Osteochondral Grafting using the Mosaicplasty Technique

Osteochondral Grafting
using the Mosaicplasty Technique
P. Christel*, G. Versier**, Ph. Landreau***, P. Djian****
Clinique Nollet, 75017 PARIS - ** Hôpital Begin, Saint-Mandé -
*** Clinique des Lilas, 93260 Les Lilas - **** Clinique Chantereine, 77177 Brou sur Chantereine - France

Treatment of cartilage and osteochondral defects has always posed difficult problems for orthopaedic surgeons (Fig. 1). Conventional methods, Pridie’s perforations, microfractures or subchondral abrasion (abrasion arthroplasty) lead to imperfect results in around 50% of cases (Fig. 2). These techniques lead to the formation of fibrocartilaginous scar tissue whose biomechanical properties are significantly inferior to those of hyaline cartilage, and which does not prevent the onset or progression of a degenerative arthropathy.

Figure 1: The problem: cartilage fracture with detached cartilage. Figure 2: Conventional treatment: excision of detached cartilage, subchondral abrasion, Pridie's perforations.

CURRENT METHODS OF REPAIRING CARTILAGE DEFECTS

In order to overcome the drawbacks inherent in traditional methods, alternative methods have been developed: osteochondral or chondrocyte allografts and autografts.
Good results for osteochondral allografts have been published. However, owing to the risk of transmission of viral diseases and uncertainties surrounding maintenance of the properties of the transplanted tissue, many authors have switched to using autografts. Various types of autograft are available:
Grafts using autologous chondrocytes involve two surgical procedures: one to remove a disc of hyaline cartilage, the other to implant the chondrocytes after they have been cultivated. This involves cumbersome and expensive biotechnology. Implantation of the cultivated chondrocytes involves an open procedure, with harvesting of a graft of periosteum sutured to the periphery of the cartilage defect, which will cover the gel containing the chondrocytes. The efficacy of this method has not yet been scientifically proved. In particular, the new cartilage has not been shown to adhere to the subchondral bone or to the periphery of the cartilage defect. In addition, it is not known whether the chondrocytes obtained following tissue culture are capable of maintaining their phenotypic expression in the long term.
Autologous periosteal grafts involve removing a section of periosteum, the size being dependent on the area to be covered. The lower face of the periosteum, which contains the stem cells capable of differentiation into chondrocytes, is turned towards the subchondral bone, which has first been prepared (dechondrified). The defect is filled with organic cement, and then the periosteum, sutured to the periphery of the cartilage. Thus this method is identical with the previous one, except for the fact that cultivated chondrocytes are not used. This method is simple, and the results that have been published on series of procedures which, although small, have been well researched, are extremely encouraging.
Large autografts come up against the problem of morbidity due to the size of the donor site area and that of re-establishing the local radius of curvature, in other words articular congruence, with the risk of secondary mechanical conflict.
Mosaic autografts (mosaicplasty): more recently, several authors have suggested using not an osteochondral autograft in a single block, but a collection of small osteochondral cylinders inserted side by side, thus making it possible to maintain the radius of curvature of the articular surface, or congruence. The advantage of this technique is the integration of the spongy element of the graft, which fuses with the spongy bed at the recipient site, and the integration of the transplanted cartilage with the adjacent hyaline cartilage by means of the fibrocartilage which forms between the various grafts, growing upward from the prepared subchondral bed (Fig. 3). This technique, known as mosaic grafting (mosaicplasty), was developed and perfected by L. Hangody in Hungary in the early 1990s. The first animal experiments were conducted in 1991, and the first human grafts were carried out in 1992. This technique has been increasingly used in Europe since 1995-1996, and it was given extensive coverage at the recent ESSKA Congress, held in Nice between 29 April and 2 May 1998.

Figure 3: Histological view following mosaicplasty in a German shepherd dog.

A - After four weeks (HE, x 20, graft on right of photo): complete osseous fusion. There is a space between the hyaline cartilage surfaces, but connecting tissue is already visible at the bottom of this space (coll. L. Hangody).

B - Six weeks after implantation (HE, x 20, graft on right of photo): the space between the cartilage is filled with fibrous connective tissue (coll. L. Hangody). C - After eight weeks (HE, x 20, graft on right of photo): the connective tissue filling the space between the hyaline cartilage obliterates the differences in curvature of the surface area (coll. L. Hangody).

Principles of osteochondral grafts using the mosaic technique
This is a technique in which, in the course of a single procedure, cylindrical osteochondral grafts are removed from a donor site and transported to holes prepared at the recipient site. The most commonly used donor sites are, in order of preference: the non-weight-bearing patellofemoral area, the medial, then lateral, rim of the femoral trochlea, and finally, the periphery of the intercondylar notch. Cylindrical holes of a depth and diameter appropriate to those of the grafts are prepared at the site of the cartilage defect and then filled with the cylindrical grafts harvested, in such a way as to restore the original curvature of the articular surface.
Experimental studies have shown that to obtain good results, at least 70% of the defective cartilage surface needs to be replaced. Histological studies have shown that around 10 weeks after the procedure, the grafted area contains 60% to 70% hyaline cartilage and 30% to 40% fibrocartilage, the latter being formed from the spongy bone prepared in advance. Identical results have been found in human biopsies taken by L. Hangody, even after 5 years.

SURGICAL TECHNIQUE

Our description is based on the primary technique, namely the graft performed on a cartilage defect in a femoral condyle. The procedure can be carried out via either arthrotomy or arthroscopy.
Arthrotomy technique
The initial route is parapatellar, medial or lateral depending on the site of the condylar lesion. First, a curette or a bistoury is used to bring back the edges of the cartilage defect to good hyaline cartilage at a right angle (Fig. 4). The base of the lesion is then abraded to obtain viable subchondral bone (Fig. 5), which encourages the production of fibrocartilage between the grafts.

Figure 4: Debridement of the cartilage defect. The edges are brought back to a right angle and any detached cartilage is removed. Figure 5: The base of the lesion is abraded to viable subchondral bone.

Under direct visualization, the first tunnel is drilled perpendicular to the surface of the condyle, using an appropriate drill bit, between 2.7 and 8.5 mm in diameter (Fig. 6). The depth of the tunnel is 15 mm for cartilage tears and 20 to 25 mm for osteochondritis dissecans. The greater the depth of the bed of the osteochondritis, the greater the depth of the hole will be.

Figure 6 : Under direct visualization, the first recipient tunnel is drilled perpendicular to the surface of the condyle.

A cylindrical graft is then harvested from the medial rim of the femoral trochlea, away from femoropatellar weight-bearing areas, with a tubular chisel of the same diameter as the drill bit used to prepare the hole at the recipient site. An essential point is that the chisel must be carefully positioned perpendicular to the cartilage surface (Fig. 7). It is driven in to a depth corresponding to that of the recipient site (Fig. 8). The chisel is then toggled, causing the graft to break free at the chisel tip (Fig. 9). The chisel, containing the graft, is then removed, and the graft can easily be ejected from the chisel with the aid of a guide and a harvesting tamp (Fig. 10).

Figure 7: The graft is harvested perpendicular to the surface of the donor site. Figure 8: The tubular chisel is driven by hammer to the desired depth, avoiding any overheating that could put the viability of the graft at risk.

Figure 9: The graft is detached from the trochlea in two perpendicular toggling movements, without rotating the chisel. The degree of movement must be carefully judged, to avoid fracturing the rim of the trochlea. Figure 10: The graft is gently ejected from the chisel, using a tamp and an appropriately sized guide.

The recipient hole is then dilated with a dilator, a tapered instrument which makes it possible to widen the upper part of the hole and thus to insert the osteochondral graft, then to make it penetrate further into the recipient hole without difficulty.
A guide tube is put in place at the opening of the recipient hole, then the cylindrical graft is inserted into the hole using a graduated harvesting tamp, enabling the graft to be inserted to the desired depth, such that the cartilage surface of the graft is level with the adjacent hyaline cartilage (Fig. 11). Ideally, the length of the osteochondral graft should match the depth of the hole and the level of the cartilage/osteochondral defect. A wide tamp can be used to adjust the level of the graft in relation to the adjacent surface (Fig. 12).

Figure 11: The graft is inserted, using a guide tube and a graduated tamp. Figure 12: The level of the graft is adjusted.

Once the first graft has been put in place, the procedure is repeated as many times as necessary, maintaining a distance of a maximum of 1 mm between the various grafts at the recipient site (Fig. 13). At the donor site, a distance of around 3 mm should be maintained between the various holes. The diameter and the number of holes to be used depend on the size of the cartilage defect and its morphology.
It should be borne in mind that the grafts should cover at least 70% of the cartilage defect (Fig. 14).

Figure 13: A second graft is inserted in the same way 1 mm from the preceding graft. Figure 14: Appearance on completion: grafts in place and donor sites.

Arthroscopic technique
This technique is feasible if the lesion is less than 2 cm in diameter and a maximum of 4 to 6 grafts are sufficient to cover it. The arthroscopic technique is a difficult one and requires a surgeon trained in the technique. The choice of portal is critical, since the instruments must be positioned perpendicular to the recipient site (Fig. 15). Several portals will sometimes be required in order to ensure perpendicular access. An intramuscular needle can be used to help to determine the site of the access portal (Fig. 16). Since the condylar defect is usually situated close to the intercondylar notch, the working portal is generally more central than the usual anteromedial or anterolateral portals. In addition, the patient’s position must permit 120° flexion of the knee.

Figure 15: The shading shows the areas of the condyle that can be grafted by arthroscopy. Figure 16: A needle is used to determine the site for the arthroscopic portal such that the instruments will be perpendicular to the area to be grafted.

The principles of the procedure are similar to those of the open procedure: the detached cartilage must be excised and the edges of the defect brought back until they are perpendicular to the subchondral surface. Using a motorized burr, the subchondral bone must be abraded until it bleeds. In the case of osteochondritis dissecans, an increased level of curettage is necessary, in order to remove the sclerous tissue of the bed of the osteochondritis. A guide tube of between 2.7 and 4.5 mm in diameter is then inserted (Fig. 17). This makes it possible to determine the number and size of the grafts required.

Figure 17: A - The guide tube is inserted for the first graft. B - Endoscopic view

The tubular chisel is then introduced via an ancillary medial parapatellar portal, in order to harvest the osteochondral grafts from the medial rim of the trochlea. It is difficult to harvest the grafts arthroscopically, and pronounced lateral rotation of the patella is required. It is essential to remember that the grafts must be harvested perpendicular to the cartilage surface, and in the event of problems, one should not hesitate to resort to a small medial parapatellar arthrotomy of 1.5 to 2 cm.

Figure 18: Grafts ready for implanting: 8 x 4.5 mm grafts and 2 x 3.5 mm grafts, between 15 and 20 mm in length.

Once the number of grafts required have been harvested (Fig. 18), the guide tube is put back in place. This tube is provided with lateral windows permitting arthroscopic monitoring of the progress of the graduated drill bit which will be used to drill the hole (Fig. 19). Once the desired depth has been achieved, the drill bit is replaced by a dilator, which is also graduated (Fig. 20). The graft can then be put in place and driven in to the desired depth, following the rules described earlier (Fig. 21).

Figure 19: Drilling of the hole.
A - The depth of the drill bit is monitored via the window in the guide tube. B - Endoscopic view

Figure 20: Insertion of the dilator.

Figure 21: A - Insertion of the graft into the dilated hole. B - Endoscopic view

The procedure is repeated as many times as necessary, following the same sequence each time: drill, dilate, deliver. Since it is not possible to harvest more than 4 grafts at most, 4 or 5 mm in diameter, from the medial rim of the trochlea, it may be necessary to supplement these by harvesting grafts from the periphery of the intercondylar notch. If this is still insufficient, the lateral rim of the trochlea will have to be harvested, but this is only feasible via arthrotomy, owing to the fact that the extensor apparatus is bent outwards. However, it should be remembered that the cartilage of the trochlea is thicker and of better quality than that of the intercondylar notch. Moreover, grafts harvested too close to the boundary of the bone tissue do not contain hyaline cartilage only.
As in the open procedure, at least 70% of the cartilage defect should be covered (Fig. 22). Grafts should be 15 mm in length for straightforward cartilage defects and 20 to 25 mm in length in the case of osteochondritis.

Figure 22: Appearance on completion, endoscopic view.

In general, use of the arthroscope makes it possible to achieve more precise delivery and positioning than the open procedure, especially as regards re-establishing congruence, since differences in the level of the cartilage are more clearly visible with the arthroscope owing to the optical magnification.
Post-operative treatment
In both cases, arthrotomy and arthroscopy, an intra-articular Redon drain is inserted for 24 hours. The patient is instructed to remain non-weight-bearing for a period of between 4 and 8 weeks, depending on the size and position of the cartilage defect. The return to weight-bearing is gradual. Throughout this time, the knee can be mobilized without putting the grafts at risk.

OTHER AREAS

With regard to the knee joint, L. Hangody has recently proposed extending this technique to cartilage defects in the tibia, patella and trochlea, excluding the degenerative lesions involved in arthrosis. The results reported are encouraging, but the place of complementary actions to realign the skeleton has yet to be determined: tibial osteotomy, transposition of tibial tuberosity, at the same time or in advance.
Osteochondral lesions of the talus also constitute an excellent indication for mosaicplasty. Not so much anterolateral lesions, always traumatic in origin, which are often small and respond well to excision/curettage via arthroscopy, but, above all, posteromedial lesions, genuine osteochondritis, often poorly tolerated. The functional prognosis is not the same for the latter lesions following excision/curettage as for anterolateral lesions.
It is not easy to gain access to posteromedial osteochondral lesions of the talus. They are not accessible, in their entirety, via an anterior surgical approach. Recourse has to be made to an osteotomy of the medial malleolus. The path of the osteosynthesis screw is prepared in advance of the osteotomy, which is deliberately carried out obliquely, in a downward and outward direction, using a fine circular crosscut saw, in order to make it possible for the tubular chisels to have an angle of attack perpendicular to the surface of the talus. When the osteotomy is performed, care needs to be taken with regard to the tendons of the posterior leg muscle and the toe flexors.
The surgical procedure is identical with that already described for the femoral condyles. The osteochondral grafts are harvested from the medial rim of the trochlea of the ipsilateral knee, either via arthroscopy or via a small arthrotomy. Overall, the morbidity rate involved in this harvesting is very low, but naturally the patient needs to be advised of this possibility and informed consent must be obtained.
Once the grafts have been inserted in the recipient site, the medial malleolus is fixed with a Maconor type intermediate screw, 4.5 mm in diameter. The patient is instructed to remain non-weight-bearing for at least one month, until the malleolar osteotomy is consolidated. Post-operative immobilization is not necessary. On the contrary, the ankle joint needs to be mobilized to enable the mosaic to be broken in and to prevent stiffening. The non-weight-bearing period may be longer in cases of extensive osteochondral defects.
The results of talus mosaics recently presented to the ESSKA by various authors are excellent overall. This may be due to the fact that straightforward osteochondral disease was involved, with no underlying mechanical problems, involving misalignment or instability of the ligaments.
The hip: the principle of mosaicplasty is attractive in respect of treatment of localized aseptic necrosis of the head of the femur. R. Jakob presented preliminary results of a few cases at the ESSKA Congress. This is, of course, an approach that must be seen as purely experimental. Grafts were harvested from the ipsilateral knee. It has proved to be particularly difficult to re-establish a radius of curvature identical with the original. R. Jakob suggests tamping spongy bone tissue in order to fill the areas between the cylindrical grafts. However, not enough time has elapsed and there is an insufficient number of cases to provide a definitive idea of the value of mosaicplasty for this indication.
The pitfalls of the technique
As a general rule, the grafts must be harvested perpendicular to the surface of the joint. This makes it possible to obtain grafts with a cartilage cap of homogeneous thickness and perpendicular to the axis of the subjacent subchondral and spongy cylinder. Grafts harvested obliquely result in specimens in which the cartilage surface is oblique in relation to the axis of the graft. When this is the case, the projecting element of the cartilage has very little osseous support and simply behaves like cartilage placed on an osseous surface. However, a certain degree of obliqueness may be sought in cases where the radius of curvature is very pronounced, as with the talus, for example. There appears to be a degree of tolerance as regards the obliqueness of the graft, but this has yet to be determined.
It is important not to damage the cartilage cap when the graft is being harvested, and if this shows a tendency to become detached from the subchondral plaque, an additional graft should be harvested.
In the event that a graft is higher than the adjacent articular surface, the spongy part of the graft must be shortened until it is flush with the articular surface.
If, on the other hand, the graft is too deep, it must be withdrawn without damaging, and a new graft of the correct dimensions must be used. When it is has been completely driven in, the osteochondral graft must not be able to sink, as then the reconstructed articular surface will be lower.
Subchondral grafts must be tapped down gently, in order to avoid bruising of the cartilage. R. Jakob has developed instruments which render it unnecessary to tap on the graft at the point when it is extracted from the harvesting chisel.
Coverage of the cartilage defect must be as complete as possible, and one should have no hesitation in adding grafts if there is too large an area between the grafts.

ANALYSIS OF THE LITERATURE

Following animal experiments conducted on dogs and perfection of the instrumentation on cadavers, in 1992 L. Hangody was the first to use osteochondral autografts in accordance with the mosaicplasty technique. His multicentre study currently involves 367 cases operated on using arthrotomy or arthroscopy, all areas combined. This study has provided an opportunity to make a comparative evaluation of the various methods that can be used to treat cartilage defects: subchondral abrasion (109 cases), microfractures (102 cases), Pridie’s perforations (78 cases), mosaicplasty (128 cases), over a 4-year period. The size of the cartilage defects ranged from 1 to 9 cm_. The results were evaluated using modified HSS and Cincinnati scores: 139 MRIs and 46 follow-up arthroscopies were performed, together with 28 biopsies. The results obtained with the mosaicplasty technique were significantly superior to those obtained with the other techniques, which result only in the formation of fibrocartilage. Moreover, arthroscopic checks showed that the donor sites had filled with fibrocartilage and remained asymptomatic insofar as they were non-weight-bearing areas (Fig. 23).

Figure 23: Biopsy of a donor site after 3 years (polarized light, x 12): the hole at the donor site is filled with fibrocartilage only (coll. L. Hangody).

L. Hangody recently published results relating to 44 patients with a cartilage defect grafted by arthroscopy alone, evaluated between 1 and 5 years later and compared with a control group treated with abrasion arthroplasty. The mean pre-operative HSS score for the mosaicplasty group was 62, 18, changing to 94, 23 at the last check. For the patients who had arthroplasty with subchondral abrasion, the mean pre-operative score was 59, 64, and the postoperative score, at the same stage, was 78, 24.
Ten patients had an arthroscopic follow-up between 12 weeks and 3_ years after the procedure. These arthroscopies showed progressive filling of the space between the grafts with fibrocartilage; the recipient site presented a hyaline aspect both in colour and in consistency, with a surface remaining congruent and with no secondary degradation.
Six biopsies were the subject of histological evaluation. This showed that fibrocartilage was forming in the harvesting holes at the donor sites (Fig. 24) and that the transplanted hyaline cartilage retained a normal histological and histochemical structure: type II collagen composition, normal fibre orientation and GAG content.

Figure 24 : A - View of the transition zone between hyaline cartilage at the periphery of the recipient site and the osteochondral graft after 5 years (alcian blue in polarized light, x 120, graft on right and intermediate fibrocartilage on left): there is no interface between the two types of cartilage (coll. L. Hangody). B - View of the transition zone between hyaline cartilage at the periphery of the recipient site and the osteochondral graft, 5 years after grafting (HE, x 120, graft on right and intermediate fibrocartilage on left): there is no interface between the two types of cartilage (coll. L. Hangody). C - Histological view of a biopsy of a grafted site after 21/2 years (red picrosirius, x 12, polarized light): the hyaline cartilage of the graft has a normal histological structure (coll. L. Hangody).

R. Jakob and his team have been using the same technique since 1995, with slightly different instrumentation. At the ESSKA Congress, R. Jakob reported his first results for the knee, covering 32 patients, all of whom were operated on via arthrotomy: 12 traumatic cartilage tears, 7 cases of osteochondritis, 7 of femoropatellar arthrosis, 4 recurrent dislocations of the patella and 2 cases of medial femorotibial arthrosis. The size of the cartilage defect ranged from 2 to 13 cm_. An average of 6 grafts (2-16) were required, using grafts at least 5 mm in diameter. Many associated actions were involved: tibial or femoral osteotomy, transposition of tibial tuberosity, trochleoplasty, LCA plasty, etc.
In general, patients were instructed to remain non-weight-bearing for 8 weeks. Pre-operative symptoms had disappeared in 98% of cases checked after 1 year. Follow-up arthroscopy confirmed L. Hangody’s findings.
The same team has also published very good results in respect of treatment of necrosis of the talus.
The grafts used were harvested from the ipsilateral knee and measured between 5 and 7 mm in diameter.
Overall, the results in the literature are encouraging. However, it is not always easy to analyse them, owing to the large number of surgical procedures involved. The time-spans involved are often short, and longer-term findings will be required before definitive conclusions can be drawn. Several key points have yet to be clarified: Does this technique need to be reserved only for traumatic cartilage lesions, or do the indications need to be expanded to include degenerative conditions? Is there an age limit? What is the minimum or maximum diameter of the grafts that can be used? What factors are involved in the prognosis? Do defective axes need to be systematically corrected? etc.

OUR SERIES

Between July 1996 and April 1998, we operated on 33 patients. Twenty-one knees and 12 ankles were treated in accordance with the technique described by L. Hangody, using the MosaicPlastyTM system distributed by Smith & Nephew laboratories.
The knee lesions involved 18 men and 3 women aged between 17 and 41. The site of the lesions was broken down into 15 medial condyles, 5 lateral condyles and 1 trochlea. There were 12 cases of osteochondritis dissecans and 9 cartilage or osteochondral tears. Six patients had already undergone one or more procedures without success in connection with the cartilage lesion. Six patients had already undergone a stabilizing ligamentoplasty with no residual instability. This is an important point, since one could not consider performing this reconstructive osteochondral surgery on a knee presenting an non-repaired lesion of the central pivot. Osteochondral grafts must be performed on a stable knee.
The size of the cartilage defect ranged from 1 to 12 cm_ and averaged 4.6 cm_. The number of grafts required per patient ranged from 1 to 12; a total of 117 grafts were used: 6 x 2.7 mm, 18 x 3.5 mm and 93 x 4.5 mm. Thus the majority constituted grafts 4.5 mm in diameter, which ranged from 15 to 25 mm in length, depending on the depth of the loss of osseous tissue. The grafts were harvested from the medial rim of the trochlea, the periphery of the intercondylar notch and, if necessary, the lateral rim of the trochlea. In the case of the trochlea graft, the cartilage defect was very large (12 cm_) and it was not possible to graft the whole area, since the lateral rim of the trochlea was itself denuded of cartilage. The grafts were harvested and delivered by arthroscopy in 6 cases and arthrotomy in 15 cases. The decision on whether to use arthrotomy or arthroscopy was based on the surface to be grafted, the number of grafts to be used and the anatomical site: 4 times out of 5, the lateral condyles had to be accessed by arthrotomy, and it was only possible to graft the trochlea by arthrotomy.
In the course of the same operation, associated actions involved in the mosaicplasty were as follows: 2 tibial osteotomies to achieve a valgus effect and 3 reconstructions of the anterior cruciate ligament.
All patients were instructed to remain non-weight-bearing for periods ranging from 4 to 8 weeks, depending on the size of the area grafted and the associated actions.
No postoperative complications occurred, and no morbidity was recorded in connection with the harvesting of grafts from the rim of the trochlea or the intercondylar notch.
Two of our cases are illustrated in Figures 25 and 26.

Figure 25: Simple case: symptomatic cartilage defect of the medial condyle discovered during ligamentoplasty of the anterior cruciate ligament.
A - The defect measures 9 x 5 mm. B - Appearance after insertion of two grafts, 4.5 and 3.5 mm.

Figure 26: Difficult case: extensive osteochondritis of the medial condyle of the right knee.
A - This osteochondritis became symptomatic at the age of 41 with separation of the sequestrum. B - Medial arthrotomy. The osteochondral defect measures 3 x 1.8 cm. The circular marks represent the tracks of guide tubes inserted to calculate the number and diameter of the grafts to be harvested. C - Appearance on completion.

To date, 3 MRIs and 6 follow-up arthroscopies have been performed, 6 months after surgery. The MRIs showed consistent cartilage coverage in two cases and detached cartilage in one case, in the 14th month after surgery (we shall come back to this case later). The follow-up arthroscopies showed that cartilage coverage was achieved from the 6th month onwards, with a slight "golf ball" effect representing the alternation of sections of hyaline cartilage and connecting fibrocartilage.
Clinical results appear to be achieved from the 6th month after surgery onwards, the time at which sporting activities can be gradually resumed. If we look at the 8 cases that date back more than 6 months, it can be said that we have achieved 3 excellent results, 4 good results and one average result.
The average result relates to a young girl who had a mosaicplasty to treat osteochondritis of the medial condyle (4 cm_), with an associated tibial osteotomy to achieve a valgus effect. Five 4.5 mm grafts and 3 2.7 mm grafts were used (= 3.18 cm_). A follow-up arthroscopy in the sixth month after surgery at the time of ablation of osteosynthesis material had shown a virtually "perfect" result (Fig. 27).

Figure 27: Follow-up check on osteochondritis after 6 months. Medial condyle of right knee, anterolateral aspect. Initial defect: 4 cm<=; 5 x 4.5 mm grafts and 3 x 2.7 mm grafts were used. Appearance very good, knee asymptomatic.

This result was clinically excellent for 14 months, with resumption of sport at a level identical with the earlier level. For no specific reason, the situation progressively deteriorated, with recurrence of medial femorotibial pains. A follow-up MRI showed partial detachment of the cartilage, confirmed by arthroscopic investigation. Calculation of the degree of coverage of the surface of the osteochondritis, however, showed that this was adequate: 3.18 / 4 cm_ = 79%. It is likely that the length of the grafts, 15 to 20 mm, was insufficient, since the depth of the area of osteochondritis was 10 mm, which meant that only 5 to 10 mm of the grafts were buried in spongy bone tissue. If one accepts that fibrocartilage alone occupies the space between the grafts, one can conceive that a certain degree of instability of the grafts was responsible for the failure. In such cases, where the diseased area is deep, it is therefore imperative to use grafts 20 to 25 mm long and with a large diameter (at least 4.5 mm), which are therefore more rigid, and perhaps to add grafts of spongy bone tamped down between the osteochondral cylinders.
Twelve talus mosaicplasties were performed in 10 men and 2 women aged between 17 and 52. They involved 9 posteromedial osteochondral lesions, 1 anteromedial lesion, 1 anterolateral lesion and 1 case of necrosis of the cupula. Two patients had already been the subject, several years earlier, of arthroscopic curettage with Pridie’s perforations, which had never really improved matters. In 8 cases out of 9, the posteromedial lesions had to be accessed via osteotomy of the medial malleolus. All the grafts were harvested, with the patient’s consent, from the medial rim of the trochlea of the ipsilateral knee, twice via arthroscopy and ten times in an open procedure. The average size of the cartilage defect to be grafted was 1.5 cm_, ranging from 0.5 to 3.75 cm_. Depending on the patient, 1 to 5 grafts were required. A total of 36 grafts were used: 7 x 2.7 mm, 13 x 3.5 mm, 15 x 4.5 mm and 1 x 6.5 mm. The grafts were delivered only once by arthroscopy, with the other 11 cases involving an arthrotomy. The patients were simply instructed to remain non-weight-bearing for 4 to 8 weeks, with no immobilization except inasmuch as most of them wore an inflatable splint.
Postoperatively, no complications arose in connection with the malleolar osteotomy. One case of algodystrophy was seen.
Two cases are illustrated in Figures 28 and 29.

Figure 28: Osteochondral lesion of the anteromedial talus, right ankle.
A - It was possible to access the lesion by anteromedial arthrotomy, with the foot in a pronounced plantar flexed position. B - Debridement of the lesion with subchondral abrasion. C - After preparation of the recipient holes.

D - Harvesting of grafts via a short medial parapatellar arthrotomy on the ipsilateral knee. E- View after insertion of grafts.

Figure 29: Osteochondral lesion of the posteromedial talus, right ankle.
A - Access via osteotomy of the medial malleolus following preparation for osteosynthesis with screw. Use of 5 grafts harvested from the medial rim of the ipsilateral femoral trochlea. B - Radiography after 14 months, at the time of ablation of the malleolar screw.

Three follow-up arthroscopies were performed after four months, at the time of ablation of the medial malleolar screw (Fig. 30). The cartilage of the grafted area was twice found to have become partially detached, in combination with non-re-establishment of articular congruence. Two MRIs were performed, at 4 and 8 months, and appeared to show satisfactory cartilage consistency. However, the images obtained were difficult to interpret.

Figure 30: Follow-up arthroscopy of a posteromedial lesion of the left talus after 5 months, at the time of ablation of osteosynthesis material. The lesion had initially been accessed via osteotomy of the medial malleolus and filled with 2 x 4.5 mm grafts and 1 x 2.7 mm graft.

With regard to clinical results, 5 patients had surgery over six months ago, and the results can be considered to be excellent in one case, good in two cases, and average in two cases. All the patients have improved in terms of pain symptomatology, but we feel, unlike L. Hangody, that the overall quality of the clinical results is less good than for the knee.
Overall, mosaicplasty of the ankle is technically more difficult than mosaicplasty of the knee. All the surgeons emphasize the difficulty of restoring the curvature of the articular surface, which is very pronounced in this area. Failure to restore articular congruence is detrimental to the result. Arthroscopic surgery of posteromedial lesions should be excluded. The instruments cannot be positioned perpendicular to the ankle surface to be grafted; thus, in the only case of a posteromedial lesion grafted by arthroscopy the surgeon was unable to graft the posterior part of the lesion. Thus it is essential to have recourse to an open procedure and to use an osteotomy of the medial malleolus.
It is preferable to harvest the grafts in an open procedure rather than by arthroscopy. In the two cases of arthroscopic harvesting, the surgeons confirmed that the cartilage cap of the grafts was oblique, which translated into detachment of the cartilage in one of the arthroscopic follow-ups.

CONCLUSIONS

We are currently at the dawning of a new, modern era for surgical treatment of cartilage defects. However, no team has currently had sufficient long-term experience to be able to wholeheartedly favour osteochondral grafts using the mosaicplasty technique. The results of both animal experiments and medium-term human trials are very interesting and show that it is worth continuing down this path. Our results are along the same lines as those reported by other authors, with few or no complications and no morbidity linked to the harvesting of grafts.
Communication needs to be developed in this field, in order to build up a common language: classification of cartilage lesions, method of evaluation and quality criteria for the results. The international Cartilage Repair Society is working along these lines. A standard evaluation form for the knee is already available and should be used by all those wishing to work in this field. There are also systems that make it possible to measure the rigidity of cartilage in situ, via arthroscopy, and thus to quantify its quality. This type of objective measurement is essential in order that we can evaluate our results, as are non-invasive imaging techniques, such as MRI, although their interpretation in respect of cartilage needs further refinement.
The authors wish to thank László Hangody for his assistance and for the histological illustrations for this article.

BIBLIOGRAPHY
Aichroth P, Burwelle RG, Laurence M (1971) An experimental study osteoarticular grafts to replace articular surfaces. J Bone Joint Surg (Br) 53: 554

Bobic V (1996) Arthroscopic osteochondral autograft transplantation in anterior cruciate ligament reconstruction : a preliminary clinical study. Knee Surg Sports Traumatol Arthrosc 3: 262

Brittberg M, Lindahl A, Nilsson A, et al (1994) Treatment of deep cartilage defects in the knee with autologous chondrocyte transplantation. N Engl J med 331: 889

Friedlaender GE, Horowitz MC (1992) Immune ersponses to osteochondral allografts : nature and significance. Orthopedics 15: 1171

Garrett JC (1986) Treatment of osteochondritis dissecans of the distal femur with fresh osteochondral allografts. Arthroscopy 2: 222

Gautier E, Mainil-Varlet, Saager C, Jakob RP (1998) Osteochondral autografts for the treatment of avascular necrosis of the talus, Book of abstracts, 8th Congress of the European Society of Sports Traumatology, Knee Surgery and Arthroscopy (ESSKA), pp 336.

Hangody L, Kish G, Kàrpàti Z, et al (1997) Autogenous osteochondral graft technique for replacing knee cartilage defects in dogs. Orthop Int 5: 1

Hangody L, Kish G, Kàrpàti Z, et al (1997) Treatment of osteochondritis dissecans of the talus: the use of the mosaicplasty technique - preliminary report. Foot Ankle Int 18: 10

Hangody L, Kish G, Kàrpàti Z, Szerb I, Udvarhelyi I (1997) Arthroscopic autogenous osteochondral mosaicplasty for the treatment of femoral condylar articular defects. A preliminary report. Knee Surg, Sports Traumatol, Arthrosc 5 : 262

Jakob RP, Mainil-Varlet P, Saager Ch, Gautier E (1998) Mosaïcplasty : the in vivo engineered approach, Book of abstracts, 8th Congress of the European Society of Sports Traumatology, Knee Surgery and Arthroscopy (ESSKA), pp 106.

Lorentzon R, Alfredson H (1998) Periosteum transplantation, Sports Medicine and Arthroscopy Review, 6 : 60

Matsusue T, Yamamuro T, Hama M (1993) Arthroscopic multiple osteochondral transplantation to the chondral defect in the knee associated with anterior cruciate ligament disruption: case report. Arthroscopy 9: 318

O’Driscoll SW, Salter RB (1986) The repair of major osteochondral defects in joint surfaces by neochondrogenesis with autogenous osteoperiosteal grafts stimulated by continuous passive motion: an experimental investigation in the rabbit. Clin Orthop 208: 131

Oateshott RD, Farine J, Pritzker KPH, et al (1988) A clinical and histological analysis of failed fresh osteochondral allografts. Clin Orthop 233: 283

Pridie KH (1959) A method of resurfacing osteoarthritic knee joint. J Bone Joint Surg (Br) 41: 618

Yamashita F, Sakakida K, Suzu F, Takai S (1985) The transplantation of an autogenic osteochondral fragment for osteochondritis dissecans os the knee. Clin Orthop 210: 43


Figure 1: The problem: cartilage fracture with detached cartilage.	Figure 2: Conventional treatment: excision of detached cartilage, subchondral abrasion, Pridie's perforations.

A - After four weeks (HE, x 20, graft on right of photo): complete osseous fusion. There is a space between the hyaline cartilage surfaces, but connecting tissue is already visible at the bottom of this space (coll. L. Hangody).

B - Six weeks after implantation (HE, x 20, graft on right of photo): the space between the cartilage is filled with fibrous connective tissue (coll. L. Hangody).	C - After eight weeks (HE, x 20, graft on right of photo): the connective tissue filling the space between the hyaline cartilage obliterates the differences in curvature of the surface area (coll. L. Hangody).


Figure 4: Debridement of the cartilage defect. The edges are brought back to a right angle and any detached cartilage is removed.	Figure 5: The base of the lesion is abraded to viable subchondral bone.


Figure 7: The graft is harvested perpendicular to the surface of the donor site.	Figure 8: The tubular chisel is driven by hammer to the desired depth, avoiding any overheating that could put the viability of the graft at risk.


Figure 9: The graft is detached from the trochlea in two perpendicular toggling movements, without rotating the chisel. The degree of movement must be carefully judged, to avoid fracturing the rim of the trochlea.	Figure 10: The graft is gently ejected from the chisel, using a tamp and an appropriately sized guide.


Figure 11: The graft is inserted, using a guide tube and a graduated tamp.	Figure 12: The level of the graft is adjusted.


Figure 13: A second graft is inserted in the same way 1 mm from the preceding graft.	Figure 14: Appearance on completion: grafts in place and donor sites.


Figure 15: The shading shows the areas of the condyle that can be grafted by arthroscopy.	Figure 16: A needle is used to determine the site for the arthroscopic portal such that the instruments will be perpendicular to the area to be grafted.


Figure 17: A - The guide tube is inserted for the first graft.	B - Endoscopic view


Figure 18: Grafts ready for implanting: 8 x 4.5 mm grafts and 2 x 3.5 mm grafts, between 15 and 20 mm in length.


Figure 19: Drilling of the hole. A - The depth of the drill bit is monitored via the window in the guide tube.	B - Endoscopic view


Figure 21: A - Insertion of the graft into the dilated hole.	B - Endoscopic view


Figure 23: Biopsy of a donor site after 3 years (polarized light, x 12): the hole at the donor site is filled with fibrocartilage only (coll. L. Hangody).


Figure 24 : A - View of the transition zone between hyaline cartilage at the periphery of the recipient site and the osteochondral graft after 5 years (alcian blue in polarized light, x 120, graft on right and intermediate fibrocartilage on left): there is no interface between the two types of cartilage (coll. L. Hangody).	B - View of the transition zone between hyaline cartilage at the periphery of the recipient site and the osteochondral graft, 5 years after grafting (HE, x 120, graft on right and intermediate fibrocartilage on left): there is no interface between the two types of cartilage (coll. L. Hangody).	C - Histological view of a biopsy of a grafted site after 21/2 years (red picrosirius, x 12, polarized light): the hyaline cartilage of the graft has a normal histological structure (coll. L. Hangody).


Figure 25: Simple case: symptomatic cartilage defect of the medial condyle discovered during ligamentoplasty of the anterior cruciate ligament. A - The defect measures 9 x 5 mm.	B - Appearance after insertion of two grafts, 4.5 and 3.5 mm.


Figure 26: Difficult case: extensive osteochondritis of the medial condyle of the right knee. A - This osteochondritis became symptomatic at the age of 41 with separation of the sequestrum.	B - Medial arthrotomy. The osteochondral defect measures 3 x 1.8 cm. The circular marks represent the tracks of guide tubes inserted to calculate the number and diameter of the grafts to be harvested.	C - Appearance on completion.


Figure 28: Osteochondral lesion of the anteromedial talus, right ankle. A - It was possible to access the lesion by anteromedial arthrotomy, with the foot in a pronounced plantar flexed position.	B - Debridement of the lesion with subchondral abrasion.	C - After preparation of the recipient holes.


D - Harvesting of grafts via a short medial parapatellar arthrotomy on the ipsilateral knee.	E- View after insertion of grafts.


Figure 29: Osteochondral lesion of the posteromedial talus, right ankle. A - Access via osteotomy of the medial malleolus following preparation for osteosynthesis with screw. Use of 5 grafts harvested from the medial rim of the ipsilateral femoral trochlea.	B - Radiography after 14 months, at the time of ablation of the malleolar screw.