ORIGINAL RESEARCH

Physiological Achilles Tendon Length and Its Relation to

Tibia Length
Claudio Rosso, MD, MSc,* Philipp Schuetz, MD, MPH, Caroline Polzer, MD,* Lukas Weisskopf, MD,
Ulrich Studler, MD, and Victor Valderrabano, MD, PhD*

Objective: The optimal intraoperative Achilles tendon length

(ATL) adjustment is crucial for the physiological functioning of
the musculotendinous unit. To date, the resting ATL and its relation
to tibia length (TL) have never been dened in healthy subjects. We
thus performed metric measurements of the ATL and TL.

Design: Case series.

between ATL, TL and body height and dened an algorithm of

ATL based on TL. The ATL and the ATL-algorithm might be
important in patients with impaired tendons such as AT ruptures.
Key Words: Achilles tendon length, MRI, tibia length, BMI, Achilles tendon length algorithm
(Clin J Sport Med 2012;22:483487)

Setting: Clinical and radiological measurements.

Participants: Fifty-two subjects placed in a 3-T magnetic
resonance imaging with the ankle in neutral position.

Main Outcome Measures: Unilateral ATL measured from the

calcaneal insertion to the beginning of the medial gastrocnemius muscle,
TL measured from the intercondylar eminence to the center of the ankle,
and qualitative tendon parameters in T2 sequences and human
parameters were noted; results were correlated with age, gender, body
height, weight, body mass index (BMI), and side of the AT and TL.
Results: The mean ATL was 180.6 25.0 mm and the mean TL
was 371.9 25.4 mm with an ATL:TL ratio of 49 5%. Achilles
tendon length correlated signicantly with body height (R2 = 38%,
P , 0.0001) and with TL (R2 = 41%, P , 0.0001) but did not
correlate with age, BMI, and side of the AT. Tibia length correlated
with body height (R2 = 83%, P , 0.0001) and in multivariate linear
regression, TL was the only independent predictor of ATL following
the algorithm, ATL (mm) = 0.6 TL (mm) 2 53 (R2 = 41%).

Conclusions: We dened a new way to measure the ATL in

a consistent way in healthy subjects and showed correlations
Submitted for publication February 20, 2012; accepted June 7, 2012.
From the *Orthopaedic Department, University Hospital Basel, University of
Basel, Basel, Switzerland; Harvard School of Public Health, Boston,
Massachusetts; Rennbahn Clinic, Muttenz, Switzerland; and Department
of Radiology, University Hospital Basel, University of Basel, Basel,
Switzerland.
Supported by a grant from the Swiss National Accident Insurance Company
(SUVA). Each author certies that he or she has no commercial
associations (eg, consultancies, stock ownership, equity interest, patent/
licensing arrangements, etc.) that might pose a conict of interest in
connection with the submitted article. In addition, each author certies
that his or her institution has approved the human protocol for this
investigation, that all investigations were conducted in conformity with
ethical principles of research and that informed consent for participation
in the study was obtained.
The authors report no conicts of interest.
Corresponding Author: Claudio Rosso, MD, MSc, Orthopaedic Department,
University Hospital of Basel, University of Basel, Spitalstrasse 21, 4031
Basel, Switzerland ([email protected]).
Copyright 2012 by Lippincott Williams & Wilkins

Clin J Sport Med
Volume 22, Number 6, November 2012

INTRODUCTION
Rupture of the Achilles tendon (AT) is a common
problem in sports medicine, with an estimated incidence of 10
to 20 per 100 000 or roughly 23 000 operative and nonoperative treatments per year in the United States alone.1 The AT,
as with other tendons in the human body, acts as a spring and
can store energy.2 This energy storage capacity is particularly
important for both running and normal walking, where the
energy storage within the AT accounts for 8% and 6%,
respectively, of the total mechanical energy.36 Central to the
ATs ability to store energy is its nominal or resting length.
It is thus crucial that the original length is restored
during surgical treatment to assure optimal function. The
potential sequelae of repairing an AT in a lengthened
condition can be decreased muscle volume,7 fatty muscle
atrophy,8 and decreased force.911 These facts emphasize the
importance of AT length (ATL) as an inuential variable in
the postoperative functional and clinical outcome after AT
repair.12 The amount of lengthening resulting in reduced function is currently unknown.
Given the importance to accurately assess and restore
ATL after rupture, this study had 2 objectives: rst, to
develop a new way to accurately and reliably establish ATL
on magnetic resonance imaging (MRI) of healthy subjects;
and second, to develop an easy formula to calculate ATL
based on human parameters like gender, age, body height,
body mass index (BMI), activity level, and tibia length (TL).

SUBJECTS AND METHODS

For this case series, 52 subjects (11 women and 41
men) were recruited. One leg was evaluated in each subject
(Table 1). Patients with a unilateral healthy leg with physiological alignment of the knee, foot, and ankle determined
clinically; without neuromuscular impairments including
muscle dystrophies; without posttraumatic injuries; and without osteoarthritis of the knee, foot, or ankle were included.
Exclusion criteria were general MRI exclusion criteria
www.cjsportmed.com |

483

Clin J Sport Med
Volume 22, Number 6, November 2012

Rosso et al

TABLE 1. Outcome Measures

Parameter
Age, y
Body height, cm
Weight, kg
BMI, kg/m2
ATL, mm
TL, mm
ATL of TL, %
Tendon morphology
Tendon sheath uid

Whole Cohort (n = 52)

48.6
176.8
82.1
26.2
180.6
371.9
53.3
0.77
0.71

8.7
8.8
12.9
3.9
25.0
25.4
5.1
0.76
0.85

Women (n = 11)
49.9
166.5
70.8
25.7
162.5
342.6
53.1
0.73
0.64

6.5
5.8
15.5
6.3
18.3
15.6
4.8
0.79
0.81

Men (n = 41)
48.2
179.6
85.1
26.4
185.4
379.7
53.4
0.78
0.73

9.3
7.2
10.4
3.0
24.5
21.5
5.2
0.76
0.87

P*
0.57
,0.0001
0.001
0.59
0.006
,0.0001
0.88
0.84
0.74

Values are given as mean SD.

ATL, Achilles tendon; BMI, body mass index; TL, tibia length.
*One-way analysis of variance (men vs women).

(eg, pacemaker, other metal not qualifying for an MRI, tattoos, and claustrophobia) and a BMI of . 40 kg/m2.
The following subject parameters were noted: gender,
age, body height, weight, BMI, and activity level according to
the Valderrabano Sports Score.13 On average, the male population was taller (179.6 7.2 cm, P , 0.0001) and heavier
(85.1 10.4 kg, P = 0.001), but there was no difference in age
(P = 0.57) or BMI (P = 0.59) between the groups (Table 1).
Imaging was carried out on a 3-T scanner (MAGNETOM Verio; Siemens Medical Solution, Erlangen, Germany).
Subjects were examined in the supine position with straight
legs, that is, the ankle dorsiexed at 90 degrees (exion
neutral position) and the knee in 0-degree exion and xated
in that position with straps and sandbags to assure reliable
side comparability. A peripheral angiography matrix coil
with 16 coil elements was used for image acquisition. Images
were obtained in the coronal plane with a single-slab T1weighted magnetization-prepared rapid acquisition gradientecho (MPRAGE) sequence (repetition time [milliseconds]/

echo time [milliseconds], 1890/3.37; inversion time [milliseconds], 1000; isotropic resolution, 1 1 1 mm). The
MRPRAGE protocol was used to cover the entire lower leg
from above the femoral condyles through the bottom of the
calcaneus allowing a 3-dimensional reconstruction with slices
of 1 mm. Axial T2-weighted fat-saturated fast spin-echo
images (5000/76; echo train length, 14; number of signals
acquired, 1; section thickness, 3 mm; intersection gap, 0 mm;
matrix size, 212 512; eld of view, 168 400 mm) of both
legs were acquired simultaneously from the musculotendinous
junction to the calcaneal insertion of the Achilles.
Quantitative and qualitative measurements were performed. Quantitative parameters such as ATL and TL were
measured in the MPRAGE sequences. The length of the AT
was measured in transverse (axial) planes by identifying the
tendon at its insertion at the calcaneus as the rst slice in
which the calcaneus bone touches the AT dening the distal
point. The proximal point was dened as the last slice where
the gastrocnemius medialis did not appear when working

FIGURE 1. Illustration of a novel way

to measure ATL. The gastrocnemius
medialis (GM) muscle inserts more
distally than the gastrocnemius lateralis (GL) as part of the MTJ (A). In (B),
the GM can still be seen. The insertion
of the GM is identified in axial slices as
the last slice not showing the muscle
(C). The distal insertion where the AT
inserts at the calcaneus is then identified (D). The ATL is then the number
of 1-mm slices between these 2 defined segments. For demonstration
purposes, T1-weighted instead of
MPRAGE images are shown here.
CALC, calcaneus; dist., distal; prox.,
proximal; SOL, soleus.

484

| www.cjsportmed.com

 2012 Lippincott Williams & Wilkins

Clin J Sport Med
Volume 22, Number 6, November 2012

from distally to proximally (Figure 1). This site is seldom

involved in AT ruptures and can consequently also be measured in a ruptured, repaired, or impaired tendon. The number
of slices between the distal and the proximal point was in turn
the amount of millimeters the AT traveled between the 2
points. The resulting length is thus the length of the AT.
The TL was measured from the lowest point of the
intercondylar eminence of the tibia (tibial spine) to the center of
the ankle formed by the tibia and the bula according to
Yoshioka et al14 (Figure 2). Measured qualitative parameters
were the degree of tendon morphology and tendon sheath uid.
The degree of tendon degeneration and tendon sheath uid
were scored according to Valderrabano et al8 on a 4-point grading scale (none = 0, mild = +, moderate = ++, severe = +++) on
axial T2-weigthed fat-saturated fast spin-echo images. Radiological parameters were measured using the DICOM Editor
(version 3.1.11; Santesoft, Athens, Greece).

Institutional Review Board and

Statistical Analysis
The local institutional review board approved this
study. Written consent was collected from all subjects. We

Physiological Measurements of Achilles Tendon Length

based our sample size on 2 considerations; rst, precision of

the length measurement and, second, strength of association
with TL. Based on estimates of ATL from the literature, an
alpha of 5% and a power of 90%, a minimum sample size of
43 subjects was required. Thus, after including 20% (n = 9) to
account for potential subject attrition or data loss, we included
52 (n = 43 + 9) healthy subjects in this study.
We used descriptive statistics including mean with
standard deviation and frequencies to describe the populations, as appropriate. A 1-way analysis of variance with
a Bonferroni post hoc analysis was used to compare age, body
height, weight, BMI, ATL, TL, percentage of ATL and TL,
tendon morphology, and tendon sheath uid.
Next, to obtain the association of ATL with different
physiological and sociodemographic parameters, we calculated Spearman correlations and performed multivariate linear
regression analysis. Analysis of residuals was used to conrm
the assumptions of linearity. With a stepwise selection
procedure (P , 0.2) we selected the most important predictors for ATL. Thereby, age, gender, body height, weight, side
of AT (left or right), and TL were used as candidate covariates. Based on this analysis, we propose an easy to use
calculation for estimation of ATL.
All reported condence intervals are 2-sided 95%
condence intervals and tests were performed at the 2-sided
5% signicance level. All analyses were performed with
STATA 9.2 (Stata Corp, College Station, Texas).

RESULTS

FIGURE 2. Tibia length was defined as the distance between

the lowest point of the intercondylar tuberosity and the part of
the tibia at the center of the ankle joint as defined by Yoshioka
et al.14
 2012 Lippincott Williams & Wilkins

The mean length of all ATs was 180.6 25.0 mm

(mean SD) (95% condence interval [CI], 173.7-187.6).
Women had a signicantly shorter ATL compared with
men (absolute difference, 22.9 mm, P , 0.01). The mean
TL was 371.9 25.4 mm and again signicantly shorter in
women compared with men (342.6 15.6 mm vs 379.8
21.5 mm, P , 0.0001). When looking at the qualitative
parameters in these asymptomatic subjects, tendon sheath
uid was graded none in 26 subjects, mild in 17 subjects,
moderate in 7 subjects, and severe in 2 subjects without
gender difference (P = 0.74). Tendon morphology was graded
none in 22 subjects, mild in 20 subjects, and moderate
in 10 subjects; none was graded severe without gender
difference (P = 0.84). As expected, ATL closely correlated
with body height (R2 = 37.9%, P , 0.0001) to a similar
extent it correlated with TL (R2 = 40.6%, P , 0.0001). The
length of the AT was 48.5 5.3% (95% CI, 47.0-50.0) of the
TL in this healthy cohort. In turn, TL correlated with body
height (R2 = 82.5%, P , 0.0001). The results are depicted in
Figure 3 and Table 1.
In multivariate analysis, using only sociodemographic
parameters (age, gender, height, weight) but not TL, we found
that height was the most important predictor for ATL with an
increase of 20.1 mm (95% CI, 12.3-29.8) per 10 cm increase
in height (R2 = 43.1%). Of note, gender, weight, and side of
AT (left or right) measurement were not signicant predictors
and did not further improve the model t in terms of R2.
When also considering TL as a candidate predictor, TL
was the only signicant predictor (P , 0.001) in the nal
www.cjsportmed.com |

485

Clin J Sport Med
Volume 22, Number 6, November 2012

Rosso et al

FIGURE 3. Achilles and TLs in millimeters. Median lines are

depicted; upper and lower ends of boxes represent 75th and
25th percentiles; whiskers represent the lowest and highest
values.

model. For each centimeter increase in TL the healthy ATL

increased by 0.6 cm (95% CI, 0.4-0.8). This model explained
a similar amount of ATL variability (R2 = 40.6%) as the
model above. Based on this model, the normal length of
a healthy AT can be estimated with the following calculation:
ATlengthinankleneutralpositionmm
0:6 Tibialengthmm 2 53:

DISCUSSION
The AT length is pivotal for a normal gait as it helps the
calf muscles to act at their optimum length,2 and restoring the
original length after AT rupture is therefore an important determinant of postoperative outcome.79 We thus investigated
a new way to measure the ATL in an MRI setup not involving
the myotendinous junction (MTJ) between the 2 gastrocnemii
into the measurements and used these ndings to establish
a formula based on TL that allows to accurately and precisely
calculate the ATL from a plain anterior-posterior x-ray.
Pang and Ying15 were the rst to measure ATL in
healthy patients by means of ultrasound. The length was
measured from the insertion of the AT on the calcaneus to
the MTJ between the tendon and the gastrocnemii muscles.
As can be seen in Figure 1A, the MTJ is not a straight line but
more M-shaped. It is therefore difcult to dene the exact
proximal insertion point. Consequently, the mean ATL was
117.7 mm with a high SD of 23.7 mm and a wide range of
58.5 to 166.2 mm emphasizing the difculty of length measurements in a sonographic setup. Compared with Pang and
Ying,15 our measurements were more consistent with a mean
length of 180.6 mm and an SD of 25.0 mm with a 95% CI of
173.7 to 187.6 mm. In our opinion, the actual values cannot
be compared due to the different ways of measuring the ATL
as we took the appearance of the gastrocnemius medialis
muscle belly and Pang and Ying took the MTJ as a proximal
point of origin. In AT ruptures, the central MTJ can be

486

| www.cjsportmed.com

involved and may consequently not be an optimal point of

measurement.
Another way to determine a more functional ATL was
published by Peltonen et al16 in which the MTJ and the calcaneal insertion were noted in sonography. The focus of their
study was to determine changes in the ATL during gait. No
value of the measured static ATL was stated.
Yoshioka et al14 had dened TL in 1989 and came up
with mean values of 353 23.7 mm in the overall cohort and
349 25.4 mm in women and 357 20.4 mm in men in
a cadaveric study of donors aged 61 to 89 years.14 These
ndings are close to our measurements of 371.9 25.4 mm
in the overall cohort and 342.6 15.6 mm in women and
379.7 21.5 mm in men. Burrows et al17 have measured TLs
in 278 subjects aged 15 to 20 years with a mean of 17.0 years
in girls and 16.6 years in boys and found lengths in the overall
cohort of 400 30 mm (range, 285-482 mm) and 385 26
mm in women and 415 27 mm in men. These measurements were higher than ours in a younger cohort with 12% of
active growth plates. Unfortunately, the points of measurement were not clearly dened in that study but it can be
assumed that the distal reference line was the tibial
plafond reference line without denition of the proximal
point. Nevertheless, our measurements are in between the
ones of Yoshioka et al14 and Burrows et al17 in a cohort aged
20 to 65 years.
In contrast to our study, Pang and Ying15 did not nd
a positive correlation between body height and tendon length
(dominant ankle, r = 0.26, P , 0.05; nondominant ankle,
r = 0.26, P , 0.05), whereas we found a highly signicant
correlation of r = 0.62 (P , 0.0001). To our knowledge, the
TL has never been correlated to body height or ATL and has
never been stated in the literature. The 2 parameters we measured by means of MRI can also be palpable in athletes. Tibia
length could be measured from a point between the medial
and lateral joint line representing an approximation of the
intercondylar eminence, the second point being the middle
of the ankle joint. Achilles tendon length could be approximated by measuring from the palpable beginning of the gastrocnemius medialis to the proximal insertion of the AT on
the calcaneus. These measurements are feasible in a clinical
setup and could give a good estimate of ATL also in side
comparison.
In our ndings, activity does not seem to be
inuencing resting ATL. The only comparable study was
performed by Hansen et al18 who showed unchanged tendon-aponeurosis displacement during maximal voluntary
contraction after approximately 9 months of habitual running. The calculated equation could serve as a useful reference to study pathological states of the AT such as
postoperative overlengthening after AT ruptures, Achilles
tendinopathy, diabetes-related states such as equinus contractures, or neuromuscular diseases.
Tendon morphology was graded moderate in 10
(19.2%) and tendon sheath uid was graded severe in 2
(3.8%) asymptomatic patients. Possibly, changes that are not
clinically relevant might be overestimated in MRI. This was
already described by Astrom et al19 and Shalabi et al20 showing alterations in MRI images in asymptomatic tendons.
 2012 Lippincott Williams & Wilkins

Clin J Sport Med
Volume 22, Number 6, November 2012

One strength of this study was to use MRI measurements with a high anatomic resolution and thin slices of 1 mm
to determine ATL. This novel measurement method can also
be used in AT status after rupture. The consistent measurements of our method to determine ATL and TL with small
SDs and tight 95% CI and the clear correlation between ATL
and body height and TL, respectively, were further strengths
of this study. We additionally came up with a way to estimate
the ATL by using the TL. Tibia length can easily be measured
on AP radiographs or on an MRI of the lower leg including
the knee and ankle and measuring the distance between the
lowest point of the intercondylar eminence of the tibia (tibial
spine) to the center of the ankle formed by the tibia and the
bula according to Yoshioka et al.14
A limitation of this study was that the ATL was dened
as the distance between the proximally dened point and the
distal insertion without taking the 3-dimensionality of the
tendon into account. The tendon could travel in an oblique
way through this eld of interest making the ATL slightly
longer. As can be seen in Figure 4 in a cadaveric setup, line
A0 is a line parallel to line A being the anatomical tibial axis.
Line c is a line parallel to the AT axis. Line b is perpendicular
to A0 . The distances are as follows: a is the ATL as measured
in our study, b is the distance between the lines A0 and c being
the distance from the calcaneal insertion to the medial gastrocnemius in the axis of the AT. Assuming that b (distance
from line A0 to line C) can measure maximally 10 mm, which
would be an anterior-posterior (a.p.) travel of 10 mm, being
a rather high a.p. travel. According to the Pythagorean theorem (a2 + b2 = c2), with our measured mean lengths of 180.6
(a), c would be maximally 180.9 mm. This would result in an
error of 0.3 mm or 0.15%.
In conclusion, we dened a new way to measure the
ATL in a consistent way showing good comparability
between healthy subjects. We found that TL was a strong
and independent predictor for ATL. Using the knowledge
of TL, ATL can be estimated using the following formula:
ATL in ankle neutral position (mm) = 0.6 TL (mm) 2 53.
Finally, the present study could help to evaluate ATL on
patient status after AT rupture and compare ATL in different
surgical techniques postoperatively.

FIGURE 4. Three-dimensionality of the AT in a cadaveric

setup. Line A0 is a line parallel to line A being the anatomical
tibial axis. Line c is a line parallel to the AT axis. Line b is
perpendicular to A0 . The distances are as follows: a is the ATL as
measured in our study, b is the distance between the lines A0
and c being the distance from the calcaneal insertion to
the medial gastrocnemius in the axis of the AT. For an exact
description, see Discussion.
 2012 Lippincott Williams & Wilkins

Physiological Measurements of Achilles Tendon Length

ACKNOWLEDGMENTS
We thank Tanja Haas for her MRI examinations and
her steady motivation. We thank Dr Marc Lottenbach and
Dr Andri Lutz of the Hospital of Fribourg, Switzerland, for
their contribution in patient acquisition and Dr Glen
Lichtwark of the University of Queensland, Australia, and
Dr Michael Ying of The Hong Kong Polytechnic University
for their valuable inputs. A special thanks to Dr Mathias
Miska for his help in ATL measurements.
REFERENCES
1. Leppilahti J, Puranen J, Orava S. Incidence of Achilles tendon rupture.
Acta Orthop Scand. 1996;67:277279.
2. Rosso C, Valderrabano V. Biomechanics of the Achilles tendon. In:
Calder JDF, Karlsson J, Maffulli N, et al., eds. Achilles Tendinopathy.
Current Concepts. Vol 1. 1st ed. Giuldford, United Kingdom: DJO Publications; 2010:2534.
3. Magnusson SP, Narici MV, Maganaris CN, et al. Human tendon behaviour and adaptation, in vivo. J Physiol. 2008;586:7181.
4. Fukunaga T, Kubo K, Kawakami Y, et al. In vivo behaviour of human
muscle tendon during walking. Proc Biol Sci. 2001;268:229233.
5. Roberts TJ. The integrated function of muscles and tendons during
locomotion. Comp Biochem Physiol A Mol Integr Physiol. 2002;133:
10871099.
6. Maganaris CN, Paul JP. Tensile properties of the in vivo human gastrocnemius tendon. J Biomech. 2002;35:16391646.
7. Rebeccato A, Santini S, Salmaso G, et al. Repair of the achilles tendon
rupture: a functional comparison of three surgical techniques. J Foot
Ankle Surg. 2001;40:188194.
8. Valderrabano V, Hintermann B, Wischer T, et al. Recovery of the posterior tibial muscle after late reconstruction following tendon rupture.
Foot Ankle Int. 2004;25:8595.
9. Leppilahti J, Lhde S, Forsman K, et al. Relationship between calf muscle size and strength after achilles rupture repair. Foot Ankle Int. 2000;21:
330335.
10. Mullaney MJ, McHugh MP, Tyler TF, et al. Weakness in end-range
plantar exion after Achilles tendon repair. Am J Sports Med. 2006;34:
11201125.
11. Bressel E, McNair PJ. Biomechanical behavior of the plantar exor
muscle-tendon unit after an Achilles tendon rupture. Am J Sports Med.
2001;29:321326.
12. Kangas J, Pajala A, Ohtonen P, et al. Achilles tendon elongation after
rupture repair: a randomized comparison of 2 postoperative regimens. Am
J Sports Med. 2007;35:5964.
13. Valderrabano V, Pagenstert G, Horisberger M, et al. Sports and recreation activity of ankle arthritis patients before and after total ankle
replacement. Am J Sports Med. 2006;34:993999.
14. Yoshioka Y, Siu DW, Scudamore RA, et al. Tibial anatomy and functional axes. J Orthop Res. 1989;7:132137.
15. Pang BS, Ying M. Sonographic measurement of achilles tendons in
asymptomatic subjects: variation with age, body height, and dominance
of ankle. J Ultrasound Med. 2006;25:12911296.
16. Peltonen J, Cronin NJ, Avela J, et al. In vivo mechanical response of
human Achilles tendon to a single bout of hopping exercise. J Exp Biol.
2010;213:12591265.
17. Burrows M, Liu D, McKay H. High-resolution peripheral QCT imagining of bone micro-structure in adolescents. Osteoporos Int. 2010;21(3):
515520.
18. Hansen P, Aagaard P, Kjaer M, et al. Effect of habitual running on
human Achilles tendon load-deformation properties and cross-sectional
area. J Appl Physiol. 2003;95:23752380.
19. Astrom M, Gentz CF, Nilsson P, et al. Imaging in chronic achilles
tendinopathy: a comparison of ultrasonography, magnetic resonance
imaging and surgical ndings in 27 histologically veried cases. Skeletal
Radiol. 1996;25:615620.
20. Shalabi A, Kristoffersen-Wilberg M, Svensson L, et al. Eccentric training
of the gastrocnemius-soleus complex in chronic Achilles tendinopathy
results in decreased tendon volume and intratendinous signal as evaluated
by MRI. Am J Sports Med. 2004;32:12861296.

www.cjsportmed.com |

487