Secondary Logo

Journal Logo


Overview of Knee Disarticulation

Stark, Gerald BSME, CP, FAAOP

Author Information
JPO Journal of Prosthetics and Orthotics: October 2004 - Volume 16 - Issue 4 - p 130-137
  • Free


Older estimates of knee disarticulation have been placed as little as 0.7% to 2%1 of the total amputee population, but more recent studies have increased that estimate to 24%,2 especially in some rehabilitation centers such as Copenhagen and Toronto.3 Although many prosthetists are familiar with the contrasting characteristics of superior function but poor cosmesis, many do not fully utilize the advantages of end-bearing and self-suspension with alignment and socket design.

Dr. Pinzur of Loyola University3 has been an advocate for the knee disarticulation level because of the combined advantages of distal end-bearing, a longer lever arm, and surgical benefits. He contends that end-bearing is more physiological than the transfemoral or transtibial levels, which provide axial support in combination with circumferential tension and transverse musculoskeletal contouring. Knee disarticulation essentially places the biomechanic support at the base femoral condyles, a skeletal loading surface, which results in a more laterally displaced fulcrum. As a result, the trimline need not encapsulate the ischium with a more proximal medial trimline, but can be lowered to at least one-third of the total limb length. Dr. Pinzur also notes the importance of the longer lever arm for better sitting balance and leverage, especially important for older and bilateral patients (Figure 1). Also, there are significant surgical advantages because there is decreased surgical blood loss, increased resistance to infection, and definitively less bony or muscular disruption. Because knee disarticulation preserves the femoral epiphysis and avoids bony overgrowth, it is often used pediatrically. With growth-inhibiting techniques, the limb length eventually shortens to accommodate standard prosthetic componentry while preserving the end-bearing properties of knee disarticulation. Self-suspension over the condyles is another advantage that is often not fully exploited. A variety of interface designs have been developed that use pads, straps, or inflatable internal bladders directly over the medial femoral condyle rather than using standard transfemoral suction socket techniques. To take advantage of these advantages, the interface design should be decidedly different than common transfemoral designs with much more proximal trimlines. Many prosthetists attempt ischial and gluteal support with global circumferential loading that inhibits end-bearing.

Figure 1.:
Transfemoral patient with short limb lengths (A) has a smaller base of support than knee disarticulation. (B) A knee disarticulation patient shows an increased lever arm for sitting balance and transferring.

One of the more obvious disadvantages of a knee disarticulation prosthesis is that it is less cosmetic because of the prominence of the femoral condyles exacerbated by the thickness of the socket and liner. This can be alleviated with single-wall lamination techniques rather than double-walled systems with a separate interface layer. Other two step surgical techniques can block off and narrow the medial–lateral profile of the condyles but negate the self-suspension characteristics. Because the knee center is approximately 12 mm proximal to the distal end, the longer limb length also needs to be accommodated with a special distal end attachment and knee componentry.


The frontal plane biomechanics for knee disarticulation are different than that of the transfemoral prosthesis if there is true end-bearing. One of the goals of the transfemoral interface is to displace the fulcrum medially at the ischium rather than the physiological fulcrum at the hip joint to decrease the overall medial weight moment. Transfemoral two-point stabilization counteracts the medial weight moment with a small medial force applied at the ischium and slightly distal at the subischial triangle. A larger stabilizing gradient force is applied laterally along the femoral shaft to keep the patient upright (Figure 2).

Figure 2.:
Transfemoral frontal biomechanics uses two-point support with a medial force at the ischium and a gradient force along the femoral shaft to resist the medial weight moment about the fulcrum at the ischium.

With knee disarticulation, the fulcrum is placed more lateral and distal under the femoral condyles. There is still a medially directed weight moment, but the force picture is altered with a smaller lateral force at the femoral condyle and a larger gradient force along the medial wall. For this reason, the shape and flaring of the medial wall against the adductors become more critical for stabilization as the lateral femoral wall in transfemoral biomechanics (Figure 3). It is important that the medial wall be loaded during impression taking to simulate this altered two-point pressure loading. A common error is to treat a knee disarticulation patient as a transfemoral patient with a high medial trimline and circumferential tension, which inhibits end-bearing. With true end-bearing ischial or gluteal loading with high proximal trimlines is not necessary.

Figure 3.:
Because the fulcrum is at the distal femur, knee disarticulation frontal biomechanics uses two-point support with a lateral force directed at the femoral condyle and a gradient force along the medial wall to resist the medial weight moment.


When taking the impression, the two goals are self-suspension and end-bearing. Older methods suspended the patient with a loosely fitted Berkeley Quad Brim (Hosmer-Dorrance Corp., Campbell, CA) with a padded stand for end-loading.4 The load distribution was balanced approximately 50% in suspension and 50% end-bearing to achieve proper soft tissue compression4 (Figure 4). Another common technique creates a more anatomic shape by simply wrapping the limb circumferentially then loading it on a stand for full end-bearing. It should be noted that additional circumferential reduction is necessary because the volume is not placed in tension during casting. To cup the bony contours of the distal femoral condyles, the adjustable stand should have a soft foam rubber endpad. The presence of the patella should be noted, as well as its stability, to be used as a possible area of suspension. In certain cases such as the Gritti-Stokes procedure,5 the patella has been blocked and pinned distally to the femur as a bony loading pad. If the patella can be moved during the impression, the patient should attempt to move the patella within the mold to provide adequate relief in the interface.4

Figure 4.:
The knee disarticulation axial support is provided primarily by end-loading and secondarily by circumferential tension. Note the concave medial and lateral profile.

During the impression taking, an anterior cut strip might be needed to remove the impression because there will be contouring over the medial condyle, which could make removal difficult. Another method is to use a felt pad secured proximal to the condyles with double-stick tape that slides out as the residual limb is removed from the impression.4 Contouring should be made proximal to the medial condyle with relief for the adductor tubercle. The lateral supracondylar area acts primarily as a counterpressure to the medial condyle modification. It should be flattened and not compress the lateral supracondylar area to avoid nerve or ligamentous impingement. The lateral supracondylar area should be flattened primarily as a counterpressure to the medial condyle modification because some surface sensitivity could be present. The proximomedial wall should then be contoured as a concave biomechanic surface and the lateral surface should be flattened along the femoral shaft. No ischial or gluteal modification is necessary unless the user cannot tolerate full end-bearing. Slight flaring proximally increases comfort and alleviates the possibility of tissue impingement when donning.4

Modification involves removal of 6 mm of circumferential tension to the condyles, which are simply smoothed. The lateral wall should be flattened and concave when viewed in the frontal plane. The medial wall should also be concave with local reduction over the medial condyle but not so much so that it is difficult to push in for donning (Figure 4). Transversely, the interface should look somewhat triangular with adequate relief for the adductor longus, posterior angled medial wall, and flattened lateral wall. The most common error is not removing enough material along the medial wall, causing gapping at the lateral trimline4 (Figure 5).

Figure 5.:
The transverse view of the knee disarticulation interface superimposed over the anatomic upper third femoral cross-section. It illustrates the flattening over the femoral shaft, posterior–lateral angle, posterior–medial angle, and femoral triangle flattening. This is generally a less dramatic shape than most ischial–ramal containment designs. (Femoral cross-section from the Visible Human Project, US National Institutes of Health, National Library of Medicine, Used with permission.)


The original molded leather lace-up liner interface was uncosmetic, cumbersome, and difficult to make, but it could be easily custom-contoured to the medial and lateral profiles. Also, it could be tightened with variable tensions per the level of the interface. Because the knee was a free motion joint set at the physiological joint center, which is slightly more anterior, there is little inherent stability. A back-check and extension strap is also required to prevent terminal impact and as an extension bias, respectively (Figure 6A).

Figure 6.:
Schematic representations of interface options. (A) Outside hinges with lace-up corset and back-check. (B) Gel liner with air expulsion valve. (C) Gel liner with external binding catch for suspension. (D) Medial door suspension. (E) Pelite “stovepipe buildup” is split medially. (F) Integrated bladders or pads can be inflated pneumatically. (G) Hypobaric socks create a distal gasket or seal. (H) Zettl growth interface7 uses a distal strap for suspension and a proximal cuff attached with adjustable length uprights.

A common type of interface and suspension for knee disarticulation is a flexible gel liner in combination with air expulsion (Figure 6B). Although a pin is not recommended for space consideration, a recent attachment uses a plastic ski binding that is glued to the liner proximally, fed thru a hole in the socket wall, and attached with a spring catch. This creates a positive lock that is engaged even if the volume changes (Figure 6C). Although the use of a gel liner simplifies construction, it increases distal end bulk and does not use the condylar suspension characteristics. As a result of aggressive circumferential tensions advocated for gel liners, patients often do not get the full end-bearing benefits because they are suspended within the interfaces.

Interfaces that use a medial door (Figure 6D) with an external strap can apply variable stability with direct condylar suspension. Because there is no extra liner, they can be constructed with a fairly thin wall and decreased distal bulk compared with other options. For this suspension to be functional, it is important that the patient be well healed with a prominent medial condyle. Like with transtibial condylar suspension, if the patient is too fleshy or muscular, it could be difficult to achieve good condylar suspension.

A forerunner to the gel liner interface was the stovepipe Pelite buildup that was adapted from Syme’s-type construction (Figure 6E). With this method, wedges are built up around the condyles to the largest circumference until the limb is more cylindrical in nature (similar to a stovepipe). Small slits are cut over the medial condyle (with holes at either end for stress relief) to help don the foam liner. This system is easy to fabricate and provides good suspension and comfortable padding, especially for more bony limbs. A disadvantage is that it increases wall thickness and bulk distally. Another adaptation is to create a split distal cup that extends to the distal third of the limb. Although less bulky, this could be uncomfortable to the patient at the transition line within the interface and create a small adductor roll.

Pneumatic pads or silicone bladders such as the Lange-Type Inner Sleeve6 (Figure 6F) have been used with some success. They provide excellent suspension localized over the condyles, but add slightly to wall thickness. The greatest disadvantage is that the fabrication is more involved with the multiple layers of silicone and laminate. Also, it is recommended that the residual limb be stabilized with respect to volume because it can be difficult to accommodate great changes in circumference.

Hypobaric socks provide a flexible distal seal that integrates partial suction suspension with and the friction relief and breathability of a sock. It can be recommended for those who cannot tolerate the contact of true suction suspension and desire push-on donning because of upper extremity involvement. Because the volume reduction is not aggressive, it is not considered full suction suspension and requires a proximal elastic suspension belt (Figure 6G) that increases the proximal bulk more than a simple transfemoral prosthesis.

One of the more novel interfaces for knee disarticulation is the Zettl growth interface7 as described in Chapter 19B of the Prosthetic Atlas (Figure 6H). It uses a distal end cup with a more proximal cuff connected by parasagittal bars that can be extended for growth compensation. Suspension is provided by a simple supracondylar strap wrapped around the femoral condyles and tightened. This is a unique solution for juveniles undergoing growth changes or patients who require a more open interface with less skin contact.


Although all categories of knee designs are available for knee disarticulation, many cannot be used because they must contend with the main issue of limb length. Previously, outside hinges were used because the anatomic knee axis is proximal to the distal end of the residuum. These knees provided little stability because they had to be placed near the physiological knee center and had no friction control. The extension bias and terminal limit were set using external back-check and elastic extension straps, respectively. Outside hinges are also restricted to exoskeletal construction unless special attachments are created to secure the uprights directly to a pylon. One exoskeletal setup of note offers a hydraulic control unit connected posteriorly to the joints with a fitting yoke.

Single-axis knees present the widest array of control mechanisms ranging from friction, pneumatic, and hydraulic to other features such as stance control. Unfortunately, there is a greater distance between the anatomic and prosthetic knee axis. Although more recent knee designs have minimized the distance from the proximal attachment to the knee axis, the prosthetist must choose between a low-profile 1-inch discrepancy with no alignment capability or 2-inch discrepancy to include a pyramid attachment. With a bulbous distal end, this often this results in an uncosmetic shortened lower shin section that swings freely when sitting.

Polycentric knees enjoy the greatest popularity for knee disarticulation because they locate the instant center more proximally and provide toe clearance during swing phase. They do this by allowing the lower shin section to tuck around the distal femur (similar to our physiological knee but with a magnified kinematic movement). Also, polycentric knee design varies greatly. Knees optimized for knee disarticulation are not necessarily considered inherently stable because their motion allows more anterior translation as the knee is flexed. This maximizes the amount of toe clearance and makes the knee extremely easy to flex, but its “triggered position” requires the patient to maintain stability with hip extensors. This is usually not an issue because knee disarticulations generally have a large amount of voluntary control as a result of their limb length and hip extensor strength.8 The instant center is found by projecting a line proximally through the polycentric links viewed in the sagittal plane (Figure 7A). As the instant center is plotted by progressively flexing the knee, the pathway or centrode is developed9 (Figure 7B). Qualitative and quantitative examination of the centrode can tell us how the knee will work during flexion.8 There are actually two centrodes, one holding the thigh section stationary and one holding the shin section stationary. The motion of one centrode over the other gives us the knee’s true motion during swing8 (Figure 7C). This analysis can be extended to tell us the toe clearance properties as well (Figure 7D). However, the centrode data are often not available or clinically practical to judge motion characteristics of polycentric knees. For that reason, certain visual cues can be used to compare the various polycentric knees. A vertical anterior link is indicative of a knee design with a more forward and proximal instant center. This makes the knee feel somewhat “triggered” and easy to flex and shortens the lever arm to control flexion because the center is closer to the hip extensors (Figure 8). A longer posterior link is also present, which pushes the centrode more forward and increases the anterior gliding as the knee is initially flexed. A posterior link that is appreciably shorter than the front link creates a knee motion that collapses distally more quickly and does not present as much toe clearance (Figure 9). Polycentric knees for knee disarticulation typically have longer links that also allow for greater anterior–posterior translation and toe clearance. This also makes it difficult to cover the prosthesis with a one-piece cosmetic cover. A common error is choosing a knee that is too stable and inhibits knee flexion late in stance or a knee that does not compensate for limb length.

Figure 7.:
(A) The instantaneous center is found by projecting the two side links proximally to find their intersection. (B) The path of the instantaneous center can be plotted as the knee is flexed. This path is called the centrode. (C) There are two centrodes, one with the thigh stationary (light) and one with the shin stationary (dark). The motion of the knee is most accurately described as one centrode rolling on the other in a cam fashion. (D) Toe clearance, important for knee disarticulation, can be plotted for various designs. The dark path indicates the toe path of a single axis knee, which effectively lengthens during swing. The light path indicates the relative shortening with data showing the exact shortening for the corresponding degree of knee flexion. (Pictures from the CENTRODE plotting software, Courtesy of Charles Radcliffe, PhD. Used with permission.)
Figure 8.:
(A) A stabilized polycentric knee centrode that is well posterior of the sagittal alignment reference line. It is smaller and parabolic in nature, showing a loss of proximal positioning for voluntary control. (B) A more “triggered” polycentric knee centrode showing proximal positioning during the first 5° of flexion preserving voluntary control. (C) A centrode showing an enlarged parabolic shape that still preserves proximal positioning as a result of its magnified size. (Centrodes from the CENTRODE plotting software, Courtesy of Charles Radcliffe, PhD. Used with permission.)
Figure 9.:
(A) A knee mechanism specifically adapted for knee disarticulation. Note the longer links for greater translational gliding and the longer posterior link that pushes the centrode more anterior. The anterior link is only slightly retroverted for a more proximal initial center that is easy to flex. (Courtesy of USMC/Seattle. Used with permission.) (B) Stabilized knee design not necessarily recommended for knee disarticulation. Note the retroverted front link that positions the centrode and initial center well posterior for greater involuntary stability. The posterior link is shorter than the anterior link describing a parabolic centrode that shows less anterior–posterior translation and relative shin shortening. (Courtesy of Otto Bock Healthcare, Minneapolis, MN. Used with permission.) (C) Knees with greater anterior–posterior gliding show a greater tendency to “wrap around” the thigh section and toe clearance. (Courtesy of Hosmer-Dorrance Corporation, Campbell, CA. Used with permission.)

By recognizing the differences with transfemoral level and using the inherent advantages of knee disarticulation, the prosthetist can optimize prosthetic use.


1.Bennet A. Limb Prosthetics, 6th ed. New York: Demos Publications; 1989.
2.Pernot HF, Winnubst GM, Cluitmans JJ, DeWitt LP. Amputees in Limberg: Incidence, morbidity and mortality, prosthetic support and functional level after one year. Prosthet Orthot Int 2000;24:90–96.
3.Pinzur M. Knee disarticulation/surgical procedures. In: Bowker J, Michael J, eds. Atlas of Limb Prosthetics, 2nd ed. St. Louis: Mosby Year Book; 1992:479–486.
4.Karolewski T. The Impression and Modification for the Knee Disarticulation Interface. Lecture Notes. Chicago: Northwestern University Prosthetic-Orthotic Center; Spring 2004.
5.Faber D, Fielding P. Gritti-Stokes (through-knee) amputation: Should it be reintroduced? South Med J 2001;94:997–1001.
6.Lange L. The Lange partial foot prosthesis. J Prosthet Orthot 1992;4:56–61.
7.Michael J. Knee disarticulation/prosthetic management. In: Bowker J, Michael J, eds. Atlas of Limb Prosthetics, 2nd ed. St. Louis: Mosby Year Book; 1992:487–499.
8.Radcliffe CW. The Knud Jansen lecture: Above knee prosthetics. Prosthet Orthot Int 1977;1:146–160.
9.Greene MP. Four bar linkage knee analysis. Orthot Prosthet 1983;37:15–24.

biomechanics; interface design; knee disarticulation; pediatric prosthesis; polycentric knee

© 2004 American Academy of Orthotists & Prosthetists