistically significant

185  differences were noted between sides for other type of contractions.
M
186  4. Discussion
d

187  The application of the PCB resulted in a significant increase in eccentric strength of the hamstrings in
te

188  the outer range for both participant groups. Specifically, eccentric strength of the injured hamstrings

189  was found to be weaker by 10% in the terminal range compared to the uninjured (contralateral)
p

190  hamstrings (Table 2) and wearing a PCB was found to improve this by an average of 10% (Table 3).
ce

191  The magnitude of change induced by the PCB in the hamstring-injured group (Supplementary Fig.

192  2A) was not significantly different from the control group (Supplementary Fig. 2B). The hamstring-
Ac

193  injured group included participants undergoing final stages of rehabilitation and those who had

194  already fully returned to sports training. This could have accounted for variations in injured

195  participants responses to the PCB resulting in an overall small effect size (d = 0.33). With the PCB,

196  there was also an increase in normalised EccH PT for the injured side of the hamstring-injured group

197  (5%) but not for the control participants.

8 
 
Page 8 of 23
198  The results of this study indicate that a PCB applied to the pelvis can affect eccentric hamstring

199  muscle performance. Previous studies have investigated the effects of other interventions, such as

200  manipulation, applied to the lumbopelvic spine on strength of the thigh muscles.5,10 In the current

201  study, there was no evidence of any change in normalised PT of ConQ and ConH contractions with

the application of the PCB in sportsmen with and without hamstring injury. Cibulka et al5 documented

t
202 

ip
203  an increase in ConH PT for injured hamstrings following manipulation of the SIJ in 10 participants,

while no significant change was found for ConQ contractions. The effect size for ConH PT in Cibulka

cr
204 

205  et al’s5 study was also small (d = 0.46) although the percentage change was equivalent to 22% for

us
206  ConH PT. Other studies reported an immediate increase in quadriceps strength up to 3% in healthy

207  individuals (n = 13)10 and 12% (n = 18)23 in individuals with anterior knee pain/patellofemoral pain

an
208  syndrome following lumbopelvic and SIJ manipulation, respectively. However, the test conditions

209  (manipulation vs. no manipulation) were not randomised in any of these studies. Irrespective of the
M
210  differences in methods and intervention used, all of these studies confirm a putative neuromotor link

211  between the pelvis and thigh muscles. The latent effects of SIJ manipulation on increased thigh

muscle strength are transient and the neuroreflexive pathway is uncertain;10 whether similar effects
d

212 

from external pelvic compression can be sustained while wearing the PCB needs further investigation.
te

213 
p

214  Various biomechanical and neurophysiologic mechanisms supporting the effects of the PCB on
ce

215  strength of hamstrings have been hypothesized.8 It is proposed that application of a PCB below the

216  anterior superior iliac spines can decrease sacral nutation24 by exerting pressure on the posteroinferior
Ac

217  aspect of the sacrum,25 potentially leading to decreased tension of the long head of biceps femoris

218  following relaxation of the sacrotuberous ligament.8 There is some evidence that a decrease in passive

219  hamstring stiffness (≈ 22%) occurs following core stability training.26 Thus, a reduction in hamstring

220  stiffness secondary to an improvement in lumbopelvic stability is plausible. Similarly, application of

221  the PCB might lead to a relative decrease in hamstring stiffness resulting from the extrinsic (reflexive)

222  and/or intrinsic (active sarcomeres and passive connective tissue) components leading to an increase

223  in the eccentric torque in the lengthened range. In addition to contractile components of the

9 
 
Page 9 of 23
224  hamstrings, the contribution of non-contractile components play an important role in resultant torque

225  generated during isokinetic testing in the lengthened ranges.8

226  As the innervation of the SIJ (L2 – S4), quadriceps (L2 - L4) and hamstrings (L5 – S2) share some

227  common nerve root levels, it is argued that altering the sensory input to one structure could possibly

t
228  influence motor output of all the structures that receive innervation from the same root levels.27 The

ip
229  effects of external pelvic compression might suppress descending inhibitory mechanisms resulting

cr
230  from nociceptors by blocking the pain gate at the spinal cord28 and, in turn, enhance the performance

231  of hamstrings in those with hamstring injury. Pressure on body parts has been reported to stimulate

us
232  descending pain inhibitory systems in the central nervous system29 and the PCB might produce similar

233  effects. The mechanisms underpinning peripheral hypoalgesic response induced in limbs following

an
234  external pelvic compression and following spinal mobilisation30 could hypothetically be similar.

235  Though spinal mobilisation and external pelvic compression have different mechanical stimuli, the
M
236  initial effects of interventions on the spine have been reported to have a sympathoexcitation mediated

237  hypolagesia specifically focused on mechanical nociception.30 This will need further validation for
d

238  external pelvic compression.

te

239  There were no statistically significant differences between groups for any of the isokinetic variables
p

240  obtained during trials without the PCB (Supplementary Table). These results agree with previous
ce

241  findings for PT and torque ratios (conventional and/or functional ratios) between hamstring-injured

242  and control participants.14,31 Sole et al. reported significant differences in outer range eccentric
Ac

243  hamstring torque between the injured and control participants of their study.3 Another study

244  concluded that hamstring-injured participants show decreased EccH PT at fast and slower velocities,

245  and decreased ConH and ConQ PTs at slower velocities than the control participants.32 The average

246  absence from sports participation due to injury was nearly 2 months in the study by Jonhagen et al.32

247  as opposed to other studies and the current investigation (range: 2-4 weeks). This reflects that their

248  participants could have had severe injuries with more functional limitations than others, thus

249  contributing to the differences in findings between the studies. However, these additional findings on

10 
 
Page 10 of 23
250  between-group differences should be interpreted after considering the fact that groups are different in

251  height, weight and BMI because the number of rugby players in the hamstring-injured group (45%)

252  was greater than the healthy group (7%) in this study. Exploring this further is neither the aim of this

253  study nor within the scope of this article.

t
ip
254  Though the mean increase in terminal range eccentric torque was 10% for the hamstring-injured

255  group, the range for this change was 44% to -13% for individual participants which indicate

cr
256  individual-specific responses to the PCB. In clinical practice, it would need to be assessed on an

us
257  individual basis whether the sportsperson with a hamstring injury responds positively to the PCB in

258  terms of symptoms and strength output, and pragmatically decide to intervene with the PCB as an

an
259  adjunct for rehabilitation. A previous study indicated that trunk stabilisation exercises decrease the

260  risk for hamstring injury recurrence.33 Whether the application of a belt as an adjunct to the exercises
M
261  might have similar effects could be evaluated. Further, wearing the belt for longer periods and effects

262  thereof at a functional level rather than just at an impairment level could be explored.
d

263  Participants in the hamstring-injured group were recruited based on their self-reported history of
te

264  hamstring injury with eligibility confirmed based on reproduction of symptoms whenever possible
p

265  during clinical examination and previous diagnosis of injury by a health professional. Only two of the
ce

266  20 participants underwent imaging investigations while the others were diagnosed and managed

267  clinically (Table 2). As this study recruited mainly community-level sportsmen, imaging was not
Ac

268  possible as part of standard care. Five participants, clinically diagnosed with hamstring injuries, were

269  classified as having minor injuries based on the period of absence from sports participation (Table 2).

270  Moreover, approximately 30% of athletes with clinically diagnosed minor or moderate hamstring

271  injuries are likely to have no MRI evidence of injury.34 Thus, our results apply to athletes with a

272  clinical diagnosis of hamstring injury without confirmation of changes on imaging.

273  Assessment of psychosocial factors (including emotional responses), functional limitations, and

274  kinesiophobia could help in understanding the influence of these factors on neuromuscular

11 
 
Page 11 of 23
275  performance of (injured) hamstrings. Being a cross-sectional and cross-over study, immediate effects

276  on neuromotor control of the lumbopelvic and hamstring muscles were assessed before and after

277  application of the PCB without considering possible psychosocial influences. However, while

278  examining baseline differences between groups these factors might be important and warrant

279  investigation in a future study.

t
ip
cr
280  5. Conclusion

281  Increased eccentric flexor torque in the lengthened range was found for sportsmen with and without

us
282  recent hamstring injuries with application of a PCB; however, the magnitude of increase was not

283  significantly different between groups. There was a deficit in eccentric torque in the lengthened range,

an
284  and concentric (peak) torque of injured hamstrings compared to uninjured hamstrings. Future studies

285  would need to confirm whether or not these residual strength deficits predispose to further injury. The
M
286  current study being a cross-sectional investigation cannot imply directly whether the PCB can be used

287  for eccentric training of the hamstrings which warrants further investigation.
d
te

288  6. Practical implications

p

289  • Injured hamstrings were found to be significantly weaker than uninjured hamstrings during
ce

290  terminal range eccentric (10%), and concentric hamstring (7%) contractions in sportsmen with

291  unilateral hamstring injury.

Ac

292  • Application of a pelvic compression belt significantly increased terminal range eccentric

293  strength of injured hamstrings by 10% in sportsmen with hamstring injury and uninjured hamstrings

294  by 5% in healthy sportsmen.

295  • Application of a pelvic compression belt did not change concentric (peak) torque of the

296  quadriceps and hamstring muscles in sportsmen with and without hamstring injury.

12 
 
Page 12 of 23
297  7. Acknowledgements

298  This study was supported by an internal grant from the School of Physiotherapy, University of Otago

299  to cover the associated research costs but there is no conflict of interest that could have influenced this

300  study.

t
ip
cr
us
an
M
d
p te
ce
Ac

13 
 
Page 13 of 23
301 

302  8. References

303  1 Yu B, Queen RM, Abbey AN et al. Hamstring muscle kinematics and activation during

t
ip
304  overground sprinting. J Biomech 2008; 41(15):3121-3126.

305  2 Thelen DG, Chumanov ES, Sherry MA et al. Neuromusculoskeletal models provide insights

cr
306  into the mechanisms and rehabilitation of hamstring strains. Exerc Sport Sci Rev 2006;

us
307  34(3):135-141.

308  3 Sole G, Milosavljevic S, Nicholson H et al. Selective strength loss and decreased muscle

an
309  activity in hamstring injuries. J Orthop Sports Phys Ther 2011; 41(5):354-363.

310  4 Mason D, Dickens V ,Vail A. Rehabilitation for hamstring injuries. Cochrane Database Syst

Rev 2007; 1:CD004575.

M
311 

312  5 Cibulka MT, Rose SJ, Delitto A et al. Hamstring muscle strain treated by mobilizing the

313  sacroiliac joint. Phys Ther 1986; 66(8):1220-1223.

d

314  6 Wingerden JPv, Vleeming A, Snijders CJ et al. A functional-anatomical approach to the

te

315  spine-pelvis mechanism: interaction between the biceps femoris muscle and the sacrotuberous
p

316  ligament. Eur Spine J 1993; 2(3):140-144.

ce

317  7 Arumugam A, Milosavljevic S, Woodley S et al. Effects of external pelvic compression on

318  form closure, force closure, and neuromotor control of the lumbopelvic spine – A systematic
Ac

319  review. Man Ther 2012; 17(4):275-284.

320  8 Arumugam A, Milosavljevic S, Woodley S et al. Can application of a pelvic belt change

321  injured hamstring muscle activity? Med Hypotheses 2012; 78(2):277-282.

322  9 Sole G, Milosavljevic S, Sullivan SJ et al. Running-related hamstring injuries: A

323  neuromuscular approach. Phys Ther Rev 2008; 13(2):102-110.

324  10 Grindstaff TL, Hertel J, Beazell JR et al. Effects of lumbopelvic joint manipulation on

325  quadriceps activation and strength in healthy individuals. Man Ther 2009; 14(4):415-420.

14 
 
Page 14 of 23
326  11 Gabbe B, Finch C, Bennell K et al. How valid is a self reported 12 month sports injury

327  history? Br J Sports Med 2003; 37(6):545-547.

328  12 Opar DA, Drezner J, Shield A et al. Acute hamstring strain injury in track-and-field athletes:

329  A 3-year observational study at the Penn Relay Carnival. Scand J Med Sci Spor 2014; In

330  Press. doi: 10.1111/sms.12159

t
ip
331  13 Bennell K, Wajswelner H, Lew P et al. Isokinetic strength testing does not predict hamstring

332  injury in Australian Rules footballers. Br J Sports Med 1998; 32(4):309-314.

cr
333  14 Brockett CL, Morgan DL ,Proske U. Predicting hamstring strain injury in elite athletes. Med

us
334  Sci Sports Exerc 2004; 36(3):379-387.

335  15 Peterson MJ, Czerwinski SA ,Siervogel RM. Development and validation of skinfold-

an
336  thickness prediction equations with a 4-compartment model. Am J Clin Nutr 2003;

337  77(5):1186-1191.
M
338  16 Liemohn W, Sharpe GL ,Wasserman JF. Criterion Related Validity of the Sit-and-Reach Test.

339  J Strength Cond Res 1994; 8(2):91-94.

17 Arumugam A, Milosavljevic S, Woodley S et al. Evaluation of changes in pelvic belt tension

d
340 

341  during 2 weight-bearing functional tasks. J Manipulative Physiol Ther 2012; 35(5):390-395.
te

342  18 Jung H-S, Jeon H-S, Oh D-W et al. Effect of the pelvic compression belt on the hip extensor
p

343  activation patterns of sacroiliac joint pain patients during one-leg standing: A pilot study.
ce

344  Man Ther 2013; 18(2):143-148.

345  19 White KK, Lee SS, Cutuk A et al. EMG power spectra of intercollegiate athletes and anterior
Ac

346  cruciate ligament injury risk in females. Med Sci Sports Exerc 2003; 35(3):371-376.

347  20 Lee MJC, Reid SL, Elliott BC et al. Running biomechanics and lower limb strength

348  associated with prior hamstring injury. Med Sci Sports Exerc 2009; 41(10):1942-1951.

349  21 Sorensen H, Zacho M, Simonsen EB et al. Joint angle errors in the use of isokinetic

350  dynamometers. Isokinet Exerc Sci 1998; 7(3):129-133.

351  22 Cohen J. Statistical power analysis for the behavioral sciences, 2nd edn, Hillsdale, NJ,

352  Lawrence Erlbaum Associates, 1988.

15 
 
Page 15 of 23
353  23 Suter E, McMorland G, Herzog W et al. Decrease in quadriceps inhibition after sacroiliac

354  joint manipulation in patients with anterior knee pain. J Manipulative Physiol Ther 1999;

355  22(3):149-153.

356  24 Sichting F, Rossol J, Soisson O et al. Pelvic belt effects on sacroiliac joint ligaments: a

357  computational approach to understand therapeutic effects of pelvic belts. Pain physician

t
ip
358  2014; 17(1):43-51.

359  25 Snijders CJ, Vleeming A ,Stoeckart R. Transfer of lumbosacral load to iliac bones and legs:

cr
360  Part 1: Biomechanics of self-bracing of the sacroiliac joints and its significance for treatment

us
361  and exercise. Clin Biomech 1993; 8(6):285-294.

362  26 Kuszewski M, Gnat R ,Saulicz E. Stability training of the lumbo-pelvo-hip complex influence

an
363  stiffness of the hamstrings: a preliminary study. Scand J Med Sci Spor 2009; 19(2):260-266.

364  27 Suter E, McMorland G, Herzog W et al. Conservative lower back treatment reduces inhibition
M
365  in knee-extensor muscles: A randomized controlled trial. J Manipulative Physiol Ther 2000;

366  23(2):76-80.

28 Wall PD. The gate control theory of pain mechanisms: A re-examination and re-statement.
d
367 

368  Brain 1978; 101(1):1-18.

te

369  29 Takeshige C, Sato T, Mera T et al. Descending pain inhibitory system involved in
p

370  acupuncture analgesia. Brain Res Bull 1992; l29(5):617-634.

ce

371  30 Wright A. Hypoalgesia post-manipulative therapy: a review of a potential neurophysiological

372  mechanism. Man Ther 1995; 1(1):11-16.

Ac

373  31 Worrell T, Perrin D, Gansneder B et al. Comparison of isokinetic strength and flexibility

374  measures between hamstring injured and noninjured athletes. J Orthop Sports Phys Ther

375  1991; 13(3):118-125.

376  32 Jonhagen S, Nemeth G , Eriksson E. Hamstring Injuries in Sprinters: The Role of Concentric

377  and Eccentric Hamstring Muscle Strength and Flexibility. Am J Sports Med 1994; 22(2):262-

378  266.

379  33 Sherry MA ,Best TM. A comparison of 2 rehabilitation programs in the treatment of acute

380  hamstring strains. J Orthop Sports Phys Ther 2004; 34(3):116-125.

16 
 
Page 16 of 23
381  34 Schneider-Kolsky ME, Hoving JL, Warren P et al. A comparison between clinical assessment

382  and magnetic resonance imaging of acute hamstring injuries. Am J Sports Med 2006;

383  34(6):1008-1015.

384  35 Arnason A, Andersen TE, Holme I et al. Prevention of hamstring strains in elite soccer: An

385  intervention study. Scand J Med Sci Spor 2008; 18(1):40-48.

t
ip
386 

cr
387   

us
388 

an
M
d
p te
ce
Ac

17 
 
Page 17 of 23
Table 1

Demographic and anthropometric data and sports participation of participants

Variable HIG (n = 20) CG (n = 29)

Age (years), mean (SD) 22.0 (1.5) 23.5 (1.5)

Anthropometric measurements, mean (SD)

t
Body weight (kg) 85.5 (14.4) 71.2 (10.9)

ip
Height (m) 1.81 (0.08) 1.76 (0.08)
BMI (kg/m2) 25.9 (3.4) 22.9 (2.7)
Body fat (%) 23.3 (3.4) 23.6 (4.4)

cr
Flexibility, mean (SD)
Sit-and-reach (cm)
23.1 (6.5) 23.8 (11.3)
Sports participation, n (%)

us
Rugby
Soccer/Football 9 (45) 2 (7)
Hockey 8 (40) 10 (35)
Ice hockey 1 (5) 4 (14)
Sprinting 0 (0) 1 (3)

an
Long distance running 1 (5) 1 (3)
Triathlon 0 (0) 2 (7)
Weight-lifting 0 (0) 1 (3)
Racquet sports 0 (0) 2 (7)
Cricket 1 (5) 2 (7)
M
Basket ball 0 (0) 2 (7)
0 (0) 2 (7)
CG, Control group; HIG, Hamstring-injured group.
d
p te
ce
Ac

19 
 
Page 18 of 23
Table 2.

Data relating to hamstring injuries

Characteristic, n (%)* Number (except where indicated)

Unilateral injury 17 (85)**

t
ip
$
Bilateral injury 3 (15)
‡
Recurrent injury 10 (50)
Imaging investigation

cr
Magnetic resonance imaging 1 (5)
Ultrasound 1 (5)
No imaging 18 (90)

us
Treatment history
Physiotherapy 19 (95)
Osteopathy 1 (5)

Injured leg

an
Preferred side 13 (57)
Nonpreferred leg 10 (43)

Muscle injured
Biceps femoris 11 (48)
M
Medial hamstring 12 (52)
€
Severity of injury
Minor 5 (25)
Moderate 6 (30)
d

Severe 9 (45)

Time since recent injury (months), mean (SD) 4.85 (3.97)

te

Time taken off from sports training due to injury (weeks), mean (SD) 3.55 (2.24)

Strength deficit of injured limb compared to uninjured limb in participants with

p

unilateral hamstring injury, mean (CI) (Nm/kg)

Terminal range EccH*** (25° - 5°) -18.63 (-36.27 to -1.00)
ce

ConQ PT† -12.96 (-35.77 to 9.85)

ConH PT† -10.12 (-18.46 to -1.78)
†
EccH PT -13.83 (-31.11 to 3.44)
Ac

ConH, concentric hamstring contraction; ConQ, concentric quadriceps contraction; CG, control group; CI,
confidence interval; EccH, eccentric hamstring contraction; PT, peak torque.
*Hamstring injury history based on self-report.
**One participant underwent an anterior cruciate ligament reconstruction surgery with bone-patellar tendon-bone
graft one year prior to the onset of hamstring injury.
$
Time of onset of injury was more than 12 months for one side of a participant.
‡
Time of these injuries ranged between 3 months and 5 years prior to the recent injury.
€
The severity of hamstring injury has been classified using the period of absence from sports participation as
minor (≤ 7 days), moderate (8 to 21 days) or severe (> 21 days).35
***Average torque normalised to bodyweight multiplied by 100.
†
Peak torque normalised to bodyweight multiplied by 100.

20 
 
Page 19 of 23
 i
cr
us
Table 3

Effects of application of the pelvic compression belt on isokinetic variables for the hamstring-injured group and control group

an
Isokinetic variables Test conditions, mean (SD) Difference between conditions within Paired t test (p value)
groups, (95% CI)

M
No belt With belt No belt – With belt No belt vs. With belt

HIG (n = 22) CG (n = 29) HIG (n = 22) CG (n = 29) HIG CG HIG CG

Terminal range EccH* 185.60 (56.40) 192.43 (46.03) 203.67 (51.20) 202.51 (43.37) -18.07 -10.08 0.003 0.044
(25° - 5°) (Nm/kg) (-29.37, -6.78) (-19.86, -0.30)

ConQ PT† (Nm/kg) 266.28 (59.93)

ed 286.26 (37.90) 272.73 (55.18) 282.90 (44.84) -6.45
(-16.56, 3.66)
3.36
(-5.82, 12.54)
0.199 0.459
pt
†
ConH PT (Nm/kg) 150.02 (42.03) 148.49 (27.20) 152.04 (42.46) 147.30 (28.98) -2.02 1.19 0.824 0.634
(-20.64, 16.60) (-3.87, 6.28)
ce

EccH PT† (Nm/kg) 227.42 (51.84) 233.13 (42.97) 238.86 (48.50) 237.73 (40.29) -11.44 -4.60 0.025 0.313
(-18.53, -4.35) (-13.78, 4.57)
Ac

Functional torque ratio 86.56 (13.14) 83.01 (21.16) 88.77 (15.29) 85.50 (16.11) -2.21 -2.49 0.413 0.312
(EccH:ConQ) (-7.66, 3.24) (-8.62, 3.65)

ConH, concentric hamstring contraction; ConQ, concentric quadriceps contraction; CG, control group; CI, confidence interval; EccH, eccentric hamstring contraction; HIG,
injured side of hamstring-injured group; PT, peak torque.
*Average torque normalised to bodyweight multiplied by 100.
†
Peak torque normalised to bodyweight multiplied by 100.

21 
 

Page 20 of 23
 i
cr
us
Supplementary Table

an
Isokinetic variables obtained without application of the pelvic compression belt for the hamstring-injured and control groups

Isokinetic variables Groups, mean (SD) Differences, (95% CI) p value

M
CG (n = 29) HIG (n = 22) CG vs. HIG CG vs. HIG‡

Terminal range EccH* 192.43 (46.03) 185.60 (56.40) 6.83 (-21.99 to 35.66) 0.636
(25° - 5°) (Nm/kg)

ConQ PT† (Nm/kg) 286.26 (37.90)

ed 266.28 (59.93) 19.98 (-9.68 to 49.65) 0.180**

pt
ConH PT† (Nm/kg) 148.49 (27.20) 150.02 (42.03) -1.52 (-22.43 to 19.39) 0.883**
ce

EccH PT† (Nm/kg) 233.13 (42.97) 227.42 (51.84) 5.71 (-20.98 to 32.40) 0.669

Functional torque ratio 83.01 (21.16) 86.56 (13.14) -3.55 (-13.87 to 6.77) 0.498
Ac

(EccH:ConQ)

ConH, concentric hamstring contraction; ConQ, concentric quadriceps contraction; CG, control group; CI, confidence interval; EccH, eccentric hamstring contraction;
HIG, Injured side of hamstring-injured participants; PT, peak torque.

*Average torque normalised to bodyweight multiplied by 100.

†
Peak torque normalised to bodyweight multiplied by 100.
‡
independent t test.
**Equal variances not assumed.

22 
 

Page 21 of 23
Figure Captions –

Fig.1. Position of the pelvic compression belt as used in the study.

t
ip
cr
us
an
M
d
p te
ce
Ac

23 
 
Page 22 of 23
Figure(s)

i
cr
us
an
M
ed
pt
ce
Ac

Fig. 1. Position of the pelvic compression belt as used in the study. Page 23 of 23