- Email: [email protected]
socket pressure profiles of the pressure cast prosthetic socket
Clinical Biomechanics 18 (2003) 237–243 www.elsevier.com/locate/clinbiomech
Stump/socket pressure profiles of the pressure cast prosthetic socket J.C.H. Goh *, P.V.S. Lee, S.Y. Chong Department of Orthopaedic Surgery, The National University of Singapore, 5 Lower Kent Ridge Rd, Singapore 119074, Singapore Received 14 March 2002; accepted 10 December 2002
Abstract Objective. The aim was to evaluate stump/socket interface pressure in amputees wearing a socket developed by a pressure casting system. Design. Five unilateral transtibial amputees wore a pressure cast socket and walked at a self-selected speed. Background. The socket produces equally distributed pressure at the stump/socket interface, deviating from the conventional belief that pressure varies in proportion to the pain threshold of different tissues in the stump. Methods. The socket was fabricated while the subject placed his stump in a pressure chamber. Pressure was applied while he adopted a normal standing position. A specially built strain gauged type pressure transducer was used for measuring pressure distribution. Pressure and gait parameters were measured simultaneously while the subjects were standing and walking. Results and conclusion. The pressure cast technique was able to provide comfortable fitting sockets. A hydrostatic pressure profile was not evident during standing or gait. Results also showed that no standard pressure profile for the pressure cast socket was observed. This was expected as no rectifications were done on the pressure cast socket. Pressure profiles at 10%, 25% and 50% of gait cycle did not correlate with the pressure profiles previously proposed. Relevance The hydrostatic theory is an attractive concept in socket design as it produces a stump/socket pressure profile that is evenly distributed. Furthermore, it is a method that is easily implemented, independent of a prosthetistÕs skill and experience and reduces manufacturing time. However, there is still controversy surrounding the efficacy of this hydrostatic theory. Ó 2003 Elsevier Science Ltd. All rights reserved. Keywords: Transtibial amputees; Socket; Stump; Pressure; Biomechanics
1. Introduction In 1965, Murdoch (1965) introduced a pressure casting (PCast) concept where fluid was used as a medium to apply uniform pressure around the stump. The motivation for the development of this concept was due to inconsistencies in producing satisfactory patellartendon-bearing (PTB) socket. Thus, the Dundee socket was developed with the intention to remove some factors related to manual dexterity during the casting process. However, an indentation to the patellar tendon was still implemented. Kristinsson (1992) used the same PCast concept by using air as a medium in the Icelandic roll on silicone
*
Corresponding author. E-mail address: [email protected] (J.C.H. Goh).
socket (ICEROSS) system. The socket shape was defined by casting plaster wrap over the stump wearing the ICEROSS silicone liner using an air pressure chamber. However, rectifications were done by adding padding over bony areas of the stump during the casting process. Casting was performed on the patient in a non-weight bearing fashion i.e. not standing, in the air pressure chamber. His belief is that a transtibial socket, designed to transfer loads primarily to limited areas of the limb such as the patellar tendon, the medial flare and condyles of the tibia, for instance, is in most cases both ineffective and uncomfortable; the most effective socket, in his view, is one that relies on the hydrostatic principle for load transfer (Kristinsson, 1992). The theory behind such sockets also known as the hydrostatic socket is PascalÕs principle of fluid dynamics. This principle states that in a fluid at rest, the fluid pressure on any surface exerts a force perpendicular to
0268-0033/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0268-0033(02)00206-1
238
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
that surface. This law also assumed to include the transmissibility of fluid pressure that states any additional pressure will be transmitted equally to every point in the fluid. To obtain a socket fitting, the prosthetist needs to incorporate a PCast system that applies equal pressure on the limb as a plaster mould is obtained, the hydrostatic socket is then produced from a mould with small rectifications done on the anterior distal tibia, fibula head and the tibia crest. This translates to a socket that is different in shape to the PTB socket. A major difference is that the hydrostatic socket is not indented proximally at the patellar tendon area and the posterior aspect of the socket (Fergason and Smith, 1999; Lee et al., 2000). Another difference is that while the PTB socket biomechanics had been defined for each of the progression phases of gait, the hydrostatic socket makes no accommodation for such dynamic forces. However, it assumes that pressure at one point will simply be transferred by the fluid principle to other accommodating soft tissues (Fergason and Smith, 1999). In this study, a PCast system was developed (Lee et al., 2000). However, no rectifications were done on the socket. While there are many advantages to the PCast concept, such as ease in implementing the technique, no rectifications required and improved prosthesis delivery time, there are still many fundamental issues unresolved. Questions on whether the hydrostatic socket would fit all patients and whether the socket developed from this concept would function according to the principle of hydrostatic pressure had been raised (Fergason and Smith, 1999; Lee et al., 2000). Therefore, the aim of this study is to investigate the stump/socket pressure distribution of five unilateral amputees wearing the PCast socket.
2. Methods 2.1. Subjects Five unilateral transtibial amputees volunteered for this study. All subjects were male and had a unilateral amputation at least three years prior to this study. The
detailed particulars of the subjects are shown in Table 1. The subjects signed an informed consent conforming to the guidelines of the ethics committee of the hospital. 2.2. PCast system The PCast system consists of the PCast tank, plaster wrap, stockinette and weighing machine. The PCast tank consists of an opening at the top that allows the subject to place his stump in. A plaster wrap cast was first applied over the stump. The subject placed his stump in the tank separated from the hydraulic fluid medium by a diaphragm. He was then requested to stand in his normal standing position without any aid. When level standing was ensured by estimating the height of the right and left anterior superior iliac spines, pressure was introduced. The contra-lateral limb was positioned over a weighing scale to ensure half of the body weight was placed on the PCast system. Fig. 1 shows a subject with his stump immersed in the PCast tank. Upon hardening, the system was depressurised and the stump with the plaster wrap was removed from the PCast tank. The plaster wrap was then removed from the stump. A positive cast was generated from the wrap cast and a socket was fabricated using traditional lamination methods without any rectifications performed on cast. 2.3. Test socket fabrication A PCast test socket was made for each subject. The test socket was fabricated with 16 pressure measurements sites. Four sites were situated on each side of the socket: the anterior, posterior, medial and lateral sides. The exception was made for subjects #3 and #5. Due to space constraints, only three sites were situated at the posterior side. The sites were openings made in the socket to allow the pressure transducers to be mounted upon so that contact can be made with the stump (refer to Fig. 2). Araldite glue was used to ensure that the socket adaptors do not shift during fabrication of the socket. The socket shape was not altered to accommodate any of these pressure transducers.
Table 1 Particulars of subjects in this study Subject
#1
#2
#3
#4
#5
Age (years) Body weight (kg) Amputation side Stump length (from mid-patella to end of stump) in cm Stump length (from mid-patella to end of tibia) in cm Reason for amputation
54 76.2 Left 14.0
62 84.0 Left 14.0
31 87.0 Right 11.0
41 75.4 Left 15.0
34 62.8 Right 12.5
13.0
10.0
8.0
14.0
10.0
Vascular disease
Vascular disease
Traumatic injuries
Traumatic injuries
Traumatic injuries
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
239
hamstrings and another site (P4) was chosen 4 cm from the distal end; this is to allow ample space for the socket connector. Another two sites were chosen (P2 and P3) evenly spaced between P1 and P4. The sites chosen for the anterior, medial and lateral sides were at the same level as the posterior side. 2.5. Pressure transducer assembly
Fig. 1. PCast system.
The pressure transducers were constructed as shown in Fig. 3. The housing is designed to minimise any cross sensitivity to shear loads. The pressure transducer assembly included a load cell model ELFM-B1-5L (Entran International, New Jersey, USA). The load cell is a sensitive diaphragm onto which miniature electrical resistance strain gauges in a full Wheatstone bridge configuration were bonded. The specifications of the load cell are given in Table 2. The complete assembly includes a cylindrical piston (of diameter 5.6 mm) that transferred normal pressure from the stump tissues to the load cell. A nylon housing which housed the load cell was locked to the transducer adaptor to keep it secured and flushed with the inner surface of the socket at all times during the test (refer to Fig. 2). The transducer assembly was similar to that described in Lee et al. (1997). Calibration of the fully assembled transducer was performed using dead weights up to 200 kPa. It was estimated that the error of the transducer was 1.25 kPa. Since the diameter of the piston was relatively small, it is reasonable to assume that the stresses were uniformly distributed over the surface of the piston.
Fig. 2. Pressure transducer assembly being mounted on a measurement site.
2.4. Location of pressure measurement sites The anterior–posterior (AP) plane was defined by the patellar tendon, the popliteal depression and the distal end. The medial–lateral (ML) plane was perpendicular to the AP plane with reference to the distal end. At the posterior side, one site (P1) was chosen 2 cm from the edge of the posterior wall that gives relief to the
Fig. 3. A cross-sectional view of pressure transducer assembly: (1) socket adaptor, (2) transducer adaptor, (3) load cell housing, (4) transducer holder, (5) piston, (6) load cell (with shielded wires), (7) stopper, (8) stopper cap.
240
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
Table 2 Load cell specifications (FS: full scale) Range Non-linearity Hysteresis Sensitivity Operating temperature Thermal sensitivity shift a
3. Results and discussion 5 lbs (25 N) 0.25% FS 0.25% FS 1.777 mV/FSa )50 to 120 °C 0.02%/°C
Unique to individual load cell.
2.6. Gait analysis The pressure transducers were connected to the VICON 370 (Vicon Motion Systems, Oxford, UK) 3D motion analysis system in conjunction with two Kistler (Kistler Instruments, Winterthur, Switzerland) force platforms. Ground reaction forces (GRFs) and pressure data were all acquired simultaneously at 250 Hz. The motion analysis system uses five infra-red cameras which tracked 17 retro-reflective markers attached to the following body landmarks: shoulder, anterior superior iliac spine, sacrum, mid-thigh, lateral knee centre, mid-tibia, heel, second metatarsal head and lateral malleolus. Markers were placed on the prosthetic socket, at positions corresponding to the lateral knee centre, mid-tibia, heel, second metatarsal head and lateral malleolus. The markers were captured at a sampling rate of 50 Hz.
In the dynamic test, pressures were only considered when the subject stepped on the force plate. Before the results were averaged, the data was first normalized to 100 per cent of gait cycle. A pressure profile was thus drawn and compared with the anticipated pressure profile drawn. The exact time of gait cycle was determined by VICON 370 3D motion analysis system and Kistler force platforms. Fig. 4 shows the static pressure profile of subject #4. Figs. 5–7 show the AP pressure profile of subject #1 with his respective knee moments and RadcliffeÕs anticipated pressure profile. Knee flexing moments were
Fig. 4. AP and ML static pressures of subject #4 (all values in kPa).
2.7. Testing procedures During data acquisition session, each subject wore stump socks without liner. Thus, the transducers measured socket–sock interface stresses, not stresses directly on the skin surface. Each socket was assembled on to an Endolite Multiflex (Blatchford, Hampshire, UK) ankle foot system. Alignment was performed by one trained prosthetist. The subjects agreed that locomotion after alignment was acceptable. A muley strap was supplied for suspension of the prostheses. The muley strap was fork-shaped design and stitched to a length of elastic webbing and metal buckle. It was used to suspend the prosthesis over the supracondyles that will prevent piston action and migration of the stump in the socket. Each subject was required to walk with the prosthesis for at least 15 min to become accustomed to the test socket. All data collection was performed on the same day without the subject removing the socket at any interval during the test. The tests were divided into static (standing) and dynamic (walking) stages. During the static test, pressure measurements were taken when the subject was in normal standing position with prosthetic limb on force plate. For the dynamic test, the subject was requested to walk a distance of 10 m at normal selfselected speed, stepping on the Kistler force plates at approximately midway through the walk. A minimum of three trials was recorded for each static and dynamic test.
Fig. 5. AP pressure profile of subject #1 at 10% of gait cycle (all pressure values in kPa unless otherwise stated).
Fig. 6. AP pressure profile of subject #1 at 25% of gait cycle (all pressure values in kPa unless otherwise stated).
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
241
shown as positive values while knee extending moments were shown as negative values. Fig. 8 shows subject #4Õs ML pressure profile at 25% of gait cycle. This study placed pressure measurement sites at locations that were based on the geometry of each subjectÕs stump, rather than specific locations. It should be noted that these locations are important, as a small shift would result in a different pressure reading. Furthermore, these test sockets are different from the subjectsÕ regular prosthesis. One subjective comment by the subjects was that there was no specific load bearing areas. This was expected, as the PCast socket has no rectifications done unlike the traditional PTB socket.
One reason a hydrostatic pressure profile was not present could be that the stump is of complex nature. It is often assumed that the stump acts like an elastic solid with low stiffness surrounding a piston: the tibia with its condyles. If this can be fitted into a containing vessel corresponding exactly to its volume, the fitting can be expected to behave like a hydrostatic system when loaded. Then, in the absence of motion, there is no shear stress; the internal state of pressure at any point is determined by applied pressure alone. Hence, the pressure at a point is the same in all directions and the pressure required to support the weight would be determined by the cross-section of the socket opening (Kristinsson, 1993). However, this goal can be reached only to a certain degree as the stump consists of soft tissues like the patellar tendon, the posterior wall and protruding bone structures like the fibula head and the distal end of the stump. The PCast system could play a part as well. The system would only be effective if it is tight. The moment it leaks, it would lose its mechanical stiffness. Though the system cannot be referred to as an effectively closed hydrostatic system, the stump/socket interface may probably be assumed to act as an elastic coupling with hydrostatic characteristics during weight bearing. Given that the posterior wall is high enough and the system does not leak past the condyles, the PCast system should be able to transfer the forces generated during weight bearing without any need for limited area loading or even conformity to the underlying skeleton (Kristinsson, 1993). Furthermore, force interactions between socket and the stump would differ at different periods during gait. Volume changes that occur during the day would also affect socket fitting. Therefore, a loss of hydrostatic stability would be inevitable at these different intervals. One way to minimise this problem is to ensure that the stump is firmly secured to the socket when a suspension system is used.
3.1. Hydrostatic pressure profile
3.2. Dynamic pressure profile
During standing, no specific pressure profile for all subjects can be seen. The AP pressure profile was inconsistent for all subjects. Subjects #1 and #4 exhibited higher pressure at the anterior proximal regions while subjects #2 and #3 exhibited high pressures at the distal ends. The ML pressure profile showed consistent high pressures at the medial proximal, probably due to adduction moment caused by the ML GRF component. The lateral side showed higher pressure at the lateral distal region with the exception of subject #2 where he exhibited higher pressure at the lateral proximal side. In all, hydrostatic pressure profile was not evident during standing. One example is subject #4Õs static pressure profiles as shown in Fig. 4.
As the pressure profiles were determined to be not uniform, the next step was to determine whether the resulting pressure profiles could be explained biomechanically. The analysis by Radcliffe (1961) and Radcliffe and Foort (1961) was based on the assumption that a below knee amputee was able to walk in a manner similar to that of a normal person. During heel contact, the line of action of the GRF acts anterior to the knee, causing the knee to extend. The hamstrings, acting to prevent knee hyperextension, would cause high pressure concentration at the anterior proximal and the posterior distal region. During midstance, the GRF would be acting posterior to the knee. In such an instance, the knee
Fig. 7. AP pressure profile of subject #1 at 50% of gait cycle (all pressure values in kPa unless otherwise stated).
Fig. 8. MP pressure profile of subject #4 at 25% of gait cycle (all values in kPa).
242
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
would buckle. However, this was resisted by action of the quadriceps and forceful extension of the hip where high pressure concentration would occur at the anterior proximal, anterior distal and the popliteal region. During toe-off where the line of action of the GRF remained posterior to the knee, the same three areas at the anterior proximal, anterior distal and popliteal region would experience high pressure concentration. In summary, Radcliffe proposed that the AP pressure profile would invert from heel strike to toe-off. From our pressure profiles, no inversion was seen from heel strike to toe-off. Taking the case of subject #1; at 10% of gait cycle, his GRF occurred anterior to his knee. At this instance, as the hamstrings act to prevent hyperextension of the knee, high pressures at the anterior proximal and the posterior distal region would be expected from RadcliffeÕs model. However, subject #1 experienced a pressure profile which was more uniform (refer to Fig. 5). At 25% of gait cycle, subject #1Õs GRF occurred posterior to his knee. As the knee has a tendency to flex, the pressure profile would start inverting as this flexing moment is controlled by the quadriceps. Subject #1 started to experience high pressure at the popliteal region as anticipated, however, high pressure occurred at the anterior proximal instead of the anterior distal (refer to Fig. 6). At 50% of gait cycle, when the GRF shifted to the anterior of his knee, high pressure at anterior proximal and posterior distal was expected. Subject #1 did exhibit high pressure concentration at the anterior proximal region (refer to Fig. 7). It should be noted that Radcliffe assumed that the GRF would shift from the anterior of the knee at heel strike to the posterior of the knee at midstance where it remained till toe-off (Radcliffe, 1961; Radcliffe and Foort, 1961). This did not happen in our subjects. It was especially demonstrated in subject #3 and #5 whose gait were not typical of transtibial walking patterns. Subject #3Õs GRF occurred anterior to the knee throughout stance, thus, his knee continually experienced an extending moment while subject #5Õs GRF occurred posterior to the knee throughout stance, resulting in a knee flexing moment. The ML component of GRF always acted towards the intact limb, therefore high pressure was expected at the medial proximal and lateral distal regions. All the subjects experienced high pressure at the medial proximal region, however, not all subjects experienced high pressure at the lateral distal region (refer to Fig. 8). This study showed that not all subjects had a pressure profile as anticipated due to the effects of GRF. This was similar to ConveryÕs study where he stated that a review of the pressure distribution for the hydrocast socket during gait did not indicate a logical explanation or agreement with the biomechanical principles proposed by Radcliffe (1961) and Radcliffe and Foort (1961). However, he did add that the relationship of the line of
action of the GRF to the socket during stance phase of gait might influence the pressure data. He further mentioned that his subjectÕs GRF always passed ahead of the socket (Convery and Buis, 1993). One reason was that the GRF was not the only single factor that determined pressure profiles, for example, at 50% of gait cycle, high pressure profiles occurred at the anterior proximal region regardless of the line of action of the GRF. High pressure would be expected at the anterior proximal region if it was the traditional PTB prosthetic socket as the design of the socket consisted of a patellar-tendon ÔbarÕ. However, even with the absence of this ÔbarÕ on the PCast socket, high pressure was demonstrated at the anterior proximal region, probably due to the patellar tendon acting on the socket when full body weight was needed for push-off. Though a consistent pressure profile was not found in this study, it should be emphasized that many other factors play an important role in determining each subjectÕs pressure profile. Considering each subjectÕs varieties for example in possessing different stump shape and other factors that include socket alignment and thigh muscle strength, all these could affect the resulting pressure profile. Furthermore, the limited number of subjects used in this study could also be another reason. However, it should still be noted that despite the different pressure profiles, the PCast technique was able to fabricate comfortable fitting sockets for these five subjects.
4. Conclusion A PCast system, using the concept of producing uniformly distributed pressure at the stump/socket interface was developed to produce the PCast socket This system requires the subject to place his stump in a pressure chamber. Pressure was then applied to the stump while he adopted a normal standing position. No rectifications were done on the PCast socket. Results show that a hydrostatic pressure profile was not evident during standing or gait. This could be a result of the complex nature of the stump or due to the different force interactions between the socket and stump at different phases of gait. As no rectifications were done on the PCast socket, no standard pressure profile could be seen. However, despite these inconsistent pressure profiles, the PCast technique was able to fabricate comfortable fitting sockets for these five subjects. In our study, our subjectsÕ pressure profiles measured were different compared with RadcliffeÕs anticipated pressure profile. One reason could be that the line of action of the GRF was not be the only single factor in affecting pressure distribution. More subjects should be done to ascertain and determine the factors affecting the resulting pressure profiles.
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
Acknowledgements This research was supported by the National Medical Research Council, Singapore. We would also like to thank Mr. Cheung Sze Kwong, Mr. Pan Seng Kie, Ms. Grace Lee, Mr. Wang Jit Beng, Mr. Hazlan Bin Sanusi and Mr. Azmee Murat for their technical assistance.
References Convery, P., Buis, A.W.P., 1993. Socket/stump interface dynamic pressure distributions recorded during the prosthetic stance phase of gait of a trans-tibial amputee wearing a hydrocast socket. Prosthetic and Orthotics International 23, 107–112. Fergason, J., Smith, D.G.S., 1999. Socket consideration for the patient with a transtibial amputation. Clinical Orthopaedics and Related Research 361, 76–84.
243
Kristinsson, O., 1992. Pressurised casting instruments. In: Proceedings of the 7th World Congress, International Society of Prosthetics and Orthotics, Chicago, USA. Kristinsson, O., 1993. The ICEROSS concept: a discussion of philosophy. Prosthetics and Orthotics International 17, 49–55. Lee, V.S.P., Solomonidis, S.E., Spence, W.D., 1997. Stump–socket interface pressure as an aid to socket design in prostheses for transfemoral amputees––a preliminary study. In: Proceedings of Institution of Mechanical Engineers, Vol. 211, Part H, pp. 167–180. Lee, P., Goh, J., Cheung, S.K., 2000. Biomechanical evaluation of the pressure cast (PCast) prosthetic socket for trans-tibial amputee. In: Proceedings of the World Congress on Medical Physics & Biomedical Engineering, Chicago, USA. Murdoch, G., 1965. The Dundee socket for below knee amputation. Prosthetic International 3, 12–14. Radcliffe, C.W., 1961. The biomechanics of below-knee prosthesis in normal, level, bipedal walking. Artificial Limbs 6, 16–24. Radcliffe, C.W., Foort, J., 1961. The patella-tendon-bearing belowknee prosthesis. Biomechanics laboratory, University of California, Berkeley, CA.
Stump/socket pressure profiles of the pressure cast prosthetic socket J.C.H. Goh *, P.V.S. Lee, S.Y. Chong Department of Orthopaedic Surgery, The National University of Singapore, 5 Lower Kent Ridge Rd, Singapore 119074, Singapore Received 14 March 2002; accepted 10 December 2002
Abstract Objective. The aim was to evaluate stump/socket interface pressure in amputees wearing a socket developed by a pressure casting system. Design. Five unilateral transtibial amputees wore a pressure cast socket and walked at a self-selected speed. Background. The socket produces equally distributed pressure at the stump/socket interface, deviating from the conventional belief that pressure varies in proportion to the pain threshold of different tissues in the stump. Methods. The socket was fabricated while the subject placed his stump in a pressure chamber. Pressure was applied while he adopted a normal standing position. A specially built strain gauged type pressure transducer was used for measuring pressure distribution. Pressure and gait parameters were measured simultaneously while the subjects were standing and walking. Results and conclusion. The pressure cast technique was able to provide comfortable fitting sockets. A hydrostatic pressure profile was not evident during standing or gait. Results also showed that no standard pressure profile for the pressure cast socket was observed. This was expected as no rectifications were done on the pressure cast socket. Pressure profiles at 10%, 25% and 50% of gait cycle did not correlate with the pressure profiles previously proposed. Relevance The hydrostatic theory is an attractive concept in socket design as it produces a stump/socket pressure profile that is evenly distributed. Furthermore, it is a method that is easily implemented, independent of a prosthetistÕs skill and experience and reduces manufacturing time. However, there is still controversy surrounding the efficacy of this hydrostatic theory. Ó 2003 Elsevier Science Ltd. All rights reserved. Keywords: Transtibial amputees; Socket; Stump; Pressure; Biomechanics
1. Introduction In 1965, Murdoch (1965) introduced a pressure casting (PCast) concept where fluid was used as a medium to apply uniform pressure around the stump. The motivation for the development of this concept was due to inconsistencies in producing satisfactory patellartendon-bearing (PTB) socket. Thus, the Dundee socket was developed with the intention to remove some factors related to manual dexterity during the casting process. However, an indentation to the patellar tendon was still implemented. Kristinsson (1992) used the same PCast concept by using air as a medium in the Icelandic roll on silicone
*
Corresponding author. E-mail address: [email protected] (J.C.H. Goh).
socket (ICEROSS) system. The socket shape was defined by casting plaster wrap over the stump wearing the ICEROSS silicone liner using an air pressure chamber. However, rectifications were done by adding padding over bony areas of the stump during the casting process. Casting was performed on the patient in a non-weight bearing fashion i.e. not standing, in the air pressure chamber. His belief is that a transtibial socket, designed to transfer loads primarily to limited areas of the limb such as the patellar tendon, the medial flare and condyles of the tibia, for instance, is in most cases both ineffective and uncomfortable; the most effective socket, in his view, is one that relies on the hydrostatic principle for load transfer (Kristinsson, 1992). The theory behind such sockets also known as the hydrostatic socket is PascalÕs principle of fluid dynamics. This principle states that in a fluid at rest, the fluid pressure on any surface exerts a force perpendicular to
0268-0033/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0268-0033(02)00206-1
238
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
that surface. This law also assumed to include the transmissibility of fluid pressure that states any additional pressure will be transmitted equally to every point in the fluid. To obtain a socket fitting, the prosthetist needs to incorporate a PCast system that applies equal pressure on the limb as a plaster mould is obtained, the hydrostatic socket is then produced from a mould with small rectifications done on the anterior distal tibia, fibula head and the tibia crest. This translates to a socket that is different in shape to the PTB socket. A major difference is that the hydrostatic socket is not indented proximally at the patellar tendon area and the posterior aspect of the socket (Fergason and Smith, 1999; Lee et al., 2000). Another difference is that while the PTB socket biomechanics had been defined for each of the progression phases of gait, the hydrostatic socket makes no accommodation for such dynamic forces. However, it assumes that pressure at one point will simply be transferred by the fluid principle to other accommodating soft tissues (Fergason and Smith, 1999). In this study, a PCast system was developed (Lee et al., 2000). However, no rectifications were done on the socket. While there are many advantages to the PCast concept, such as ease in implementing the technique, no rectifications required and improved prosthesis delivery time, there are still many fundamental issues unresolved. Questions on whether the hydrostatic socket would fit all patients and whether the socket developed from this concept would function according to the principle of hydrostatic pressure had been raised (Fergason and Smith, 1999; Lee et al., 2000). Therefore, the aim of this study is to investigate the stump/socket pressure distribution of five unilateral amputees wearing the PCast socket.
2. Methods 2.1. Subjects Five unilateral transtibial amputees volunteered for this study. All subjects were male and had a unilateral amputation at least three years prior to this study. The
detailed particulars of the subjects are shown in Table 1. The subjects signed an informed consent conforming to the guidelines of the ethics committee of the hospital. 2.2. PCast system The PCast system consists of the PCast tank, plaster wrap, stockinette and weighing machine. The PCast tank consists of an opening at the top that allows the subject to place his stump in. A plaster wrap cast was first applied over the stump. The subject placed his stump in the tank separated from the hydraulic fluid medium by a diaphragm. He was then requested to stand in his normal standing position without any aid. When level standing was ensured by estimating the height of the right and left anterior superior iliac spines, pressure was introduced. The contra-lateral limb was positioned over a weighing scale to ensure half of the body weight was placed on the PCast system. Fig. 1 shows a subject with his stump immersed in the PCast tank. Upon hardening, the system was depressurised and the stump with the plaster wrap was removed from the PCast tank. The plaster wrap was then removed from the stump. A positive cast was generated from the wrap cast and a socket was fabricated using traditional lamination methods without any rectifications performed on cast. 2.3. Test socket fabrication A PCast test socket was made for each subject. The test socket was fabricated with 16 pressure measurements sites. Four sites were situated on each side of the socket: the anterior, posterior, medial and lateral sides. The exception was made for subjects #3 and #5. Due to space constraints, only three sites were situated at the posterior side. The sites were openings made in the socket to allow the pressure transducers to be mounted upon so that contact can be made with the stump (refer to Fig. 2). Araldite glue was used to ensure that the socket adaptors do not shift during fabrication of the socket. The socket shape was not altered to accommodate any of these pressure transducers.
Table 1 Particulars of subjects in this study Subject
#1
#2
#3
#4
#5
Age (years) Body weight (kg) Amputation side Stump length (from mid-patella to end of stump) in cm Stump length (from mid-patella to end of tibia) in cm Reason for amputation
54 76.2 Left 14.0
62 84.0 Left 14.0
31 87.0 Right 11.0
41 75.4 Left 15.0
34 62.8 Right 12.5
13.0
10.0
8.0
14.0
10.0
Vascular disease
Vascular disease
Traumatic injuries
Traumatic injuries
Traumatic injuries
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
239
hamstrings and another site (P4) was chosen 4 cm from the distal end; this is to allow ample space for the socket connector. Another two sites were chosen (P2 and P3) evenly spaced between P1 and P4. The sites chosen for the anterior, medial and lateral sides were at the same level as the posterior side. 2.5. Pressure transducer assembly
Fig. 1. PCast system.
The pressure transducers were constructed as shown in Fig. 3. The housing is designed to minimise any cross sensitivity to shear loads. The pressure transducer assembly included a load cell model ELFM-B1-5L (Entran International, New Jersey, USA). The load cell is a sensitive diaphragm onto which miniature electrical resistance strain gauges in a full Wheatstone bridge configuration were bonded. The specifications of the load cell are given in Table 2. The complete assembly includes a cylindrical piston (of diameter 5.6 mm) that transferred normal pressure from the stump tissues to the load cell. A nylon housing which housed the load cell was locked to the transducer adaptor to keep it secured and flushed with the inner surface of the socket at all times during the test (refer to Fig. 2). The transducer assembly was similar to that described in Lee et al. (1997). Calibration of the fully assembled transducer was performed using dead weights up to 200 kPa. It was estimated that the error of the transducer was 1.25 kPa. Since the diameter of the piston was relatively small, it is reasonable to assume that the stresses were uniformly distributed over the surface of the piston.
Fig. 2. Pressure transducer assembly being mounted on a measurement site.
2.4. Location of pressure measurement sites The anterior–posterior (AP) plane was defined by the patellar tendon, the popliteal depression and the distal end. The medial–lateral (ML) plane was perpendicular to the AP plane with reference to the distal end. At the posterior side, one site (P1) was chosen 2 cm from the edge of the posterior wall that gives relief to the
Fig. 3. A cross-sectional view of pressure transducer assembly: (1) socket adaptor, (2) transducer adaptor, (3) load cell housing, (4) transducer holder, (5) piston, (6) load cell (with shielded wires), (7) stopper, (8) stopper cap.
240
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
Table 2 Load cell specifications (FS: full scale) Range Non-linearity Hysteresis Sensitivity Operating temperature Thermal sensitivity shift a
3. Results and discussion 5 lbs (25 N) 0.25% FS 0.25% FS 1.777 mV/FSa )50 to 120 °C 0.02%/°C
Unique to individual load cell.
2.6. Gait analysis The pressure transducers were connected to the VICON 370 (Vicon Motion Systems, Oxford, UK) 3D motion analysis system in conjunction with two Kistler (Kistler Instruments, Winterthur, Switzerland) force platforms. Ground reaction forces (GRFs) and pressure data were all acquired simultaneously at 250 Hz. The motion analysis system uses five infra-red cameras which tracked 17 retro-reflective markers attached to the following body landmarks: shoulder, anterior superior iliac spine, sacrum, mid-thigh, lateral knee centre, mid-tibia, heel, second metatarsal head and lateral malleolus. Markers were placed on the prosthetic socket, at positions corresponding to the lateral knee centre, mid-tibia, heel, second metatarsal head and lateral malleolus. The markers were captured at a sampling rate of 50 Hz.
In the dynamic test, pressures were only considered when the subject stepped on the force plate. Before the results were averaged, the data was first normalized to 100 per cent of gait cycle. A pressure profile was thus drawn and compared with the anticipated pressure profile drawn. The exact time of gait cycle was determined by VICON 370 3D motion analysis system and Kistler force platforms. Fig. 4 shows the static pressure profile of subject #4. Figs. 5–7 show the AP pressure profile of subject #1 with his respective knee moments and RadcliffeÕs anticipated pressure profile. Knee flexing moments were
Fig. 4. AP and ML static pressures of subject #4 (all values in kPa).
2.7. Testing procedures During data acquisition session, each subject wore stump socks without liner. Thus, the transducers measured socket–sock interface stresses, not stresses directly on the skin surface. Each socket was assembled on to an Endolite Multiflex (Blatchford, Hampshire, UK) ankle foot system. Alignment was performed by one trained prosthetist. The subjects agreed that locomotion after alignment was acceptable. A muley strap was supplied for suspension of the prostheses. The muley strap was fork-shaped design and stitched to a length of elastic webbing and metal buckle. It was used to suspend the prosthesis over the supracondyles that will prevent piston action and migration of the stump in the socket. Each subject was required to walk with the prosthesis for at least 15 min to become accustomed to the test socket. All data collection was performed on the same day without the subject removing the socket at any interval during the test. The tests were divided into static (standing) and dynamic (walking) stages. During the static test, pressure measurements were taken when the subject was in normal standing position with prosthetic limb on force plate. For the dynamic test, the subject was requested to walk a distance of 10 m at normal selfselected speed, stepping on the Kistler force plates at approximately midway through the walk. A minimum of three trials was recorded for each static and dynamic test.
Fig. 5. AP pressure profile of subject #1 at 10% of gait cycle (all pressure values in kPa unless otherwise stated).
Fig. 6. AP pressure profile of subject #1 at 25% of gait cycle (all pressure values in kPa unless otherwise stated).
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
241
shown as positive values while knee extending moments were shown as negative values. Fig. 8 shows subject #4Õs ML pressure profile at 25% of gait cycle. This study placed pressure measurement sites at locations that were based on the geometry of each subjectÕs stump, rather than specific locations. It should be noted that these locations are important, as a small shift would result in a different pressure reading. Furthermore, these test sockets are different from the subjectsÕ regular prosthesis. One subjective comment by the subjects was that there was no specific load bearing areas. This was expected, as the PCast socket has no rectifications done unlike the traditional PTB socket.
One reason a hydrostatic pressure profile was not present could be that the stump is of complex nature. It is often assumed that the stump acts like an elastic solid with low stiffness surrounding a piston: the tibia with its condyles. If this can be fitted into a containing vessel corresponding exactly to its volume, the fitting can be expected to behave like a hydrostatic system when loaded. Then, in the absence of motion, there is no shear stress; the internal state of pressure at any point is determined by applied pressure alone. Hence, the pressure at a point is the same in all directions and the pressure required to support the weight would be determined by the cross-section of the socket opening (Kristinsson, 1993). However, this goal can be reached only to a certain degree as the stump consists of soft tissues like the patellar tendon, the posterior wall and protruding bone structures like the fibula head and the distal end of the stump. The PCast system could play a part as well. The system would only be effective if it is tight. The moment it leaks, it would lose its mechanical stiffness. Though the system cannot be referred to as an effectively closed hydrostatic system, the stump/socket interface may probably be assumed to act as an elastic coupling with hydrostatic characteristics during weight bearing. Given that the posterior wall is high enough and the system does not leak past the condyles, the PCast system should be able to transfer the forces generated during weight bearing without any need for limited area loading or even conformity to the underlying skeleton (Kristinsson, 1993). Furthermore, force interactions between socket and the stump would differ at different periods during gait. Volume changes that occur during the day would also affect socket fitting. Therefore, a loss of hydrostatic stability would be inevitable at these different intervals. One way to minimise this problem is to ensure that the stump is firmly secured to the socket when a suspension system is used.
3.1. Hydrostatic pressure profile
3.2. Dynamic pressure profile
During standing, no specific pressure profile for all subjects can be seen. The AP pressure profile was inconsistent for all subjects. Subjects #1 and #4 exhibited higher pressure at the anterior proximal regions while subjects #2 and #3 exhibited high pressures at the distal ends. The ML pressure profile showed consistent high pressures at the medial proximal, probably due to adduction moment caused by the ML GRF component. The lateral side showed higher pressure at the lateral distal region with the exception of subject #2 where he exhibited higher pressure at the lateral proximal side. In all, hydrostatic pressure profile was not evident during standing. One example is subject #4Õs static pressure profiles as shown in Fig. 4.
As the pressure profiles were determined to be not uniform, the next step was to determine whether the resulting pressure profiles could be explained biomechanically. The analysis by Radcliffe (1961) and Radcliffe and Foort (1961) was based on the assumption that a below knee amputee was able to walk in a manner similar to that of a normal person. During heel contact, the line of action of the GRF acts anterior to the knee, causing the knee to extend. The hamstrings, acting to prevent knee hyperextension, would cause high pressure concentration at the anterior proximal and the posterior distal region. During midstance, the GRF would be acting posterior to the knee. In such an instance, the knee
Fig. 7. AP pressure profile of subject #1 at 50% of gait cycle (all pressure values in kPa unless otherwise stated).
Fig. 8. MP pressure profile of subject #4 at 25% of gait cycle (all values in kPa).
242
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
would buckle. However, this was resisted by action of the quadriceps and forceful extension of the hip where high pressure concentration would occur at the anterior proximal, anterior distal and the popliteal region. During toe-off where the line of action of the GRF remained posterior to the knee, the same three areas at the anterior proximal, anterior distal and popliteal region would experience high pressure concentration. In summary, Radcliffe proposed that the AP pressure profile would invert from heel strike to toe-off. From our pressure profiles, no inversion was seen from heel strike to toe-off. Taking the case of subject #1; at 10% of gait cycle, his GRF occurred anterior to his knee. At this instance, as the hamstrings act to prevent hyperextension of the knee, high pressures at the anterior proximal and the posterior distal region would be expected from RadcliffeÕs model. However, subject #1 experienced a pressure profile which was more uniform (refer to Fig. 5). At 25% of gait cycle, subject #1Õs GRF occurred posterior to his knee. As the knee has a tendency to flex, the pressure profile would start inverting as this flexing moment is controlled by the quadriceps. Subject #1 started to experience high pressure at the popliteal region as anticipated, however, high pressure occurred at the anterior proximal instead of the anterior distal (refer to Fig. 6). At 50% of gait cycle, when the GRF shifted to the anterior of his knee, high pressure at anterior proximal and posterior distal was expected. Subject #1 did exhibit high pressure concentration at the anterior proximal region (refer to Fig. 7). It should be noted that Radcliffe assumed that the GRF would shift from the anterior of the knee at heel strike to the posterior of the knee at midstance where it remained till toe-off (Radcliffe, 1961; Radcliffe and Foort, 1961). This did not happen in our subjects. It was especially demonstrated in subject #3 and #5 whose gait were not typical of transtibial walking patterns. Subject #3Õs GRF occurred anterior to the knee throughout stance, thus, his knee continually experienced an extending moment while subject #5Õs GRF occurred posterior to the knee throughout stance, resulting in a knee flexing moment. The ML component of GRF always acted towards the intact limb, therefore high pressure was expected at the medial proximal and lateral distal regions. All the subjects experienced high pressure at the medial proximal region, however, not all subjects experienced high pressure at the lateral distal region (refer to Fig. 8). This study showed that not all subjects had a pressure profile as anticipated due to the effects of GRF. This was similar to ConveryÕs study where he stated that a review of the pressure distribution for the hydrocast socket during gait did not indicate a logical explanation or agreement with the biomechanical principles proposed by Radcliffe (1961) and Radcliffe and Foort (1961). However, he did add that the relationship of the line of
action of the GRF to the socket during stance phase of gait might influence the pressure data. He further mentioned that his subjectÕs GRF always passed ahead of the socket (Convery and Buis, 1993). One reason was that the GRF was not the only single factor that determined pressure profiles, for example, at 50% of gait cycle, high pressure profiles occurred at the anterior proximal region regardless of the line of action of the GRF. High pressure would be expected at the anterior proximal region if it was the traditional PTB prosthetic socket as the design of the socket consisted of a patellar-tendon ÔbarÕ. However, even with the absence of this ÔbarÕ on the PCast socket, high pressure was demonstrated at the anterior proximal region, probably due to the patellar tendon acting on the socket when full body weight was needed for push-off. Though a consistent pressure profile was not found in this study, it should be emphasized that many other factors play an important role in determining each subjectÕs pressure profile. Considering each subjectÕs varieties for example in possessing different stump shape and other factors that include socket alignment and thigh muscle strength, all these could affect the resulting pressure profile. Furthermore, the limited number of subjects used in this study could also be another reason. However, it should still be noted that despite the different pressure profiles, the PCast technique was able to fabricate comfortable fitting sockets for these five subjects.
4. Conclusion A PCast system, using the concept of producing uniformly distributed pressure at the stump/socket interface was developed to produce the PCast socket This system requires the subject to place his stump in a pressure chamber. Pressure was then applied to the stump while he adopted a normal standing position. No rectifications were done on the PCast socket. Results show that a hydrostatic pressure profile was not evident during standing or gait. This could be a result of the complex nature of the stump or due to the different force interactions between the socket and stump at different phases of gait. As no rectifications were done on the PCast socket, no standard pressure profile could be seen. However, despite these inconsistent pressure profiles, the PCast technique was able to fabricate comfortable fitting sockets for these five subjects. In our study, our subjectsÕ pressure profiles measured were different compared with RadcliffeÕs anticipated pressure profile. One reason could be that the line of action of the GRF was not be the only single factor in affecting pressure distribution. More subjects should be done to ascertain and determine the factors affecting the resulting pressure profiles.
J.C.H. Goh et al. / Clinical Biomechanics 18 (2003) 237–243
Acknowledgements This research was supported by the National Medical Research Council, Singapore. We would also like to thank Mr. Cheung Sze Kwong, Mr. Pan Seng Kie, Ms. Grace Lee, Mr. Wang Jit Beng, Mr. Hazlan Bin Sanusi and Mr. Azmee Murat for their technical assistance.
References Convery, P., Buis, A.W.P., 1993. Socket/stump interface dynamic pressure distributions recorded during the prosthetic stance phase of gait of a trans-tibial amputee wearing a hydrocast socket. Prosthetic and Orthotics International 23, 107–112. Fergason, J., Smith, D.G.S., 1999. Socket consideration for the patient with a transtibial amputation. Clinical Orthopaedics and Related Research 361, 76–84.
243
Kristinsson, O., 1992. Pressurised casting instruments. In: Proceedings of the 7th World Congress, International Society of Prosthetics and Orthotics, Chicago, USA. Kristinsson, O., 1993. The ICEROSS concept: a discussion of philosophy. Prosthetics and Orthotics International 17, 49–55. Lee, V.S.P., Solomonidis, S.E., Spence, W.D., 1997. Stump–socket interface pressure as an aid to socket design in prostheses for transfemoral amputees––a preliminary study. In: Proceedings of Institution of Mechanical Engineers, Vol. 211, Part H, pp. 167–180. Lee, P., Goh, J., Cheung, S.K., 2000. Biomechanical evaluation of the pressure cast (PCast) prosthetic socket for trans-tibial amputee. In: Proceedings of the World Congress on Medical Physics & Biomedical Engineering, Chicago, USA. Murdoch, G., 1965. The Dundee socket for below knee amputation. Prosthetic International 3, 12–14. Radcliffe, C.W., 1961. The biomechanics of below-knee prosthesis in normal, level, bipedal walking. Artificial Limbs 6, 16–24. Radcliffe, C.W., Foort, J., 1961. The patella-tendon-bearing belowknee prosthesis. Biomechanics laboratory, University of California, Berkeley, CA.