Osteochondral Grafting of Articular Cartilage Injuries 

Updated: Sep 26, 2017
Author: Andrew Turtel, MD; Chief Editor: Thomas M DeBerardino, MD 



Partial- and full-thickness cartilage injuries, as well as osteochondral pathology in weightbearing joints, have produced deleterious effects in knees in both the short and long term. The decreased capacity of damaged articular cartilage to heal or regenerate has contributed measurably to these effects. Surgeons, therefore, are challenged to search for ways to overcome this inadequacy in order to reestablish normal joint function in the face of trauma or disease.[1, 2, 3, 4, 5, 6]

Lesions can be traumatic or degenerative (arthritic). Whether a lesion is traumatic or degenerative is called into question more often with large lesions. It is an issue of relevance, in that all of the techniques discussed are recommended for posttraumatic lesions. Obviously, many of the procedures are performed for degenerative lesions.

Peripheral containment of the lesion is another factor when subchondral bone is markedly involved. Ideally, the periphery of the defect should be rigid enough to contain the outermost grafts. This is intuitive and important from a mechanical stability standpoint; without adequate press fit, unconstrained grafts may loosen in the early postoperative period when range of motion is started, even without weightbearing.

In the past, articular cartilage lesions have been treated by subchondral bone abrasions or drilling at the site of focal damage with procedures popularized by Pridie and Johnson.[7, 8] For osteochondral lesions, bulk autografts and allografts[9, 10] have been used. However, these generally are reserved for massive (>10 cm2) lesions.[11, 12] These procedures have evolved to modern-day techniques, but to date, no single procedure has gained universal acceptance. Both small and large articular surface lesions continue to pose challenges to surgeons.

Attempts to restore weightbearing hyaline cartilage via clinical techniques of joint resurfacing have been described. Although elderly patients can benefit from total joint replacement surgery when singular lesions or global arthrosis has affected the joint, younger patients have higher rates of failure with these procedures. Therefore, it would be advantageous to resurface symptomatic chondral and osteochondral defects to relieve the pain of those lesions and halt the progression of degenerative arthrosis. With the available reports, it appears that osteochondral grafting is efficacious for restoring weightbearing joints.[13, 14, 15]

Although biologic resurfacing may not be an appropriate first-choice procedure for patients with these problems, a large population of patients with articular surface lesions exists in whom simple debridement has failed to alleviate symptoms. Within this population, many patients are too young to consider a total joint replacement. Others simply refuse total joint replacement (regardless of age), though joint-surface incongruity or defects due to cartilage lesions have left them handicapped. With disability derived solely from articular disorders of the patellofemoral joint, trochlear replacement systems may be an option in a limited number of instances.[3, 16, 17]

When an unexpected chondral or osteochondral lesion is found during surgery or when simple debridement of damaged tissue does not suffice, a limited number of procedures appear to be available. Techniques such as microfracture,[18] popularized by Steadman,[19, 20] and autologous chondrocyte transplantation have shown some promise.[21] However, the former actually does not recreate a hyaline cartilage surface. The latter requires two procedures, is dependent upon an outside laboratory, is very expensive, and requires an arthrotomy.

Thus, transplants of autogenous or allogeneic osteochondral plugs have become popular for the following reasons:

  • They offer the chance at true hyaline cartilage resurfacing
  • They can be performed in a single procedure
  • They are performed with reusable equipment
  • They do not require outside laboratory assistance

However, unlike microfracture, osteochondral grafts are not always amenable to arthroscopic technique and may require an arthrotomy.[22]

Hangody helped promote the use of small-diameter osteochondral cylinders to resurface damaged chondral surfaces.[23, 24, 25, 26, 27, 28] His inspiration came from the noted longevity of the wooden mosaic walkways on the shores of Lake Balaton in Hungary. In Japan, Matsusue began using multiple autogenous osteochondral pegs, expanding on the work of Yamashita, who used autogenous shell autografts obtained from the noncontact areas of the femoral condyles.[29, 30]

Clinical trials began in 1992 in Hungary with instrumentation created for procurement and insertion of grafts after years of study in horses and dogs. Originally, procedures involved an open technique, with subsequent modifications to include equipment for arthroscopic techniques.

For patient education resources, see the First Aid and Injuries Center, as well as Knee Pain and Knee Injury.


The technique and science for osteochondral grafting continues to evolve, as do the indications for its use. Hangody made early suggestions for patient selection in order to maximize the chance for success. This included limiting surgery to focal lesions and patients younger than 45 years who are in good physical condition. In addition, preeducation regarding science of the grafts, informed consent on the possibility of finding an unknown lesion intraoperatively, and postoperative protocols were stressed.

Although an absolute age cutoff might seem reasonable, especially in social healthcare systems that challenge quality-of-life disabilities, other factors should be considered. Certainly, a 40-year-old patient with global arthrosis is less of a candidate than is a 60-year-old patient with a symptomatic small focal traumatic lesion. Therefore, as long as a bony healing response can be expected, a wide age range is acceptable for surgical indication.

In theory, this technique could be used for any joint surface. However, practical considerations have limited its early use to a small number of joint surfaces. The talus of the ankle has been approached in an open fashion, both with and without malleolar osteotomies. In addition, resurfacing of the shoulder and elbow has been reported.

The knee joint, because of its size and varied pathology, is the most readily approached with this technique.[18] Femoral condyles can be approached by an open or an arthroscopic technique. The retropatellar area and trochlea groove necessitate an open approach because perpendicular access to the patella usually can be obtained arthroscopically. An exception may be the knee with a patella that is sufficiently lax to allow displacement and eversion with a smaller incision. Retrograde techniques currently are being examined in various laboratories.

As already indicated, the tibia presents a unique difficulty. Because direct perpendicular access is not possible with either an open or an arthroscopic approach, an indirect retrograde method can be used. Care must be taken to obtain oblique donor grafts that match the angle of the recipient tunnel surface angle. This is a very technically demanding approach to the problem. Retrograde fill of the defect with plug(s) and elevation of the ipsilateral collateral ligament with a piece of bone are options to enable tibial access for graft transplantation.


The most obvious contraindication is global arthrosis. This does not necessarily mean chondral disease in two or three compartments; focal lesions in two or more areas of the knee may be amenable to the technique. However, where secondary changes exist (eg, osteophytes, joint space narrowing), the efficacy of the procedure is thought to be decreased.

Certainly, it is not appropriate to address the articular surface abnormality in a vacuum. Associated mechanical malalignment or instability must be addressed to maximize the long-term success of this procedure. Osteotomy for malalignment and/or ligament reconstruction for instability optimizes the mechanical milieu in which any cartilage transfer takes place. In situations where mechanical issues cannot be addressed, this must be thought of as a contraindication. Finally, tumor, synovial disease, and any other factor that would make a patient a poor candidate for delicate and complicated surgery should be strongly considered before the decision is made to proceed with this procedure.

Technical Considerations


The anatomy of the knee is reflective of its function in ambulation. Knee stability and pain-free range of motion are important in maintaining daily function. Most commonly, overuse, age, and traumatic injuries cause structural damage to the knee that may limit its function. Therefore, a thorough understanding of the anatomy of the knee is essential to properly diagnosing and treating knee pathology.

The femur is the longest and strongest bone in the human body. The proximal end forms the head of the femur, which projects anterosuperomedially to articulate with the acetabulum. The distal end is wider and forms a double condyle that articulates with the tibia and patella. The tibia articulates with the distal lateral and medial femoral condyles. The patella articulates anteriorly to the femoral condyles in the region of the intercondylar fossa (trochlear groove).

The tibia lies distal to the femur and medial to the fibula. The proximal end consists of medial and lateral condyles, an intercondylar area, and the tibial tuberosity that articulates with the medial and lateral condyles of the femur. Distally, the tibia articulates with the ankle. The distal and proximal ends of the tibia articulate with the fibula. In addition, the shaft of the tibia and fibula are connected with an interosseous membrane to form a syndesmosis joint.

The fibula does not articulate with the femur or patella. Furthermore, the fibula is not directly involved in weight transmission.

The patella is the largest sesamoid bone in the human body. This bone is flat, proximally curved, and distally tapered; however, the shape can vary. The posterior patella articulates with the femur, but the apex sits proximal to the line of the knee joint. The tendon of the quadriceps femoris completely encompasses the patella.

For more information about the relevant anatomy, see Knee Joint Anatomy, as well as Technique.

Best practices

Both partial- and full-thickness hyaline cartilage defects have well-documented progressions of degenerative pathology. Cartilage is avascular and therefore has virtually no potential to heal. Existing lesions tend to progress in severity, altering the biomechanics, rheostosis, and nutrition of the articular surfaces. These can predispose the joint to further degeneration and progressive symptomatology.[31]

For these reasons, interest and activity have increased in replenishing articular surfaces on the basis of the hypothesis that subchondral pain and joint degeneration will be thwarted. Even today, it is clear that symptoms derived from some lesions can be eradicated by techniques discussed herein. Whether or not these procedures influence future degenerative change is not clear.

This raises the issue of whether lesions not known to cause symptoms should be considered for treatment. If chondral repair is needed to reduce pain, why does debridement of chondral lesions often result in pain abatement? Should the type, depth, dimensions, or other specific lesion attributes determine the surgical action?

Provided that an answer of sorts is available for focal chondral defects, what lesions can be addressed and fixed? Certainly, global compartment arthrosis (severe joint-space narrowing or collapse, osteophyte formation, and/or subchondral cyst formation) is not amenable to cartilage resurfacing at this time. Conversely, a small (< 1 cm2) chondral lesion is a reasonable target for a chondral repair operative technique, especially if the apparent cause of significant symptoms has not been relieved by lesser therapies.

The primary dilemmas are larger lesions: Should they be operated upon, and which of the available procedures should be used?

Steadman claimed success with the microfracture technique on lesions having a diameter of up to 3 cm.[19, 20] Pedersen similarly reported success with lesions of this large dimension using chondrocyte transplantation techniques. With respect to osteochondral grafting methods (eg, mosaicplasty, osteoarticular transfer surgery), large lesions indeed are a dilemma.

The amount of tissue available for transfer is the only limitation for chondral lesions with minimal subchondral bone loss. Lesions of the femoral condyle up to 8.5 cm2 have been filled by up to 19 cylindrical osteochondral plugs, measuring 4.5-6.5 mm in diameter (see the image below). However, 4 cm2 appears to be the upper limit for lesions in which reasonable results can be expected. Hangody considers lesions larger than this to be salvage situations.

Lesions of the femoral condyle up to 8.5 cm2 have Lesions of the femoral condyle up to 8.5 cm2 have been filled by up to 19 cylindrical osteochondral plugs measuring 4.5-6.5 mm in diameter. However, 4 cm2 appears to be the upper limit for lesions in which reasonable results can be expected.

For osteochondral defects, a limit may be approached relative to width and length of lesions. Additionally, depth of bony involvement becomes a factor. Because the technique involves placing osteochondral cylindrical plugs into recipient holes on the basis of a press fit, a finite depth of lesion crater can offer sufficient stability for the cancellous bone plug. Defects deeper than 10 mm appear to compromise this stability. In these situations, a primary procedure of bone grafting may provide a secondary osteochondral grafting with a better chance at mechanical survival. No data support this depth limit at present.

Apart from the contentious issue of whether donor sites contribute morbidity and/or degenerative progression of the knee, there are the not insignificant issues of how to gain enough harvest to fill the defect and, further, to fill the defect with reasonable congruence. The more Herculean efforts demanded by chondrocyte transplantation and mosaicplasty-type procedures are justified, in theory, by their potential for restoring a surface with hyaline cartilage.[32] The microfracture technique results in a fibrocartilage surface but has reported efficacy in symptom relief, with longevity of maintained relief for nearly a decade in some instances.

Possibly, some of the drawbacks in the use of osteocartilaginous grafts for large lesions, especially donor-site morbidity and scarcity of available graft, might be alleviated with the use of allograft osteocartilaginous plugs. A respectable survival rate of chondrocytes has been demonstrated. Furthermore, plugs may be harvested from areas similar to the recipient sites (eg, femoral condylar grafts for the femoral condyle).

The vast majority of cartilage repair procedures are performed for lesions of the femur and the patellofemoral articulation. The tibia rarely is the recipient of these procedures, predominantly because of its inaccessibility and the relative infrequency of obviously traumatic lesions on the plateau. The tibia is inaccessible to all but the microfracture technique; osteocartilaginous grafts would require an oblique insertion (with an oblique harvest). Hangody has performed such procedures, but they are extraordinarily labor-intensive.[23, 24, 25, 26, 27, 28] Oblique allografts might lessen the burden.

Technically, a chondrocyte transplantation procedure upon the tibia would be very difficult to perform. The development of matrices, laden with chondrocyte, growth factors, and cytokines, representing induction, conduction, and a vehicle may threaten current techniques of cartilage repair.[33, 34]


Periprocedural Care

Preprocedural Planning

Patients usually present with mechanical complaints. These symptoms are sometimes difficult to distinguish from more common meniscal symptoms. Routine imaging studies, such as radiography and magnetic resonance imaging (MRI) usually are performed to define the lesion, as with any intra-articular knee problem.[35]

When osteochondral injuries are present, they usually are easily distinguishable on MRI. More difficult is the situation in which only a chondral injury is present; these often are missed on routine MRI. More often than not, these lesions are first detected at the time of the arthroscopy; a high index of suspicion is needed preoperatively so that this can be discussed at that time, and surprise at the time of surgery can be eliminated.

Once the size and location of the lesion have been accurately determined, a decision is made to perform the transplant in either an open or an arthroscopic fashion. Generally, patellofemoral articulation is approached with an open procedure. For femoral condyles, there is more latitude in the decision-making process, and the decision is based on many factors. First and foremost is the surgeon's familiarity with the procedure. For the individual surgeon, initial procedures should be done via an open technique unless the surgeon has extensive lab experience. This is true even for cases that appear to be straightforward.

Intimate familiarity with instrumentation is critical, and on a first-time basis, an open procedure allows more accurate recipient and donor site preparation because the surgeon has total perspective of the instrumentation for both the recipient hole and donor harvesting. The instrumentation is relatively large and cannot be seen easily in its entire circumference arthroscopically. This and the fact that extreme flexion angles (which close down the anterior capsule) are occasionally necessary make the arthroscopic procedure technically demanding. Even an experienced arthroscopist may have orientation problems using the transplant equipment for the first time, and this may prove detrimental.

The second critical factor is the size of the lesion. Femoral condyle defects larger than 1.5 cm in diameter or lesions in which more than half of the lesion is posterior to the center of the weightbearing surface should be approached via an open technique. Gaining perspective arthroscopically is more difficult with larger lesions, making it difficult to place multiple transplants accurately enough to recreate proper contour. For lesions posterior to the weightbearing area, the flexion angle needed makes visualization difficult, and the patella may become an obstacle.

Femoral condyle lesions smaller than 1.5 cm2 are thought to be appropriate for an arthroscopic approach when the surgeon has sufficient experience with the procedure. In reality, these are arbitrary lesion measurements, but until more data concerning the efficacy of the procedure are available, this relative lesion size seems to have become commonly accepted.

Patient Preparation

Surgical technique begins with patient positioning. After induction of general or regional anesthesia, a tourniquet is placed high on the thigh. Tourniquet use is not mandatory; however, it may be advantageous later. A leg holder is suggested but not mandatory. Epinephrine may be used in the irrigation fluid at the discretion of the surgeon.

As in routine arthroscopic cases, the leg must be capable of flexion to 120° so that the majority of the femoral condyle and any symptomatic lesions can be visualized. It is important to achieve this flexion in both a varus and a valgus stressed position, depending on the compartment where the pathology is found.

Usually, arthroscopy is performed to visualize the defect if this was not done previously.



Approach Considerations

The use of large autogenous osteochondral fragments and patellar grafts has been reported, but results have been mixed, and concern exists regarding donor-site morbidity from such large grafts.

There is growing interest in the concept of smaller and more uniform cylindrical grafts, obtained locally, that can be implanted into prepared recipient sites in the lesion. Although there are some technical differences between the various commercially marketed techniques in which cylindrical osteochondral plugs are transferred, the overall concepts are similar. The technique can be used in specific situations of deficit size and location in properly selected patients.[36]

The biology of these grafts is well documented in animal and clinical studies. Strong data support the ability of cancellous bone plugs to heal, whether the recipient holes have been drilled, trephined, or cored (see the first image below). Biopsy studies also have shown the ability of transplanted cartilage to survive if placed in a mechanically advantageous position (see the second image below). 

Strong data support the ability of cancellous bone Strong data support the ability of cancellous bone plugs to heal, whether the recipient holes have been drilled, trephined, or cored.
Biopsy studies have shown the ability of transplan Biopsy studies have shown the ability of transplanted cartilage to survive if placed in a mechanically advantageous position.

The advantage of small grafts is obvious. If the hyaline cartilage could survive, there would be a great expectation of resurfacing damaged areas with more congruence than when larger autografts are used.

Questions that persist revolve around the following:

  • The fate of the interstices between the cylindrical grafts
  • The importance of their size in what material forms in this area
  • The relative inability to resurface the entire defect area (see the image below)
  • The concept of "robbing Peter to pay Paul" by taking cartilage from a healthy part of the knee and placing it into a damaged site
The relative inability to resurface the entire def The relative inability to resurface the entire defect area is a persisting concern.

This is an exciting time for biologic resurfacing of weightbearing joints. Osteochondral transplants, in comparison with some of the other technical procedures available, have many advantages and few reported drawbacks. The goal is to resurface defects with hyaline cartilage in a one-step procedure. This procedure offers this opportunity without the need for support labs or additional costs.

True double-blind comparisons with a sufficient number of patients and lengthy follow-up time may never be possible. In this age of quickly adopting new and better procedures, zeal to repair these defects must be tempered by the lack of true understanding of whether patients are improving. With the available reports, it appears that osteochondral grafting is an efficacious procedure to restore these surfaces. However, controversy persists as to its place in the temporal scope of care of these patients.[2, 3, 4, 33, 34]

Numerous tissue-engineering studies for articular cartilage injuries are ongoing worldwide. Some of these studies have demonstrated that bioengineered cartilage tissue can regenerate when implanted in patients with cartilage injuries.[37, 38, 39, 40, 41]

Open Approach

After arthroscopy, an arthrotomy is made over the involved area. A medial or lateral anterior sagittal arthrotomy can be used, as well as a transvastus or subvastus approach for lesions on the medial femoral condyle.

For defects of the femoral condyles, the incision should be long enough distally to allow viewing of the lesion with the knee flexed, and it should extend proximally to allow viewing of the superior aspect of the trochlea (where donor grafts can be obtained) with the knee in extension. For the patellofemoral joint, the incision should be long enough to rotate the patella 90° on a longitudinal axis.

Cartilage lesions are debrided sharply back to a circumferentially stable articular cartilage. Abrasion arthroplasty of the exposed subchondral bone is then carried out (see the image below). This is performed even though the surface will be resurfaced to encourage the interstices between the grafts to form a fibrocartilage grout or seal between the native cartilage and the grafts.

After identification of the lesion, all cartilage After identification of the lesion, all cartilage is removed down to the subchondral bone. The edges of the lesion are taken back to areas of well-attached hyaline cartilage. Abrasion of exposed subchondral bone.

Hangody showed that at 8 weeks postoperatively, the areas between the cartilage interfaces seal with fibrocartilage that is generated from the abraded subchondral area (see the first image below).

Hangody has shown that at 8 weeks postoperatively, Hangody has shown that at 8 weeks postoperatively, the areas between the cartilage interfaces seal with fibrocartilage that is generated from the abraded subchondral area.

The lesion is measured in an attempt to estimate the number and sizes of grafts that will appropriately fill the lesion. An instrument with a known size (generally supplied in the instrumentation) allows accurate measurement of the lesion (see the second image below). Controversy exists regarding whether larger or smaller grafts should be used. In either case, perpendicular access is critical.

The lesion is measured in an attempt to estimate t The lesion is measured in an attempt to estimate the number and sizes of grafts that will appropriately fill the lesion. An instrument with a known size (generally supplied in the instrumentation) allows for accurate measurement of the lesion.

The appropriate graft size for a given lesion is a matter of debate. Some believe that multiple smaller diameter (2.7 mm, 3.5 mm, and 4.5 mm) grafts should be used. Others believe that larger (>5 mm diameter) but fewer grafts should be used. No data conclusively favor one opinion over another; thus, the choice appears to be personal.

In either case, after the defect is sized, the sizes and number of grafts needed are estimated. The estimate is based on a combination of measurement and experience. Regardless of whether larger or smaller grafts are used, the first grafts obtained and placed are the larger of those chosen. After these are placed, the smaller grafts can be used to fine-tune any smaller gaps remaining.

Once the size of the grafts and the number of each size graft are determined, harvesting begins (see the image below). Typically, harvest sites include the superior trochlear ridge and the intercondylar notch area. Considerable debate exists regarding where the best hyaline cartilage for grafting can be harvested. The periphery of the supracondylar ridge is the most commonly used area, in that it has relatively thick hyaline cartilage, is relatively nonweightbearing, and is easily accessible in both  open and arthroscopic techniques.

Once the size of the grafts and the number of each Once the size of the grafts and the number of each size graft are determined, harvesting begins. Typically, harvest sites include the superior trochlear ridge and the intercondylar notch area.

Some reports have suggested use of the medial, rather than the lateral, side out of concern regarding earlier and greater patella contact on the lateral side during early flexion. The intercondylar area is useful as well, in that it can be approached arthroscopically, though fewer grafts are available because of the decreased surface area. (Previous reference has been made to questions regarding the quality and shape of the cartilage in this area.) With perpendicularity again being of utmost importance, appropriately sized grafts are harvested.

The various commercially available instrumentations have subtle differences, but the end result is that cylinders of osteochondral grafts are obtained. The devices used are specially designed tubular chisels, which allow harvesting of a core of hyaline cartilage, subchondral bone, and cancellous bone.

Care must be taken to obtain appropriate-length grafts for the defects being addressed. For chondral lesions, grafts generally are 15 mm in length, whereas for osteochondral defects, slightly longer (20 mm) grafts are needed. Grafts that are too short compromise the surface area of the press fit and are not stable enough. Longer (>20 mm) grafts generally are unnecessary. For the patella, slightly shorter grafts are used owing to its thickness, depending on the facet being resurfaced.

Commercial devices each have a particular way to break or cut the medullary bone base so that the core can be removed. Each has its own recommendations and admonitions for its use, which should be studied in detail.

While inside the harvesting device, the base of the graft is either cut or broken in a controlled manner. The extractor should not be spun to remove the harvester before the base is broken, because the graft may become loose in the device, making removal difficult. The grafts are removed from the device by tamping on the cancellous side to avoid damaging the hyaline cartilage.

After the graft is removed, it is inspected for damage. The hyaline cartilage thickness and its appropriate adherence to its subchondral bone are noted. In addition, the angle that the articular cartilage makes with the long axis of the cylinder is examined. Ideally, it should be perpendicular to this axis. The graft is placed in isotonic sodium chloride solution–soaked gauze.

The recipient site now is prepared. The periphery is the best place to start, and attention is directed to the surrounding surfaces, the radius of curvature, and the donor graft that has been obtained. With these factors in mind, an appropriate recipient hole is created, so that when filled with the donor graft, it will recreate the surface intended. These holes can be drilled, trephined, or cored, depending on the technique being used. The depth of holes varies. For chondral lesions, about 15 mm is needed, and for osteochondral lesions, 20 mm is necessary. The hole is inspected.

Some techniques call for impacting the internal cancellous bone of the created tunnel in order to remove any impediment to placement of the graft. The graft then is tamped gently into the recipient hole (see the image below). It need not bottom out, in that it is a circumferential press that creates stability. Once this is done, successive grafts are harvested and placed.

Graft insertion. Graft insertion.

Grafting is started at the periphery of the lesion, closest to the major part of the weightbearing area. The sequence of progression is to fill this major weightbearing periphery and work toward the center. As noted above, regardless of the commercial system being used, the larger of the grafts are used initially, followed by smaller grafts that fill in any gaps left by the larger grafts.

Depending on the number of grafts needed, spacing of the grafts must be planned. The grafts should not lie directly side by side, because stability will be compromised. Approximately 0.5-1 mm of bone is left between each graft. This ensures a solid wall for the press fit.

In addition, care must be taken so that convergence does not cause the grafts to hit each other, which will damage the cancellous bone of its base and affect its stability and ability to be fully seated. Occasionally, this cannot be avoided, because the radius of curvature changes too rapidly. If convergence happens at depth, this should not pose a problem.

In short, grafts should be inserted more for their tangency to the articular surface than to avoid deep convergence. Recipient holes generally are drilled 2-3 mm past the length of the individual donor grafts. The sequence of recipient hole creation and recipient hole filling is important because as recipient hole filling affects subsequent holes. Avoid the urge to save time by creating all of the holes first with the thought of filling them later.

It is acceptable to obtain multiple donor grafts if the size and numbers needed are adequate. As the defect is filled from the periphery, grafts progress from larger to smaller, enabling fine-tuning of the amount of coverage achieved.

Finally, the surface is examined to ensure that the grafts are at the proper depth. At first, it may be prudent to leave the grafts too proud (~1 mm) rather than too deep; extraction, though possible, can be difficult and risks damaging the grafts. The knee is taken through a final range of motion. Routine closure is done in layers. Use of a drain is optional.

Arthroscopic Approach

The arthroscopic approach is technically challenging. Both location and lesion size determine whether the experienced surgeon chooses this route. Again, perpendicular access and portal placement are critical. Generally, portals are slightly more central than usual because the approach to the main weightbearing areas points more centrally than expected. After debridement and subchondral abrasion, the lesion is measured for size.

If the working portal being used does not appear to be perpendicular, knee flexion can be altered or a spinal needle can be used to reassess proper portal placement and perpendicular position. Multiple viewing angles are used to be sure that the measuring device is flush on the lesion from edge to edge to accurately gauge its size, to determine the number and size(s) of grafts needed, and to assess the direction that is perpendicular to the surface. This is more difficult arthroscopically, and experience with the open technique is beneficial here.

Donor grafts are obtained from either the supracondylar ridge or the intercondylar notch. The medial trochlea is easier to approach when the scope is used. As the knee inflates with fluid, the patella naturally moves laterally away from the medial ridge. The lateral side is used for the intercondylar notch. This site is useful when only a few grafts are needed. The supracondylar ridge can be approached with the arthroscopic donor instrumentation via a portal or small open incision. As in the open procedure, care must be taken to ensure that the harvesting device is perpendicular to the articular surface.

Generally, the ipsilateral inferior portal can be used to visualize the harvest sites. In taking multiple transplants arthroscopically, it is important to remember that the previous donor site must be visualized while subsequent grafts are taken. If the previous donor site cannot be seen, there is a risk of breaking into that site. Therefore, taking grafts in succession is suggested, going away from the visualization portal. In this way, the edge of the donor site can be seen and protected from being broken through.

Either a cylindrical chisel or a drill creates recipient holes. Care must be taken because significant changes of curvature radius necessitate marked changes in the approach angle of the device creating the recipient hole. Perform multiple visual checks before committing to coring or drilling. With some instrumentation devices, a dilator is used to smooth and compact the recipient canal of cancellous bone. This is helpful in arthroscopic procedures, in that irrigation fluid can cause narrowing of the canal secondary to cancellous bone swelling. In addition, debris can find its way into the canal and inhibit graft insertion.

The grafts are impacted into place with specialized tamps, and the next recipient hole is created if necessary, as in the image below. Wounds are closed in routine fashion, and a compression dressing is placed on the knee.

Graft insertion. Graft insertion.

Postoperative Care

Controversy exists regarding postoperative protocols for these procedures. Some use fewer and larger grafts and recommend shorter (2- to 3-day) nonweightbearing periods. For most techniques and surgeons who have performed them, especially for larger lesions, when smaller grafts were used, a more cautious attitude toward weightbearing was initially adopted (6-8 weeks), but the approach has since become more aggressive (2-3 weeks).

Patients in Hangody's early series were encouraged to remain nonweightbearing for approximately 6-8 weeks to allow for cancellous bone healing. This period is not arbitrary; excessive compressive pressure may tend to force the transplants into a more recessed position prior to cancellous healing.

In his original animal studies with dogs in 1991, Hangody showed early 4-week healing of cancellous bone cylinders. However, when the animals were allowed to bear weight immediately after surgery, approximately one third of the grafts in weightbearing areas showed subsidence. None of the grafts in nonweightbearing areas showed any sign of subsidence. This prompted the suggestion of nonweightbearing for 6-8 weeks in early clinical practice.

Since 1994, in the largest series to date, Hangody has revised this suggestion to 2-3 weeks of nonweightbearing, with a slow progression through partial weightbearing on to full weightbearing over the following 2-3 weeks. Larger lesions probably necessitate a more conservative approach postoperatively because more grafts generally are used to resurface a larger area. At present, not enough follow-up information is available to determine what is needed for each size and depth of lesion. Smaller lesions may need less respite from weightbearing, but this area of discussion is still open to debate.

When surgery is performed on the trochlear groove, the retropatellar area, or both, weightbearing is allowed while a knee immobilizer is worn. This decreases the contact pressures at the patellofemoral joint. The immobilizer can be removed for sedentary activity and range-of-motion exercises during the initial 3-week postoperative period, followed by progressive unsupported weightbearing.


Donor-site morbidity remains a concern for osteochondral transplantation. Hangody recognized a 3% complication rate, which included excessive postoperative bleeding (with the larger of his grafts) and donor-site pain. In the long term, he has not noted any deleterious effects of graft harvest.

Arthroscopic follow-up of the graft sites shows filling of the defects with connective tissue at depth and a fibrocartilage cap at the surface. In these situations, no evidence of degeneration is present at the site of or the opposite side of the joint. Work is under way on developing plugs to fill these defects so as to reduce the incidence of immediate postoperative bleeding and further promote filling of the defect.

Another concern is that the various techniques call for harvesting of osteochondral grafts from nonarticular areas. However, it has been demonstrated that these areas are indeed contact- and stress-bearing.

In a study specifically targeting donor sites, significant contact stresses were recorded from 0º to 110º of knee motion. This was done by creating donor site defects by obtaining round osteochondral plugs (8 mm in diameter) from some of the recommended sites. The lateral superior trochlear area above the femoral condyles and medial and lateral intercondylar notch areas were used as areas of harvest. Unfortunately, no data were reported from one of the more favored locations of harvest, the medial femoral condyle periphery of the patellofemoral joint.

No long-term studies demonstrate whether articular contact contributes to degenerative changes at these donor sites. However, data exist confirming increased stress concentration at the rim of weightbearing osteochondral defects in smaller lesions than these. The long-term clinical significance of these findings is unknown. Certainly, the utmost care must be taken to ensure that the areas of harvest are remote from major weightbearing areas.

Work currently is under way to assess bone cores and articular surfaces via magnetic resonance imaging (MRI). This will be helpful, in that the techniques learned should help improve the diagnosis of cartilage injuries and assess transplants.

During this assessment, an interesting observation has been made. Significant metallic debris has been noted at the sites of implantation of osteochondral plugs. The cause has not been definitively established; however, some of the devices that use disposable instruments for implantation are thought to be at potential risk for shedding metallic debris. The decreased tough material of disposable instruments comes into question; reported cases of outright device failure and collapse of coring devices may represent a more catastrophic failure, with more subtle failure and debris generation going unnoticed at the time of surgery.

The significance or long-term problems related to this debris is unknown, but its occurrence must be considered in the performance of postoperative MRI.

Computed tomography (CT) arthrography has been suggested as a potentially useful alternative to radiography or MRI for postoperative evaluation of osteochondral allograft transplants of the distal femur; however, it does not appear to predict functional outcome.[42]