Damian Mc Cormack,

International Fellow in Paediatric Orthopaedic Surgery,

Atlanta Scottish Rite Hospital,

Georgia, USA.

There are currently over 100 different prosthetic knee designs available to the orthopaedic surgeon. The following is an account of relevant gait biomechanics and a functional classification of these knee mechanisms. It is important to match a component which offers the optimum function and reliability to the amputee.

Saggital Plane Biomechanics of Gait.

Prosthetic knees offer no special function in the coronal or transverse planes beyond simple stability. All their functional distinctions occur in the saggital plane and therefore it is necessary to review saggital plane biomechanics of normal 4 and then "prosthetic" gait.

The knee joint, both biological and prosthetic, must facilitate 1) weight acceptance, 2) single limb support and 3) limb advancement 5.

The stability of a prosthetic knee is most important during these three periods and in particular during stance. Because stance phase begins and ends with periods of double limb support, maximum stability is required during single limb support, which constitutes less than half (40%) of the gait cycle.

The ideal prosthesis should mimic the alignment and gait characteristics of the normal limb during each of the phases of the gait cycle.

Initial Contact:

This begins the stance phase and initial double support. The hip is flexed, the knee extended and the ankle dorsiflexed to neutral. Prosthetic components can replicate this position easily.

Loading Response:

This occurs as the body carries forward and includes elements of shock absorption, weight-bearing stability and preservation of forward motion. The knee gradually flexes, using the heel as a rocker, until the forefoot contacts the floor. This action is simulated by the use of a cushion heel or leaf spring heel offering appropriate resistance. This shock absorbing effect maintains knee stability but tends to retard forward progression. Most knee prostheses require full knee extension for stability and therefore raise the center of gravity during stance phase. Only a few more sophisticated designs allow stable weight bearing in flexion.


The body progresses forward, using the ankle as a rocker, into single leg support, with a stable limb and stationary foot. Solid ankle-feet prostheses with a flexible or dynamic response keel simulate this transition, as do some multiaxial articulated prostheses. This period of single limb support demands maximum knee stability, which is provided prosthetically in several ways.

Terminal Stance:

The body progresses forward over the forefoot rocker, as the heel rises, to complete single limb stance. The hip extension increases as the limb lags behind the body and the knee flexes a little in preparation for swing. Prosthetic designs which do not allow this knee flexion under load delay swing phase and thus impede forward motion.


This is the second period of double support in stance phase. The ankle plantar flexes, the knee flexes and hip extension decreases in preparation for swing. Only those prosthetic knees which allow flexion during load bearing will simulate this response at the knee and only dynamic response feet simulate ankle plantar flexion by the rebound of a plastic keel spring2.

Initial Swing:

Increasing hip and knee flexion advance the limb as the ankle gradually dorsiflexes toward neutral. Many prostheses cannot adequately control heel rise during the initial swing phase.


The ankle continues to neutral as the hip continues to flex. Gravity takes the knee into extension. Most prosthetic designs simulate this well unless there is excessive heel rise in initial swing which delays forward progression of the shank.

Terminal Swing:

Terminal swing begins with a vertical tibia and ends when the foot touches the floor. Limb advancement is controlled by knee extension. Many prostheses cannot adequately decelerate the shin to prevent terminal impact as full extension is reached. This disrupts swing phase mechanics and increases energy costs.

In terms of saggital plane dynamics the ideal knee mechanism would;

1. be stable enough to accept weight in early stance

2. absorb shock and allow smooth forward progression of the body via controlled knee

flexion during weight acceptance

3. support body weight in midstance with a flexed knee

4. begin flexion during single limb weight bearing late in terminal stance

5. respond instantly in the swing phase to a faster pace and to variable cadences.

A Functional Classification of Knee Mechanisms.

Michael 3 has proposed a simple functional classification of prosthetic knee mechanisms. The five types are classed according to their major functional characteristics. They are

1. Constant Friction prostheses

2. Stance Control prostheses

3. Polycentric knees

4. Manual Locking prostheses

5. Fluid Controlled devices.

1. Constant Friction Prostheses

This design group ( "single axis" prosthesis) is the oldest historically and consists of a simple axle connecting the thigh and shank segments. These prostheses are relatively inexpensive and simple to manufacture. Modern versions, such as that manufactured by Otto Bock, have an adjustable friction cell and spring loaded extension assist to improve swing phase function.

Constant friction knees are best for level ground walking at constant speed but demand sufficient hip power to prevent the knee from buckling. More athletic amputees find this simple design too restrictive.

Biomechanically the Constant friction prosthesis gait resembles that of a patient with a flail leg ( eg polio victim). The requirement in both is to keep the ground reaction line in front of the knee from initial contact through midstance in order to maintain a stable extended knee joint. This ground reaction line should pass behind the knee in terminal stance to ensure ease of knee flexion. Therefore the optimal setting in a constant friction prosthesis maintains the ground reaction line within the above parameters.

If a patient lacks hip power and cannot maintain an extended knee in early stance the prosthesis may be adjusted into " hyperextension", by moving the knee center backwards. However this makes knee flexion more difficult during swing phase. The patient must fully unload the knee in order to flex it and this creates the characteristic delayed and abrupt knee flexion on entering the swing phase.

The Constant Friction prosthesis provides only a single fixed cadence during swing phase and therefore if a patient increases his or her walking speed the heel will rise excessively and prolong the swing phase. This encourages the patient to extend the contralateral stance phase by excessively plantarflexing the ankle. In other words he vaults over his prosthesis, not because it is too long, but because it is prolonging swing phase on that side. If this patient tries to run he or she hops off the biological leg as this effect is exaggerated. This was the gait pattern demonstrated by Terry Fox, a now famous amputee who attempted to jog across Canada several years ago.

The final problem with the Constant Friction knee is its tendency to give way on declines and on uneven ground.

Stance Control prostheses.

This knee prosthesis uses a weight activated braking mechanism which adds resistance to bending during stance only. This consists of a spring loaded brake bushing which binds when loaded during stance but is released during swing. The amount of "friction lock" is adjustable. However the brake tends to wear over time and no such device can support full body weight in extreme flexion. The amputee must also delay knee flexion until the device is fully unloaded during swing and this produces an inefficient gait. The device must be fully unloaded before sitting down. This makes it virtually impossible for a bilateral amputee to use Stance Control prostheses. Boimechanically this knee type best suits the elderly patient with poor hip control. Despite the need for periodic maintenance the Stance Control prosthesis remains very popular.

Polycentric Knees

These complex designs comprise multiple centers of rotation. Many have four pivot points and are referred to as " 4 bar linkage" devices. Essentially this consists of paired anterior and posterior , superior and inferior hinges linked together. Mechanically the summation of the potential polycentric rotations will determine an instantaneous center of rotation peculiar to a particular device. The stability in polycentric devices is described in terms of " and stability". stability is determined by the distance that the instant center of rotation is behind the ground reaction line. The greater the distance the greater the inherent stability of the device during stance, just as for the above two types of device. The distance that the instant center of rotation is above the joint line determines the amount of voluntary control the patient has over the prosthesis and is referred to as the stability 1.

Most Polycentric Knees have their instant centers of rotation quite proximal and posterior for greater stability. Their stability is inherent in their design and not dependent on a brake bushing like the Stance Control device discussed above.

The instant center of rotation moves forward quickly in the swing phase, thus unlocking the joint and facilitating flexion but still offering excellent stance phase stability which allows load bearing during flexion. The polycentric knees shorten slightly during flexion thus adding additional toe clearance during midswing.

A specific modification of the polycentric knee is available for the knee disarticulation patient, which has long linkage bars placed below the joint line. This offers cosmetic but not mechanical advantage.

Manual Locking prostheses

This device offers ultimate stability but is seldom required and produces an uncosmetic and energy - consuming gait pattern. It is useful for the manual laborer who demands stability in the limb. The remote release cable requires a free hand to release it prior to sitting; bilateral device require both hands. The patient falls into the chair with sudden release of both prostheses. Manual locking devices are rarely used.

Fluid Controlled Devices.

These devices utilize a fluid ( silicone oil) or gas filled piston which offers automatic hydraulic or pneumatic cadence control respectively. Fluid filled hydraulic devices are stronger. The device allows the amputee to vary their cadence at will. These devices produce the most normal gait parameters. They are relatively heavy and expensive.

All five device types may be incorporated within a prosthesis with a soft skin like covering ( Endoskeleton) or may be left " exposed" as an Exoskeleton. The exoskeleton "bionic " look seems to have caught the imagination of the American public at least.

Many of the more recent knee prosthesis designs are hybrids which combine some of the properties of the above groups. Otto Bock, for instance, produce a titanium polycentric device which incorporates a mini hydraulic unit for swing phase control. Blatchford, U.K, have produced " bouncy" knees which control knee flexion during stance. Several "intelligent " knees are now available which incorporate microprocessors!


Orthopaedic surgeons must remain familiar with the biomechanics of pathological gait and prosthetic design if they are to correctly assess and prescribe for their amputees. The ever growing prosthetic industry has moved beyond the manufacture of simple medical devices and is developing ever more sophisticated prostheses for recreational extremists. Just as in the field of arthroplasty it behooves the orthopaedic surgeon to maximize the biological setting and prescribe the simplest and most suitable device rather than bend to the pressures of prosthetic commercialism.

Selected References

  1. Green MP. Four bar knee linkage analysis. Orthot Prosthet 1983; 37: 15- 24.
  2. Michael JW. Component selection criteria: Lower limb disarticulations. Clin Prosthet Orthot. 1988: 12; 99-108.
  3. Michael J W. Overview of prosthetic feet. In Green WB (ed): Instructional course lectures. 1990; Vol 39: Chicago, A.A.O.S. 367-372.
  4. Perry J. Gait Analysis- Normal and Pathological Function.Thorofare, NJ, Slack, 1992
  5. Radcliff CW: The Knud Jansen lecture- above knee prosthetics. Prosthtic Orthot Int . 1977; 1: 146 - 160.