Content uploaded by Karin Grävare Silbernagel
Author content
All content in this area was uploaded by Karin Grävare Silbernagel on Feb 23, 2014
Content may be subject to copyright.
Alexander Gustavsson
Camille Neeter
Pia Thomee
´
Karin Gra
¨
vare Silbernagel
Jesper Augustsson
Roland Thomee
´
Jon Karlsson
A test battery for evaluating hop performance
in patients with an ACL injury and patients
who have undergone ACL reconstruction
Received: 14 February 2005
Accepted: 9 August 2005
Published online: 9 March 2006
Springer-Verlag 2006
Abstract The purpose of this study
was to develop a test battery of hop
tests with high ability to discriminate
(i.e. high test–retest reliabili ty, sen-
sitivity, specificity and accuracy)
between the hop performance of the
injured and the uninjured side in
patients with an ACL injury and in
patients who have undergone ACL
reconstruction. Five hop tests were
analysed: three maximum single hop
tests and two hop tests while devel-
oping fatigue. Fifteen healthy sub-
jects performed the five hop tests on
three separate occasions in a test–
retest design. Thirty patients, mean
11 months after an ACL injury and
35 patients, mean 6 months after
ACL reconstruction were tested.
ICC values ranged from 0.85 to 0.97
for the five hop tests, indicating that
all the tests had high test–retest
reliability. Sixty-seven percent to
100% of the healthy subjects had
normal symmetry (i.e. <10% side-
to-side difference) in the five hop
tests. Abnormal symmetry in the five
hop tests ranged from 43 to 77% for
patients with an ACL injury and
from 51 to 86% for patients who
had undergone ACL reconstruction
respectively. The three tests with the
highest ability to discriminate hop
performance were chosen for the test
battery; they were the vertical jump,
the hop for distance and the side
hop. The test battery revealed a high
level of sensitivity and accuracy in
patients with an ACL injury (87 and
84%) and in patients who had
undergone ACL reco nstruction (91
and 88%), when at least one of the
three tests was classified as abnor-
mal. To summarise, the test battery
consisting of both maximum single
hop performances: the vertical jump
and the hop for distance and hop
performance while developing fati-
gue: the side hop, produced high
test–retest reliability, sensitivity and
accuracy. Further, the test battery
produced higher values compared
with any of the three hop tests indi-
vidually revealing that only one out
of ten patients had restored hop
performance 11 months after an
ACL injury and 6 months after ACL
reconstruction. It is concluded that
this test battery showed a high abil-
ity to discriminate between the hop
performance of the injured and the
uninjured side both in patients with
an ACL injury and in patients who
have undergone ACL reconstruc-
tion.
Keywords Anterior cruciate
ligament Æ Knee Æ Rehabilitation Æ
Hop test
DOI 10.1007/s00167-006-0045-6
Knee Surg Sports Traumatol Arthrosc
(2006) 14: 778–788
SPORTS MEDICINE
A. Gustavsson Æ C. Neeter
P. Thomee
´
Æ K. Gra
¨
vare Silbernagel
J. Augustsson Æ R. Thomee
´
Æ J. Karlsson
Department of Orthopaedics,
Sahlgrenska University Hospital,
Go
¨
teborg University, Go
¨
teborg, Sweden
A. Gustavsson Æ C. Neeter
P. Thomee
´
Æ K. Gra
¨
vare Silbernagel
J. Augustsson Æ R. Thomee
´
Sportrehab – Physical Therapy and Sports
Medicine Clinic, Go
¨
teborg, Sweden
A. Gustavsson (&)
Department of Orthopaedics, Lundberg
Laboratory for Human Muscle Function
and Movement Analysis, Sahlgrenska
University Hospital, 413 45 Go
¨
teborg,
Sweden
E-mail:
alexander.gustavsson@orthop.gu.se
Tel.: +46-31-426891
Fax: +46-31-416816
Introduction
One difficult challenge in the rehabilitation of anterior
cruciate ligament (ACL) injury and reconstruction is to
determine when it is safe to return to strenuous physical
activities. Single-leg hop tests are commonly used to
evaluate functional performance after an ACL injury
[1–7] or reconstruction [8–19]. To some de gree, hop tests
have been able to discriminate between the hop perfor-
mance of the injured and uninjured side in patients with
an ACL injury and between controls and patients [1, 2,
4, 7, 10]. There is also some evidence that single-leg hop
tests are important when trying to predict whether pa-
tients will have future difficulty in terms of recurrent
knee instability after an ACL injury [3].
The ratio between the involved and uninvolved leg
has been the most frequently reported criterion for
determining normal or abnormal hop test scores [1–6,
8–10, 13, 15, 17–21]. It has been suggested that the
normal ratio in healthy subjects is greater than or equal
to 85% [1, 21] or 90% [10, 20].
Several hop tests are described in the literature,
including various single-leg hop tests for distance, time
[1–6, 8–11, 13–21] and height [1 , 13, 15, 20, 21]. The
evaluation of functional outcome after an ACL injury
is often limited to only one test, the single-leg hop for
distance [9, 14, 17, 18]. However, the reported sensi-
tivity for detecting functional limitations associated
with ACL deficiency with the single-leg hop for dis-
tance test is relatively low, ranging from 38 to 52%
[4, 5, 7]. Noyes et al. [5] combined two horizontal hop
tests, a single-leg hop for dista nce and a timed hop in
order to increase the sensitivity. Itoh et al. [4] argued
that hop tests should involve more twisting and cutting
movements and evaluated the figure-of-eight hop test,
the up–down hop test and the side hop test. When
these three tests and the single-leg hop for dista nce test
were combined in a test battery, a sensitivity of 82%
was obtained. Sports injuries often tend to occur at the
end of a sporting event, when a participant is fati gued
[22–24]. None of the tests recommended by either
Noyes et al. [5] or Itoh et al. [4] is, however, per-
formed under fatigued conditions. Augustsson et al. [8]
reported an improved sensitivity level when testing hop
performance under fatigued conditions. All patients in
the study by Augustsson et al. [8] were classified,
11 months after ACL reconstruction, as having normal
hop capacity when performing a single leg hop for
distance non-fatigued. After a fatiguing quadriceps
muscle exercise only one-third of the patients were
classified as having normal hop performance.
The purpose of this study was to develop a test battery
of hop tests with a high ability to discriminate (i.e. high
test–retest reliability, sensitivity, specificity and accuracy)
between the hop perfo rmance of the injured and the
uninjured side in patients with an ACL injur y and in
patients who have undergone ACL reconstruction.
We hypothesise that a test battery evaluating different
hop qualities, i.e. maximum single hop performance, as
well as hop performance while developing fatigue, will
increase the opportunity to detect discrepancies in hop
performance (i.e. increase the test sensitivity) compared
with using only a single hop test.
Materials and methods
Subjects
Three groups of subjects participated in this study,
healthy individuals, patients with an ACL injury and
patients who had undergone ACL reconstruction.
The patients physi cal activity levels prior to their
ACL injury and at the test occasion were documented
using the Tegner score [25]. The Tegner score is an
activity grading scale, where activities of daily living,
recreation, competitive sports and work are graded
numerically from 1 to 10. One represents the least
strenuous knee activity and ten is hard strenuous knee
activity such as rugby and international soccer. The
score was modified in the year 2000 but this version has
not been published. The modified version was used in
the present study with the permission of the authors. To
complement the Tegner score with the subjects’ intensity
and frequency in participation in physical activity a four
grade physical activity scale was constructed, using a
validated score for elderly people [26] as a model. After
discussions in an expert group consisting of experienced
physical therapists and orthopaedic surgeons good face
validity of the new physical activity scale was assured.
On the physical activity scale the subjects made their
own judgment on how vigorously and frequent they
participated in physical activity at the present time as
well as prior to their knee injury. The four grades in this
physical activity scale were:
1. Hardly any physical activity at all.
2. Light physical activ ity a few hours a week.
3. Somewhat strenuous physical activity 2–3 h a week.
4. Hard strenuous physical activity on a regular basis.
A convenience sample of nine male and six female
healthy subje cts were recruited. The subjects had no
history of back, hip, knee or ankle dysfunction and were
tested on three separate occasions in a test–retest design.
Thirty patients, 18 males and 12 females, with an
ACL injury were tested. An interval of 4–44 months
elapsed between the index knee injury and the test
occasion. The inclusion criteria consisted of a positive
anterior drawer and Lachman’s test, performed by
779
experienced orthopaedic surgeons and verified by the
patient’s physiotherapist, history of knee injury and
subjective giving way, no acute pain or swelling and no
prior surgical procedu re in either leg.
Thirty-five patients, 25 males and 10 females, who
had undergone ACL reconstruction, were tested. The
patients were included in this study if they met the
following criteria, no acute pain or swelling and no
prior surgical procedure in either leg. The operated
patients were tested 6 months after ACL reconstruc-
tion. Descriptive data for all subjects are presented in
Table 1.
All patients in this study were recruited from the
same physiotherapy clinic and underwent supervised
physiotherapy according to a criterion based protocol.
The patients trained two to three sessions per week
(initially combined with a home based program). Wet-
vest training and functional group training were offered
as additional sessions. Full immediate weight bearing
was allowed post operatively. Early range of motion
exercises was encouraged. Both closed and open kinetic
chain exercises were used to restore muscle strength of
lower extremity muscles. Also neuromuscular training
(balance, proprioception and plyometric exercises) were
incorporated with gradually increased load and com-
plexity.
Prior to any testing, written informed consent forms
were signed by all the participants. Approval for the
study was obtained from the Human Ethics Committee
at Go
¨
teborg University, Sweden.
Procedure
The healthy subjects performed five single-leg hop tests
in both legs at three different test sessions in a test–retest
design. An interval of 3–13 days elapsed between test
occasion 1 and 2 and an interval of 3–19 days between
test occasion 2 and 3. Subjects were asked not to par-
ticipate in strenuous physical activities the day before
testing. For the healthy subjects, the order of the tests
and the leg that was first tested were randomised,
whereas the test order was pre-determined for the pa-
tients. The patients performed the tests in the same order
as they are presented. The practice and test trials were
performed using the uninjured leg first, followed by the
injured leg. The patients were thoroughly familiarised at
the physiotherapy clinic prior to the test session. All the
tests were supervised by the same test leader. Verbal
encouragement was used and athletic footwear were
standardised.
Before testing, the subjects completed a warm-up
consisting of 5 min of stationary cycling, two times 10
squats, two times 10 toe rises and warm-up jumps.
The following single-leg hop tests were used: (1) ver-
tical jump, (2) hop for distance, (3) drop jump followed
by a double hop for distance, (4) square hop and (5) side
hop. The tests were chosen on the basis of hop tests
commonly described in the literature [1–6, 8–21], as well
as clinical experience, and were designed to reflect various
hop qualities. The tests should also be easy to administer
in a clinical setting. The subjects performed three to five
practice trials followed by three maximum approved trials
for the vertical jump, the hop for distance and the drop
jump, followed by a double hop for distance. However, if
the subjects increased their hop performance in all three
hops, additional hops were performed until no increase
was seen. The square hop and the side hop were tested
once. Three minutes of rest were used between each hop
test. The best trial for each leg in each test was used for
data analysis. The hop tests were videotaped for sub-
sequent control of test procedures.
Table 1 Descriptive characteristics of healthy subjects and patients
Age (years) Height
(cm)
Weight
(kg)
Tegner
before
injury
PAS
before
injury
Tegner at
test occasion
PAS at test
occasion
Weeks
injury/op-test
Healthy subjects
Male (n=9) 29±5 181±6 84±10
Female (n=6) 26±4 168±8 61±6
All subjects (n=15) 28±4 175±9 75±14
ACL injury
Male (n=18) 28±7 181±8 91±17 7.5 (2.0) 4.0 (1.0) 4.0 (2.2) 2.5 (1.0) 50±48
Female (n=12) 36±8 167±5 71±12 4.5 (3.5) 3.0 (1.0) 3.0 (1.0) 3.0 (1.0) 45±35
All subjects (n=30) 31±9 175±9 83±18 7.0 (3.0) 3.0 (1.0) 4.0 (1.2) 3.0 (1.0) 48±43
ACL reconstruction
Male (n=25) 27±8 180±8 81±11 8.0 (2.0) 3.0 (1.0) 4.0 (3.5) 3.0 (1.5) 28±2
Female (n=10) 27±7 167±5 60±5 7.0 (2.5) 3.5 (1.0) 3.0 (1.0) 3.0 (1.0) 28±2
All subjects (n=35) 27±7 177±10 75±13 8.0 (2.0) 3.0 (1.0) 4.0 (2.0) 3.0 (1.0) 28±2
All values are expressed as means (±SD), except for Tegner activity score and Physical Activity Scale (PAS), which is expressed as median
values (interquartile range)
780
Hop tests
Vertical jump (Fig. 1a)
The vertical jump test was performed as a counter-
movement jump. The starting position was an upright
position with the hands place d behind the back. The
subjects quickly bent their knee as much as desired and
then immediately jumped upwards, attempting to max-
imise the height jumped. A computerised system (Mus-
cleLab, Ergotest Technology) using a field of infrared
light (approximately 10 mm above the floor), serving as
a ‘‘contact mat’’, made it possible to measure the flight
time. The system then converted the flight time into
jump height in centimetres.
Hop for distance (Fig. 1b)
The subjects stood on the test leg and then hopped as far
as possible and landed on the same leg. Free leg swing
was allowed. The hands were placed behind the back.
The subjects were instructed to perform a controlled,
balanced landing and to keep the landing foot in place
(i.e. no extra hops were allowed) until (2–3 s) the test
leader had registered the landing position. Failure to do
so resulted in a disqualified hop. The distance was
measured in centimetres from the toe at the push-off to
the heel where the subject landed.
Drop jump followed by a double hop for distance
(Fig. 1c)
The starting position was standing on the leg to be tested
on a box, at a height of 30 cm, with the hands behind the
back. Forty-fi ve centimetres in front of the box, a strip of
tape marked the starting line. The subjects jumped down
on one leg, without crossing or touching the starting line,
and then immediately performed two one-legged maxi-
mum hops forward. Th e subjects were instructed to
Fig. 1 Five single leg hop tests
were used: a vertical jump,
b hop for distance, c drop jump
followed by a double hop for
distance, d square hop and
e side hop
781
perform a controlled, balanced landing and to keep the
landing foot in place (i.e. no extra hops were allowed)
until (2–3 s) the test leader had regi stered the landing
position. Failure to do so resul ted in a disqualified hop.
The distance was measured in centimetres from the
starting line to the heel where the subject landed.
Square hop (Fig. 1d)
The subj ects stood on the leg to be tested, with their
hands behind their back, outside a 40·40 cm square
marked with tape on the floor. A 10 cm frame was also
marked around the square with tape. For the right leg,
the subjects were instructed to jump clockwise in and out
of the square as many times as possible during a period of
30 s. The numb er of successful jumps performed, with-
out touching the taped frame, was recorded. Touching
the taped frame was recorded as an error and, if more
than 25% of the jumps had errors, a second trial of 30 s
was performed after a 3-min rest period. For the left leg,
the subject performed the test in a counter-cl ockwise
mode. This test was modified from O
¨
stenberg et al. [21].
Side hop (Fig. 1e)
For the side hop test, the subjects stood on the test leg,
with their hands behind their back, and jumped from side
to side betw een two parallel strips of tape, placed 40 cm
apart on the floor. The subjects were instructed to jump
as many times as possible during a period of 30 s. The
number of successful jumps performed, without touching
the tape, was recorded. Touching the tape was recorded
as an error and, if more than 25% of the jump s had
errors, a second trial of 30 s was performed after a 3-min
rest period. This test was modified from Itoh et al. [4].
Statistical analysis
Test–retest analysis was determined by an intraclass
correlation coefficient (ICC) with a 95% confidence
interval (CI) [27]. The methodological error in percent
was calculated according to the formula [28]:
[e/m] 100 = % difference between test
e the square root of [
P
(d
2
)/2n]
d difference between tests
n number of subjects
m mean value.
The mean and standard deviation were calculated for
the jump tests.
Wilcoxon’s signed rank test was used for dependent
values and the Mann–Whitney U test was used for
independent values. The Spearman Rank was used to
study the relationship between dependent variables. The
level of significance was set at P £ 0.05.
The lower limb symmetry index (LSI) was calculated
to determine whether a side-to-side leg difference was
classified as normal or abnormal. The LSI is defined as
the ratio of the involved limb score and the uninvolved
limb score expressed in per cent (involved/unin-
volved · 100 = LSI). In this study, an LSI greater than
or equal to 90% [10, 20] was classified as normal.
Sensitivity, specificity and accuracy were calculated
for the five hop tests and for the test battery using the
following definitions.
Sensitivity (= number of patients classified as
abnormal/total number of patients) expresses the per-
centage probability that the tests would demonstrate an
abnormal LSI in the patients [29].
Specificity (= number of healthy subjects classified as
normal/total number of healthy subjects) expresses the
percentage probability that the tests would demonstrate
a normal LSI in the normal subjects [29].
Accuracy (= number of patients classified as
abnormal + number of healthy subjects classified as
normal/total number of patients and healthy subjects) is
defined as the percentage probability that the tests
would demonstrate a normal LSI in the normal subjects
and an abnormal LSI in the patients [29].
A factor analysis with a principal axis factoring,
varimax rotation, and for two factors was applied for
evaluation of the structure of the test battery.
Results
Test–retest reliability for the five hop tests
Significant differences were found between test occasion
1 and 2 for male subjects and for male and female
subjects combined in the hop for dista nce (P=0.02) and
the square hop (P=0.001). A significant difference was
also found between test occasion 1 and 2 for all subjects
in the side hop (P=0 .01). No significant differences were
found between test occasion 2 and 3 in any of the five
hop tests, as shown in Table 2.
The ICC, 95% CI and methodological error in per-
cent and in absolute values are presented in Table 2.In
the five tests, the ICC between test occasion 2 and 3
ranged from 0.85 to 0.97 and the methodological error
ranged from 3 to 6%.
Comparison of hop performance between healthy male
and female subjects
A comparison between healthy male and female sub-
jects’ hop performance in the five hop tests on test
782
occasion 3 is presented in Table 3. Significant differ-
ences were found between males and female s in the
vertical jump (P=0.03), the hop for distance (P=0.03)
and the drop jump followed by a double hop for
distance (P=0.01) but not in the square hop and the
side hop. No significant difference was, however, found
in any of the tests when comparing the absolute values
for side-to-side difference in males and females. Fur-
ther analyses were therefore conducted, regardless of
gender.
Hop pe rformance in patients
A significant difference between the injured and unin-
jured side was found in all the hop tests except for the
square hop tests in patients with an ACL injury and in
all the tests in patients who had undergone ACL
reconstruction (Table 4).
Comparison of hop performance between healthy
subjects and patients
Patients with an ACL injury and patients who had
undergone ACL reconstruction had a significantly larger
side-to-side difference in all the tests except for the
square hop compared with healthy subjects (Table 5).
The results of the square hop test did not reveal a sig-
nificant diff erence between healthy subjects and patients
when comparing the side-to-side difference.
A normal LSI for the five hop tests ranged from 67 to
100% in the healthy subjects, from 23 to 57% in the
patients with an ACL injury and from 14 to 49% in the
Table 2 Test–retest reliability between test occasion 1 and 2 (Test
1–2
) and test occasion 2 and 3 (Test
2–3
)
Test P-value ICC 95% CI Methodological error
% cm or jumps
Vertical jump (cm)
Male (n=9) Test
1–2
0.08 0.83 0.61–0.93 8 1.4
Test
2–3
0.85 0.95 0.88–0.98 4 0.7
Female (n=6) Test
1–2
0.72 0.95 0.84–0.99 5 0.7
Test
2–3
0.82 0.97 0.90–0.99 4 0.6
All subjects (n=15) Test
1–2
0.11 0.89 0.79–0.95 7 1.2
Test
1–2
0.94 0.97 0.93–0.98 4 0.6
Hop for distance (cm)
Male (n=9) Test
1–2
0.02 0.91 0.78–0.97 3 4.2
Test
2–3
0.20 0.86 0.67–0.95 3 4.6
Female (n=6) Test
1–2
0.35 0.88 0.64–0.96 3 4.7
Test
2–3
0.08 0.98 0.93–0.99 2 2.1
All subjects (n=15) Test
1–2
0.02 0.94 0.88–0.97 3 4.3
Test
2–3
0.07 0.95 0.90–0.98 3 3.7
Drop jump followed by a double hop for distance (cm)
Male (n=9) Test
1–2
0.07 0.80 0.54–0.92 3 9.6
Test
2–3
0.41 0.85 0.65–0.94 3 9.6
Female (n=6) Test
1–2
0.43 0.67 0.19–0.89 6 15.0
Test
2–3
0.13 0.91 0.73–0.97 3 8.3
All subjects (n=15) Test
1–2
0.07 0.85 0.72–0–93 4 12.1
Test
2–3
0.13 0.93 0.85–0.96 3 8.7
Square hop (jumps)
Male (n=9) Test
1–2
0.00 0.64 0.18–0.87 13 6.9
Test
2–3
0.09 0.78 0.50–0.91 5 3.1
Female (n=6) Test
1–2
0.06 0.55 0.01–0.85 12 6.9
Test
2–3
0.48 0.89 0.67–0.97 5 3.2
All subjects (n=15) Test
1–2
0.00 0.58 0.25–0.78 13 6.9
Test
2–3
0.08 0.85 0.71–0.93 5 3.1
Side hop (jumps)
Male (n=9) Test
1–2
0.06 0.72 0.39–0.88 9 4.5
Test
2–3
0.61 0.78 0.51–0.91 5 3.0
Female (n=6) Test
1–2
0.10 0.87 0.62–0.96 14 5.5
Test
2–3
0.58 0.95 0.83–0.98 8 3.5
All subjects (n=15) Test
1–2
0.01 0.87 0.75.0.94 10 4.8
Test
2–3
0.95 0.93 0.87–0.97 6 3.2
P-values, intraclass correlation coefficients (ICC), 95% confidence intervals (CI) and the methodological errors are given in percent and in
absolute values
783
patients who had undergone ACL reconstruction
(Fig. 2).
Sensitivity, specificity and accuracy of the five hop tests
Table 6 shows the sensitivity, specificity and accuracy of
the five hop tests. Sensitivity ranged from 43 to 77% in
the patients with an ACL injury. Sensitivity in the pa-
tients who had undergone ACL reconstruction ranged
from 51 to 86%. Specificity ranged from 67 to 100% in
the five hop tests. Accuracy ranged from 58 to 80% in
the patients with an ACL injury. Accuracy in the pa-
tients who had undergone ACL reconstruction ranged
from 56 to 86%.
Correlation between the five hop tests
The correlation in healthy subjects and patients ranged
from r
s
=0.70 to r
s
=0.94 (P=0.01) between the three
maximum hop tests (vertical jump, hop for distance and
drop jump followed by a double hop for distance). The
correlation in healthy subjects and patients ranged from
r
s
=0.72 to r
s
=0.92 (P=0.01) between the two endur-
ance hop tests (square hop and side hop). The correla-
tion between the three maximum hop tests and the two
endurance hop tests ranged from r
s
=0.29 to r
s
=0.57
(P=0.05) in the healthy subjects and from r
s
=0.62 to
r
s
=0.88 (P=0.01) in the pa tients.
Factor analysis on the five hop tests
The factor analysis on the healthy subjects as wells as the
factor analysis on the patients divided the five hop tests
in factor one: maximal hop tests (vertical jump, hop for
distance and drop jump followed by a double hop for
distance) and facto r two: endurance hop tests (square
hop and side hop).
The test battery
The following three tests were chosen for the test bat-
tery: (1) the vertical jump, (2) the hop for distance and
(3) the side hop.
Among healthy subjects, 3 of 15 (20%) were classified
as abnormal in at least one of the three tests in the test
battery. Among patients with an ACL injury ,12of30
(40%) obtained abnormal LSI values in all three tests
and 4 of 30 (13%) obtained normal values in all three
tests. Twenty-six of 30 (87%) patients with an ACL in-
jury obtained abnormal LSI values in at least one of the
three tests. Among patients who had undergo ne ACL
reconstruction, 19 of 35 (54%) obtained abnormal LSI
values in all three tests and three of 35 (9%) obtained
normal values in all three tests. Thirty-two of 35 (91%)
patients who had undergone ACL reconstruction ob-
tained abnormal values in at least one of the three tests.
The sensitivity of the test battery, i.e. identifying a
patient as abnormal when at least one of the three tests
produced an abnormal LSI value, was 87% in the pa-
tients with an ACL injur y and 91% in the patients who
had undergone ACL reconstruction. The accu racy of the
test battery, i.e. identifying a healthy subject as normal
when all three tests produced a normal LSI value and
identifying a patient as abnormal when at least one of
the three tests produced an abnormal LSI value, was
found to be 84 and 88% respectively.
The square hop was not chosen due to its inability to
discriminate between the side-to-side difference in heal-
thy subjects and patients (Table 5). The square hop also
showed the lowest specificity (67%) and acc uracy (58
and 56%) of the five hop tests . The correlation with the
side hop ranged from r
s
=0.72 to r
s
=0.92 (P=0.01) for
Table 4 Means, standard
deviations (SD) in the five hop
tests in patients with ACL
injury and patients who have
undergone ACL reconstruction
(ACL recon.)
P-values between the injured
and the uninjured sides are
given
Test ACL injury (n=30) P-value ACL recon (n=35) P-value
Injured Uninjured Injured Uninjured
Vertical jump (cm) 11.5±5.4 14.5±5.4 0.05 13.3±5.0 17.5±4.6 <0.00
Hop for distance (cm) 115±39 135±29 <0.04 128±28 148±23 <0.00
Drop jump followed by
a double hop for distance (cm)
225±83 270±55 <0.03 253±56 297±48 <0.00
Square hop (jumps) 38±18 46±15 NS 49±17 57±12 <0.03
Side hop (jumps) 28±17 39±16 <0.01 39±16 49±13 <0.00
Table 3 Means, standard deviations (SD) in the five hop tests on
test occasion 3 for the right and left leg
Test All subjects
(n=15)
Male
(n=9)
Female
(n=6)
P-value
Vertical jump (cm) 16.7±3.7 18.1±3.3 15.0±3.7 <0.03
Hop for distance (cm) 151±16 160±11 137±13 <0.01
Drop jump followed
by a double hop
for distance (cm)
304±34 320±27 281±30 <0.01
Square hop (jumps) 62±7 62±6 60±9 NS
Side hop (jumps) 50±13 55±6 41±16 NS
P-values between males and females among healthy subjects are
given (NS=not significant)
784
all subjects, indicating a strong relationship between the
two endurance hop tests.
The drop jump followed by a double hop for distance
was not chosen due to its high correlation with the sin-
gle-leg hop for distance (r
s
>0,80, P=0.01), accuracy
similar to that of the hop for distance in the patients
with an ACL injury (62 vs. 69%) and identical to that
for the patients who had undergone ACL reconstruction
(74%). We therefore concluded that the drop jump fol-
lowed by a double hop for distance and the hop for
distance were equal and measured the same ability.
Discussion
The principal finding in this study was that a test bat-
tery, consisting of the vertical jump, the hop for distance
and the side hop tests, had a high ability to discriminate
between the hop perfo rmance of the injured and the
uninjured side both in patients 11 months after an ACL
injury and in patients 6 months after ACL reconstruc-
tion. The test battery may help in the process of deciding
whether and when patients can safely return to strenu-
Table 6 Results for sensitivity, specificity and accuracy in the five hop tests for healthy subjects (n=15), patients with ACL injury (n=30)
and patients who have undergone ACL reconstruction (n=35)
Vertical jump Hop for distance Drop jump followed by
a double hop for distance
Square hop Side hop
Sensitivity (%)
ACL injury 67 53 43 53 77
ACL recon. 86 63 63 51 69
Specificity (%)
Healthy subjects 87 100 100 67 87
Accuracy (%)
ACL injury 73 69 62 58 80
ACL recon. 86 74 74 56 74
P-values for comparisons between healthy subjects and patients with ACL injury and ACL reconstruction (ACL recon.) are given
Table 5 Means, standard deviations (SD) for side-to-side difference in the five hop tests
Test Healthy subjects (n=15) ACL injury (n=30) P-value ACL recon. (n=35) P-value
Vertical jump (cm) 1.2±1.4 3.0±2.5 <0.01 4.3±2.8 <0.01
Hop for distance (cm) 6±4 21±24 <0.05 20±11 <0.01
Drop jump followed by a
double hop for distance (cm)
8±7 45±62 <0.01 44±32 <0.01
Square hop (jumps) 4±3 7±12 NS 8±9 NS
Side hop (jumps) 4±3 11±9 <0.05 10±9 <0.05
47%
57%
47%
23%
37% 37%
49%
67%
87%
100%100%
87%
33%
34%
14%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Vertical jump Hop for distance
Drop jump followed
by a double hop for
distance
Square hop Side hop
Healthy
ACL injury
ACL recon.
Fig. 2 The percentage of
healthy subjects (n=15) and
patients with ACL injury
(n=30) and ACL reconstruc-
tion (ACL recon.)(n=35) clas-
sified as normal in the five hop
tests
785
ous physical activities after an ACL injury or recon-
struction. The test battery has, however, to be studied in
prospective randomised trails to be able to draw definite
conclusions on its actual ability to contribute to the
decision making process.
Single-leg hop tests are commonly used to study knee
function in patients with an ACL injury and are de-
signed to reflect the demands of a high level of physical
activity [1–6, 8–21]. In the literature it is argued that by
using a variety of hop tests, different hop qualities can be
evaluated and thereby increase the opportunity to detect
discrepancies in hop performance (i.e. increase the test
sensitivity) [1, 4, 5, 16]. The importance of hop perfor-
mance testing of patients after ACL reconstruction in a
fatigued state has also been advocated [8]. In the present
study three different maximum hop tests and two
endurance hop tests were used.
The difference between the three maximum hop tests
and the two endurance hop tests is reflected by the
correlation between the two types of hop tests, indicat-
ing a fair to moderate [30] degree of relationship
(r
s
=0.29–0.57) in the healthy subjects. O
¨
stenberg et al.
[21] reported similar correlations between the square
hop and the vertical jump (r=0.12) and the single-leg
hop for distance (r=0.38). The results indicate that the
two types of hop tests measure different aspects of
jumping ability, which is further confirmed by the factor
analysis on the five hop tests. Patients with an ACL
injury had, however, a strong relationship (r
s
=0.82–
0.88) between the two types of hop tests. One explana-
tion could be that the patients with ACL injury in this
study are not as homogeneous a group as the patients
who had undergone ACL reconstruction, especially in
terms of the interval between the index injury and the
test occasion and in the diversity of the functional defi-
cits.
During the two endurance hop tests, square hop and
side hop, the patients perfo rmed as many jumps as
possible during a period of 30 s, thereby demanding
knee stability while developing muscle fatigue. The side
hop test used by Itoh et al. [4] was modified not only by
increasing the test duration from less than 2 to 30 s but
also by increasing the distance the patients were required
to jump from 30 to 40 cm. This could explain the higher
sensitivity of 77% for the side-hop test found in the
present study compared with the 44% in the study by
Itoh et al. [4]. The reported sensitivity in these studies
should, however, be compared with caution, due to
different populations in terms of the time between the
injury and the test occasion, for example. In this study, a
sensitivity of 53% at a 90% LSI cut-off score in the
single-leg hop for distance in the patients with an ACL
injury was noted, which is in accordance with Noyes
et al. [5] and Itoh et al. [4], who observed a sensitivity of
52 and 42%, respectively, at an 85% LSI cut-off score.
In a study by Barber et al. [1], many of the healthy
subjects scored outside the normal lower limb symmetry
range in the vertical jump and specificity was as low as
48% at 90% LSI). This is considerably different from
the specificity of 87% in the present study, which is a
more acceptable result. One explanation for this differ-
ence in results could be due to the fact that Barber et al.
[1] used the jump-and-reach method, which has in an-
other previous study showed poor reliability [coefficient
of variation (CV) of 9.9%] [20].
The ICC values ranged from 0.85 to 0.97 for the five
different hop tests, indicating that all the tests had high
reliability. The importance of familiarisation needs to be
mentioned, as a significant increase in performance was
seen between test occasion 1 and 2 in the hop for dis-
tance and in the square hop. The difference may be the
result of a learning process. The patients therefore
practiced and were thoroughly familiarised at the
physical therapy clinic before the test occasions.
The test–retest result in the vertical jump (ICC 0.97)
is in agreement with those reported by Wilson et al. [31]
and Manske et al. [32], who obtained ICC values of 0.97
and 0.98 using other apparatus. Several authors have
described and tested the reliability of the single-leg hop
for distance in healthy subjects [ 7, 14, 32] and patients
after an ACL injury [10, 14]. ICC values ranging from
0.89 and 0.97 are reported [14, 32]. These results agree
with the ICC value of 0.95 in this study.
The square hop and the side hop were generally
performed only once by the healthy subjects. Despite
this fact, the tests showed high reliability for both the
square hop (ICC 0.85) and the side hop (ICC 0.93).
Three of the patients, who had errors in more than 25%
of the jumps, in one of the two tests, made a second
attempt. To minimise the effect of fatigue, 3 min of rest
were given before the second attempt. The number of
trials was limited in order to control the fatigue effects
that were induced and to reduce the total time of the test
battery.
The results of the present study revealed no differ-
ence between males and females in the side-to-side
comparison in any of the tests. Patients can therefore
be evaluated on an individual basis, regardless of
gender. This is also in accordance with previous studies
[1, 4, 20].
An LSI of both 85 and 90% has been advocated to
assess whether a hop test is normal or abnormal [1, 3, 8,
10, 20, 21 ]. Fitzgerald et al. [3] described a decision-
making scheme for returning patients with an ACL in-
jury to a high level of physical activity. Patients suc-
cessfully returning to pre-injury levels of activity had a
mean hop test score of 95%, compared with the mean of
85% in the patients who failed rehabilitation. On the
basis of this, an LSI of 90% was chosen as the cut-off
score in this study.
786
The square hop test used by O
¨
stenberg et al. [21] was
modified in the present study. In order to force the pa-
tients to perform more demanding jumps, the volume of
the square was increased. Although good reliability
(ICC 0.85) was obtained in the square hop test, this test
was not able to discriminate healthy subjects from pa-
tients. It was therefore concl uded that the test was not
optimal when it came to evaluating patients with an
ACL injury or patients who had undergone ACL
reconstruction. A new test, the drop jump followed by a
double hop for distance, was designed for additional
demands. This test is characterised by high intensity with
regard to deceleration and acceleration forces. Even
though the intention was to design a new, more
demanding hop test, this was not supported by the re-
sults. The drop jump followed by a double hop for
distance showed a strong correlation with the more
established hop for distance test. The sensitivity and
accuracy were not in favour of the newly designed test.
To summarise, a test battery consisting of the vertical
jump, the hop for distance and the side hop produced
the most information. It could also be regarded as
advantageous in the clinical setting to use a test battery
comprising three tests, instead of four or five tests.
It is worrying to note that only one out of ten patients
were classified as normal (i.e. having >90% hop
performance on the injured side com pared with the
uninjured side) 11 months after an ACL injury and
6 months after ACL reconstruction, as it is now
customary to allow a return to full sports activities
6 months after ACL reconstruction [33, 34], with some
authors advocating a return to sports as early as
4 months post-operatively [35, 36].
Conclusion
The test battery consisting of both maximum single hop
performances: the vertical jump and the hop for distance
and hop performance while developing fatigue: the side
hop, produced high test–retest reliabili ty, sensitivity and
accuracy. Further, the test battery produced higher
values compared with any of the three hop tests indi-
vidually, revealing that only one out of ten patients had
restored hop performance 11 months after an ACL in-
jury and 6 months after ACL reconstruction. It is con-
cluded that this test battery showed a high ability to
discriminate between the hop performance of the injured
and the uninjured side both in pa tients with an ACL
injury an d in patients who have undergone ACL
reconstruction.
Acknowledgments This study was supported by grants from the
Swedish Centre for Research in Sports and Research and The
Local Research and Development Council for Gothenburg and
Southern Bohusla
¨
n Sweden.
References
1. Barber SD, Noyes FR, Mangine RE,
McCloskey JW, Hartman W (1990)
Quantitative assessment of functional
limitations in normal and anterior cru-
ciate ligament-deficient knees. Clin
Orthop 255:204–214
2. Eastlack ME, Axe MJ, Snyder-Mackler
L (1999) Laxity, instability, and func-
tional outcome after ACL injury: copers
versus noncopers. Med Sci Sports Exerc
31:210–215
3. Fitzgerald GK, Axe MJ,
Snyder-Mackler L (2000) A decision-
making scheme for returning patients to
high-level activity with nonoperative
treatment after cruciate ligament
rupture. Knee Surg Sports Traumatol
Arthrosc 8:76–82
4. Itoh H, Kurosaka M, Yoshiya S,
Ichihashi N, Mizuno K (1998) Evalua-
tion of functional deficits determined by
four different hop tests in patients with
anterior cruciate ligament deficiency.
Knee Surg Sports Traumatol Arthrosc
6:241–245
5. Noyes FR, Barber SD, Mangine RE
(1991) Abnormal lower limb symmetry
determined by function hop tests after
anterior cruciate ligament rupture. Am
J Sports Med 19:513–518
6. Rudolph KS, Axe MJ, Snyder-Mackler
L (2000) Dynamic stability after ACL
injury: who can hop? Knee Surg Sports
Traumatol Arthrosc 8:262–269
7. Tegner Y, Lysholm J, Lysholm M,
Gillquist J (1986) A performance test to
monitor rehabilitation and evaluate
anterior cruciate ligament injuries. Am J
Sports Med 14:156–159
8. Augustsson J (2004) Ability of a new
hop test to determine functional deficits
after anterior cruciate ligament recon-
struction. Knee Surg Sports Traumatol
Arthrosc 12:350–356
9. Jerre R, Ejerhed L, Wallmon A, Kartus
J, Brandsson S, Karlsson J (2001)
Functional outcome of anterior cruciate
ligament reconstruction in recreational
and competitive athletes. Scand J Med
Sci Sports 11:342–346
10. Juris PM, Phillips EM, Dalpe C,
Edwards C, Gotlin RS, Kane DJ (1997)
A dynamic test of lower extremity
function following anterior cruciate lig-
ament reconstruction and rehabilita-
tion. J Orthop Sports Phys Ther 26:184–
191
11. Keays SL, Bullock-Saxton J, Keays AC,
Newcombe P (2001) Muscle strength
and function before and after anterior
cruciate ligament reconstruction using
semitendinosus and gracilis. The Knee
8:229–234
12. Liu-Ambrose T, Taunton JE,
MacIntyre D, McConkey P, Khan KM
(2003) The effects of proprioceptive or
strength training on the neuromuscular
function of the ACL reconstructed
knee: a randomized clinical trail. Scand
J Med Sci Sports 13:115–123
787
13. Pantano KJ, Irrgang JJ, Burdett R,
Delitto A, Harner C, Fu FH (2000) A
pilot study on the relationship between
physical impairment and activity
restriction in persons with anterior cru-
ciate ligament reconstruction at long-
term follow-up. Knee Surg Sports
Traumatol Arthrosc 9:369–378
14. Paterno MV, Greenberger HB (1996)
The test–retest reliability of a one legged
hop for distance in young adults with
and without ACL reconstruction. Iso-
kinet Exerc Sci 6:1–6
15. Petschnig R, Baron R, Albrecht M
(1998) The relationship between isoki-
netic quadriceps strength test and hop
tests for distance and one-legged verti-
cal jump test following anterior cruciate
ligament reconstruction. J Orthop
Sports Phys Ther 28:23–31
16. Risberg MA, Ekeland A (1994) Assess-
ment of functional tests after anterior
cruciate ligament surgery. J Orthop
Sports Phys Ther 19:212–217
17. Ross MD, Irrgang JJ, Denegar CR,
McCloy CM, Unangst ET (2002) The
relationship between participation
restrictions and selected clinical mea-
sures following anterior cruciate liga-
ment reconstruction. Knee Surg Sports
Traumatol Arthrosc 10:10–19
18. Sernert N, Kartus J, Ko
¨
hler K, Stener
S, Larsson J, Eriksson B, Karlsson J
(1999) Analysis of subjective, objective
and functional examination tests after
anterior cruciate ligament reconstruc-
tion. A follow up of 527 patients. Knee
Surg Sports Traumatol Arthrosc 7:160–
165
19. Wilk KE, Romaniello WT, Soscia SM,
Arrigo CA, Andrews JR (1994) The
relationship between subjective knee
scores, isokinetic testing, and functional
testing in the ACL-reconstructed knee. J
Orthop Sports Phys Ther 20:60–71
20. Risberg MA, Holm I, Ekeland A (1995)
Reliability of functional knee tests in
normal athletes. Scand J Med Sci Sports
5:24–28
21. O
¨
stenberg A, Roos E, Ekdahl C, Roos
H (1998) Isokinetic knee extensor
strength and functional performance in
healthy female soccer players. Scand J
Med Sci Sports 8:257–264
22. Dugan SA, Frontera WR (2000) Muscle
fatigue and muscle injury. Phys Med
Rehabil Clin N Am 11:385–403
23. Feagin JA Jr, Lambert KL, Cunning-
ham RR, Anderson LM, Riegel J, King
PH, Van Genderen L (1987) Consider-
ation of the anterior cruciate ligament
injury in skiing. Clin Orthop 216:13–18
24. O
¨
stenberg A, Roos H (2000) Injury risk
factors in female European football. A
prospective study of 123 players during
one season. Scand J Med Sci Sports
10:279–285
25. Tegner Y, Lysholm J (1985) Rating
system in the evaluation of knee liga-
ment injuries. Clin Orthop 198:43–49
26. Grimby G (1986) Physical activity and
muscle training in the elderly. Acta Med
Scand 711(Suppl.):233–237
27. Shrout PE, Fleiss JL (1979) Intraclass
Correlations: uses in assessing rater
reliability. Psychol Bull 86:420–428
28. Dahlberg G (1940) Statistical methods
for medical and biological students.
Allen and Unwin, London
29. Altman DG (1991) Practical statistics
for medical research. Chapman & Hall,
London, pp 410–411
30. Portney LG, Watkins MP (2000)
Foundations of clinical research: appli-
cations to practice, 2nd edn. Prentice-
Hall, New Jersey, pp 494–495
31. Wilson GJ, Newton RU, Murphy AJ,
Humphries BJ (1993) The optimal
training load for the development of
athletic performance. Med Sci Sports
Exerc 11:1279–1286
32. Manske RC, Smith B, Wyatt F (2003)
Test–retest reliability of lower extremity
functional tests after a closed kinetic
chain isokinetic testing bout. J Sports
Rehabil 12:119–132
33. MacDonald PB, Hedden D, Pacin O,
Huebert D (1995) Effects of an acceler-
ated rehabilitation program after ante-
rior cruciate ligament reconstruction
with combined semitendinosus-gracilis
autograft and a ligament augmentation
device. Am J Sports Med 23:588–592
34. Shelbourne KD, Nitz PA (1991) The
O’Donoghue triad revisited. Combined
knee injuries involving anterior cruciate
and medial collateral ligament tears.
Am J Sports Med 19:474–477
35. De Carlo M, Shelbourne KD, Oneacre
K (1999) Rehabilitation program for
both knees when the contralateral
autogenous patellar tendon graft is used
for primary anterior cruciate ligament
reconstruction: a case study. J Orthop
Sports Phys Ther 29:144–153
36. Howell S, Taylor M (1996) Brace-free
rehabilitation, with early return to
activity, for knees reconstructed with a
double-looped semitendinosus and
gracilis graft. J Bone Joint Surg Am
78:814–825
788
