Clinical Features of Patellar Tendinopathy and Their

Implications for Rehabilitation

Kay M. Crossley,1 Kaviraj Thancanamootoo,1 Ben R. Metcalf,1 Jill L. Cook,2 Craig R. Purdam,3 Stuart J. Warden1,4
1
Centre for Health, Exercise and Sports Medicine, School of Physiotherapy, The University of Melbourne, Victoria, Australia
2
Musculoskeletal Research Centre, School of Physiotherapy, La Trobe University, Victoria, Australia
3
Department of Physical Therapies, Australian Institute of Sport, Belconnen, Australian Capital Territory, Australia
4
Department of Physical Therapy, School of Health and Rehabilitation Sciences, Indiana University, Indianapolis, Indiana

Received 16 November 2006; accepted 12 March 2007

Published online 27 April 2007 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jor.20415

ABSTRACT: This study investigated the clinical features of patellar tendinopathy (PT), with focus
on individuals with unilateral and bilateral PT. A cross-sectional study design was employed to
compare individuals with unilateral (n ¼ 14) or bilateral (n ¼ 13) PT and those without PT (control,
n ¼ 31). Features assessed included thigh strength (normalized peak knee extensor torque) and
flexibility (sit-and-reach and active knee extension), calf endurance (heel-rise test), ankle flexibility
(dorsiflexion), alignment measures (arch height and leg length difference), and functional measures
(hop for distance and 6 m hop test). Groups were matched for age and height; however, unilateral and
bilateral PT had greater mass with a higher body mass index (BMI) than control. Also, bilateral PT
performed more sport hours per week than both unilateral PT and control. Unilateral PT had less
thigh strength than control and bilateral PT, whereas bilateral PT had more thigh flexibility than
control and unilateral PT. Both unilateral and bilateral PT had altered alignment measures
compared to control. Features that predicted symptoms in PT were lower thigh flexibility and
strength, whereas those that predicted function were higher thigh strength and lower ankle
flexibility. These findings indicate that unilateral and bilateral PT represent distinct entities, and
that thigh strength appears particularly important in PT as it predicted both symptoms and
function in PT. ß 2007 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop
Res 25:1164–1175, 2007
Keywords: anterior knee pain; jumper’s knee; sports medicine; tendinitis; tendinosis

INTRODUCTION pain, focal patellar tendon tenderness, and intra-

tendinous imaging changes.2 It is a common
Overuse conditions of the patellar tendon have condition encountered in clinical practice, with a
historically been labeled as jumper’s knee or prevalence of 14% in elite athletes depending on
patellar tendinitis. However, these terms are sport.3 However, the true clinical significance of
misnomers as the condition is found in many who PT lies in the morbidity that it causes. Disability
do not participate in jumping sports,1–3 and associated with PT is often prolonged and in many
histopathological studies have consistently shown instances causes the premature cessation of an
the underlying pathology to be degenerative athletic career.6
(tendinosis) rather than inflammatory (tendini- Despite being a significant clinical problem,
tis).2,4 As tendinosis refers to a distinct histo- little is known about the pathogenesis of PT. This
pathological finding that can only be assessed has restricted treatment options, with manage-
following invasive tissue biopsy, the term patellar ment presently being more of an art than a science.7
tendinopathy (PT; tendino- ¼ tendon; -pathy ¼ Conceptually, as with most overuse conditions, the
disease) has been advocated clinically to describe development of PT results from the interplay of two
overuse conditions of the patellar tendon.5 PT is groups of factors—extrinsic and intrinsic factors.8
characterized by activity-related anterior knee Extrinsic factors refer to factors in the environment
or external to the individual that influence the
Correspondence to: Stuart J. Warden (Telephone: 1-317-278-
8401; Fax: 1-317-278-1876; E-mail: [email protected])
likelihood of sustaining PT. These factors are
ß 2007 Orthopaedic Research Society. Published by Wiley Periodicals,
most commonly indicted in the pathogenesis of
Inc. PT, with the single most frequently reported

1164 JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007

 CLINICAL FEATURES OF PATELLAR TENDINOPATHY 1165

causative factor being mechanical overload. How- current knee pain, or knee pain or injury in the past
ever, the presence of PT is not consistent 2 years that interfered with training or caused them
across individuals exposed to equivalent loading to seek professional advice. To be classified as having
levels, suggesting that intrinsic factors also PT, participants met the additional eligibility criteria
of: knee pain on at least one of jumping/landing, running
contribute.
or changing directions; pain on palpation of the
Intrinsic factors refer to processes internal to
patellar tendon;15 score of less than 80 points on the
the individual that influence their response to Victorian Institute of Sport Assessment (VISA) scale;16
external mechanical loading. Intrinsic factors can symptoms sufficient to affect exercise and activity for at
range from those that can only be derived using least 6 months; and have a confirmed hypoechoic
specialized biomechanics equipment (such as lesion within the patellar tendon on ultrasonography.17
three-dimensional joint kinematics and kinetics) The study was approved by the Human Research Ethics
to factors that can easily be assessed in standard Committee at the University of Melbourne and all
clinical practice. The later are the focus of the participants provided written informed consent.
current study given their more widespread applic-
ability, and include features such as anthropo- Outcome Measures
metric characteristics, thigh muscle strength and All measures were obtained by the same two examiners
flexibility, calf strength, ankle flexibility, foot arch (KT and BRM). Test–retest reliability was assessed
height, lower limb alignment, and functional prior to the study in a subset of 10 healthy individuals
capacity. While preliminary studies have demon- assessed twice (7–10 days apart) by one investigator
strated potential roles for some of these factors in (BRM). Standard demographic [sex, age, height,
the development of PT,9–14 the data are far from mass, body mass index (BMI), leg dominance, sport
hours per week] data were obtained at inclusion. Leg
conclusive. In addition, clinical practice and pre-
dominance was defined as the leg the participant
liminary research9,10,13 suggests that individuals
predominantly jumps with. After familiarization, out-
who develop unilateral PT may differ from those come measures were performed three times per side
with the condition bilaterally (bilateral PT), and (with the exception of the calf endurance test which was
that males and females may also differ. performed once per side) and the mean value recorded
In order to confirm the preliminary research and (with the exception of normalized peak knee extensor
clinical suspicions, and facilitate a more scientifi- torque and hop for distance for which maximal values
cally based approach to the rehabilitation of were recorded).
individuals with PT, this study aimed to: (1)
identify clinical features of individuals with uni- Thigh Strength and Flexibility
lateral and bilateral PT which are potentially
Normalized Peak Knee Extensor Torque
modifiable through rehabilitation; and (2) investi-
gate the influence of sex on these clinical features. Thigh strength was measured isometrically using a Kin
Further study aims were to: (a) establish the Com dynamometer (Kin Com 5.30; Chattecx, Chatta-
repeatability of the chosen clinical measures; (b) nooga Inc., TN). The setup was individualized for each
determine the symmetry in clinical features participant, including aligning the lateral femoral
epicondyle with the lever arm axis of rotation. Partici-
between limbs of PT participants; and (c) identify
pants were stabilized and performed two submaximal
whether scores on the clinical measures
and one maximal warm-up test at 608 of knee flexion,
predict symptoms and function in participants followed by three maximal isometric tests (5 s duration
with PT. with 20 s rest between). Peak torque (Nm) was recorded
and normalized for mass (Nm.kg1).
METHODS
Study Design and Subjects Sit-and-Reach Test
A cross-sectional study design was used to compare Flexibility of posterior structures (knee, hip, and lower
clinical features in individuals with unilateral (n ¼ 14) back) was assessed by having the participant sit with
or bilateral (n ¼ 13) PT and those without PT (control their knees extended and feet flat against the vertical
group, n ¼ 31). General inclusion criteria included: aged wall of a sit-and-reach box. They stretched forward as

18 years; ability to speak English; and participation in far as possible and held the position for 1 s, while
competitive basketball, netball, volleyball, soccer, or maintaining one hand on top of the other. A measuring
tennis at least once per week. General exclusion criteria tape indicated where the tip of the middle finger
included: previous patellar tendon surgery; injections in reached (to the nearest 0.5 cm) relative to zero (aligned
the knee in the past 6 months; clinical evidence of other with the vertical wall of the sit-and-reach box; i.e., foot
knee pathology; and inability to walk without a limp. position). A positive measurement indicated that
Further to this, controls were excluded if they had the fingers extended beyond the toes, while a negative

DOI 10.1002/jor JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007

1166 CROSSLEY ET AL.

measurement indicated that the individual could not get

their fingers to their toes.

Active Knee Extension Flexibility

Hamstring length was assessed by having the partici-
pant maximally extended their knee while lying supine
with their hip at 908 flexion (Fig. 1). A gravity
inclinometer (Acuangle, Isomed, Portland, OR) on the
anterior tibial border recorded tibial inclination relative
to vertical. A negative value indicated inability to reach
the vertical and, thus, lower posterior thigh flexibility.
An individual who could reach full knee extension
without restriction was instructed to maintain knee
extension and slowly flex their hip as far as possible
(resulting in a positive angle of the tibia with respect to
the vertical).

Calf Endurance
Heel-Rise Test
Calf-muscle performance was assessed using a single-
leg heel-rise test. Heel-rise height was established by
having the participant maximally plantarflex their
ankle and a horizontal string was placed at the height
of the dorsal ankle, just inferior to the distal aspect of
the anterior tibia (Fig. 2). During the test, participants
maximally raised their heel off the floor, touching the

Figure 2. Assessment of calf endurance.

string before returning to the floor at a rate of 1 heel rise

every 2 s (paced by a metronome). Their knee was
maintained in extension and no more than 2% of whole
body mass could be placed through their contralateral
arm on the wall (measured with a Nicholas Manual
Muscle Tester). Five practice heel rises were performed
followed by a 1 min rest prior to the actual test. The
participant was asked to perform as many heel rises as
possible. Repetitions were counted each time the foot
contacted the string. The test was terminated if the
participant leaned forward with a force greater than 2%
body weight, the participant’s knee flexed, or the
participant failed to contact the string during three
consecutive heel rises.

Ankle Flexibility
Ankle Dorsiflexion Range
A line was placed on the floor at right angle to the wall
(floor line) and continued up the wall at right angle to
the floor (wall line). The foot was positioned such that
the bisection of the heel and second toe were aligned
on the floor line (Fig. 3A). The distance of the foot to the
wall was initially self-selected by the participant. The
Figure 1. Assessment of active knee extension. participant lunged forward to touch the center of their

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007 DOI 10.1002/jor

 CLINICAL FEATURES OF PATELLAR TENDINOPATHY 1167

Alignment Measures
Arch Height
Arch height was measured as the distance from the
inferior aspect of the navicular tubercle to the floor. This
distance was measured during maximal weight bearing
using a digital vernier caliper (Fig. 3B). Individuals
stood on a step, with one foot on the lower step and
the other on the higher step. Both feet were aligned in
the sagittal plane such that the center of the heel
was aligned with the second toe. Participants were
instructed to shift their weight such that the majority
was placed over the lower foot. Arch height of the lower
foot was measured.

Leg Length Differences

Leg length was measured in supine as: (a) anterior–
superior iliac spine to the tip of the medial malleolus
(medial measure); and (b) lateral prominence of the
greater trochanter to the lateral malleolus (lateral
measure). The length of the asymptomatic (or least
symptomatic) leg was subtracted from the length of the
symptomatic (or most symptomatic) leg. For the control
group, differences were obtained by subtracting the
length of the nondominant leg from the dominant leg.

Functional Measures
Hop for Distance
The participant hopped on one leg as far as possible from
a standing start after a submaximal warm-up. The test
was repeated three times for each leg and the best score
was identified and recorded.

Six-Meter Hop Test

The participant hopped on one leg as fast as possible
over a distance of 6 m using large forceful one-legged
hopping movements. The time taken to complete this
task was recorded.

Pain and Disability

Pain and disability measures assessed symptom inten-
sity and quantified functional capacity. Symptom inten-
sity was assessed using an 11-point numerical pain scale
Figure 3. Assessment of (A) ankle dorsiflexion range and (NPS; with 0 ¼ no pain and 10 ¼ worst possible pain).
(B) arch height. The usual (NPS-U) and worst (NPS-W) pain on the
subject-nominated most painful activity were recorded.
These scales are valid, reliable, and responsive in
patella against the wall line while keeping the heel the assessment of anterior knee pain.18 Functional
on the ground and knee in line with the second toe. If capacity was quantified using the VISA score, a
participants could or could not touch their patella on the validated measure of knee function in athletes with
wall line, their foot was progressively moved further PT.16 The scale ranges from 0 to 100, with 100 indicating
from or closer to the wall, respectively. At the maximum full, pain-free function and a score below 80 being
lunge point, a gravity inclinometer was placed on the associated with symptomatic PT.1 The VISA scale has
anterior border of the tibia (15 cm below the tibial excellent short-term test–retest and intertester relia-
tuberosity) and the angle relative to the vertical bility (both, r > 0.95), as well as good short-term (1 week)
recorded. stability (r ¼ 0.87).16

DOI 10.1002/jor JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007

1168 CROSSLEY ET AL.

Data Analyses and Statistical Evaluation To determine if the clinical measures were predictors
of pain (NPS-W and NPS-U) and function (VISA) in the
All data were analyzed using SPSS version 14.0
PT group, stepwise multiple linear regressions were used
(Norusis/SPSS Inc., Chicago, IL). Group differences
(criteria for entry, p < 0.05, and removal, p < 0.10). To
(control vs. unilateral PT vs. bilateral PT) in ratio
reduce the number of potential predictors only one
demographic data (age, height, mass, BMI, and sport
measure of posterior thigh flexibility (active knee exten-
hours per week) were assessed using one-way factorial
sion), one measure of leg length difference (leg length
analyses of variance (ANOVA), with Fisher’s protected
difference measured medially), and one functional test
least-significant difference (PLSD) used for post hoc
(hop for distance) were included. To further lessen the
analyses. Mean differences and 95% confidence inter-
number of predictor variables, bivariate correlations
vals (95% CI) were obtained for each post hoc compar-
were performed between the measures, and symptom
ison. Group differences (control vs. unilateral PT vs.
and function scores. Only clinical measures that were
bilateral PT) in nominal demographic data (dominant
correlated with symptoms or function at a significance
leg and sex) were assessed using chi-squared (w2)
level of p < 0.20 were accepted for inclusion in the linear
analyses. Unpaired t-tests were used to compare
regression.
symptom duration, VISA, NPS-U, and NPS-W between
unilateral and bilateral PT. Intraclass correlation
coefficients (ICC) were generated for the repeatability RESULTS
data and standard error of measurement (SEM) calcu- Demographic Data
lated for each measure. Paired t-tests were used to
establish symmetry in clinical features between the Demographic data are shown in Table 1. Age,
symptomatic (or most symptomatic) and asymptomatic height, number of female-to-male participants, or
(or least symptomatic) limbs within unilateral and side of leg dominance did not differ between
bilateral PT. Group differences in clinical features were groups (all p ¼ 0.16–0.98). However, there were
assessed using two-way factorial analyses of variance significant group differences for mass, BMI, and
(ANOVA), with condition (control vs. unilateral PT vs. sport hours per week (all p < 0.02). Both unilateral
bilateral PT) and sex (male vs. female) as the indepen- and bilateral PT had greater mass than control
dent variables. Mass was used as a covariate for the
(unilateral PT vs. control: mean differences, 9.6 kg,
heel-rise test, arch height, hop for distance, and 6 m hop
95% CI, 1.1–18.2 kg; bilateral PT vs. control: mean
test analyses. In the event of a nonsignificant ANOVA
interaction, the main effect for each variable was differences, 11.2 kg, 95% CI, 2.5–20.0 kg; all
explored, with Fisher’s protected least-significant dif- p < 0.03). As height did not differ between groups,
ference (PLSD) used for post hoc analyses of significant the greater mass in unilateral and bilateral PT
condition main effects. Mean differences and 95% CI resulted in these groups having higher BMIs
were obtained for each post hoc comparison. than control (unilateral PT vs. control: mean

Table 1. Demographic Data for Each Condition

Control Unilateral PT Bilateral PT

Variable (n ¼ 31) (n ¼ 14) (n ¼ 13)
Age (years) 24  6 26  7 28  8
Height (m) 1.77  0.88 1.78  0.9 1.76  0.9
Mass (kg)* 71  11 80  16{ 82  14{
BMI (kg.m2)* 22.5  2.4 25.2  4.0{ 26.2  3.5{
Sport (hrs.wk1)* 3.3  1.8 3.6  2.1 6.7  9.3{
Dominant leg (R:L) 27:4 12:2 11:2
Sex (F:M) 11:20 4:10 4:9
Symptom duration (years) — 2.8  3.0 5.0  3.2
VISA — 60  15 58  20
NPS-U — 5.6  2.0 5.5  3.0
NPS-W — 6.9  2.5 7.0  2.5

Abbreviations: PT, patellar tendinopathy; BMI, body mass index; VISA, Victorian Institute of
Sport Assessment (100 represents full function); NPS-U, numerical pain scale (10 represents
maximal pain) for the usual pain in the preceding week; NPS-W, numerical pain scale (10
represents maximal pain) for the worst pain in the preceding week.
Values are mean  SD (except for dominant leg and sex, which are represented by number of
cases).
*Indicates significant differences between conditions (p < 0.02, one-way factorial ANOVA).
{
Indicates significant difference to control (p < 0.03, Fisher’s PLSD).
{
Indicates significant difference to both control and unilateral PT (p < 0.02, Fisher’s PLSD).

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007 DOI 10.1002/jor

 CLINICAL FEATURES OF PATELLAR TENDINOPATHY 1169

differences, 2.7 kg.m2, 95% CI, 0.6–4.7 kg.m2; (Fig. 4). The most symptomatic and least sympto-
bilateral PT vs. control: mean differences, matic legs in the bilateral PT subgroup recorded
3.7 kg.m2, 95% CI, 1.6–5.7 kg.m2; all p  0.01). similar scores. Interestingly, in unilateral PT the
There were no differences between unilateral and symptomatic and asymptomatic legs also scored
bilateral PT for mass (p ¼ 0.75) or BMI (p ¼ 0.40); similarly.
however, the bilateral PT group reported more
sport hours per week than both unilateral Clinical Features of Unilateral and Bilateral
PT (mean differences, 3.2 h.wk1, 95% CI, 0.6– Patellar Tendinopathy, and the Influence of Sex
5.8 h.wk1; p ¼ 0.02) and control (mean differ- There were no interactions between condition
ences, 3.4 h.wk1, 95% CI, 1.2–5.7 h.wk1; (control vs. unilateral PT vs. bilateral PT) and
p < 0.01). Sport hours per week did not differ sex (male vs. female) for any clinical measure (all
between unilateral PT and control (p ¼ 0.81), and p > 0.05). However, there were significant sex
there no differences between unilateral and bilat- (male vs. female) effects for a number of variables,
eral PT for symptom duration, VISA, NPS-U, or regardless of condition. Females were more flex-
NPS-W (p ¼ 0.08–0.97). ible than males in the sit-and-reach test (mean
difference, 7.5 cm; 95% CI, 1.5–13.4 cm), and a
similar trend towards females was noted for the
Repeatability of Clinical Measures active knee extension test (mean difference, 7.28;
Clinical measures utilized in this study were 95% CI, 0.28–14.68). In contrast, females per-
repeatable as indicated by ICC values greater formed significantly worse on the hop for distance
than 0.80, with the exception of the heel-rise test test, producing a shorter hop than males (mean
(Table 2). The heel-rise test demonstrated moder- difference, 0.32 m; 95% CI, 0.52–0.14 m).
ate repeatability. However, this test appears to Similarly, females took longer to complete the 6 m
have a strong learning effect, and because the hop test than males (mean difference, 0.3 s; 95%
main study was cross-sectional in nature it was CI, 0.1–0.5 s).
still included in the testing battery. There were significant main effects for condition
(control vs. unilateral PT vs. bilateral PT) for a
number of clinical measures, irrespective of sex
Symmetry of Clinical Measures in Unilateral and (Table 3). Specifically, normalized peak knee
Bilateral Patellar Tendinopathy
extensor torque, active knee extension, arch
There were no between leg differences within height, and leg length difference measured
unilateral or bilateral PT for any clinical measure both medially and laterally all differed between

Table 2. Test–Retest Reliability of Clinical Tests

Clinical Measure Time 1* Time 2* ICC SEM

Thigh strength and flexibility
Normalized peak knee extensor 2.07  0.59 2.26  0.06 0.85 0.16
torque (Nm.kg1)
Sit-and-reach (cm) 2.5  7.9 3.4  7.2 0.93 2.00
Active knee extension (8) 30  15 30  13 0.84 6
Calf endurance
Heel-rise test (n) 31  13 44  9 0.57 8
Ankle flexibility
Dorsiflexion (8) 46  6 46  6 0.98 1
Alignment measures
Arch height during maximal weight 40  6 40  7 0.95 1.5
bearing (mm)
Leg length (measured medially) (m) 0.85  0.06 0.85  0.06 1.00 0.3
Leg length (measured laterally) (m) 0.86  0.06 0.85  0.07 0.99 0.6
Functional Measures
Hop for distance (m) 1.31  0.39 1.3  0.32 0.94 0.92
6 m hop test (s) 2.48  0.49 2.57  0.60 0.85 0.22
Abbreviations: ICC, Intraclass correlation co-efficient; SEM, standard error of measurement.
*Values are mean  SD.

DOI 10.1002/jor JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007

1170 CROSSLEY ET AL.

Figure 4. Symmetry in clinical measures between the symptomatic (or most symptomatic) and
asymptomatic (or least symptomatic) limbs within the unilateral PT and bilateral PT subgroups.
All p > 0.05 (paired t-test).

conditions (all p < 0.05). In addition, both of the range, leg length difference, and hop for distance
functional measures had trends towards a main (Table 4). These measures were subsequently
effect for condition (hop for distance, p ¼ 0.08; 6 m included in the stepwise multiple linear regres-
hop test, p ¼ 0.07). Post hoc analyses of condition sion. Those that remained in the significant final
main effects (Table 3) demonstrated unilateral PT models as predictors of increasing both NPS-U
to have less normalized peak knee extensor (p < 0.001) and NPS-W (p < 0.001) were lower
strength than both control and bilateral PT (all active knee extension flexibility and lower normal-
p < 0.05). In contrast, it was bilateral PT who were ized peak knee extensor torque (Table 5). These
more flexible than both control and unilateral PT models explained 45% and 48% of the variance in
during the active knee extension test (all p < 0.05). NPS-U and NPS-W, respectively.
Both unilateral and bilateral PT exhibited lower For the measure of function (VISA scale),
arches and greater leg length differences than the clinical measures of normalized peak knee
controls, with the symptomatic (or most sympto- extension torque, active knee extension range,
matic) leg being on average longer (all p < 0.05). ankle dorsiflexion range, and hop for distance were
Although the main effect for neither functional entered into the stepwise linear regression as
measure quite reached significance, unilateral PT they correlated with VISA at a significance of
exhibited lower performance on the two hopping p < 0.20 (Table 4). The final prediction model for
tests than both bilateral PT and control (all increasing function included higher normalized
p < 0.05). peak knee extension torque and lower ankle
dorsiflexion range, which predicted 36% of the
variance in the VISA score (Table 5).
Clinical Predictors of Symptoms and Function in
Patellar Tendinopathy
Bivariate correlations between the clinical mea- DISCUSSION
sures, and the two symptom measures (NPS-W
and NPS-U) and the functional measure (VISA) in This study investigated the clinical features of PT
PT are shown in Table 4. The clinical measures using a battery of clinically reliable and applicable
which correlated at a significance of p < 0.20 with measures. Predictable sex differences were found,
both NPS-U and NPS-W were normalized peak with males exhibiting greater functional ability
knee extension torque, active knee extension and females greater flexibility. However, there

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007 DOI 10.1002/jor

 Table 3. Effect of Condition on Clinical Measures

DOI 10.1002/jor
Condition (Mean  SD)* Significant Post Hoc Pairwise Comparisons{

Unilateral
Clinical Measure Control Unilateral PT Bilateral PT Control–Unilateral PT Control–Bilateral PT PT–Bilateral PT
Thigh strength & flexibility
Normalized peak 2.23  0.49 1.72  0.61 2.29  0.54 0.49 (0.13–0.86) 0.56 (1.00–0.13)
knee extensor
torque (Nm.kg1){
Sit-and-reach test (cm) 1.8  10.1 2.2  10.7 9.6  10.4 7.8 (14.2–1.4) 9.9 (17.4–2.5)
Active knee extension 24.6  12.9 23.4  11.0 8.1  10.9 15.9 (23.8–8.0) 14.8 (24.0–5.6)
(8)>{
Calf endurance
Heel-rise test (n)§ 25  12 26  11 26  11
Ankle flexibility
Dorsiflexion (8) 48.1  5.7 47.9  6.1 49.2  7.9
Alignment measures
Arch height during 50.4  5.9 42.1  8.5 43.1  7.0 8.4 (3.6–13.3) 7.3 (2.4–12.3)
maximal weight
bearing (mm){,§
Leg length difference 4.5  8.5 14.8  8.9 11.1  8.8 9.2 (14.4–4.0) 0.67 (1.21–0.13)
(measured
mediatly) (mm){
Leg length difference 8.1  14.2 16.9  14.8 23.6  14.6 8.7 (17.5–0.1) 1.7 (2.6–0.8)
(measured
laterally) (mm){
Functional measures
Hop for distance (m)§ 1.60  0.31 1.37  0.30 1.54  0.29 0.22 (0.03–0.42)
6 m hop test (s)§ 2.1  0.4 2.4  0.4 2.1  0.5 0.3 (0.5–0.0) 0.3 (0.0–0.6)

Abbreviation: PT, patellar tendinopathy.

Sit-and-reach, higher value is more flexible; active knee extension, higher value is more flexible; leg length difference, symptomatic (or most symptomatic) leg minus asymptomatic
(or least symptomatic leg) leg (in UPT and BPT) or dominant minus non dominant leg (in control).
*Values indicate mean  SD.
{
Values indicate mean difference (95% confidence intervals) for significant Fisher’s PLSD post hoc tests (p < 0.05).
{
Indicates significant main effect for condition (p < 0.05, two-way factorial ANOVA).
§
Indicates mass included as a covariate (mass corrected means are presented).
CLINICAL FEATURES OF PATELLAR TENDINOPATHY

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007

1171
1172 CROSSLEY ET AL.

Table 4. Bivariate Correlations between Clinical Measures, and Measures of

Symptoms and Function within the PT Group as a Whole (Unilateral and Bilateral
PT Combined)

Symptoms Function

Variable NPS-U NPS-W VISA

Demographic
Age 0.02 0.05 0.24
Gender 0.15 0.25 0.02
Height 0.07 0.07 0.22
Weight 0.10 0.08 0.09
BMI 0.18 0.15 0.05
Sport hours per week 0.04 0.11 0.07
Thigh strength & flexibility
Normalized peak knee extensor torque 0.63* 0.63* 0.40*
Active knee extension 0.27{ 0.31{ 0.27{
Calf endurance
Heel-rise test 0.12 0.10 0.21
Ankle flexibility
Dorsiflexion 0.12 0.06 0.35{
Alignment measures
Arch height during maximal weight bearing 0.03 0.04 0.19
Leg length difference (measured medially) 0.27{ 0.27{ 0.20
Functional measures
Hop for distance 0.64* 0.65* 0.47*
Abbreviations: PT, patellar tendinopathy; NPS-U, numerical pain scale for the usual pain in
the preceding week; NPS-W, numerical pain scale for the worst pain in the preceding week; VISA,
Victorian Institute of Sport Assessment; BMI, body mass index.
Values are correlation coefficients (r).
*Indicates significant bivariate correlation for inclusion into the stepwise linear regression
( p < 0.05).
{
Indicates significant bivariate correlation for inclusion into the stepwise linear regression
(0.05 < p < 0.20).

were no interactions between sex and condition, having a higher BMI. Unilateral PT and control
indicating that the clinical features of PT were not did not differ in sports hours per week; however,
influenced by sex. PT participants had greater both groups performed less hours than bilateral
mass than control, which contributed to the former PT. There were no asymmetries between legs in

Table 5. Regression Coefficients for Variables in the Final Model of the Linear Regression for Symptoms and
Function within the PT Group as a Whole (Unilateral and Bilateral PT Combined)

Dependent
Variable Predictor Variables B SE B b p 95% CI for B
Symptoms
NPS-U Normalized peak knee extensor 2.49 0.56 0.65 <0.001 3.64 to 1.34
torque (Nm.kg1)
Active knee extension (8) 0.06 0.03 0.31 <0.05 0.12 to 0.00
NPS-W Normalized peak knee extensor 2.33 0.51 0.65 <0.001 3.38 to 1.27
torque (Nm.kg1)
Active knee extension (8) 0.06 0.03 0.35 <0.05 0.12 to 0.01
Function
VISA Normalized peak knee extensor 15.14 4.47 0.60 <0.01 5.91 to 24.37
torque (Nm.kg1)
Ankle dorsiflexion (8) 1.34 0.43 0.52 <0.01 2.23 to 0.46

Abbreviations: PT, patellar tendinopathy; NPS-U, numerical pain scale for the usual pain in the preceding week; NPS-W,
numerical pain scale for the worst pain in the preceding week; VISA, Victorian Institute of Sport Assessment.

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007 DOI 10.1002/jor

 CLINICAL FEATURES OF PATELLAR TENDINOPATHY 1173

individuals with either unilateral or bilateral PT; established inhibitory influences on muscle,21 and
however, there were group differences between tendon pain during strength testing in unilateral
conditions. Unilateral PT had reduced thigh PT may have compromised the test. In addition, it
strength compared to both bilateral PT and is possible that muscle atrophy associated with
control. In contrast, bilateral PT had increased the long duration of symptoms in unilateral PT
thigh flexibility compared to both unilateral PT may have contributed to the differences. However,
and control, while both unilateral and bilateral PT bilateral PT did not differ from control in thigh
had lower arch height and increased leg length strength despite having equivalent pain and longer
difference compared to control. Clinical features symptom duration than unilateral PT, indicating
that predicted symptoms in PT participants were that pain and disuse atrophy may not completely
lower thigh flexibility and strength, whereas those explain our findings. Overall, the clinical ramifica-
that predicted function were higher thigh strength tion of the positive relationship between thigh
and lower ankle flexibility. muscle strength, and symptoms and function in PT
The results of the current study indicate that is that both tendon pain and muscle function need
individuals with unilateral and bilateral PT repre- to be addressed in the rehabilitation of individuals
sent distinct clinical entities. Unilateral and bilat- with PT.
eral PT differed from one another and controls in Muscular tightness is considered an important
thigh strength and flexibility. These observations factor in the development of lower limb overuse
support previous findings.9,10,13 For instance, Cook injuries, including PT.22 Accordingly, thigh muscle
and colleagues9,10,13 showed females with unilat- flexibility was both a discriminative clinical featu-
eral PT to be less flexible than both controls and re of PT and predictor of symptoms. Bilateral
bilateral PT, while males with bilateral PT had PT had increased posterior thigh flexibility than
greater hip girth and a larger waist-to-hip ratio both unilateral PT and control, suggesting that
than both unilateral PT and control. However, increased flexibility may contribute to the develop-
these previous studies are limited as PT symptoms ment of bilateral PT. However, posterior thigh
in participants were not assessed, with group flexibility in PT was negatively correlated with
allocations being based solely on ultrasonography pain, indicating that increased flexibility was
findings. The presence of a hypoechoic lesion on associated with reduced pain. This finding is in
ultrasonography is not predictive of current PT accordance with Witvrouw and colleagues,14 who
symptoms,17 and does not appear to predict the found thigh flexibility to be a prospective determin-
likelihood that an individual will become sympto- ing factor in PT development. Similarly, Cook and
matic in the future.19,20 Consequently, the current colleagues9 demonstrated athletes with abnormal
study provides the strongest evidence to date that patellar tendon morphology to have reduced thigh
unilateral and bilateral PT represent different flexibility. Grouped, these findings suggest that
subgroups of PT, confirming clinical suspicions interventions aimed at improving thigh flexibility
and highlighting the need for these subgroups to be may facilitate reductions in PT symptoms and be
considered separately in both clinical and research an important component of PT rehabilitation. The
settings. mechanisms for this contribution are currently
Thigh strength measures differed across condi- hypothetical, but may include changes in knee joint
tions, and represented a predictor of symptoms and mechanics. For instance, posterior thigh tightness
function in PT. Unilateral PT had less thigh results in an increase in knee flexion during
strength compared to bilateral PT and control, stance,23 a predictive variable of PT.24
consistent with a previous report,10 and the Arch height differed between groups, with both
reduced thigh strength in unilateral PT potentially unilateral and bilateral PT having lower arch
explains their reduced performance on functional height during weight bearing compared to control.
testing. Supporting this, higher thigh strength was Foot morphology has long been considered an
predictive of both reduced pain and improved important contributor to lower limb mechanics
function in PT. These findings support the belief due to the position of the feet at the ground
that strength training is important in the rehabi- interface, and has been implicated in the develop-
litation of PT. Muscle strength also may be an ment of knee overuse injuries.22 This suggests that
important contributor to PT, but the current arch height should be addressed in the rehabilita-
results cannot confirm this due to its cross- tion of individuals with PT, possibly through the
sectional design. While unilateral PT had reduced modification of extrinsic (shoes or orthoses) or other
thigh strength that compromised function, it is not intrinsic (function of the extrinsic and intrinsic foot
possible to establish cause or effect. Pain has well- muscles) factors. This is simplistic for a number of

DOI 10.1002/jor JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007

1174 CROSSLEY ET AL.

reasons. Arch height neither correlated to or forces to the patellar tendon–quadriceps com-
predicted symptoms or function in PT. Thus, arch plex.29 Supporting this, previous studies have
height modification may not necessarily produce found ankle dorsiflexion range to differentiate
desired changes in pain or function. Also, morphol- between those with and without abnormal patellar
ogy is not the sole feature of the foot that is tendon morphology,12 and that individuals with PT
considered responsible for knee overuse conditions. have altered ankle joint mechanics.30 While ankle
Foot morphology is typically assessed statically, as dorsiflexion range was not found to be a discrimi-
in the current study. While this provides a native clinical feature in the current study, inter-
quantitative clinical measure, it does not represent estingly it was found to be a significant predictor
how the foot functions dynamically.25–27 What of function in PT. However, it was negatively
appears more important in terms of injury devel- correlated with function such that decreases in
opment is the amount and duration of foot motion dorsiflexion range were actually coupled with
during dynamic loading. Unfortunately, quantita- increases in function. This finding requires further
tive clinical measures of this motion are not widely exploration as it directly opposes the general
available. hypothesis that a reduction in dorsiflexion range
Further discriminative features of PT included is a risk factor for PT.
sport hours per week, mass, BMI, and leg length A number of limitations of the current study
difference. Every attempt was made to match warrant discussion. The most obvious is its cross-
controls to those with PT in terms of sport hours sectional design which limits the ability to estab-
per week. However, it proved difficult to recruit lish cause and effect. For instance, it is not possible
controls that completed as many hours as bilateral to determine whether differences between condi-
PT. While this may represent a study limitation, it tions preexisted and contributed to the develop-
is also an important observation that is consistent ment of PT or whether they developed subsequent
with the theory that mechanical overload is a key to symptom onset. Despite this, the study does
component in PT development. Individuals with PT provide useful information in terms of clinical
had greater mass and BMI than controls, consis- features that deserve consideration during rehabi-
tent with previous data.11,13 Increased mass theo- litation of PT and presents evidence for features
retically places increased load on the patellar that need to be studied prospectively. A further
tendon. Whether this is causative of PT is not clear limitation was the low number of individuals
as tendon damage is a function of tendon stress, recruited with unilateral and bilateral PT. As a
which is influenced not only by load but also tendon result, it was not viable to perform separate linear
size. In addition, prospective data did not find a regressions on the unilateral and bilateral PT
difference in mass between those who did and did groups to establish whether clinical features that
not develop PT.14 Leg length differences were also predicted symptoms and function differed between
greater in individuals with PT than controls. Leg subgroups. A final limitation was the assessment of
length discrepancy can generate gait deviations a limited number of primarily static clinical
and may contribute to the development of lower features. A number of potentially important fea-
limb musculoskeletal injuries.28 Thus, it repre- tures were not assessed due to time restraints,
sents an important feature to assess in individuals including strength and range measures of the hip,
presenting with PT. However, the decision of and knee extensor endurance. These need consid-
whether to intervene is not so clear as the eration in future studies. Similarly, it is acknowl-
discrepancy may represent a coincident rather edged that dynamic assessments (such as gait
than causal finding and typically a discrepancy assessment) may provide useful information; how-
of >20 mm is considered the intervention break- ever, these typically require specialized equipment
point. In the current study, unilateral and bilateral to obtain quantitative outcomes. This equipment is
PT had leg length discrepancies of 14.8  8.9 mm not typically available in clinical practice, and we
and 11.1  8.8 mm, respectively. While signifi- wanted to explore the features of PT using easily
cantly different from discrepancies observed in performed and clinical applicable tests.
control (4.5  8.5 mm), these values are below those In summary, the results of the current study
considered to be functionally important. indicate that individuals with unilateral and
Ankle dorsiflexion range and calf endurance bilateral PT share common clinical features; how-
were not discriminative clinical features of PT in ever, they differ for thigh strength and flexibility.
the current study. These measures are believed to This suggests that unilateral and bilateral PT
be risk factors for PT, with reductions in either represent distinct clinical entities and should be
potentiating the proximal transmission of impact considered separately in rehabilitation. When

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007 DOI 10.1002/jor

 CLINICAL FEATURES OF PATELLAR TENDINOPATHY 1175

treating individuals with PT, thigh strength 13. Malliaras P, Cook JL, Kent PM. 2007. Anthropometric risk
appears particularly important as it was reduced factors for patellar tendon injury among volleyball players.
in individuals with unilateral PT, and predicted Br J Sports Med 41:259–263.
14. Witvrouw E, Bellemans J, Lysens R, et al. 2001. Intrinsic
symptoms and function in PT. risk factors for the development of patellar tendinitis in an
athletic population. A two-year prospective study. Am J
Sports Med 29:190–195.
ACKNOWLEDGMENTS 15. Cook JL, Khan KM, Kiss ZS, et al. 2001. Reproducibility
and clinical utility of tendon palpation to detect patellar
tendinopathy in young basketball players. Br J Sports Med
The authors gratefully acknowledge the assistance of 35:65–69.
Dr. Zoltan S. Kiss and Ms. Elin Wee for invaluable 16. Visentini PJ, Khan KM, Cook JL, et al. 1998. The VISA
assistance in examining participants. This study was score: an index of severity of symptoms in patients with
partly supported by funding from the Australian Sports jumper’s knee (patellar tendinosis). J Sci Med Sport 1:22–
Commission. 28.
17. Warden SJ, Kiss ZS, Malara FM, et al. 2007. Comparative
accuracy of magnetic resonance imaging and ultrasono-
graphy in confirming clinically diagnosed patellar tendino-
REFERENCES pathy. Am J Sports Med 35:427–436.
18. Crossley KM, Bennell KL, Cowan SM, Green S. 2004.
1. Cook JL, Khan KM, Harcourt PR, et al. 1997. A cross Analysis of outcome measures for persons with patellofe-
sectional study of 100 athletes with jumper’s knee moral pain: which are reliable and valid? Arch Phys Med
managed conservatively and surgically. The Victorian Rehabil 85:815–822.
Institute of Sport Tendon Study Group. Br J Sports Med 19. Cook JL, Khan KM, Kiss ZS, et al. 2001. Asymptomatic
31:332–336. hypoechoic regions on patellar tendon ultrasound: a 4-year
2. Khan KM, Bonar F, Desmond PM, et al. 1996. Patellar clinical and ultrasound followup of 46 tendons. Scand J
tendinosis (jumper’s knee): findings at histopathologic Med Sci Sports 11:321–327.
examination, US, and MR imaging. Victorian Institute of 20. Malliaras P, Cook J, Ptasznik R, Thomas, S. 2006.
Sport Tendon Study Group. Radiology 200:821–827. Prospective study of change in patellar tendon abnormality
3. Lian OB, Engebretsen L, Bahr R. 2005. Prevalence of on imaging and pain over a volleyball season. Br J Sports
jumper’s knee among elite athletes from different sports: a Med 40:272–274.
cross-sectional study. Am J Sports Med 33:561–567. 21. Hopkins JT, Ingerson CD. 2000. Arthrogenic muscle
4. Maffulli N, Testa V, Capasso G, et al. 2004. Similar inhibition: a limiting factor in joint rehabilitation. J Sport
histopathological picture in males with Achilles and Rehabil 9:135–169.
patellar tendinopathy. Med Sci Sports Exerc 36:1470– 22. Murphy DF, Connolly DA, Beynnon BD. 2003. Risk factors
1475. for lower extremity injury: a review of the literature. Br J
5. Khan KM, Cook JL, Kannus P, et al. 2002. Time to Sports Med 37:13–29.
abandon the ‘‘tendinitis’’ myth. BMJ 324:626–627. 23. Whitehead CL, Hillman SJ, Richardson AM, et al.
6. Kettunen JA, Kvist M, Alanen E, Kujala UM. 2002. Long- 2006. The effect of simulated hamstring shortening on
term prognosis for jumper’s knee in male athletes. A gait in normal subjects. Gait Posture. Published online:
prospective follow-up study. Am J Sports Med 30:689–692. doi:10.1016/j.gaitpost.2006.07.011
7. Khan KM, Maffulli N. 1998. Tendinopathy: an Achilles’ 24. Richards DP, Ajemian SV, Wiley JP, Zernicke RF. 1996.
heel for athletes and clinicians. Clin J Sport Med 8:151– Knee joint dynamics predict patellar tendinitis in elite
154. volleyball players. Am J Sports Med 24:676–683.
8. Warden SJ, Brukner P. 2003. Patellar tendinopathy. Clin 25. Cashmere T, Smith R, Hunt A. 1999. Medial longitudinal
Sports Med 22:743–759. arch of the foot: stationary versus walking measures. Foot
9. Cook JL, Kiss ZS, Khan KM, et al. 2004. Anthropometry, Ankle Int 20:112–118.
physical performance, and ultrasound patellar tendon 26. Nachbauer W, Nigg BM. 1992. Effects of arch height of the
abnormality in elite junior basketball players: a cross- foot on ground reaction forces in running. Med Sci Sports
sectional study. Br J Sports Med 38:206–209. Exerc 24:1264–1269.
10. Gaida JE, Cook JL, Bass SL, et al. 2004. Are unilateral and 27. Nigg BM, Cole GK, Nachbauer W. 1993. Effects of arch
bilateral patellar tendinopathy distinguished by differ- height of the foot on angular motion of the lower
ences in anthropometry, body composition, or muscle extremities in running. J Biomech 26:909–916.
strength in elite female basketball players? Br J Sports 28. Gurney B. 2002. Leg length discrepancy. Gait Posture 15:
Med 38:581–585. 195–206.
11. Lian O, Refsnes P-E, Engebretsen L, Bahr R. 2003. 29. Cook JL, Khan KM, Maffulli N, Purdam C. 2000. Overuse
Performance characteristics of volleyball players with tendinosis, not tendinitis. Part 2: Applying the new
patellar tendinopathy. Am J Sports Med 31:408–413. approach to patellar tendinopathy. Physician Sports Med
12. Malliaras P, Cook JL, Kent P. 2006. Reduced ankle 28:31–46.
dorsiflexion range may increase the risk of patellar tendon 30. Richards DP, Ajemian SV, Wiley JP, et al. 2002. Relation
injury among volleyball players. J Sci Med Sport 9:304– between ankle joint dynamics and patellar tendinopathy in
309. elite volleyball players. Clin J Sport Med 12:266–272.

DOI 10.1002/jor JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2007