Knee Hip Pain
5 East 98th Street
Box 1188
New York, NY 10029
Phone: 212-241 2914
Fax: 212 - 534-6202

Get Information on Dr. Grelsamer's Books Here!
This section has been adapted from Chapter 13 of The Patella - A Team Approach, by Ronald Grelsamer and Jenny McConnell, Aspen Publishers 1998 (currently distributed by Pro-Ed at

For some patients, surgery is the conservative option. Non-operative treatment has either failed or can be expected to fail. This is occasionally the case for people whose patella exhibits considerable malalignment.

The following text is somewhat technical. Don't try this at home.

In 1959, Cotta in Germany performed a literature review and found 139 procedures pertaining to the patella. This was before any of today’s commonly-used procedures had been described! Whenever there are many operations for one problem, it is a strong indication that none of these operations approaches perfection. The surgeon must first determine exactly what condition he or she is operating for. Categories include chronic malalignment, acute dislocations, chondral lesions, and combinations of these conditions.

The common platitude is that surgery is indicated when non-operative treatment has failed. In principle, this is true. However, this orthopaedic cliché should be amended to read "appropriate non-operative treatment". There is more to non-operative treatment than attaching weights to the ankle or placing the patient in an exercise machine (the "electric chair").

Surgical procedures have been arbitrarily divided into proximal and distal procedures. All seek to somehow transform a tilted and/or lateral-tracking patella into a less tilted, centrally tracking patella. Proximal re-alignments involve surgical manipulation of the lateral retinaculum, the medial retinaculum, the vastus lateralis, the vastus medialis obliquus and any combination thereof. Like balancing a marionette, it is a question of tightening certain "strings" and giving others more slack. The term distal re-alignment denotes (by convention) a transfer of the tibial tuberosity. Certain procedures fall into neither category: prosthetic re-surfacing of the patella or trochlea for example.

There is still confusion and controversy with respect to nomenclature. To some surgeons, a "lateral release" (see below) is not considered a re-alignment or a reconstruction (as in ‘we will perform a lateral release and then go to a formal reconstruction if this does not work’). To others (including this author) any procedure which changes the position, tracking or articular contact area of the patella qualifies as a reconstruction and a re-alignment. The difference in philosophy stems from the fact that some surgeons consider the lateral release to be a small procedure to be done arthroscopically while "reconstructions" are done with a formal skin incision. The problem with this nomenclature is that, in my opinion, it can lead the surgeon to think of the lateral release as a minor, relatively risk-free procedure to be done nearly on a casual basis while other procedures are major and therefore require more planning, more discussion with the patient and more concerns with respect to complications. In my experience, such a distinction between the lateral release and other procedures is not warranted.

Lateral Retinacular Release

The lateral retinacular release has been an integral part of just about all re-alignment procedures and since the 1970’s has often been carried out as an isolated operation. In fact, the isolated lateral retinacular release is now probably the most commonly performed re-alignment procedure. It is such a technically simple procedure that one has to wonder why it was not popularized sooner. The answer, though, I believe is quite simple: until the 1970s, orthopedists concentrated on the unstable patella which slipped out of the trochlea. The focus was on the subluxating or dislocating patella. Patients who were operated on felt the patella slip in and out and/or had episodes of frank dislocation out of the trochlea requiring reduction by a doctor. Though Merchant from California published a report on the lateral release for just those cases as far back as 1974, most surgeons at the time felt that for unstable patellae a more significant reconstruction is required. Ficat from southwestern France was the first to popularize the concept of the tilted patella. This patella could be perfectly centered in the trochlea yet be malaligned by virtue of tilting in the axial plane. In the non-operated knee, this tilt is always lateral, i.e. the lateral border of the patella dips down while the medial border is raised (like an airplane turning). This tilting can be readily detected clinically and radiographically. Tilting of the patella is always associated with increased tightness of the lateral retinaculum.

Tilting of the patella was hypothesized by Ficat to cause excessive pressure on the lateral facet of the patella thus leading to what in English has been translated to the "excessive lateral pressure syndrome". It stands to reason that the treatment for this syndrome is simple release of the lateral retinaculum. Technically this means surgical separation of the lateral retinaculum from the lateral border of the patella. When done through a large incision it is as quick as closing a scissors. When done through a small incision or arthroscopically it is already more challenging, but, relative to other procedures, still relatively simple. This has certainly contributed to its meteoric success.

Alternate mechanism of action: Fulkerson et al. have noted that patients with patellar pain can have fibrosis of the nerves within the lateral retinaculum. Division of the lateral retinaculum can have a denervating effect on the patella and some have suggested that this plays a significant role in the pain relief obtained after a lateral release.

Indications: The lateral release can be carried out as an isolated procedure or in combination with others. There is little controversy with respect to combining the lateral release with other procedures. The controversy begins when one discusses the lateral release as the only form of surgical re-alignment. In favor of this approach is the fact that the procedure is relatively straight-forward (no special instructional course required), it can often be done in an ambulatory surgery setting, and there are indeed articles in the orthopaedic literature supporting it. Against it are the fact that although it is simple it still involves a trip to the operating room and an anesthetic, it still requires a full course of physical therapy, it can still lead to significant complications, and there are equally numerous articles to suggest that the isolated lateral release does not work well in all patients with patellar pathology. Even if it does work well, will it hold up in the long run?

Contra-indications: Patients with no malalignment cannot be expected to gain much from this procedure.

Surgical technique and variations: Many variations have been described. Deep to the lateral retinaculum lies the synovium of the joint, the smooth, moist layer which secretes synovial fluid. In performing a lateral release, some surgeons choose to leave the synovium unopened while others choose to release the synovium along with the retinaculum. The lateral retinaculum is usually released from approximately the superior pole of the patella (proximal to this one encounters the muscle fibers of the vastus lateralis) down to a variable distance. At the very least, the release is carried out to the level of the inferior pole but it can be carried out all the way to the tibial tuberosity. All of these variations qualify as a lateral release. Exactly how far the release should go is also a subject of controversy. A "turn up" sign has been described whereby the lateral release is carried out to the point where the patella can be "turned up" 90 degrees. This was recommended to guard against an insufficient release, but in some patients this amount of release is excessive.

The lateral release is said to be an "open" release when an incision is made at the lateral border of the patella. The incision is usually near the superior pole of the patella so as to directly visualize the supero-lateral geniculate vessels which located transversely across the incision at the level of the superior pole of the patella just deep to the retinaculum. These vessels are responsible for one of the more common peri-operative complications of this procedure which is a hemarthrosis (blood in the joint). The risk of a hemarthrosis can probably be decreased by direct visualization and cauterization of the vessels. Interestingly, a hemarthrosis can occur even with this approach. Presumably, vessels can go into spasm at the time of surgery - only to open up and bleed in the post-operative period. For this reason, it is not unreasonable to put a drain in the knee. The incision can be closed with a subcuticular (so-called "plastic surgery") closure. This will avoid the "railroad track" look. It may however spread in time as vertical incisions about the knee are wont to do. According to traditional teaching horizontal incision can give better cosmesis, and it remains an acceptable surgical approach. However, it may need to be longer to give adequate visualization proximally and distally. Moreover a study which specifically compared the cosmesis of long transverse incisions and long parapatellar incisions found triple the number of cosmetically good results in the parapatellar group.

The lateral release is said to be an "arthroscopic" release when it is effected via one or more of the arthroscopy portals. These arthroscopic releases can be done from inside out or outside in. When done from outside in a hooked instrument is passed between the skin and the retinaculum. For example, a small hook at the end of a long thin instrument is passed from the antero-lateral portal up to the proximal portion of the retinaculum. The hook is pushed into the joint and pulled distally, thus cutting the retinaculum from approximately the superior to the inferior pole of the patella. This can be done with a metallic hook or can be done with cautery. When done with cautery, the overlying skin is at risk for being burned, a risk which can be decreased by injecting fluid between the skin and the retinaculum. When done from inside out the same hooked instrument is inserted into the joint and pushed out of the synovium and retinaculum at a distance away from the entry point. Again this can be done with a metallic hook or a cautery. A holmium laser has also been used for this procedure. With a wavelength of 2.1 microns it readily goes through the soft tissues about the knee. It has the ability to cauterize small vessels and on some occasions can cauterize the superior geniculate vessels. In my experience it does not do so on an absolutely consistent basis. Advances in the delivery of laser energy may change this.

Regardless of the technique used I think it is important to cut the retinaculum with as few cuts as possible. One fell swoop with a Mayo-type scissors is probably best. Multiple small cuts with a small instrument may be more traumatic and may increase the risk of reflex sympathetic dystrophy (an unproved supposition on my part).

The arthroscopic approach has the advantage of minimizing scars. This has an obvious cosmetic benefit which has to be weighed against the factors listed above. As an aside, it is important to note that the patient overly concerned about a knee scar may in fact not have enough disability to warrant surgery. What constitutes appropriate concern as opposed to excessive concern is a matter for the orthopaedic team to determine.

Complications: In addition to a hemarthrosis (described above), other complications can occur. Excessive release -especially proximally into the vastus lateralis itself - can lead to medial subluxation of the patella. When carrying out the proximal part of the release, it is possible to drift towards the midline and section part of the quadriceps tendon. This is clearly not desirable especially if the medial retinaculum release is divided as the initial part of a "proximal re-alignment". The quadriceps tendon is placed at risk for rupture.

Reflex sympathetic dystrophy is a risk associated with any surgical procedure.

Medial Plication

The medial retinaculum and medial patellofemoral ligament provide a passive medial restraint to lateral displacement of the patella and the vastus medialis obliquus (VMO) is the major dynamic stabilizer of the patella. It is natural to try to increase the pull on the inner (medial) aspect of the kneecap.

The simplest procedure in this category is the medial plication as popularized in the United States by Hughston. The medial retinaculum is divided as it inserts into the patella. It is then re-attached more medially onto the patella itself in a "pants-over-vest " fashion.

In addition to plicating the medial retinaculum, the VMO can be dissected free by dissecting along its insertion and then dissecting along its inferior border. The muscle can then be pulled in the direction of its fibers as far as one wishes onto the supero medial portion of the patella. The inferior fibers can also be sutured more distally to "horizontalize" the VMO. Dissection of the VMO changes the rehabilitation considerably since vigorous contraction of the muscle could conceivably pull the VMO out of its new position.

*Note: A lateral release is carried out in conjunction with either of the above medial procedures.

Indications: Moderate or severe patellar tilt, lateral translation of the patella, subluxation (sensation of giving way).

Contra-Indications: The presence of arthritis on the medial half of the patella.

Complications: Since the origin of the VMO is posterior to the plane of the patella, concerns have been voiced with respect to creating a posterior vector on the patella. If we pull or push the medial portion of the patella posteriorly are we not going to increase pressure on the patella and increase pain? The answer to this is unknown. In defense of the procedure, the patella is "pulled down" only as much as is needed to bring it back to a normal position.

Numbness: Sensation about the knee comes to a large extent from branches of the saphenous nerve. This nerve sweeps around the medial aspect of the knee and fans around anteriorly. The skin incision interrupts some of these branches and numbness lateral to the incision ensues. This numbness can resolve in time and usually does not interfere with every day function. Patients undergoing patellar re-alignment should be willing to trade some pain for an element of numbness which is not to say that - all things being equal- the surgeon shouldn’t try to minimize it. In my experience, patellar patients rate pain relief and cosmesis above sensation.

Tibial tuberosity transfer (Elmslie-Trillat)

The concept of displacing the tibial tuberosity to affect patellar tracking goes back to the 19th century. The Swiss Roux described displacing the patellar tendon in 1888 (see below). He was also one of the first to publish the observation that patellar dislocation is a problem that occurs near extension (slight flexion). People do not dislocate their patella going up and down stairs, but they readily dislocate with social dancing.

Elmslie popularized the procedure in Great Britain and it was further popularized in the Western world by Albert Trillat from Lyon, France, and by Cox in the USA. By transferring the tibial tuberosity medially, one decreases the quadriceps (Q) angle. Accordingly one decreases the bowstring effect of the quadriceps mechanism, and, most significantly the tendency of the patella to move laterally when the quadriceps contracts.

Indications: The presence of an increased quadriceps (Q) angle is the prime indication for the procedure since the main effect of the operation is to decrease the Q angle.

Contra-indications: 1) An open apophysis at the tibial tuberosity. Premature closure of this apophysis can ensue with a resultant recurvatum deformity. 2) Ligamentous laxity, in combination with an out-toeing gait. Medial transfer of the tibial tuberosity decreases the internal rotation moment about the knee. As a result, the tibia is free to externally rotate even more than before and the patient can end up with a Charlie Chaplin gait (not a crowd pleaser).

Complications: Loss of fixation is the most obvious complication though not the most serious. Excessive medialization. Some surgeons will medialize the tibial tuberosity even in the presence of a normal Q angle. This leads to abnormally low Q angle. There is some disagreement here as some investigators have disapproved of this. For example, Insall has stated that the quadriceps line of pull "should never be angled even minimally to the medial side". Conceptually I feel that this is correct since the main goal of surgery is to restore normal anatomy and function. But in some patients medialization of the line of pull may be the best way to unload painful lateral lesions.


Maquet procedure: Anterior elevation of the tibial tuberosity (the site of the patellar tendon insertion) is referred to as the Maquet procedure. It can be argued that this is not a realignment procedure as it corrects neither the medial-lateral position nor the tilt of the patella. In 1963, Paul Maquet from Liège, Belgium was one of the first to describe the procedure. It is based on the concept that articular pressure (and therefore pain) can be diminished if the patella is elevated off the trochlea. Much has been published on the subject as a result of many in vitro and some in vivo studies. Despite this, there is no agreement on the theoretical foundations, the exact surgical technique nor the clinical results of the procedure. The procedure differs conceptually from the operations described in prior paragraphs to the extent that it makes no effort to create normal mechanics. In fact, it specifically changes the mechanics to diminish pain during activities of daily living.

The Maquet procedure is usually combined with a lateral release. Sorting out the benefits of the lateral release from those of the tuberosity elevation is therefore difficult.

Indications: Overall, it would appear that indications are decreasing. Most controversial would be the use of this procedure for "patellar pain" without arthritis.

Contra-indications: Raising the tibial tuberosity places some increased tension on the overlying skin. Unhealthy skin of any kind is therefore a relative contra-indication for this type of operation. Because the tuberosity is raised and becomes more prominent, kneeling can remain painful after surgery. Patients whose jobs or hobbies involve much kneeling are at particular risk for this. Global patello-femoral arthritis as noted above is probably not a good indication for the procedure.

Complications: Wound healing is the complication traditionally associated with this operation. Maquet himself recognized this problem in his original works. Healing problems can be minimized by either using a lesser elevation or by using a long shingle. Early range of motion is probably not a good idea either. With the patient in a cast or immobilizer this is usually not an issue.

Subsidence of the graft with loss of the tuberosity elevation is a complication particular to this procedure. The graft has to be sufficiently strong to avoid being crushed. Pressure on the graft has to be great enough to keep the graft in place, but not so strong as to crush the graft or cause it to subside into the underlying cancellous bed. Tendinitis. Pain at the patellar tendon insertion has been reported. It may be due to altered mechanics, an altered angle between the tendon and the tuberosity, or both.

The Elmslie-Trillat osteotomy and the Maquet osteotomy represent simple one-plane mobilizations of the patellar tendon insertion (tibial tuberosity): the former medializes, the latter elevates ("anteriorizes" or "ventralizes"). The tibial tuberosity can also be osteotomized in such a way as to provide a measure of both medialization and elevation. The simplest way of combining elevation with medialization is to angle one straight cut. The cut is the same as for an Elmslie-Trillat except that it is angled in an antero-medial to postero-lateral direction. Thus when the tuberosity is displaced medially, it is also displaced anteriorly. In the United States this procedure has been popularized by J. P. Fulkerson, MD and called the:

Antero-medialization (AMZ, osteotomy ("Fulkerson procedure" in the USA)

The advantage of this osteotomy is that with one cut both medialization and elevation can be obtained. No bone graft is needed (though it can be added for increased elevation). The disadvantage is that one loses correction relative to one-plane osteotomies. For example, a 1 cm displacement along a cut angled 45 degrees leads to a half cm of medialization and a half cm of vertical displacement. A fortuitous byproduct of this osteotomy is that screws placed perpendicular to the osteotomy are aimed away from the popliteal vessels.

This osteotomy most likely unloads the disto- lateral aspect of the patella and loads its supero-medial aspect.

Indications: for the antero-medialization (AMZ) osteotomy. Broadly speaking, this osteotomy is indicated whenever both elevation and medialization are desired.

Contra-indications: Poor skin anteriorly is a relative contra-indication. Elevations obtained with this procedure tend not to be as high as with the pure (2.5cm) Maquet elevation, and as such may be safer in this setting. Supero-medial articular cartilage lesions contra-indicate the AMZ osteotomy, as does the presence of arthritis in the trochlea.

Complications: These tend to be the same as those listed for the Elmslie-Trillat and the Maquet. Wound healing is somewhat less of a problem than with the standard Maquet since elevations are not as considerable. Loss of fixation has been known to occur, as have tibia fractures.


Distal transfer of the tibial tuberosity. In situations where the patella is riding high (patella alta) consideration can be given to lowering it. This can be done by osteotomizing the tibial tuberosity, freeing it up on three sides (medial, lateral, and distal), removing a portion of the tuberosity at its lower end, and transferring the remainder of the tuberosity distally into the underlying bony bed. Unfortunately, this procedure is often confused with the Hauser operation. Developed by Emil Hauser from Chicago to address patellar dislocation, this operation transferred the patella distally, medially, and - by virtue of the triangular shape of the tibia- posteriorly. The purpose of the operation was to cinch down an extensor mechanism viewed as being too lax and too lateralized. The operation was a success with respect to correction of dislocation but a failure with respect to long term function: in the long run the tightened extensor mechanism led to patellofemoral arthritis. Serious consideration should given to transferring the tibial tuberosity of patients with a Hauser procedure who demonstrate symptoms pertaining to the patella.

Just about any medial soft tissue structure can (and has) been used to pull the patella medially. In addition to the VMO, the pes tendons (semitendinosis and gracilis) have been released proximally, woven through a patellar tunnel and re-attached at the level of their insertion on the upper medial tibia. For example, the semitendinosus can be woven about the patella as a static medializing force in a procedure attributed to Galeazzi in 1921. This can be of use in the pediatric population where one is reluctant to work on the tibial tuberosity for fear of causing premature closure of the apophysis (with resultant recurvatum [back-kneed] deformity).

A portion of the patellar tendon can be utilized to stabilize the patella. The medial third of the tendon can be detached with a wafer of bone and transferred medially.

The lateral third of the tendon can be detached and re-attached medially. When it is passed under the remaining portion of the tendon, the procedure is referred to in the United States as the Roux-Goldthwait operation. In other locales it is simply called the Goldthwait procedure. The procedure described by Joel Goldthwait in 1904 was developed to address dislocation and what we would today call an increased Q angle. In 1904 the concept of patellar tilt had not yet been introduced and the possible effects on patellar tilt were therefore not a concern. The Boston Medical and Surgical Journal in which Goldthwait’s article appeared also informs us that the maximum temperature in Boston the week of February 11th 1904 was 29° F (now there was a Journal!). It was developed to address patellar dislocation at a time when the concept of patellar tilt did not exist. The procedure is effective in medializing the resultant forces on the patella, but it increases patellar tilt. Complications: Roux’s name is often used in different contexts. As is common in the field of patellofemoral disorders, the same term or eponym has different meanings in different parts of the world. Thus, it is instructive to review exactly what Roux wrote. He described the use of an osteotome to peel the patellar tendon off the tibial tuberosity. The tendon was then transferred to a pre-roughened portion of upper medial tibia and fixed with two smooth nails placed through the skin. [This article is truly a classic: often quoted, rarely read].

Patellar thinning operations have been described whereby the subchondral bone is removed along with the remaining articular cartilage. The term "spongialization" has been used by some to describe this removal of all subchondral bone down to cancellous ("spongy") bone.

Triangular wedges of bone (based dorsally) have been taken out in order to decrease venous pressure. Complete excision of the diseased medial portion of the patella (a form of partial patellectomy) was described by Sachs but Weaver obtained no good results in 12 knees.

On occasion the surgeon may feel that the trochlea itself is the problem. This is the case when the trochlea is dysplastic, i.e. relatively flat. This dysplasia is quite variable. In extreme cases the trochlea can even present an area of convexity giving the trochlea a "fried egg" profile. These are among the most difficult cases to surgically address. Trochleoplasties have been attempted in which a wedge of bone is placed within one or both condyles in an attempt to crank up the condyle(s) and provide a deeper trochlear groove. This is actually one of the oldest patellofemoral procedures, going back to Albee’s 1915 report. Unfortunately, the articular cartilage of the patella then no longer matches that of the trochlea and early arthritis has been reported with this approach. Masse’s trochleoplasty involved scooping out bone from under the cartilage but also led to poor results - perhaps due to the direct blows applied to the cartilage in an attempt to keep it affixed to its new subchondral bed. Kobayashi made use of both patellar and trochlear wedges. Most recently, D. Dejour and his team have used Masse’s approach with a more gentle approach to cartilage fixation.

Patello-Femoral Replacement

When a joint is severely worn out and severely painful it is reasonable in this day and age to seek to "replace" it. A patellofemoral replacement is a knee replacement involving only the kneecap and opposing trochlea. The rest of the knee is completely spared. Only a small percentage of people with knee arthritis are candidates for this operation, but in this group I have been very, very pleased with the results.

Cartilage Transplantation

A biological approach to the worn out articular cartilage is to transplant a source of cells which can develop into cartilage. This represents the holy grail of arthritis surgery and has been tried a number of ways. Cartilage can be harvested from the patient, cultured in vitro and re-implanted. The results of this approach for kneecap problems remain indeterminate.

Patellectomy & Patello-Femoral Replacement

Removal of the patella is tantamount to a resection arthroplasty of the patellofemoral joint. Its appeal lies in its perceived relative simplicity compared to other procedures. Moreover, by simply removing the patellar half of the painful patellofemoral joint it can be reasoned that the pain will disappear. On the down side, there is always the concern that the patient will develop thigh weakness or an extensor lag due to the loss of the patella’s lever function. "Patellar pain" is a complex issue, and, interestingly enough, removal of the patella does not automatically eliminate the pain. It appears that a patellectomy can work well in certain patients but not predictably so. The last word on the best technique and specific indications has not been written.

MPFL reconstruction

Reconstruction of the medial patello-femoral ligament (MPFL). This procedure has been designed for the patient whose kneecap dislocates (completely slips out of the trochlear groove). Operations go in and out of favor, and this one is currently in vogue. Since the MPFL helps keep the kneecap in its groove, it is logical to repair or reconstruct this ligament when the kneecap has dislocated. But repair only makes sense if the ligament was normal to begin with. In many patients it is stretched out even before it tears, so simple repair doesn't create a normal MPFL. Reconstruction implies that the surgeon is creating a ligament from scratch, and this means taking a ligament from some other part of the knee or using a synthetic material. I tend to recommend this to patients who’ve undergone a simpler procedure that has failed.