614 THE JOURNAL OF BONE AND JOINT SURGERY
K. R. Grob, MD
M. S. Kuster, MD, PD, PhD
Department of Orthopaedic Surgery, Kantonsspital, CH 9001 St Gallen,
S. A. Higgins
D. G. Lloyd, PhD
Department of Human Movement and Exercise Science, The University of
Western Australia, Perth, Australia.
Department of Human Sciences, Wako University, Tokyo, Japan.
Correspondence should be sent to Dr K. R. Grob.
©2002 British Editorial Society of Bone and Joint Surgery
Lack of correlation between different
measurements of proprioception in the knee
K. R. Grob, M. S. Kuster, S. A. Higgins, D. G. Lloyd, H. Yata
From Kantonsspital, St. Gallen, Switzerland
urrent methods of measurement of proprioceptive
function depend on the ability to detect passive
movement (kinaesthesia) or the awareness of joint
position (joint position sense, JPS). However, reports
of proprioceptive function in healthy and pathological
joints are quite variable, which may be due to the
different methods used. We have compared the
validity of several frequently used methods to quantify
Thirty healthy subjects aged between 24 and 72
years underwent ﬁve established tests of
proprioception. Two tests were used for the
measurement of kinaesthesia (KT1 and KT2). Three
tests were used for the measurement of JPS, a passive
reproduction test (JPS1), a relative reproduction test
(JPS2) and a visual estimation test (JPS3).
There was no correlation between the tests for
kinaesthesia and JPS or between the different JPS
tests. There was, however, a signiﬁcant correlation
between the tests for kinaesthesia (r = 0.86). We
conclude therefore that a subject with a given result in
one test will not automatically obtain a similar result
in another test for proprioception. Since they describe
different functional proprioceptive attributes,
proprioceptive ability cannot be inferred from
independent tests of either kinaesthesia or JPS.
J Bone Joint Surg [Br] 2002;84-B:614-8.
Received 6 April 2000; Accepted after revision 4 April 2001
Proprioception is the sum of kinaesthesia and joint position
sense. Kinaesthesia is deﬁned as the awareness of joint
movement and is dynamic. Joint position sense (JPS) is
restricted to the awareness of the position of a joint in space
and is a static phenomenon. Proprioception can also be
deﬁned as the cumulative neural input to the central nerv-
ous system from specialised nerve endings called mechano-
receptors. These are located in the joint capsules,
ligaments, muscles, tendons, and skin. Some of these recep-
tors (for example, Pacinian corpuscles) are stimulated in
the initial and terminal stages of the range of movement of
joints as well as during rapid changes in velocity and
direction (kinaesthesia). On the other hand the Rufﬁni end
organ-like receptors and Golgi tendon organ-like receptors
have been associated with a response to the relative posi-
tion of muscles and joints (joint position sense). However,
in the literature the terms kinaesthesia, joint position sense
and proprioception are often used synonymously.
Depending on the type of proprioceptive test used, dif-
ferent results have been observed in the same subject
groups. For example, Barrett
showed that JPS was sig-
niﬁcantly improved by reconstruction of the cruciate liga-
ments. MacDonald et al,
however, found that in patients
with reconstruction of the anterior cruciate ligament (ACL)
proprioception did not improve when measured by a kin-
aesthesia test. Barrett, Cobb and Bentley
improvement in JPS after total knee replacement while
Skinner et al
were unable to ﬁnd any such improvement
using a different JPS test and a kinaesthesia-based test.
In a series of studies Skinner et al
used the same testing
device for a healthy control group and the results differed
by as much as 100%.
Various authors have made deﬁnite
conclusions about overall proprioception using only either
JPS tests or kinaesthesia tests.
Because of these
contradictory results, the question arises as to whether
overall proprioceptive ability can be ascertained by a single
method. Our aim therefore was to compare frequently used
tests of proprioception and to evaluate the correlation
Patients and Methods
Thirty healthy volunteers (12 women, 18 men) with a mean
age of 41 years (
13.5; range 24 to 72) participated in the
study. None had a history of injury to the lower limb or
vestibular or neuromuscular disorders. The subjects were
tested using their dominant leg for ﬁve different tests of
In all ﬁve tests visual clues were eliminated using a
blindfold. For JPS1 and KT1 the subject’s leg was allowed
to hang freely over the side of the table at a distance of 5 to
10 cm proximal to the popliteal fossa. Soft-cast compres-
sion splints (soft scotch-cast; 3M Healthcare, St Paul,
Minnesota) were ﬁtted above and below the knee. Elastic
bandages were applied over the soft scotch-cast as tightly
as possible to neutralise cutaneous sensation (Fig. 1). For
JPS2, JPS3 and KT2 the subject’s leg was ﬁtted in an
adjustable splint (Fig. 2). To neutralise the slight vibration
615LACK OF CORRELATION BETWEEN DIFFERENT MEASUREMENTS OF PROPRIOCEPTION IN THE KNEE
VOL. 84-B, N
. 4, MAY 2002
Photograph of the testing device used for JPS1 and KT1: (1) motor; (2) motor control box; (3) vibrator; (4)
compression splint; (5) wire; (6) on-off switch of the control box.
Photograph of the testing device used for JPS1, JPS3 and KT2: (1) motor; (2) motor control box; (3) vibrator;
(4) adjustable splint; (5) wire; (6) electrogoniometer; (7) on-off switch of the control box.
created by the testing motor a vibration device was ﬁtted on
the splint for tests JPS1, KT1 and KT2. In a pilot study we
also found that patients very often could tell whether
movement occurred by the slight vibration of the starting
stepper motor rather than the movement itself. In order to
eliminate this clue the vibrator was used which is a recom-
mended procedure for measurement of proprioception.
Furthermore, the present results compare well with those in
the literature and we therefore can conﬁdently conclude
that the slight vibrations at the splint (and not at the skin)
did not signiﬁcantly affect the results and conclusions.
Additionally, in KT1 and KT2 tests the auditory clues were
eliminated by using industrial earmuffs with music
Joint position sense test 1 (JPS1). This test measured the
ability to reproduce passively the position of the lower leg
using a slow-speed motor (Proprioception 2000; Automated
Motion Systems, Perth, Western Australia) for reposition-
ing (Fig. 1).
Starting at a free-hanging position of 90°,
the motor moved the subject’s leg at a rate of 12°/s to three
randomly selected angles between 60° and 80°, 30° and 55°
and 5° and 25° of ﬂexion. The leg was held in this position
for a few seconds and the subject was asked to concentrate
on its position. The knee was returned to the starting
position and then moved again by the motor at a speed of
5°/s. When the subject thought that the leg was in the same
position as before, he or she stopped the motor by using the
on-off switch of the motor control box. The absolute angu-
lar error was measured. This procedure was repeated nine
times, three times for each angle, and a mean value of
angular error was recorded for each subject.
Joint position sense test 2 (JPS2). This test measured
the ability of a subject to reproduce the knee angle using
a hand-held electrogoniometer as a visual analogue mod-
Starting at a free-hanging position of
90°, the examiner moved the dominant leg by raising the
splint to angles of 17°, 34°, 50° and 67° of flexion. An
electrogoniometer was attached to the splint and recorded
the position of the leg to an accuracy of 0.05°. The
subject was instructed to concentrate on the location of
the leg while maintaining the position for a few seconds.
The leg was then returned to its original position. The
subject then represented the perceived angle of flexion by
the hand on a visual analogue electrogoniometer (Fig. 3).
This procedure was repeated 12 times, three times for
each angle, and the mean angular error was recorded for
Joint position sense test 3 (JPS3). This test measured the
reproduction of a knee angle using the contralateral leg as
The subject’s splinted leg was moved to four
angles of 17°, 34°, 50° and 67° of ﬂexion. After a few
seconds the leg was returned to its original position. The
subject then was asked to place the contralateral leg in the
same position as the tested leg before. An electrogoni-
ometer was attached to both legs to measure the angular
error. This procedure was repeated 12 times, three times for
each angle, and the mean angular error was recorded for
Kinaesthesia test 1 (KT1). The threshold of detection of a
passive leg movement was measured.
wound up by the slow-speed motor, was attached to the
compression splint and moved the subject’s leg (Fig. 1).
From a starting position of 60° of knee ﬂexion, and with
the tension to counter gravity already applied, the serv-
ometer slowly pulled the subject’s leg into extension at
0.5°/s. The onset of the servomotor movement had a ran-
dom delay which varied between 5 and 60 s after the
subjects were prepared for the task. The subject was given
a control box with an on-off switch to stop the motor when
a change was perceived in the position of the tested leg.
The linear movement of the wire was measured and con-
verted to angular deﬂection.
This procedure was repeated
ten times and the mean threshold angle to detect passive
movement was recorded for each subject.
Kinaesthesia test (KT2). This test was similar to KT1 in
that the awareness of movement of the joint was measured.
In contrast to KT1, the subject’s leg was ﬁtted into an
adjustable and not a compression splint (Fig. 2).
tionally, an electrogoniometer was attached to the splint
and recorded the position of the leg with an accuracy of
0.05°. This procedure was repeated ten times and the mean
threshold angle was recorded for each subject.
Statistical analysis. Descriptive statistics (mean and stand-
ard deviation) were used to determine the performance of
all subjects in each proprioceptive test. Pearson correlation
tests were used to determine the relationships between the
measures of proprioception. In addition, subjects were
ranked according to their performance in each of the tests
and rank-correlation coefﬁcients of the different tests were
616 K. R. GROB, M. S. KUSTER, S. A. HIGGINS, D. G. LLOYD, H. YATA
THE JOURNAL OF BONE AND JOINT SURGERY
Hand-held electrogoniometer used for
The conﬁdence intervals, the mean and the maximum and
minimum angular error for the JPS tests and for threshold
angles to detect passive movement are given in Table I. The
mean threshold angles from KT1 and KT2 were signiﬁ-
cantly different (p = 0.01).
Table II presents the correlation between the different
proprioception tests. There was no correlation between the
different tests for JPS or between the kinaesthesia and the
JPS tests, but there was a signiﬁcant correlation (r = 0.86)
between the two tests for kinaesthesia, KT1 and KT2. The
subjects were then ranked according to their performance in
each test and the correlation between the ranks of the
different tests established (Table III). The only signiﬁcant
correlation of the ranks again was between the two kinaes-
The absolute values for kinaesthesia and JPS tests in our
study were similar to those previously reported. For exam-
ple, in the JPS test which used visual estimation (JPS2),
recorded a mean angular error of 2.5° and Warren
10.6°; our value was 9.7°. For KT1 our value was
2.07° and for KT2 1.38° which is within the reported range
There are various tests for kinaesthesia and JPS in the
literature all of which have conclusions regarding the over-
all proprioceptive ability.
Our main ﬁndings
were that there is no signiﬁcant correlation between the
kinaesthesia and JPS tests, or between the different JPS
tests. This suggests that there is no single test which
quantiﬁes proprioception. Instead, each assesses one facet
of proprioception only and does not represent the overall
The lack of correlation between the kinaesthesia and the
JPS tests agrees with a study by Fridén et al.
evaluated proprioception in 16 patients at different times
after acute injury to a knee ligament by using two tests for
JPS and one for kinaesthesia. Loss of proprioception was
found in the kinaesthesia test but not in the JPS tests. This
lack of correlation may also explain several contradictory
results in the current literature. For example, Barrett
that proprioception was signiﬁcantly improved by recon-
struction of the cruciate ligaments after using a joint posi-
tion visual estimation test, but MacDonald et al
such improvement using a kinaesthesia test. Skinner et al
measured proprioception in a group of patients before and
after total knee replacement using a JPS and a kinaesthesia
test. They were unable to ﬁnd a difference in propriocep-
tion between knees with osteoarthritis and those with a
prosthesis. On the other hand, Barrett et al
used a joint
position visual estimation test and found that knee replace-
ment improved joint proprioception.
No conclusions can be drawn as to which test is the most
sensitive for detecting possible changes in proprioception.
Some investigators prefer the threshold tests to the JPS
tests because previous results for kinaesthesia measure-
ments are more constant than for JPS tests.
a loss of proprioception after injury to the ACL was found
in all six studies which used kinaesthesia tests,
but in only three of eight studies employing JPS
Likewise, our study showed a good correla-
tion only between the two threshold tests. This could be
because the difference in the testing device between these
two tests was minimal.
Nevertheless, even the small methodological change
such as using a tightly-ﬁtted compression splint (KT1) to
diminish cutaneous sensory information signiﬁcantly inﬂu-
enced the threshold angle. Hence, information from other
sensory pathways may explain contradictory ﬁndings such
as a reported threshold angle of 2.4° for elderly patients
after knee arthroplasty
and 2.7° for young dancers.
when apparently the same tests were used by different
authors opposing results may be explained by minor meth-
odological changes. For example, Barrack et al
signiﬁcantly higher kinaesthesia threshold angle in subjects
with injuries to the ACL while Wright et al
difference after ACL injury using the ‘same’ proprioceptive
Proprioception depends on the cumulative neural input to
the central nervous system from mechanoreceptors
617LACK OF CORRELATION BETWEEN DIFFERENT MEASUREMENTS OF PROPRIOCEPTION IN THE KNEE
VOL. 84-B, N
. 4, MAY 2002
Table I. Summary of the results (degrees) for the JPS and kinaesthesia
JPS1 JPS2 JPS3 KT1 KT2
Number of cases 30 30 30 30 30
Minimum 4.30 4.10 3.30 1.08 0.78
Maximum 24.44 25.30 26.50 3.72 2.58
Mean 13.75 9.27 10.71 2.07 1.38
95% conﬁdence interval (upper) 15.71 10.82 12.72 2.33 1.54
95% conﬁdence interval (lower) 11.79 7.73 8.70 1.80 1.23
Standard deviation 5.24 4.14 5.37 0.71 0.41
Table II. Intercorrelation (r) between all the tests
JPS1 JPS2 JPS3 KT1 KT2
JPS2 -0.20 1
JPS3 0.16 0.49 1
KT1 0.18 0.13 0.30 1
KT2 0.07 0.24 0.32 0.86 1
Table III. Intercorrelation (r) between the ranks of the
JPS1 JPS2 JPS3 KT1 KT2
JPS2 -0.25 1
JPS3 0.30 0.20 1
KT1 0.20 0.14 0.33 1
KT2 0.13 0.18 0.31 0.87 1
located in the muscles, ligaments, joint capsules, tendons
and skin. Most JPS tests also involve efferent pathways.
The lack of correlation between different JPS tests may be
due to some of the ﬁnal measurements being dependent
upon both the function of the knee being tested and also on
the sensory input from the other leg (JPS3) and the visual-
cognitive-spatial ability (JPS2). The JPS1 was the only JPS
test which was independent of secondary sensory inputs,
suggesting that JPS1 may be the most objective JPS test.
Finally, we only used healthy subjects. Most researchers
use these tests to distinguish between normal and patho-
logical conditions. Our conclusions may not apply to patho-
logical conditions and this needs further investigation.
Our study and others have assessed the conscious per-
ception of joint position and/or movement. However, the
symptom of instability suffered by patients after joint injury
does not reﬂect the conscious perception of where the joint
is in space, but the reﬂex control of its activity by its
Therefore the measurement of JPS or
kinaesthesia is not necessarily the right approach to reveal
a neurological defect after joint injury. Tests of muscle co-
ordination such as standing balance may be of more value
in assessing instability after ligament injury.
We conclude that there remains no comprehensive meth-
od for measuring proprioception. The results of studies
which use only either JPS tests or kinaesthesia tests must be
interpreted with care. Furthermore, the terms propriocep-
tion, kinaesthesia and JPS should not be used
No beneﬁts in any form have been received or will be received from a
commercial party related directly or indirctly to the subject of this
1. Voight ML, Hardin JA, Blackburn TA, Tippett S, Canner GC. The
effects of muscle fatigue on the relationship of arm dominance to
shoulder proprioception. J Orthop Sports Phys Ther 1996;23:348-52.
2. Barrett DS. Proprioception and function after anterior cruciate recon-
struction. J Bone Joint Surg [Br] 1991;73-B:833-7.
3. MacDonald PB, Hedden D, Pacin O, Sutherland K. Proprioception
in anterior cruciate ligament-deﬁcient and reconstructed knees. Am J
Sports Med 1996;24:774-8.
4. Barrett DS, Cobb AG, Bentley G. Joint proprioception in normal,
osteoarthritic and replaced knees. J Bone Joint Surg [Br]
5. Skinner HB, Barrack RL, Cook SD, Haddad RJ Jr. Joint position
sense in total knee arthroplasty. J Orthop Res 1984;1:276-83.
6. Barrack RL, Skinner HB, Brunet ME, Cook SD. Joint kinesthesia
in the highly trained knee. J Sports Med Phys 1984;24:18-20.
7. Barrack RL, Skinner HB, Buckley SL. Proprioception in the
anterior cruciate deﬁcient knee. Am J Sports Med 1989;17:1-6.
8. Co FH, Skinner HB, Cannon WD. Effect of reconstruction of the
anterior cruciate ligament on proprioception of the knee and the heel
strike transient. J Orthop Res 1993;11:696-704.
9. Skinner HB, Wyatt MP, Hodgdon JA, Conrad DW, Barrack RL.
Effect of fatigue on joint position sense of the knee. J Orthop Res
10. Attﬁeld SF, Wilton TJ, Pratt DJ, Sambatakakis A. Soft-tissue
balance and recovery of proprioception after total knee replacement. J
Bone Joint Surg [Br] 1996;78-B:540-5.
11. Cash RM, Gonzalez MH, Garst J, Barmada R, Stern SH. Proprio-
ception after arthroplasty: role of the posterior cruciate ligament. Clin
12. Clark P, MacDonald PB, Sutherland K. Analysis of proprioception
in the posterior cruciate ligament-deﬁcient knee. Knee Surg Sports
Traumatol Arthrosc 1996;4:225-7.
13. Jerosch J, Prymka M. Knee joint proprioception in patients with
posttraumatic recurrent patella dislocation. Knee Surg Sports Trauma-
tol Arthrosc 1996;4:14-8.
14. Jerosch J, Prymka M, Castro WHM. Proprioception of knee joints
with a lesion of the medial meniscus. Acta Orthop Belg
15. Jerosch J, Schmidt K, Prymka M. Propriceptive Fëhigkeit bei
Patienten mit retropatellarem Knieschmerz unter besonderer Ber-
ücksichtigung der Beeinﬂussung durch eine elastische Kniebandage.
16. Lephart SM, Giraldo JL, Borsa PA, Fu FH. Knee joint propriocep-
tion: a comparison between female intercollegiate gymnasts and con-
trols. Knee Surg Sports Traumatol Arthrosc 1996;4:121-4.
17. Petrella RJ, Lattanzio PJ, Nelson MG. Effect of age and activity on
knee joint proprioception. Am J Phys Med Rehabil 1997;76:235-41.
18. Sharma L, Pai Y-C, Holtkamp K, Rymer WZ. Is knee joint
proprioception worse in the arthritic knee versus the unaffected knee
in unilateral knee osteoarthritis? Arthritis Rheum 1997;40:1518-25.
19. Simmons S, Lephart S, Rubash H, Pifer GW, Barrack R. Proprio-
ception after unicondylar knee arthroplasty versus total knee arthro-
plasty. Clin Orthop 1996;331:179-84.
20. Warren PJ, Olanlokun TK, Cobb AG, Bentley G. Proprioception
after knee arthroplasty: the inﬂuence of prosthetic design. Clin Orthop
21. Carter ND, Jenkinson TR, Wilson D, Jones DW, Torode AS. Joint
position sense and rehabilitation in the anterior cruciate ligament
deﬁcient knee. Br J Sports Med 1997;31:209-12.
22. Jerosch J, Prymka M. Knee joint proprioception in normal volun-
teers and patients with anterior cruciate ligament tears, taking special
account of the effect of a knee bandage. Arch Orthop Trauma Surg
23. Kaplan FS, Nixon JE, Reitz M, Rindﬂeish L, Tucker J. Age-related
changes in proprioception and sensation of joint position. Acta Orthop
24. Corrigan JP, Cashman WF, Brady MP. Proprioception in the
cruciate deﬁcient knee. J Bone Joint Surg [Br] 1992;74-B:247-50.
25. Lephart SM, Kocher MS, Fu FH, Borsa PA, Harner CD. Proprio-
ception following anterior cruciate ligament reconstruction. J Sports
26. Perlau R, Frank C, Fick G. The effect of elastic bandages on human
knee proprioception in the uninjured population. Am J Sports Med
27. Fridén T, Roberts D, Zätterström R, Lindstrand A, Moritz U.
Proprioception after an acute knee ligament injury: a longitudinal
study on 16 consecutive patients. J Orthop Res 1997;15:637-44.
28. Fridén T, Roberts D, Zätterström R, Lindstrand A, Moritz U.
Proprioception in the nearly extended knee: measurements of position
and movement in healthy individuals and in symptomatic anterior
cruciate ligament injured patients. Knee Surg Sports Traumatol
29. Harter RA, Osternig LR, Singer KM, et al. Long term evaluation of
knee stability and function following surgical reconstruction for anter-
ior cruciate ligament insufﬁciency. Am J Sports Med
30. Wright SA, Tearse DS, Brand RA, Gabel RH. Proprioception in the
anteriorly unstable knee. Iowa Orthop J 1995;15:156-61.
31. Lephart SM, Pincivero DM, Giraldo JL, Fu FH. The role of
proprioception in the management and rehabilitation of athletic inju-
ries. Am J Sports Med 1997;25:130-7.
32. Beard DJ, Kyberd PJ, Fergusson CM, Dodd CAF. Proprioception
after rupture of the anterior cruciate ligament: an objective indication
of the need for surgery? J Bone Joint Surg [Br] 1993;75-B:311-5.
33. Walla DJ, Albright JP, McAuley E, et al. Hamstring control and the
unstable anterior cruciate ligament deﬁcient knee. Am J Sports Med
34. Kuster MS, Grob KR, Kuster M, Wood GA, Gächter A. The
beneﬁts of wearing a compression sleeve after ACL reconstruction.
Med Sci Sports Exerc 1999;31:368-71.
618 K. R. GROB, M. S. KUSTER, S. A. HIGGINS, D. G. LLOYD, H. YATA
THE JOURNAL OF BONE AND JOINT SURGERY