Many prosthetists have recognized the problem of the abducted residual femur and have attempted to use prosthetic solutions to try and resolve the problem.9,12,13 In recent years new prosthetic socket designs have been introduced to improve wearer comfort and gait. Some of the newer socket concepts claim that by making the mediolateral dimension narrow and the anteroposterior dimension larger, and incorporating the ischium within the brim of the socket, they can hold the residual femur in a more adducted position and allow the amputee to walk in a more normal fashion.
It has been stated that the quadrilateral above-knee sockets create a lateral displacement of the center of gravity with an accompanying side lurch, due to the inherent design. Critics state that this is in part a result of the relatively larger mediolateral versus anteroposterior dimension, which allows abduction of the femur in the socket.9,12 Consequently, the patients may walk with a wide based gait and lurch to the amputated side, which increases energy expenditure.
The proponents of the ischial containment sockets offer minimal subjective evidence and no objective data of improved patient function and comfort, especially if the factors of improved knee components and energy-storing feet are also considered. No tests of better functional restoration or gait improvement have been conducted, and no data on long-term acceptability or use of these sockets has been provided. In addition, no objective evidence is provided that the femur is truly "held" in an adducted position in the newer sockets, especially during walking and load transfer after heel contact. Observation of amputees wearing different designs of the ischial containment sockets do not show drastic and lasting improvements in the gait of these amputees.
The purpose of this study was to determine the position of the residual femur in the above-knee prosthetic sockets of various types, to highlight the reasons for malalignment of the residual femur, and to recommend procedures and methods that can restore the anatomical position of the stump, both statically and dynamically. This should then allow for easier and better prosthetic fitting, improved function, and a more energy efficient gait.
Fifty above-knee amputees were studied and divided into two groups according to the type of socket used, either quadrilateral or ischial containment. The ischial containment group included CAT/CAM (Contoured Adducted Trochanter/Controlled Alignment Method), NSNA (Normal Shape, Normal Alignment) sockets and narrow medial-lateral ischial containment sockets. The quadrilateral group included patients with hard sockets and flexible sockets. The criteria for inclusion in the study were that the patients have mature stumps and be everyday prosthetic users. The position of the stump out of the prosthesis was checked with the patient standing to see if it was in abduction or adduction. The femur was also palpated to determine its distal end and approximate position within the stump. Anteroposterior radiographs were taken of the femurs from hip to knee, with the stump in the prosthetic socket and the patient standing with the weight equally distributed on both legs and the feet 2" apart (Figure 1 and Figure 2 ). The pelvis was checked to be level. The anatomical alignment of the normal and the amputated femur was measured on the radiograph. The length of the amputated femur was measured from the middle of the lesser trochanter. The overall socket alignment and the angle of socket axis were measured by taking the midportion of the socket on the radiograph at three different levels and taking a straight line through this. The angle to the perpendicular was then measured.
An analysis of variance was done, comparing the femur positions in the different sockets. A biomechanical model of the adductors was developed to help explain the problem of the abducted femur.
The 50 patients were divided into two groups according to the type of prosthetic socket worn. These groups were comparable for age, sex, and cause of amputation (Table 1 ). There were more males than females in the study.
The anatomical axis of the normal femur was the same in both groups of patients and averaged from 80° to 100° of adduction (Table 2 ). The anatomical axis of the amputated femurs had a mean of 2° of abduction in the quadrilateral and ischial containment socket groups. This was on average, 12° from the normal anatomical axis. In neither of the two groups was the residual femur consistently in any one position. As shown in Table 2 , the position of the residual femur in the quadrilateral sockets varied from 8° of adduction to 12° of abduction, while in the ischial containment sockets the femur position varied from 8° of adduction to 14° abduction. The configuration of the prosthetic socket did not appear to influence the position of the femur in the socket as some patients in quadrilateral sockets had more femoral adduction than any patient in an ischial containment socket (Figure 3 , Figure 4a and Figure 4b) . The average length of the residual femur was the same in the two groups of patients. Socket axis did vary between the two groups and was noted to be neutral in those patients with quadrilateral sockets and from neutral to 9° adduction in the ischial containment socket group.
An analysis of variance of the anatomical position of the residual femurs in the two sets of patients revealed no statistically significant difference in the abduction angles of the amputated femurs between quadrilateral and ischial containment sockets. The biomechanical study of the adductors of the thigh showed that adductor magnus provides over 70% of the adduction moment to the normal leg, and that this mechanical advantage is lost in distal third above-knee amputations and in higher levels.
It is widely accepted that above-knee amputees have an increased energy expenditure of approximately 65% above normal for level walking. Even those patients who have remained physically active exhibit an increased oxygen consumption.5,14 After amputation, older patients with little or no physical reserve may lose the ability to walk again because of the increased energy requirements. The reduced economy of walking results in part from an increase in the displacement of the center of gravity of the body because of a functional malalignment of the amputated femur in the prosthetic socket.
In normal two legged stance the mechanical axis of the lower limb runs from the center of the femoral head through the center of the knee to the midpoint of the ankle.3,4,10,15 Due to the slight genu valgum, this mechanical axis forms an angle of varus as measured against the vertical axis of the body. The tibiofemoral angle is the angle formed by the anatomical axes of femoral and tihial bones, and averages 3° (Figure 5) . The femoral shaft axis forms an angle with the vertical axis which measures 9° These three alignments show the femur to be in adduction and allow the hip stabilizers and hip abductors to function normally and reduce the lateral motion of the center of mass. This results in a smoother and more energy efficient gait.
In most above-knee amputees, however, this set of alignments is disturbed since the residual femur no longer has its natural anatomical alignment, and the femoral shaft axis is held in an abducted position as compared to the nonamputated contralateral limb. The abduction of the above-knee amputation stump and its associated problems of side lurch and high energy consumption has been recognized by many prosthetists. The deviation of the residual femur from its natural anatomical position was noted to be the cause of the problem, but this was attributed to the design and shape of the prosthesis, most commonly the quadrilateral socket, and, therefore, a prosthetic solution for this problem was sought. The newer ischial containment sockets are an attempt to bring the residual femur back into its normal alignment, restore the function, and enable the amputee to walk with a narrow base.
However, the results of our observations and measurements as presented above do not support or confirm this claim. The abducted position of the femur in the ischial containment sockets, as well as the large variation of the position of the residual femur, and the lack of objective evidence about the functional superiority of these sockets suggest the problem is not primarily a prosthetic one, and therefore not amenable to a purely prosthetic solution.
In conventional above-knee amputations, the patients usually lose the major portion of the adductor attachment. Only the adductor magnus insertion has an attachment to the medial distal third of the femur.2,8 Once this attachment is lost, the femur swings into abduction because of the relatively unopposed action of the abductor system. This occurs at the time of surgery and the surgeon then sutures the residual adductors and the other muscles around the femur while the stump is in the abducted position.
Although above-knee amputations may be done for a variety of pathological problems, the standard surgical texts only highlight the anatomy that is encountered during the performance of the operation.1,6,7,11 No emphasis is placed on the importance of myodesis in the stabilization of the residual femur to restore the balance of forces for dynamic control of the stump. Many authors advocate myoplasty as a suitable alternative to myodesis, but fail to mention that many active patients may have discomfort from the unanchored muscle mass.1,6,7,11
A myoplasty involves suturing muscle to muscle over the end of the bone without anchoring the muscle mass to the distal bone. A better anchorage for the sectioned muscle is a myodesis which generally allows for better muscle stabilization and appears to allow for better control of the residual limb.
The higher the amputation, the greater the loss of adductor power, and the greater the mechanical advantage of the unopposed tensor fasciae-latae with its long attachment through the fasciae-latae tendon and inter-muscular septum to the linea aspera (Figure 6 and Figure 7 ). This loss of position of the femur is not seen in patients who have a through knee amputation, as the balance between the adductors and abductors is maintained.
Therefore, in an above-knee amputation we propose that the best way to maintain the normal anatomical axis of the femur is for the surgeon to actively overcorrect the alignment at the time of the definitive above-knee amputation. The higher the amputation, the greater the need for the overcorrection. As the attachments of the adductor muscles are lost, their moment arm becomes shorter (Figure 8 ). Thus, a smaller mass of adductor muscle will have to generate a larger force to hold the femur in its normal position. In conventional above-knee amputations, the residual adductor muscles invariably have been reattached to the femur with the femur in an abducted position, and they are therefore unable to generate sufficient force to pull the femur over and maintain the natural anatomical position. The muscle envelope surrounding the femur cannot be displaced by external forces. Once the femur is in an abducted position within the muscle compartment, no amount of lateral pressure by the prosthesis will alter the existing configuration. In the majority of cases that we examined the residual femur was not in the normal anatomically adducted position. The socket configuration did not affect the position of the femur within the socket. We consider the prosthetic socket analogous to an orthosis, and just as an orthosis cannot correct a deformity, so the prosthetic socket is unable to correct the abduction deformity of the residual femur.
The success of the prosthetic fitting, i.e., the optimal restoration of function and comfortable ambulation, depends on the anatomical alignment and dynamic functioning of the above-knee amputation stump. Successful prosthetic fitting starts at the time of surgery. The amputation procedure should be planned and executed based on the biomechanics of the residual limb. Even though the authors have asserted that myodesis is important, if performed with the femur already in abduction, the benefits will not be noticed. Myodesis of the adductors should be done with the residual femur in a more adducted position. This should allow for a more anatomical fitting of the above-knee socket and will also reduce the lurching gait of the above-knee amputee.
This study was unable to substantiate claims that the ischial containment sockets are able to hold the residual femur in adduction in the socket. We believe that socket configuration does not influence the position of the residual femur and that a proper surgical procedure for above-knee amputations is essential for functional restoration and success of prosthetic fitting. Although the narrow mediolateral socket concept has some merits, the anatomical alignment of the femoral bone should be achieved by proper myodesis of the adductor muscles at the time of surgery. lf the amputation is performed with regards to the biomechanics of the residual limb, any type of properly fabricated prosthesis has the potential to provide functional restoration and comfortable ambulation.
Frank A. Gottschalk, M.D., FRCS Ed.. FCS(SA) Orth., is with the Division of Orthopaedic Surgery, University of Texas Southwestern Medical Center, Dallas, Texas and the STAMP Program (Special Team for Amputation, Mobility and Prosthetics), Veteran's Administration Medical Center, Dallas, Texas, and Dallas Rehabilitation Institute, Dallas, Texas.
Sohrab Kourosh, Ph.D., P.E., is with the Division of Orthopaedic Surgery, University of Texas Southwestern Medical Center, Dallas, Texas and the STAMP Program (Special Team for Amputation, Mobility and Prosthetics), Veteran's Administration Medical Center, Dallas, Texas, and Dallas Rehabilitation Institute, Dallas, Texas.
Melvin Stills, C.O., is with the Division of Orthopaedic Surgery, University of Texas Southwestern Medical Center, Dallas, Texas and the STAMP Program (Special Team for Amputation, Mobility and Prosthetics), Veteran's Administration Medical Center, Dallas, Texas, and Dallas Rehabilitation Institute, Dallas, Texas.
Bruce McClellan, C.P.O., is with the Division of Orthopaedic Surgery, University of Texas Southwestern Medical Center, Dallas, Texas and the STAMP Program (Special Team for Amputation, Mobility and Prosthetics), Veteran's Administration Medical Center, Dallas, Texas, and Dallas Rehabilitation Institute, Dallas, Texas.
Jim Roberts, P.A., is with the Division of Orthopaedic Surgery, University of Texas Southwestern Medical Center, Dallas, Texas and the STAMP Program (Special Team for Amputation, Mobility and Prosthetics), Veteran's Administration Medical Center, Dallas, Texas, and Dallas Rehabilitation Institute, Dallas, Texas.