chapter 9:Patellar Subluxation

Operative Treatment

The number of different techniques for patellar realignment and stabilization reported in the literature is astonishing. In 1959, Cotta (20) numbered 137 surgical methods directed at solving the problem of the unstable patella. Many of these methods have not enjoyed great success, but are still practiced today. Some are relatively minor variations or combinations of procedures. Wilbur (21) further subdivided this multitude of surgical procedures.

In essence, however, correction of patellar subluxation is not extremely complicated in most cases. Occasionally, the surgeon may encounter a very shallow trochlea or unusual dysplasia of the extensor mechanism such that stabilization of the subluxating patella is difficult, but most patients with less prominent subluxation causing functional limitation will respond to simple lateral release or lateral release and medial capsule imbrication. Less commonly, a more extensive proximal and distal realignment may be necessary to compensate for more substantial dysplasia or malalignment.


The technique of lateral retinacular release in patients with subluxation is similar to that described in the chapter on tilt‑compression. Because studies (22) have indicated that lateral release alone will not always correct subluxation, the surgeon must be careful to determine if something more than lateral release is necessary in the patient with patellar subluxation. Femoral nerve stimulation by percutaneous electrode may help the surgeon assess patellar tracking under anesthesia. Lateral retinacular release becomes particularly desirable in the patient with tilt and subluxation, because the tilt component may respond particularly well to lateral release. Schonholtz (23) noted that most patients with subluxation or dislocation did well after lateral release. Nonetheless, if there is residual patellar subluxation after lateral retinacular release, particularly if such subluxation is confirmed by femoral nerve stimulation under anesthesia, an Elmslie‑Trillat procedure or medial imbrication may be indicated selectively. Brief (24) emphasized the fact that lateral release alone fails to correct the patellar hypermobility that often accompanies subluxation. Occasionally, and particularly if there is significant patellar arthrosis (Outerbridge Grades 3 to 4), a surgeon may prefer to include anteromedial tibial tubercle transfer to unload the patella and improve patellar alignment (see Chapter 13). In a cadaver model, lateral retinacular release will correct patellar tilt, but does not always correct subluxation (25). The best candidate is one with lateral compression syndrome, tilt laterally, tight lateral retinaculum, and without significant chondrosis. Subluxation alone is not a reliable indication for doing a lateral release, in my opinion.


Representing a relatively simple surgical procedure, and sufficient for many patients, capsulorrhaphy consists of weakening the lateral retinaculum (simple release) and strengthening the medial retinaculum (imbrication). Insall et al (26) reported experience with a technique of capsulorrhaphy that avoids a suture line in the medial retinaculum in a position where it might cause conflict with the medial femoral condyle on full flexion. It is important that the lateral release be sufficient, extending from the tibial tubercle, across the entire lateral retinaculum, and releasing the vastus lateralis obliquus of Hallisey.


Several authors have described attempts to reinforce the weakened medial retinacular structures by grafting a variety of materials, including fascia lata, preserved skin (27), and nylon, to mention a few. These techniques have mostly been abandoned, having never gained wide popularity. Hauser (28) has extensively reviewed these procedures.


Valgus osteotomy has been reported by Heywood (29) as being ineffective for control of recurrent dislocation of the patella. He noted five recurrences in seven patients who had osteotomies. Osteotomy may be necessary because of functional tibial or femoral overloading in one compartment, but it is not indicated as a primary treatment for recurrent subluxation of the patella.


Tendons may be transferred into the extensor mechanism either from above or below, and methods for some of these tendon transfers are summarized in Figure 9.14. Advancement of the vastus medialis has been widely recommended in conjunction with medial imbrication, particularly for the skeletally immature patient. Gracilis, semitendinosus, and sartorius tendons have all been used in a variety of ways. When the tendon is inserted from above, it is designed to act as a dynamic component, and when the tendon is inserted from below, the intended effect is tenodesis. Baker et al (30) revived the original semitendinosus tenodesis described by Galliazzi in 1921 and reported good and excellent results in 81% of 53 cases. We have not found this procedure to be necessary and are concerned that it may be insufficient regarding isometry. Lateral release and careful medial imbrication, using the central quadriceps tendon to anchor the vastus medialis obliquus, may be appropriate for some difficult cases that require surgery before epiphyseal closure. In such cases, any tibial tubercle procedure can result in anterior epiphyseal closure and genu recurvatum or distal migration of the tubercle with continued growth of the tibia. Such complications must be avoided.

Krogius (31) described a technique that has become popular in Europe. The skin incision is medial parapatellar. A strip of medial capsule with attachment of the vastus medialis is developed, creating a medial retinacular defect. A lateral retinacular release is then carried out, and the medial defect is closed. The myofascial strip that has been developed is then sutured into the lateral defect that has been created by the medial plication. The author reported good results with this technique. In our experience, however, lateral retinacular release and medial imbrication without excising any medial capsule has proved satisfactory in the majority of skeletally immature patients. This procedure, however, requires considerable attention to detail, and the surgeon must be careful not to overtighten the medial capsule, because this will increase medial facet loading and lead to articular degeneration. I prefer to use sutures to anchor the vastus medialis obliquus to the central quadriceps tendon (CQT) above the patella and at the CQT‑patella junction. Galliazzi's procedure (29) may be helpful in the skeletally immature patient (Fig. 9.15), but I have not found this procedure to be necessary. One must consider the possibility that this technique could create aberrant stresses on the patella, depending on the degree of knee flexion.


If patellectomy is indicated for the arthritic sequelae of malalignment, and the patella has been surgically centered and stable, then simple patellectomy can be appropriate. However, it must be remembered that the extensor apparatus without a patella can subluxate and dislocate as well. Benoist and Ramadier (32) have underscored this observation with the report of their results of only nine good and excellent results in 19 patients with simple patellectomies for recurrent subluxation and dislocation, whereas there were 17 good and excellent results of 19 in patellectomy associated with transplantation of the tibial tubercle. In the case of patellofemoral arthrosis in which a patellectomy is contemplated, this transplantation may include anterior displacement to facilitate the extensor moment arm. West and Soto‑Hall (33) have reported on 31 of 33 patients who had patellectomies with good or excellent results. Heywood (29) was more pessimistic, reporting 10 unsatisfactory results in 29 patients with patellectomies and only one excellent result. In general, however, the surgeon should avoid patellectomy in patients with patellar subluxation (34) unless arthrosis is severe and other alternatives, such as anteromedial tibial tubercle transfer, cannot give relief (Chapter 13). Centralization and proper alignment of the extensor mechanism are imperative before performing patellectomy.


Roux (35) in 1888 was the first to report tibial tubercle transplantation coupled with lateral release and medial plication in the treatment of recurrent patellar dislocation. His single case report was apparently dramatically effective. Hauser (28), in 1938, apparently unaware of Roux's single case report, reviewed the literature for the procedures popular at that time and reported four cases of medial tibial tubercle transplantation‑all with successful outcomes. In each case, lateral retinacular release and medial plication were carried out. Since Hauser's original report, a variety of modifications have been reported. At long‑term follow‑up (average, 8 years), Barbari et al (36) found that less than 50 percent of patients were free of pain following a Hauser procedure. Hauser's operation was widely popular in the United States but has now become contraindicated in the majority of patients because it places the tibial tubercle posterior and actually increases contact pressure on the patella (37) (Fig. 9.16). MacNab (7), reporting an average 10‑year follow‑up on 10 patients, mostly after medial tibial tubercle transplantation, attributed the lack of good results to preoperative patellar cartilage damage, and it seems likely that load may have been transferred in many cases onto significant articular lesions given the known occurrence of medial lesions in patients with patellar instability and pain . Loff and Friedebold (38) found a direct correlation between preoperative joint status and postoperative results. Crosby and Insall (39) reported not only a recurrent dislocation rate of 20% after tibial tubercle transfer in 69 patients but also a high incidence of late osteoarthritis. This late osteoarthritis was most likely due to the medial patella overloading that occurs with the Hauser procedure Although there was a 25% incidence of redislocation after 12 proximal soft‑tissue procedures, there was no late osteoarthrosis seen.. More recently, Juliusson and Markhede (40) noted that only 12% of 40 patients whose knees underwent a Hauser procedure showed objectively satisfactory results. This is clearly unacceptable.

The observation by Goutallier and Debeyre's (41) on the detrimental biomechanical effect of posteromedial transfer of the tibial tubercle by the Hauser technique (Fig. 9.17) may account for poor late results after the Hauser procedure. Posterior placement of the tibial tubercle has no place in the management of the patella with incipient or apparent patellar arthrosis.

Henry and Craven (42) reported an approach to realigning the distal extensor mechanism by transferring the patellar tendon medially with a "thin flake" of the tibial tuberosity. A "dovetail" osteotomy as described by Southwick et al (43) might be used in this way also. Again, however, these procedures typically involve a posteriorization of the tibial tubercle and resulting articular overload at the medial patella, leading to late arthritis inevitably in many patients.

Goldthwait (6) reported a technique whereby the patellar tendon is split and the lateral half detached from the tibial tubercle, passed behind the medial half, and sutured to the medial capsule. The net result is to change the resultant vector of traction of the patellar tendon. This procedure has never gained widespread popularity and carries significant risk of adversely affecting patellar contact pressure distribution by pulling down the lateral facet.


When there is minimal arthrosis and proximal realignment alone is insufficient, distal realignment of the extensor mechanism, without anteriorization, is most appropriate (44‑46).

Subluxation may be persistent despite lateral retinacular release (22), particularly as lateral release is an inappropriate procedure for patella instability.. Nonetheless, patients with patellar subluxation have been shown to improve in some cases after lateral retinacular release alone (47). Although medial imbrication will provide all of the additional balance that is needed in many patients, distal realignment (medialization of the tibial tubercle) may be necessary in some patients (skeletally mature) to assure appropriate patellar tracking. The surgeon must recognize that excessive medial imbrication to support the patella medially can overload the medial facet and aggravate an already degenerating patella. Therefore, distal realignment becomes an important consideration in extensor mechanism realignment for subluxation. Nonetheless, the author (JPF) does not recommend tibial tubercle transfer when  a careful medial imbrication can control subluxation. In most cases, tobay this will be done arthroscopically.

The Roux‑Elmslie‑Trillat procedure, as reviewed by Cox (48), is the preferred distal realignment technique because it provides pure medialization without transferring the tibial tubercle posteriorly. The technique of Goutallier and Debeyre (41) is similar in concept (Fig. 9.18). Also, if there is Outerbridge Grades 3 to 4 lateral facet degeneration, the traditional Trillat procedure can be modified to provide anteromedial tibial tubercle transfer by making the osteotomy oblique, as described in 1983 (49) (see Chapter 13). Shelbourne et al (50) have pointed out that correction of congruence by the Elmslie‑Trillat procedure will not deteriorate with time. Wootton et al (51) reported satisfactory results in the knees of 68 patients following an Elmslie‑Trillat procedure for patellar dislocation or subluxation.

The Trillat procedure may be performed through an anterior or anterolateral incision approximately 10 to 15 cm long and extending from the midlateral patella to approximately 5 cm distal to the tibial tubercle. The lateral retinaculum is incised as described in the section on lateral retinacular release. The patella is inspected, and a decision is then possible as to whether the Trillat procedure or an anteromedial tibial tubercle transfer is more appropriate (Fig. 9.19). If there is distal or lateral facet breakdown, anteromedialization may be indicated.

If the decision is to do a Trillat procedure, the anterior compartment musculature is reflected posteriorly by subperiosteal elevation, and a flat osteotomy is made deep to the tibial tubercle such that a bone pedicle (approximately 5 to 7 cm long and tapered anteriorly) is formed. The bone pedicle is hinged distally and rotated in a medial direction, maintaining good bone‑bone contact until the desired amount of patellar tendon medialization is achieved. The bone pedicle is fixed then with two screws, preferably into the posterior tibial cortex as in the anteromedial tibial tubercle transfer. Cox (48) reported that only 7% of his patients experienced subluxation after this procedure.

Postoperatively, a Hemovac drain is used for 24 hours, and the patient remains on crutches for approximately 6 weeks, but knee motion is started immediately after surgery, assuming secure fixation of the bone pedicle.

In the final analysis, what is most important is to correct accurately the specific extensor mechanism malalignment noted preoperatively and to avoid any surgery that might aggravate the situation in any way. As a general rule, the surgeon should do the least surgery possible to achieve realistic correction of patellar tracking and restoration of satisfactory patellar articular cartilage contact stress distribution.



Inside Chapter 9: