PATELLAR INSTABILITY IN TOTAL KNEE ARTHROPLASTY
James B. Stiehl, MD
Midwest Orthopaedic Biomechanical Laboratory
Columbia St.Mary’s Hospital
Milwaukee, Wisconsin
Address Correspondence To:
James B Stiehl, MD
575 W River Woods Parkway, #204
Milwaukee, Wisconsin53212
414-961-5678
Fax: 414-961-6788
Email:
Introduction:
Instability of the patella may occur after total knee arthroplasty and has been identified either with or without prosthetic resurfacing of the patella. Subluxation is more common than dislocation, but the incidence of symptomatic instability leading to reoperation is low ranging from 0.5% to 0.8% in several referral centers.22,25 In a recent multicenter study of LCS mobile bearing total knee arthroplasty, only 6 (2.3%) of 259 revisions related to patellar instability and of the overall group this accounted for a revision rate of 0.1% at average of 5.7 years followup.27 The etiology of this problem is multifactorial but can be most commonly related to a host of problems including component malposition, soft tissue imbalance, trauma, and excessive valgus alignment of the knee. In addition, problems related to surgical technique such as tibial or femoral internal rotation, increasing patella-patellar implant composite thickness, and increasing femoral component dimension can cause patellar instability. This review will explore each of these issues and describe methods of treatment.
Etiology:
Component Malposition:
One of the most common causes of patellar instability is component malposition at the time of surgical implantation.1,2,12,15,17 Any tendency to place the femoral or tibial components in abnormal internal rotation in the transverse plane can predispose to patellar subluxation essentially by increasing the Q angle at the knee joint.(Fig 1) Much has been written about correct positioning of these components, either to increase the understanding of the normal anatomical relationships or to advance methods that enable the surgeon to place the components correctly.
Surgeons have performed the distal femoral resection based on anatomical landmarks which guide the amount of normal femoral external rotation needed to match the normal proximal tibial slope of about 93 degrees in the knee joint and maintain the normal femoral position.5 Originally, the posterior condylar axis was the resection plane for making the posterior condylar cuts but this tended to place the femoral component in internal rotation leaving a trapezoidal flexion gap particularly if the proximal tibial cut was perpendicular to the long axis of the tibia. Surgeons then attempted to external rotate the posterior condylar cutting jig a fixed amount of 3 to 5 to give the appropriate femoral external rotation and a balanced symmetrical flexion gap. Berger, et.al. found the relation of the posterior condylar axis to the surgical transepicondylar axis (point of lateral epicondyle to the sulcus of the medial epicondyle) to average 3.5º for males and 0.3º for females which was a highly significant statistical difference. However, the clinical angle using the prominence of the medial epicondyle was 4.7º for males and 5.2º for females. Significantly, the variance could range from 1º to 9.3º.4 Mantas, et.al. found in normal femurs that the range of posterior condylar axis reference to the transepicondylar axis ranged from 0.1º to 9.7º.19
A more reliable method has been to use the transepicondylar axis which has been shown to be more consistent compared to the posterior condylar axis. Stiehl, et.al. has shown that the transepicondylar axis has a perpendicular relationship anatomically to the long axis of the tibial shaft.27(Fig 2) As the transepicondylar axis closely approximates and parallels the axis of knee flexion, Insall, et.al. advocated cutting the posterior condyles parallel to this axis will place the femoral intercondylar groove in the normal anatomical position.15,25 Similiarly, the Whiteside line drawn from the femoral intercondylar groove down to the center of the femoral notch is virtually perpendicular to the transepicondylar axis and also represents a suitable technique. Fehring compared the Insall method with the measured resection method of distal femur resection using a fixed posterior condylar reference guide finding that the measured resection technique resulted in rotational errors of at least 3º in 45% of knees.9 Similiarily, Olcott and Scott found that the transepicondylar reference most readily determined a balanced flexion space while using 3º rotation off the posterior condyles was least consistent.21
A more subtle problem with femoral component placement is the location of the anatomical femoral groove of the patient compared to that of the prosthetic component. Eckhoff, et.al. have shown that this groove in the anatomical specimen tends to be 2.5 millimeters lateral on average to the anatomical midplane of the distal femur based on the dimensions of the condyles and position of the femoral notch but could be up to 8 millimeters in outliers.7(Fig 3) Most contemporary implants have a symmetrical design, such that condyles are roughly equal in dimension, and the placement of the prosthetic femoral groove bisects these condyles. By placing the femoral prosthesis in what may appear to be the anatomically correct or symmetrical position in relation to the femoral notch, the surgeon may actually be medializing the femoral intercondylar groove. A technical solution to this problem it to mark the patient’s femoral intercondylar groove during preparation and then attempting to match this position with that of the prosthesis. (Fig 4) This may cause some overhang of the lateral femoral condyle, but that would be preferable to abnormally increasing the Q angle.
Tibial component placement can also be problematic, especially if the surgeon attempts to cover the tibial on the medial side. This can lead to placing the tibial component into relative internal rotation to the correct axis of the patient’s knee leading to lateral placement of the tibial tubercle and aggravation of the Q angle. Eckhoff, et.al. have shown variability of this knee axis and there is a subgroup of osteoarthritic patients where there is actually external version of the tibial in relation to the femur.8 This problem tends to magnify the amount of varus deformity that may accrue with disease. Several authors favor centering the tibial component on the medial third of the proximal tibial tubercle and this usually works. Another method is to draw a line from the patient’s femoral intercondylar groove down to the proximal tibia and matching that position throughout the procedure. That can be particularly useful with methods that use a sloped cut of the distal tibia, as making the cut out of plane may aggravate alignment errors. The best way however, is to carefully trial the implants on insertion make certain that the tibial articulation is perfectly midline and not rotated onto the anterior surface of the medial tibial insert.
Patellar component malposition usually reflects technical error in cutting the patella. The effort is to make a symmetrical hockey-puck shaped structure that has equal thickness top to bottom and side to side.(Fig 5) Under resection of the lateral facet or the distal pole will lead to tightness of the lateral retinaculum and a tendency to sublux laterally.13,23 Grace and Rand were clearly able to show increasing patellar thickness or stuffing the joint could lead to patellar lateral instability.11 Another problem is centering the patellar dome on the midpoint of the patellar remnant and not shifting the dome to the most medial edge.18 In the normal patella which averages 37 millimeters in width, the average medial facet is 14 millimeters while the average lateral facet is about 23 millimeters. Lateral dome placement increases the Q angle and also aggravates lateral retinacular tension. Another option to solve this issue is to use an eccentrically shaped dome or an anatomical mobile bearing device such as the LCS which appropriately matches patellar anatomy.(Fig 6) Finally, a poorly studied issue of the patellar articulation is the shape of the prosthetic intercondylar groove in relation to the anatomical distal femur. No doubt, earlier designs were more “boxy” in shape and tended to add metal thickness to the intercondylar groove, in effect stuffing the joint. Recent total knee designs have ameliorated this problem by designing deeper prosthetic intercondylar grooves, which will likely be demonstrated in dramatic reduction of patellar complications in long term studies. The LCS mobile bearing prosthesis had this concept as an initial design goal and has realized a patellar component survivorship of 98.5% at 14 years followup.26(Fig 7)
Limb Malalignment:
Several authors have noted the potential of a valgus knee to predispose to patellar instability. Not uncommonly, these knees have a chronically subluxed or dislocated patella. Often there is lateral retinacular tightness associated with the long standing deformity, and if not addressed with release can cause subluxation. The lateral condyle in a valgus deformity may be smaller than normal in dimension. If the posterior condylar axis is used for distal femoral resection, there may be a tendency to internally rotate the femoral component, and increase the Q angle as noted above. With the use of appropriate instrumentation, especially intramedulary femoral guides, the chance of creating a postoperative valgus deformity is diminished and this problem may be even further reduced with the advent of computer assisted navigation of alignment that has shown to be even more precise.
Soft Tissue Imbalance:
Lateral retinacular tightness remains a subtle cause of patellar instability, but usually does not result in a clinical problem by itself. A number of authors have clearly shown that surgical technique such as basing the femoral resection on the posterior condylar axis creates femoral internal rotation which will lead to the need for lateral release. Therefore, if a patient has patellar instability, the surgeon must look for other causes or problems than simply tightness of the lateral retinaculum. A chronically dislocated or subluxed patella is the exception, however, and retinacular tightness can be the primary source in this case. Unfortunately, these cases may be complicated also by medial retinacular weakness or atrophy, and every attention to the details of reconstruction are needed. In these unusual cases, attention to every detail is needed such as using a lateral parapatellar approach, carefully aligning the implants, and then positioning the implants to optimize patellar tracking with the final need to reef or reconstruct the stretched medial soft tissues.
Other Causes:
Medial retinacular weakness or disruption is possible after total knee arthroplasty and may result from an expanding hematoma, inadequate surgical closure, overintensive physical therapy, or injury.14,15 Surgical repair will be needed if medial laxity is documented with patellar subluxation. Rare causes of patellar instability have resulted from surgical misadventure, such as placing the right femoral component in the left knee, and placing a ridged anatomical patella component such that the ridge was parallel to the transverse joint plane and not verticle.6,10
Clinical Findings:
The hallmark of patellar instability is anterior knee pain aggravated stressful activites such as stair climbing or rising from a chair. With instability, there can be dramatic giving way or buckling of the knee while subluxation will cause the sensation of the knee slipping out of place. Palpation of the extensor mechanism throughout passive and active range of motion will reveal defects in continuity. Also, the patient may be able to identify areas of localized tenderness. Dislocation or subluxation may be detected by palpating the patella through the range of motion. Provocative maneuvers such as attempting to sublux the patella laterally during active flexion can also elicit pain or apprehension. Also patella alta should be observed as this may make the patella somewhat higher in the femoral groove. Finally, there are those cases who have asymptomatic lateral subluxation without any objective findings except vague medial knee pain. This problem should be identified with radiographic views as many patellar implants are subject to accelerated polyethylene wear with this abnormal articulation.
Radiographic Findings:
Radiographic evaluation of the patella relies primarily on the lateral view and the sunrise or Merchant’s view of the patella. The symmetry of the patellar cut and thickness of the patella composite is apparent and may be compared with the opposite normal patella. The infrapatellar view will demonstrate the position of the patellar component in relation to the trochlear sulcus, either centralized, tilted, subluxed, or dislocated. Tilt can be defined as medial or lateral depending on the relation to the femoral condyles. Subluxation can be measured as displacement from the center of the prosthetic femoral intercondylar groove. Component rotation can be determined with computed tomography scans through the knee joint. Berger, et.al. have identified four scans needed for this determination which are scans through the medial and lateral epicondyles, the tibial plateau immediately below the tibial base plate, the tibial tubercle, and through the tibial insert.3 The femoral component rotation is determined by measuring the angle formed by a line drawn through the medial and lateral epicondyles and the line connecting the posterior flanges of the implant. Tibial component rotation is determined by first finding the geometric center of the proximal tibia and then superimposing this point onto the image with the tibial tubercle. A line is drawn to the highest point of the tubercle and this becomes the tibial tubercle axis. This axis is then placed on the image with the tibial insert. A line is drawn perpendicular to the posterior surface of the tibia insert and when compared with the tibial tubercle axis, becomes the amount of tibial rotation. According to this method, the normal amount of tibial internal rotation is 18.
Treatment:
Prior to any surgical intervention, conservative methods should be tried including quadriceps exercises, bracing, and avoiding activities that may aggravate instability symptoms. With time, scarring of the retinacular tissues can often lead to resolution of symptoms. With chronic instability symptoms or frank dislocation, surgical intervention is necessary. Careful assessment of all possible prosthetic causes is done such as component malrotation, limb malalignment or soft tissue problems about the patella. Component malposition will require revision. Soft tissue problems may require proximal realignment, which is done with a lateral retinacular release and medial vastus advancement.14,24,25 Distal bone realignment of the tibial tubercle may also be done in combination with the proximal realignment. Numerous methods have been described for the distal realignment including a modified Trillat procedure or using a fairly long osteotomy as described by Whiteside.28 Great care must be taken to insure that an adequate piece of bone (at least 8 centimeters) is taken and that apposition and fixation is optimal. Wound complications, rupture of the patellar tendon, and fracture of the bony remnant are inherent risks with this approach, and must not be undertaken lightly.
Results:
Insall described a method of long lateral release from inside the joint outwards, and then imbricated the medial vastus retinaculum over at least 50 to 75% of the width of the quadriceps tendon. (Fig 8) Merkow, et..al. reported no recurrences using this technique in 12 cases, but noted one case of skin necrosis and one patella fracture after this approach.20 Grace and Rand reported the results of 25 knees with 14 having a proximal realignment of which 4 recurred.11 Of nine knees treated with both proximal and distal realignment, no recurrent dislocations were noted but two sustained distal patellar tendon ruptures. Two other cases had revision of components, and one of these had further subluxation. Kirk, et.al. used a modified Trillat procedure in 15 cases of patellar dislocation and noted no recurrences or problems with the patellar ligament.16
Case Report #1
A 54 year old morbidly obese female underwent total knee arthroplasty using a mobile bearing total knee arthroplasty. Component position and limb alignment were deemed optimal during the procedure and the patella component tracked well with the “no thumbs” technique. Nonetheless, postoperatively, she developed symptomatic subluxation with buckling episodes and chronic anterior knee pain. This was confirmed with a Merchant’s view of the patella.(Fig 9) At reoperation, the patella again was noted to track normally on passive motion. A proximal realignment was done with lateral release and medial imbrication. The knee was then held in a knee immobilizer for six weeks followed by rehabilitation. Within four months, symptoms had recurred to the original level. At this point, a combined proximal and distal realignment was done. The distal alignment was a long tibial crest fragment measuring 1.5 cm by 8 cm which was fixed with screws.(Fig 10) The leg was immobilized again for six weeks followed by rehabilitation. After four years followup, there has been no recurrence of symptoms.
Case Report #2
A 69 year old female was underwent revision of a failed unicondylar arthroplasty done for osteoarthritis. Following the revision, the tibial base plate was noted to be in 9 of varus and there was frank dislocation of the patella with chronic instability symptoms.(Fig 11) A proximal realignment was attempted with early recurrence of dislocation. A computed tomograph was done of the knee revealing a femoral component internal rotation of approximately 9.(Fig 1) This compared to the normal rotation of the TEA to the posterior condylar axis of 3.(Fig 12) The tomogram through the knee prosthesis revealed the patella to be perched on the lateral femoral condlyle. Revision arthroplasty was needed finding marked internal femoral rotation, apparently done to match the severe varus placement of the tibial tray. Another finding that is typical of this error is that internally rotating the femoral component this dramatically left the flexion space on the lateral side lax by 12 to 14 millimeters, functionally exaggerating the genu valgum.(Fig 13) All these problems could be resolved with satisfactory placement of revision implants, though the knee was immobilized in extensions for three weeks to allow for healing of the Insall medial retinacular realignment.(Fig 14)