Medical Engineering & Physics 32 (2010) 760–765

Contents lists available at ScienceDirect

Medical Engineering & Physics

journal homepage: www.elsevier.com/locate/medengphy

The patellar tendon bar! Is it a necessary feature?

N.A. Abu Osman a,∗ , W.D. Spence b , S.E. Solomonidis b , J.P. Paul b , A.M. Weir c
a
Department of Biomedical Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia
b
Bioengineering Unit, University of Strathclyde, Glasgow, UK
c
Westmarc, Southern General Hospital, Glasgow, UK

a r t i c l e i n f o a b s t r a c t

Article history: The purpose of this investigation was to vary the load on the patellar tendon bar and to study the subse-
Received 14 September 2009 quent effect this has on the pattern of the pressure distribution at the stump–socket interface. Ten male
Received in revised form 27 April 2010 subjects from the Southern General Hospital in Glasgow, UK participated in this study. Measuring sys-
Accepted 29 April 2010
tems utilising strain gauge and electrohydraulic technologies were designed, developed and constructed
to enable pressure measurements to be conducted. One transducer, the patellar tendon (PT) transducer,
Keywords:
was attached to the patellar tendon bar of the socket such that the patellar tendon bar was capable of
Trans-tibial
being translated by ±10 mm towards or away from the tendon. The results of this study showed that the
Pressure distribution
Patellar tendon bar
position of the patellar tendon bar had no significant effect on the pressure distribution around the socket
Stump indicating that it is an unnecessary feature, which, we propose, may be eliminated during manufacture
Pressure of a trans-tibial socket.
Patellar tendon bearing socket © 2010 IPEM. Published by Elsevier Ltd. All rights reserved.

1. Introduction to each subject and must be carefully assessed and implemented.

The interface pressure distribution between the stump tissues
of a trans-tibial amputee and the socket determines the comfort
of the subject, and therefore must be considered in the design of
the socket. An understanding of the residual limb anatomy and
the biomechanical principles involved resulted in the development
of the popular Patellar tendon bearing (PTB) trans-tibial socket
[1].
Most trans-tibial sockets are hand-cast and custom fabricated
on an individual basis by a prosthetist. The main steps during fab-
rication of PTB sockets are (a) obtaining the negative mould of the
stump, (b) producing the positive mould of the stump, and recti-
fying the cast, (c) socket forming and (d) socket finishing. During
fabrication of the PTB socket steps (a) and (b) do not duplicate
the stump shape due to the rectification involved (Fig. 1). Further
details and manufacturing procedures for the PTB socket can be
found in Radcliffe and Foort [1].
The concept of loading the “pressure tolerant” areas and reliev-
ing load from the “pressure sensitive” areas was based on logical
biomechanical principles. Whilst many amputees have successfully
worn the PTB socket over the decades, a high degree of skill is nec-
essary to take a wrap cast and rectify the positive mould in order
to produce a satisfactory socket. Also, the amount of pressure a
subject can tolerate in any one localised area is highly subjective
∗ Corresponding author. Tel.: +60 3 79674581; fax: +60 3 79674579.
E-mail address: azuan@um.edu.my (N.A. Abu Osman).
It has been proposed that a convenient area to transfer some of
the load between the stump and the socket is the patellar tendon
area, which can withstand relatively high loads without caus-
ing pain or tissue damage [2]. This group fitted an experimental
prosthesis with a two-directional load cell at the patellar ten-
don indent (which was cut out of the hard shell), attached to an
adjustable mounting part of a frame. The system was anchored
distally to the modular shank of the prosthesis [2]; suggested that
7 mm indentation was the optimum position. These authors car-
ried out gait analysis tests on two bellow knee amputees. They then
performed a theoretical analysis of their results and using an opti-
mization technique claimed that the optimum indentation is 7 mm
beyond the position established by the prosthetist during casting
of the stump. However, the methodology used in that publication
is not convincing and no proof of the optimization procedure was
given.
In our study we did not exceed an indentation of more than
4 mm from the original position established by the prosthetist dur-
ing casting, as most subjects that participated felt uncomfortable,
beyond 4 mm.
Kristinsson [3] described the use of the Icelandic Roll on Silicone
Socket (ICEROSS) (or silicone liner). During the casting process,
the subject’s limb is held in full extension with the subject seated.
Build-ups of silicone padding are added over the bony areas of the
stump, such as the anterior-distal aspect of the tibia and the fibu-
lar head. The silicone liner is rolled onto the stump and a plaster
wrap cast is applied without applying localised pressure. Placing
a pressure container over the stump allows air to be pressurised

1350-4533/$ – see front matter © 2010 IPEM. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.medengphy.2010.04.020
 N.A. Abu Osman et al. / Medical Engineering & Physics 32 (2010) 760–765 761

Fig. 2. Patellar tendon transducer mounted on a moveable patellar tendon bar of a

trans-tibial socket.

2. Methods

Fig. 1. PTB negative mould; finger pressure being applied to the plaster wrap to 2.1. Subjects
produce a “pre-compressed” state in the tissues of the patellar tendon.

Ten trans-tibial amputees participated in this study on a vol-

and thus the cast is formed to the stump. He claimed that in most
cases, the traditional concept of load transfer to areas such as the
patellar tendon, medial flare and condyles of the tibia, is ineffec-
tive and uncomfortable. A number of investigators [4–13] have
shown that the pressures within the PTB socket are, as expected,
not uniformly distributed. In general, peak pressures taken from
walking subjects yield values of around 300 kPa recorded at the
site of the patellar tendon bar. In a review paper by Mak et al. [14]
the maximum peak pressure reported on the patellar tendon bar
was stated to reach about 400 kPa [6], however the measurements
conducted in the last 15 years showed that the maximum interface
pressure for PTB sockets during walking was usually below 220 kPa
[15,16].
To date, analyses of a variety of parameters, either using
experimental measurements or finite element studies, have been
conducted on the socket geometry, liner material and ground reac-
tion force direction [10,17,18]. However no published report has
been found which shows the effect of varying the load (through
the depth of indentation) on the patellar tendon bar on the pattern
of pressure distribution at the stump–socket interface. As far as the
authors of this paper are aware, only one early report, Mizrahi et
al. [2], published data on the relationship between the geometry of
the PT insertion and gait pattern.
The current study set out to investigate the effect varying the
load on the patellar tendon bar has on the pattern of pressure
distribution at the stump–socket interface. The main objectives
of the study are twofold, firstly, is there any correlation between
varying the load on the patellar tendon and the pressure distribu-
tion at other sites in the socket? Secondly, is the patellar tendon
bar rectification really needed in the production of a trans-tibial
socket?
untary basis, and gave written informed consent. All were male
and had undergone amputation at least 5 years prior to this study.
Ethical approval was granted from the Ethical committees of the
University of Strathclyde and the Southern General Hospital NHS
Trust UK. Detailed subject information is given in Table 1.
2.2. Transducers
Three types of force/pressure transducers were used in this
study, the Patellar Tendon (PT) transducer, Entran® -based trans-
ducers and an electrohydraulic transducer.
2.3. Patellar tendon (PT) transducer
At the Department of Bioengineering, University of Strath-
clyde, a specially designed and manufactured Patellar Tendon
(PT) transducer was developed which allowed the patellar bar of
the prosthesis (which was trepanned out from the socket) to be
translated towards or away from the tendon by 10 mm allowing
the measurement of pressure on the patellar tendon to be made
(Figs. 2 and 3) [20]. To accurately determine the centre of the patel-
lar tendon bar, the centre of the tendon was marked on the stump,
transferred to the negative cast, then to the positive cast and finally
to the socket. A metal insert for attaching the PT transducer was
positioned on the positive cast before lamination of the socket. The
centre portion of the socket within the insert was cut out using a
special jig to produce a 25 mm cut out. The cut out of the patellar
tendon bar was bonded to the patellar tendon holder face by a layer
of epoxy. Thus, it was possible to incorporate the device into other
PTB prostheses by simply changing the patellar tendon bar holder.
The annular space between the patellar tendon sensing surface

Table 1
Characteristics of ten male test subjects.

Subject no. Age Height (m) Mass (kg) Reason of amputation Amputated side Stump length (mm) Mobility gradea

1 95 1.78 95 Trauma Right 177 4a

2 73 1.91 80 PVD Right 125 4c
3 58 1.69 72 Trauma Left 145 6
4 53 1.74 95 PVD Left 123 4a
5 56 1.78 76 PVD Left 150 6
6 61 1.78 108 PVD Left 120 5
7 49 1.85 133 Trauma Right 145 4a
8 34 1.75 67 Trauma Right 130 4c
9 47 1.80 89 Other Left 148 4b
10 61 1.68 64 PVD Right 130 5
a
Mobility grades in prosthetic rehabilitation (devised by Stanmore & Harold Wood Disability Service Centres).
762 N.A. Abu Osman et al. / Medical Engineering & Physics 32 (2010) 760–765

Fig. 3. Schematic diagram of the custom designed transducers. Patellar tendon transducer: (1) 25.0 mm Ø PT bar sensing surface, (2) strain gauged piston, (3) patellar tendon
bar holder, (4) socket mount, (5) metal insert, (6) laminated socket (7) housing, (8) ball bearing, (9) piston push rod, (10) displacement control knob and (11) screw thread.
Current view: PT transducer in neutral position (the original position as cast by the prosthetist), Strain gauged piston (2) with PT bar sensing surface may be translated
towards or away from the tendon.

and the socket was 1.5 mm. The socket was made to accommodate
the PT transducer flush with the interior socket wall. The trans-
ducer is able to give quantitative information about normal force.
A schematic diagram for the PT transducer and its components are
shown in Fig. 3. A full description of the design of this transducer
is given in Abu Osman et al. [20]. The complete PT transducer has
a mass of 102 g.

2.4. Entran® -based transducer

The transducer used to measure pressure at various locations

around the stump was a piston in a housing device incorporat-
ing a load cell (Entran® International, NJ, USA model ELFM-B1-5L).
The complete assembly of the Entran® -based device is described in
detail in Lee et al. [19].

2.5. Electrohydraulic transducer

A custom-built electrohydraulic transducer was used to

Fig. 4. Electrohydraulic transducer. measure the pressure at the distal end of the stump. The electro-
hydraulic transducer consisted of a 28 mm diameter; 2 mm thick
oil filled PVC bag connected via a PVC tube to a strain gauged
diaphragm transducer (Fig. 4) [20]. This was thin enough to be

Fig. 5. Marking on the positive cast.

 N.A. Abu Osman et al. / Medical Engineering & Physics 32 (2010) 760–765 763

the transducer sites were located on these lines at 50 mm intervals

longitudinally (Fig. 5).
Obviously the number of transducer sites available without
adjacent transducers impinging on each other was dependent on
the physical size of the socket (or stump). However, over the ten
subjects tested an average of 16 transducer sites were available
for pressure measurement. Owing to a limitation of the number
of recording channels available with the data acquisition system
it was only possible to take measurements from 11 ‘Entran® ’ sites
simultaneously. In order to record the pressure from the other sites
additional trial runs on the same subject had to be undertaken.
For every subject the pressure at all available sites were recorded.
Therefore sites that were not monitored were sealed with blank
transducer plugs to ensure that the pressure difference across the
socket wall was maintained.

2.6. Calibration

Fig. 6. (a) Sample transducers’ locations on a trans-tibial socket. (b) Annotations
used throughout the study for the trans-tibial socket transducer locations.
inserted between the distal end of the stump and the socket where
the Entran® -based transducer could not be installed.
The sites of interest for transducer location were the same for
all experimental sockets for each individual subject with the cen-
tre of the patellar tendon acting as a datum. The cast was marked
longitudinally at 45 degrees segments using a ‘dividing head’ and
Calibration of the electrohydraulic transducer was performed
in a pressure vessel by applying uniform pressure on the trans-
ducer. The PT and ‘Entran® ’ transducers were calibrated using the
dead weight method. The calibration results show all transducers
displaying a linear output with all cross-effects during calibration
taken into account. Under laboratory bench conditions the trans-
ducer showed: the hysteresis error from the PT transducer was
found to be 1.85% full scale output (FSO), 400 kPa for the pressure
[20].

Fig. 7. (a and b) Sample results of subject 1 and subject 9. Variation in pattern on the other areas of the socket when the patellar tendon bar was compressed (+) or relieved
(−). Insert shows the change in PT bar location from the neutral position (N). DE = distal end.
764 N.A. Abu Osman et al. / Medical Engineering & Physics 32 (2010) 760–765
2.7. Experimental sockets
So-called ‘Hard’ sockets, i.e. without liners, incorporating
machined metal inserts were manufactured, using standard lam-
inating techniques, for each subject. All subjects were fitted with
check sockets prior to production of the experimental socket to
ensure that the experimental socket was a total contact socket
and was comfortable. The metal inserts allowed the transducers
to be placed within the socket wall such that the sensing sur-
face was flush with the inside surface of the socket. The sites of
transducer location were as shown in Fig. 6. The total mass of
each experimental prosthesis, with all transducers and blanking
plugs in situ, was approximately 3.0 kg. A TEC ProLink (TEC Inter-
face Systems, MN, USA) suspension sleeve in conjunction with
a simple one-way valve was used for suspension. This method
of suspension was simply by observation found to successfully
eliminate pistoning action between the stump and the socket dur-
ing walking. The socket did not appear to move axially relative
to the rest of the limb and there was no pistoning sound heard
whilst the subject walked using the experimental prosthesis (con-
firmed that no sound was heard by the researchers, prosthetist and
subjects).
2.8. Experimental procedure
The sampling rate used was 50 Hz and data were recorded for
approximately 15 s for each walking trial at the subject’s self-
selected walking speed. Data were collected simultaneously using
a strain gauge amplifier, model DAQN-Bridge (Dewetron, Austria)
with PCMCIA Lab View DaqCard 700 (National Instruments, Austin,
TX) for data acquisition and a Dell Inspiron PIII for data storage
with Lab View version 6.1 as analyzing software. For the locomo-
tion test, the subject was asked to walk a distance of 7 m on a
level walkway. At least three trials were recorded for each test.
Data were obtained with the patellar tendon bar of the prosthesis
in the neutral position (the original position as cast and rectified
by the prosthetist), compressed by 2 and 4 mm and relieved by
2 mm. In addition to the measurement obtained from the patellar
tendon bar, the other test sites containing transducers were sam-
pled simultaneously. All data collection, from each subject, was
performed on 1 day with the amputee wearing the prosthesis at
all times.
3. Results
The results presented are focused on the correlation of the patel-
lar tendon bar pressure with those at other test sites. All results
were firstly normalized to 100% of the gait cycle. For all ten sub-
jects the results clearly indicate that the patellar tendon bar of the
PTB socket bears significantly more pressure than the other sites
Fig. 8. Results of subject 1 to subject 10. Percentage (%) increase (+) or decrease
(−) of pressure for each 2 mm of bar translation relative to the neutral position for
subject #1 to #10 with the PTB socket. Bottom data series: neutral to minus 2, centre
data series: neutral to plus 2, top data series: neutral to plus 4.
around the socket. The range of maximum pressure recorded at
the patellar tendon bar for all subjects, during walking was approx-
imately from 203 to 230 kPa (Fig. 7a and b). Altering the indentation
by 2 mm from the neutral position to minus 2, plus 2 mm or plus
4 mm decreased or increased the pressures at the patellar tendon by
the percentage shown in Fig. 8. This percentage increase/decrease
pressure was found to be subject dependent. From 10 subjects, the
average percentage increase of pressure at the patellar tendon from
neutral to plus 2, and neutral to plus 4 were 25% and 49%, respec-
tively. Whilst the average percentage decrease of pressure from
neutral to minus 2 was 92%.
4. Discussion
In this study, the effect of varying the load on the patellar ten-
don bar on the pattern of pressure distribution at the stump–socket
interface has been investigated. Data were collected with the patel-
lar tendon bar of the prosthesis in the neutral position (the original
position as cast and rectified by the prosthetist), compressed 2 mm
and compressed 4 mm and relieved by 2 mm and also the other
test sites containing transducers were sampled simultaneously. The
data provided interesting results which showed that, as the patellar
tendon area of the prosthesis was compressed, the pressure at that
site increased quite significantly for all subjects, but the amount of
pressure varied from subject to subject as would be expected. The
pressures recorded at the test sites from the Entran® -based trans-
ducers did not show any particular variation in pattern when the
patellar tendon bar was compressed or relieved, i.e. it had no sig-
nificant effect on the other areas of the socket (Fig. 7a and b). These
results, by virtue of the fact that positive pressures were recorded
from all sites, tend to support the hypothesis that they were total
contact sockets.
Looking at the interface pressures, repeatable characteristics
were evident in the data from different steps at all sites. In this
study (excluding the pressure at the patellar tendon), the results
from 10 subjects show that M1 sites (mid-way between prox-
imal and distal of the medial aspect) and anterior distal sites
recorded greater pressure than those at other sites. These were
in the range of 65–72.3 kPa and 63–110 kPa, respectively. Sanders
et al. [11] found pressure maxima at anterior and posterior sites
(including patellar tendon) and anterior medial sites ranging from
52.2 to 223.8 kPa. In this study, analysis of 10 subjects showed
pressures in PTB sockets are in the range of: 4–125 kPa (ante-
rior sites), 17.5–89.5 kPa (posterior sites), 4–109.8 kPa (medial
sites), 21.7–83.3 kPa (lateral sites) and 12–35 kPa (distal stump
end).
The range of maximum pressure recorded at the patellar tendon
bar are in agreement with the literature, which states the maximum
interface pressures for PTB sockets during walking to be approxi-
mately 300 kPa at the neutral position (the original position as cast
and rectified by the prosthetist).
Although the experimental prostheses were heavier when all
transducer sites were fitted with transducers (by approximately
1.5 kg) than the subjects’ normal prostheses, the tests were con-
ducted with little negative comment from the subject group, except
for a universal notice of a “slight increase in the weight of the
prosthesis”.
Subjective subject feedback indicated that the patellar tendon
bar could be indented a further 4 mm from the original rectified
position without causing pain. However the subjects unanimously
preferred the relief position. This study of 10 subjects therefore
supports the supposition of [3], which stated that the concept of
load transfer to areas such as the patellar tendon, medial flare and
condyles of the tibia is ineffective and uncomfortable, certainly as
far as the patellar tendon area is concerned.
 N.A. Abu Osman et al. / Medical Engineering & Physics 32 (2010) 760–765
5. Conclusions
Three types of pressure measuring transducer were successfully
designed, manufactured and utilised in the study of stump/socket
interface pressure measurement in trans-tibial prostheses. One
transducer allowed the patellar tendon depth to be varied and data
relating to the three orthogonal forces acting on the patellar tendon
to be recorded at each position. One transducer, the electrohy-
draulic transducer, allowed measurements to be obtained from the
distal end of the stump. The third type of transducer allowed the
pressure at various locations around the socket to be measured.
The transducers enabled data to be collected from 10 trans-
tibial subjects during standing and walking the length of a 7 m walk
away. Analysis of the data revealed peak pressures to be located at
the site of the patellar tendon only when the patellar tendon bar
was indented. This peak pressure varied from subject to subject but
always increased as the indentation increased. The position of the
patellar tendon bar preferred by all subjects resulted in a peak pres-
sure at the patellar tendon site which was often not significantly
higher than other sites around the socket, and again was subject
dependent.
Varying the depth of the patellar tendon bar had no effect on the
pressure distribution around the remainder of the socket. The sub-
ject felt no detrimental effect (quite the opposite) by the removal
of the patellar tendon bar and it is therefore concluded that it is an
unnecessary feature which may be eliminated during the manufac-
ture of a trans-tibial prosthetic socket.
Acknowledgements
This research is supported by the Department of Bioengineer-
ing, University of Strathclyde, UK and Department of Biomedical
Engineering, Faculty of Engineering, University of Malaya, Malaysia.
Assistance from Stephen Murray, Ian Tullis, John Maclean, David
Robb, Robert Hay and Willie Tierney, is gratefully acknowledged.
Conflict of interest
The authors have no known affiliations that present a conflict of
interest.
References
[1] Radcliffe CE, Foort J. The patellar tendon bearing below knee prosthesis. Berke-
ley: Biomechanics Laboratory, University of California; 1961.
765
[2] Mizrahi J, Susak Z, Bahar A, Seliktar R, Najenson T. Biomechanical evaluation of
an adjustable patellar tendon bearing prosthesis. Scand J Rehabil Med (Supple-
ment) 1985;12:117–23.
[3] Kristinsson. The ICEROSS concept: a discussion of a philosophy. Prosthet Orthot
Int 1993;17:49–55.
[4] Burgess EM, Moore AJ. A study of interface pressures in the below-knee
prosthesis (physiological suspension: an interim report). Bull Prosthet Res
1977;10(28):58–70.
[5] Rae JE, Cockrell JL. Interface pressure and stress distribution in prosthetic fitting.
Bull Prosthet Res 1971;10(15):64–111.
[6] Meier RH, Meeks ED, Herman RM. Stump-socket fit of below-knee prosthe-
ses: comparison of three methods of measurement. Arch Phys Med Rehab
1973;54(12):553–8.
[7] Pearson JR, Holmgren G, March L, Oberg K. Pressures in critical regions
of the below-knee patellar-tendon-bearing prosthesis. Bull Prosthet Res
1973;10(19):52–76.
[8] Katz K, Susak Z, Seliktar R, Najenson T. End-bearing characteristics of
patellar-tendon-bearing prostheses—a preliminary report. Bull Prosthet Res
1979;10(32):55–68.
[9] Pijkeren TV, Naeff M, Kwee HH. A new method for the measurement of nor-
mal pressure between amputation residual limb and socket. Bull Prosthet Res
1980;10(33):31–4.
[10] Silver-Thorn MB, Childress DS. Use of generic, geometric finite element model
of the below-knee residual limb and prosthetic socket to predict inter-
face pressures. In: Proceedings of the Seventh World Congress ISPO. 1992.
p. 272.
[11] Sanders J, Daly C, Burgess E. Interface shear stresses during ambulation with a
below-knee prosthetic limb. J Rehabil Res Dev 1992;29(4):1–8.
[12] Williams RB, Porter D, Roberts VC, Regan JF. Triaxial force transducer for
investigating stresses at the stump/socket interface. Med Biol Eng Comput
1992;30:89–96.
[13] Goh JCH, Lee PVS, Chong SY. Static and dynamic pressure profiles of a patellar
tendon bearing socket. Proc Inst Mech Eng H 2003;217:121–6.
[14] Mak AFT, Zhang M, Boone DA. State-of-the-art research in lower limb pros-
thetic biomechanics socket interface: a review. J Rehabil Res Dev 2001;38(2):
161–74.
[15] Sanders J, Daly C, Burgess E. Clinical measurement of normal and shear stress
on a transtibial stump; characteristics of wave-form shapes during walking.
Prosthet Orthot Int 1993;17:38–48.
[16] Zhang M, Turner-Smith AR, Tanner A, Roberts VC. Clinical investigation of the
pressure and shear stress on the transtibial stump with prosthesis. Med Eng
Phys 1998;20(3):188–98.
[17] Mak AFT, Hong ML, Chan C. Finite element models for analyses of stresses
within above knee stump. In: Proceedings of the Seventh World Congress ISPO.
1992. p. 147.
[18] Sanders J, Daly C. Interface pressure and shear stresses: sagital plane angular
alignment effects in three transtibial amputee case studies. Prosthet Orthot Int
1999;23:21–9.
[19] Lee VSP, Solomonidis SE, Spence WD. Stump-socket interface pressure as an
aid to socket design in prostheses for transfemoral amputees—a preliminary
study. Proc Inst Mech Eng H 1997;211:167–80.
[20] Abu Osman NA, Spence WD, Solomonidis SE, Paul JP, Weir AM. Transduc-
ers for the determination of the pressure and shear stress distribution at the
stump–socket interface of trans-tibial amputees, Proc Inst Mech Eng B: J Eng
Manuf; in press.