*Polyclinique le Languedoc - NARBONNE, **S.M.E. - Mauguio

The recent arrival on the market of several knee replacements with mobile bearings has drawn renewed attention to a complication that is well recognized, albeit rare, in fixed-bearing TKRs: dislocation.

A comparative analysis of the new designs involves a number of factors, such as the implant’s degree of congruency, its degrees of freedom, and implant stress levels.

One very important aspect is the dislocation potential of the mobile bearing. The risk of dislocation is difficult to assess. As the designers of the Tri CCC mobile-bearing prosthesis, we have been particularly interested in this problem. The literature on the subject is as yet scanty, and follow-up to date has been short. This is why we thought it useful to propose a classification of the different types of dislocation; to investigate the factors leading to the different patterns; and to establish criteria for the assessment of the dislocation risk, using known or novel measures that describe the stability of a prosthesis or of a mobile bearing.


Background : dislocation in fixed-bearing TKRs

In the 70s, John Insall designed the Total Condylar prosthesis, a tricompartmental surface replacement that allowed the anterior and posterior cruciate ligaments to be sacrificed. It was one of the first resurfacing TKRs. Initially, the model was not an unqualified success: Insall realized very soon that the clinical results were suboptimal, because of gradual posterior subluxation of the tibia leading to loss of ROM, diminished quadriceps extension strength, and a feeling of instability when descending stairs or rising from a chair.

Figure 1
1a- normal position
1b - anterior tibiofemoral dislocation

Together with Burstein, Insall next designed a posterior-stabilized system, which is still the “gold standard.” This design involved a cam between the condyles of the femoral component, and a spine arising from the tibial component.

This surface geometry soon proved effective in terms of tibiofemoral kinematics, and overcame the problems associated with the earlier design, by improving the functional outcome (greater ROM, stability when negotiating stairs or arising from sitting, greater quadriceps strength).

There was, however, a complication that had not been encountered before: tibiofemoral dislocation. As the intercondylar cam on the femoral component pushed forwards, it could, under certain circumstances (flexion, loss of bearing surface contact), ride anteriorly over the tibial spine and dislocate (Fig. 1).

Many colleagues used to employ these posterior-stabilized devices, and may have observed this complication in their patients.

The problem was also described in a large number of papers (Table 1).

Dislocation Rates Reported With Insall-Burstein I and Insall-Burstein II Knees





Incidence of Dislocations

Dislocation Rate (%)

Incidence of Dislocations

Dislocation Rate (%)

Galinat et al., 1988





Cohen et al., 1991





Striplin & Robinson, 1992





Cohen & Constant, 1992





Lombardi et al., 1993











Dislocation involves a loss of congruency between the femoral and tibial implant components, with the femur riding over the stationary tibia. This is why this complication is seen in both fixed-bearing and mobile-bearing devices.

Biomechanics of dislocation

The biomechanical conditions under which a posterior-stabilized prosthesis with a fixed or a mobile bearing will dislocate are similar. With forced flexion - when the patient arises from a low seat or puts on his or her shoes - the femur is pushed forwards and may rise sufficiently high to ride over the tibia and dislocate forwards, to finish up in front of the tibial spine designed to provide posterior stabilization.

Causes of dislocation

Ligamentous instability

Figure 2.  Ligament balancing (Insall)

Insall was the first to stress the importance of ligament balancing (Fig. 2), to ensure good ligament tension in flexion and in extension.

In all probability, inadequate ligament balancing is also to blame for implant dislocation, since the complication can be prevented by balanced tension in the ligaments.

The gaps created by bone resection must be symmetrical, and, above all, equal in flexion and in extension. A polyethylene insert of appropriate thickness will ensure that the ligaments are well balanced, and will keep the implant components in sufficiently close contact to prevent dislocation.

Design of the posterior-stabilization spine

Instability in flexion will be tolerated by different implant designs with a greater or lesser degree of forgivingness, depending on the design of the posterior-stabilization system of the implant concerned.

In order to resist anteroposterior stresses, the posterior edge of the tibial spine should be vertical. That was the pattern adopted by Insall and Burstein.

Figure 3.  Three tibial-spine designs.

Some devices have an oblique spine. This pattern offers less resistance to the advancing femur, and the forward movement of the femur will continue throughout flexion. In fact, the spine will act as a springboard and give added impetus to the femur on its anteriorly dislocating course (Fig. 3).

A measure of stability: the jump distance (or Dislocation Safety Factor)

For maximum safety in flexion, the top of the tibial spine should be as high above the bottom of the femoral cam as possible. This Dislocation Safety Factor (DSF) may be calculated by measuring what has also been called the “jump distance” - the distance by which the femoral component would have to rise in flexion so as to clear the tibial component and dislocate forwards (Fig. 4).

Figure 4.  DSF or jump distance.
ormal position | Dislocated position

This parameter has been studied in detail in primary Insall-Burstein I and II prostheses. The DSF was found to be greatest (14.5 mm) at 70 degrees of flexion, and to decline rapidly with further flexion, to reach ca. 8 mm at 120 degrees.1

The double paradox of the Insall-Burstein TKR


While the safety margin has to be high at maximum flexion, where the risk of dislocation is greatest, the DSF actually decreases between 70 and 120 degrees of flexion.

This is due to the fact that the Insall-Burstein system of posterior stabilization was designed with the emphasis on roll-back. To this end, the cam was given an oval cross-section, which will make the femur roll back gradually as the knee is flexed. Past 90 degrees of flexion, the cam will present its most unfavourable diameter, and the DSF will, therefore, decrease.


The IB II has a higher dislocation rate than the IB I.

This is due to the design policy of increasing the maximum flexion provided by the IB II as compared with that provided by the IB I. In order to obtain this greater ROM, the cam of the posterior-stabilization system was raised.


A special pattern: implants with a third condyle

Some designs involve a posterior-stabilization system with an additional articulation in the centre of the joint. In this pattern, the intercondylar portion is cylindrical, and articulates with a central well in the tibial component. Different devices have greater or lesser congruency of this articulation.

This central articulation is intended to take some of the stress imposed on the knee, by providing an additional contact surface. The third articulation also provides posterior stabilization.

In one fixed-bearing TKR with a spur-shaped third condyle, the DSF is constant past 30 degrees and up to 120 degrees of flexion. The actual value of the DSF was found to be 13 mm (Fig. 6a).

Figure 6.  DSFs measured in three different TKR designs:

6a - DSF = 13 mm past 30° of flexion;

6b - DSF = 12.6 mm throughout;

6c- Tri CCC: DSF = 14 mm throughout.

Another model, with a rotating platform and a cylindrical third condyle which engages a tibial spine early in flexion, the DSF was found to be 12.6 mm throughout the flexion arc (Fig. 6b).

In the Tri CCC, which has anteroposterior, mediolateral, and rotatory mobility, the DSF is 14 mm throughout the ROM (Fig. 6c).

The DSF or jump distance is a parameter that may be calculated as well as measured; it constitutes a safety margin to prevent tibiofemoral dislocation of the TKR. It should be borne in mind that this dislocation may occur in fixed-bearing as well as in mobile-bearing devices. The safety margin is also related to surgical technique, since it is at surgery that the required ligament balancing has to be performed, with symmetrical and equal gaps in flexion and extension.


Can PCL-sparing devices resolve the problem?

Figure 7.  Subluxation of a non-congruent PCL-sparing TKR.

To obviate the risk of dislocation, posterior cruciate ligament-sparing implants (in which the ligament itself, rather than a posterior-stabilization system, ensures stability) suggested themselves.

The experience of the first few years following the introduction of PCL-sparing TKRs, however, showed these devices to have other drawbacks: in particular, they were very prone to abnormally rapid PE wear, which could, in itself, give rise to subluxation and instability. Some manufacturing techniques may have worsened the rate of this complication; also, attempts to obtain near-physiological kinematics have resulted in point contact in the articulating surfaces, with rapid wear (Fig. 7).

In the recent generation of PCL-sparing implants, attempts have been made to overcome these problems. However, the market trend (for which read: the surgeons’ preference) is still in favour of posterior-stabilized TKRs.




This form of dislocation involves the loss of contact of the three components that make up a mobile-bearing device.

In this case, the femur remains stationary in relation to the tibia, while the mobile-bearing element dislocates. The mechanism is one of rotating-platform or meniscal-bearing extrusion, rather than of actual dislocation of the knee joint.

This is a well-recognized complication in the “gold standard” meniscal-bearing device, the Oxford TKR designed by O’Connor and Goodfellow.

These authors published the results of their first series,2 which showed a 7% rate of mobile-meniscus extrusion. The incidence decreased after selection had been restricted to patients with intact anterior and posterior cruciates.


Biomechanics of dislocation

The mechanism of extrusion is the same regardless of whether the implant has two separate meniscal bearings or one single rotating platform.

Instability in flexion

Figure 8.  Mobile-bearing extrusion.Loss of contact | Sliding | Extrusion

Stretching of the ligaments in flexion leads to loss of contact between the femoral condyle and the mobile bearing element.

Under these circumstances, posterior loading may push the bearing forwards and flip it up (Fig. 8).

Role of the PCL

The PCL may act as a stabilizer of the mobile bearing (by holding the joint surfaces together); equally, however, it may help to extrude the bearing.

Good PCL tension, in a TKR that has good congruency in flexion, will stabilize the mobile bearing element (Fig. 9).

Figure 9.  PCL retention and good congruency in flexion.

Conversely, inappropriate PCL tension (an excessively tight or excessively slack PCL), in a TKR that is non-congruent in flexion, will encourage mobile-bearing dislocation (Fig. 10).

The adverse effects of excessive PCL tightness are well known. They are frequently observed at the trial stage of knee arthroplasty using fixed-bearing PCL-sparing implants, since the trial bearing is placed on, rather than held in, the tibial tray, and, hence, not stabilized. As the knee is put into flexion, anterior lift-off or tilting will occur.

Figure 10.  PCL retention and incongruency in flexion.


There are two ways in which the pattern may be remedied:

1.       by releasing the PCL;

2.       by increasing the posterior slope of the tibial bone cut.

Posterior-stabilized mobile-bearing TKRs have a lesser risk of anterior lift-off,


firstly, because the contact zone between the femoral condyle and the PE component is less far posteriorly situated;


secondly, because the resection of the PCL will relax the posterior compartment; and


thirdly, because some of the stress will be directed obliquely at the posterior-stabilizing spine, holding the bearing down against the tray (Fig. 11).

Figure 11.  Principle of the Insall-Burstein knee.

Figure 12.  Tri CCC: effect of third condyle on the resultant force vector.


This stabilizing effect is enhanced if the posterior-stabilizing spine is weight-bearing, as is the case in TKRs with a third condyle and a congruent central tibial bearing surface (Fig. 12).

Posterior cam effect

If there is contact in flexion between the mobile-bearing element and a posterior osteophyte, the bearing may be levered off the tibia and made to lift off anteriorly (Fig. 13). This is why, at surgery, care must be taken to ensure that all osteophytes are removed. This contact pattern may be encouraged by a posterior drawer in flexion, as seen when the PCL is slack or when the posterior-stabilization system is not working properly (or absent).

Figure 13.  Posterior cam effect.


Influence of mobile- bearing constraint systems on the dislocation potential

Some designers have adopted a policy of unrestrained movement of the mobile bearing, which is allowed to slide freely on a rimless surface (Oxford TKR). Many, however, have opted for controlled mobility of the bearing element. To this end, a variety of mechanisms have been devised: non-captive pegs, inverted pegs, captive pegs, keyways, rails, etc.

Each system has its own effect on mobile-bearing stability, and its own disadvantages.

Non-captive pegs

Figure 14.  Angle of dislocability.

One or two more or less cylindrical pegs arising from the superior surface of the tibial tray are used to control the anteroposterior, mediolateral, and rotatory mobility of the mobile bearing.

Extrusion may occur if the bearing lifts off and comes off the peg (Fig. 14).

For a comparison of the different TKR models involving a non-captive peg, it is preferable to use an angle measurement, rather than the height of the peg.

Figure 15.  Angle of dislocability in two TKRs:

Angle of dislocability = 14° | Tri CCC: angle of dislocability = 37°

The angle of dislocability is the minimum angle at which the bearing will tip off the peg and be extruded forwards. The higher the angle, the safer the system will be (Fig. 15).


In these models, the mobile-bearing element slides in a track cut into the tibial tray (Fig. 16), which controls the movement of the bearing, keeping it on a track governed by the (straight or curved) pattern of the keyway. While anterior lift-off is ruled out by this design, anterior extrusion is still possible.3-5 Also, the captive nature of the design puts the PE component at greater risk of fatigue fracture. This complication has been reported in mobile bearings running in tracks,6 while trackless models do not appear to be affected.

Figure 16.  Dovetail keyway.


Inverted peg

Figure 17.  Inverted cone.

Any peg standing proud of the tibial tray’s superior surface will be limited in height by the thickness of the mobile-bearing element. Designs involving an “inverted peg” or cone do not suffer from this limitation, since the bearing has a vertical, more or less tapering extension that sits inside the stem of the tibial tray (Fig. 17).

This pattern could be of importance for implant safety, since the peg can be made long enough to protect the bearing element against dislocation by extrusion.

However, rotatory dislocation would still be possible.

This form of dislocation involves the rotation of the mobile bearing, which turns 90 degrees in relation to the femoral condyle and the tibial tray, which remain stationary. Rotatory dislocation is brought about by insufficient congruency in flexion between the femur and the mobile bearing. Dislocation is encouraged by instability in flexion, and may occur when the knee is being hyperflexed.

The incidence of rotatory dislocation has been variously reported as being between 0.8 and 2%.4,7

Figure 18.  Rotatory dislocation.

This form of dislocation cannot occur in posterior-stabilized mobile-bearing TKRs with positive engagement of the cam-and-spine mechanism.

It is, however, likely to occur in PCL-sparing devices, if the PCL is incompetent (as a result of intraoperative damage or subsequent stretching), and in PCL-sacrificing models that are not posterior-stabilized.

The concept of a “rotatory DSF” or “rotatory jump distance” could be used as a measure for the assessment of a mobile-bearing TKR’s inherent rotatory stability (Fig. 18).

Captive pegs

Figure 19.  Stress transmission to the bearing element.

Some TKRs have a captive peg designed to prevent dislocation by extrusion and by anterior lift-off.

This design principle may not, however, be entirely beneficial.

With predominantly posterior loading, the bearing’s inherent tendency to lift off anteriorly will overstress the PE in critical areas. This increased stress could cause cold flow, and, in the medium term, lead to the bearing seizing in one position (anterior translation) (Fig. 19).

A similar problem may arise with eccentric varus or valgus loading (Fig. 20).

Also, repetitive PE stress may result in wear debris formation, which may ultimately give rise to osteolysis.

It should also be borne in mind that the PE component may suffer fatigue fracture following long-term exposure to excessively high stress levels.

Figure 20.  Stress transmission to the tibial tray.

A captive peg may also lead to stress being transferred to the implant-bone (or implant-cement) interfaces. With eccentric loading, the peg will counteract the tendency of the mobile bearing to lift off; however, this action will produce shear stress, which will be transmitted to the interface. Interface stress, in turn, will adversely affect the fixation of the implant (in particular, that of the tibial tray), and may, in the long term, lead to implant loosening.


Mobile-bearing TKRs were designed to reduce PE wear, and with a view to functional improvement, especially in the patellofemoral joint.

There are now many different models, which are, however, difficult to compare.

Numerical parameters (DSF or jump distance, angle of dislocability) are very useful; however, they should be considered within the overall context of the features offered by a given implant.

In particular, attention should be paid to the risk of wear related to the mobile-bearing constraint mechanism: a highly constrained (or captive) design may look like a good idea in terms of implant safety; however, it may result in rapid PE wear and/or early fracture.

In a comparison of different models, many other factors must be considered as well: the metals and metallurgical processes used, the surface properties of the materials,  PE manufacturing and sterilization techniques, the primary and secondary fixation of the implant components, the accuracy and precision of ligament balancing in extension and flexion provided by the instrumentation, etc.

Orthopaedic surgeons will need to be thoroughly familiar with the different assessment criteria for mobile-bearing TKRs, so as to be able to effect comparisons, while waiting for a sufficiently long follow-up period to have elapsed to allow a full evaluation of each individual type of implant.

(Transl KRMB)



1- KOCMOND J.H., DELP S.L. and STERN S.H., Stability an Range of Motion of Insall-Burstein Condylar Prostheses, a computer simulation study. The J. of Arthroplasty, 1995, Vol.10 No. 3 : 383-388.

2- GOODFELLOW J.W., Characteristics of the Oxford Knee. Basic concepts of surface replacements. Int Orthop. 1993, 17(Supplement) : 10-11.

3- BERT J.M., Dislocation / Subluxation of meniscal bearing elements after New Jersey low contact stress total knee arthroplasty. Clin. Orthop. 1990, 254 :211-215.

4- BUECHEL F.F. and PAPPAS M.J., New Jersey low contact stress knee replacement system : ten year evaluation of meniscal bearings. Orthop. Clin. North Am. 1989, 20 : 147-177.

5- LEMAIRE R., RODRIGUEZ A., GILLET P. et HUSKIN J.P., Les complications spécifiques des prothèses totales de genou à surface d'appui mobile, CR Congrès AOLF, Louvain-la-Neuve, 1998 : 232-233.

6- STIEHL J.B., Comparison of long-term results with cruciate substituting or sparing mobile bearing cementless total knee arthroplasty. Orthop. Trans., 1996, 20 :928.

7- HUANG C.H., LEE Y.M., SU R.Y. and LAI J.H., Clinical results of the New Jersey low Contact stress knee arthroplasty with two to five years follow-up. J. Orthop. Surg., 1991, 8 : 295-303.