Introduction

Since its introduction in Sweden more than 10 years ago, autologous chondrocyte implantation (ACI) has generated substantial interest in the orthopaedic community (1) . The high level of interest in ACI is likely due to the troublesome nature of the articular cartilage lesions that it is designed to treat. Once destroyed, articular cartilage will not regenerate, and lesions may progress to osteoarthritis. Furthermore, these lesions frequently result in pain, swelling, and mechanical symptoms such as locking and catching, dramatically reducing a patient's quality-of-life. The basic science investigation of the mechanisms of healing and repair started at our institution in 1992, and the first patients were treated in March 1995. To date, the author has treated 160 patients. This article reviews data on 101 patients who underwent treatment between March 1995 and November 1998, with 59 patients having greater than 6 months follow-up.

Science Background and Time Course of Healing

Several basic science investigations have elucidated the mechanisms and time course of healing of autologous chon-drocyte implantation. In a rabbit study evaluating autologous chondrocytes labeled with tritiated thymidine, the in-vitro cultured chondrocytes were noted to be responsible for a large portion of the repair tissue in-vivo 2 . Further studies deter-mined that the repair tissue from ACI was superior to periosteum alone. Canine studies from our institution have demon-strated that there are several stages to the healing process (4-6) . These stages include (1) the proliferative stage (0-6 weeks), which is characterized by a primitive cell response with tissue fill of the defect; (2) the stage of transition during which a type II collagen framework is produced along with the proteogly-cans that form the matrix. The tissue is not firm nor well inte-grated, and is "milkable" when probed with an arthroscopic nerve hook; and (3) a remodeling and maturation phase that occurs over time, lasting as long as two years, as matrix proteins crosslink and stabilize in large aggregates, and the colla-gen framework reorganizes to integrate into the subchondral bone and form arcades of Benninghoff. By 4-6 months, the tis-sue has usually firmed up, is no longer "milkable" but has a "putty-like" consistency. It is well integrated to underlying bone and adjacent cartilage. Patients will start to experience good symptom relief during this period. However, the process of tissue maturation that begins during the remodeling stage continues long after this point. Excessive activity during this remodeling stage may cause repair tissue degeneration. Hence, the concept of a time course of healing is critical dur-ing ACI rehabilitation.

Practical Algorithm for Cartilage Repair

The author's approach is to use arthroscopic debridement of unstable cartilage flaps for lesions less than 1 cm 2 in low demand patients and a gentle abrasion technique in high demand patients. For lesions of 1-2 cm squared , a marrow stimulation technique (abrasion, drilling or microfracture) or an osteochon-dral grafting technique (Mosaicplasty TM ,OATS TM , or CORR TM ) may be used in high demand patients. For lesions that are larger than 2 cm squaredor have failed alternative treatment options, ACI would be used as a primary treatment. For lesions greater than 10 cm squared where the articular cartilage loss and morphology of the condyle is distorted, fresh osteoarticular allograft is most likely to succeed. The rationale for this algorithm is discussed elsewhere (7-10) .

Indications for ACI

Autologous chondrocyte implantation has been approved by the FDA for the treatment of a symptomatic full thickness chondral injury of the weight-bearing femoral articular surface in a physiologically young patient who is compliant with the rehabilitation protocol. (11) The results of ACI for the treatment of chondral injuries of the patella and tibia have not been as good as those of the femoral weight bearing condyles and trochlea. Results of ACI for treatment of osteochondritis dis-secans (OCD) have also been successful . (12) The author has also used ACI successfully as a revision chondral surgery for chronic, symptomatic lesions unresponsive to alternative treat-ments, and as a first line treatment for symptomatic defects greater than 2cm squared .

ACI is not FDA approved as a treatment for osteoarthritis, that is bipolar chondral injuries with radiographic evidence of joint space narrowing with weight bearing. Therefore, preoperative weight bearing radiographic evidence of joint space narrowing of >50%, osteophyte formation, subchondral bony sclerosis or cyst formation are all considered contraindications to ACI. Rosenberg 45° bent PA views are helpful to assess tibia-femoral loss of cartilage in flexion .(13) Axial alignment weight bearing and merchant views are necessary to assess tibiofemoral or patellofemoral malalignment. MRI does not have high sensitivity (14) in determining the extent of a chondral injury or subtle chondromalacia changes. Normal radiographs with arthroscopic assessment are the gold standard for determining whether a symptomatic patient is a candidate for ACI.

Arthroscopic Assessment and Cartilage Biopsy for Cell Culturing Arthroscopic assessment of the joint and possible biopsy for articular cartilage culturing requires a careful and system-atic evaluation of the articular surfaces with an arthroscopic probe to determine the extent of grade III and IV chondroma-lacia (CM) of the symptomatic lesion. The opposing tibial/patellar articular surface must be probed throughout to ensure that the meniscus is intact, the articular surface is healthy, and any chondromalacia is no greater than superficial fissuring (Grade II CM). A 70-degree arthroscope is useful when assessing the patella. The anterior to posterior length of the femoral condyle lesion should be assessed, as this must be technically accessible at the time of open arthrotomy for periosteal suturing. The quality and thickness of the surrounding articular cartilage will determine whether healthy cartilage will be available for periosteal suturing or whether an uncontained chondral injury will require suturing through synovium or small drill holes through the bone.

Figure 1 Arthroscopic biopsy of full thickness articular cartilage from intercondylar notch anterior to sulcus terminalis

If a lesion is considered appropriate for ACI, then a biopsy site for cartilage procurement must be determined. The lateral intercondylar notch may be biopsied, (author's preferred site, see Figure 1). Approximately 200-300mg of articular cartilage is required for enzymatic digestion and cell culturing. This is roughly a cartilage surface 5mm wide by 1cm in length containing 200,000-300,000 cells. These cells will be enzymatical-ly digested and grown to approximately 12 million cells per 0.4cc of culture media per implantation vial. Following in vitro expansion of cells some 3-5 weeks later, a suitable number and volume of cells will be grown to accommodate the defect size required. At this time, second stage open implantation may occur.

Surgical Implantation of Autologous Chondrocytes

The steps in open implantation include arthrotomy, radical lesion debridement, (leaving subchondral bone intact) periosteum procurement, periosteum fixation, periosteum watertight integrity testing, autologous or allogeneic fibrin glue sealant, chondrocyte implantation, wound closure, and rehabilitation. The technique has been described in detail previously (1,15,16) .

Surgical Correction of Factors Predisposing to Chondral Injury

Several predisposing factors to chondral injuries must be assessed so that they may be either corrected in a staged or concomitant fashion with ACI. Tibiofemoral malalignment, patellofemoral malalignment, and ligamentous or bone insufficiency must be assessed prior to definitive cartilage cell reimplantation.

When varus or valgus malalignment is associated with a medial or lateral condyle injury respectively, then a corrective osteotomy is paramount to the success of chondral implantation. This can be done either in a staged or concomitant fashion. If corrective osteotomy is done concomitantly, stable fixation must be obtained at the time of osteotomy so that continuous passive motion (CPM) and early active range of motion may be pursued immediately postoperatively. Otherwise, a staged reconstruction should be performed.

Patello-femoral mal-tracking combined with a trochlear or patellar chondral injury requires careful pre-operative assessment with physical examination and CT or MRI imaging tech-niques. Tibial tubercle osteotomy combined with soft tissue realignment to ensure proper tracking is key to successful graft healing. Congenital trochlea dysplasia is an uncommon factor contributing to patello-femoral maltracking. Preoperative CT scan demonstrating flattening of the convex superior trochlear capturing entry point best assesses this problem. Treatment is by surgical trochleoplasty combined with patellar realignment as necessary .(16)

Cartilage repair in the face of anterior cruciate ligament (ACL) insufficiency may jeopardize a newly regenerating cartilage graft. Staged or concomitant ACL reconstructive surgery should be performed with the goal of preventing shear forces and instability episodes from damaging a healing graft. ACL rehabilitation is modified to exclude closed chain resisted strengthening exercises (leg presses or squats) until 9 months after combined surgery to prevent excessive compressive load to the chondral repair site.

When bony deficiency is present such as after an osteochondral fracture or osteochondritis dissecans, the depth of the bony lesion should be assessed preoperatively through conventional radiography, tomography and/or arthroscopy. Osteochondritis dissecans defects, on average, are 6-8mm deep including cartilage and bone. These often do well using ACI alone without bone grafting. However, defects greater than 1- 2cm deep require preliminary bone grafting and healing prior to cartilage resurfacing. This may be performed arthroscopically or by open technique. An interval of 6-9 months is required before second stage articular resurfacing to allow the cancellous bone graft to incorporate. In this way, a new ‘subchondral bone plate' is formed, and minimal (if any) bleeding occurs prior to chondrocyte implantation. In addition, a patient occasionally becomes asymptomatic after bone grafting secondary to fibrocartilage repair of the overlying chondral defect, thus obviating the need for further surgery.

Clinical Series

Following approval by the human ethics committee, a prospective evaluation of ACI-treated patients was undertaken in March 1995. Patients were implanted in the manner described above with ex-vivo cultured chondrocytes injected beneath a periosteal patch secured with re-absorbable sutures and fibrin sealant (1, 15). Post-operative rehabilitation included non-weight bearing and use of continuous passive motion for 6-8 hours/day for 6 weeks, followed by progression to full weight bearing at 4 months. Patients were restricted from in-line impact activities (e.g., running) for 9-12 months, and cut-ting sports for at least 14-18 months.

	Figure2a Simple Category Case: 31 year old male with right knee, lateral (left photo) and medial (right photo) femoral condyle grade IV lesions following trauma while hiking (arthroscopic photos) (lateral femoral condyle defect 27mm long by 24mm wide, medial femoral condyle defect 24mm long by 21 mm wide).
	Figure2b Second look arthroscopy photographs of same knee two years post-operative-ly, of lateral (left) and medial (right) femoral condyles. Asymptomatic individ-ual able to participate in recreational sports. Cincinnatti Knee score 10/10, five years post operatively.

Demographic data, prior surgical history, defect characteristics, and baseline evaluation including completion of 4 validated rating scales was collected. Follow-up evaluation was completed using these same 4 metrics at 6, 12, 18, and 24 months, and at yearly intervals thereafter. Standardized rating scales included the SF-36 (Short Form 36 item quality-of-life questionnaire), WOMAC (Western Ontario and MacMaster Universities Osteoarthritis Index), KSS (Knee Society Score), and the modified Cincinnati Knee Rating System. Data were collected by a research assistant independent of the operating surgeon using standardized case report forms, and statistical analysis was conducted by an unbiased third party (AACT-Abt Associates Clinical Trials, Cambridge, MA.

Treatment Classification Patients were divided into the following groups: (8,10) Simple: isolated lesions to the weight bearing femoral condyles, with no other surfaces exhibiting > Outerbridge I-II chondromalacia. (Figure 2) Complex: uni-polar single or multi-focal lesions of the femoral condyles, tibial plateau, trochlea, or patella, (excluding ‘kissing lesions'), may require concomitant osteotomy, liga-ment reconstruction, and/or staged bone grafting. The joint is otherwise without advanced chondromalacia. (Figure 3)
	Figure 3 Complex Category Case: 44 year old male with severe anterior knee pain unable to work, considering total knee replacement. (Left) Microsuturing periosteal cover at the time of open arthrotomy, prior to autologous chondrocyte implantation to left knee trochlea, (defect size 33 mm wide by 25 mm long). (Right) Second look arthroscopy at two-year post-op. Note appearance of trochlear ACI graft demonstrating complete fill, mechanically firm and well integrated to underlying bone and adjacent cartilage. Patient remains asymptomatic and employed, five years later.

Figure 4 42 year old male, prior total lateral meniscec-tomy. Initially presented with a ten year his-tory of persistent lateral weight-bearing pain. Patient is now one year post operatively, pain free. (a) Preoperative weight-bearing radiograph of right knee. Note normal lateral joint space and varus mechanical alignment

Salvage: patients with radiographic or arthroscopically defined osteoarthritis including osteophyte formation (Ahlback stage 0), or joint space narrowing, (<50% Ahlback stage I), and those with bipolar or "kissing" lesions or generalized chondro-malacia >= Outerbridge II. (Figure 4)

	(b) Intraoperative appearance of lateral compartment right knee	Demographics and Case Mix 59 patients (116 lesions; 41 male, 18 female, mean age 35) with a minimum of one year follow-up were evaluated. There were seven patients with simple defects, 30 complex, and 22 with salvage lesions. Three of seven patients with simple defects and more than half of patients with complex and sal-vage lesions had failed previous attempts at surgical repair. The mean size of the lesions was 3.3cm squared for simple lesions and 5.1cm squared for complex and salvage lesions. In the complex group there were nine patients with osteo-chondritis dissecans, ten requiring high tibial osteotomy (HTO) for varus malalignment, three who required Maquet tib-ial tubercle elevation, four who underwent Fulkerson anterior-medial tubercle advancements, and six who had ACL recon-struction. Among the salvage cases, 23 patients required osteotomy, (6 HTO, 7 HTO with Maquet osteotomy, 10 Fulkerson osteoto-my), 2 patients required ACL reconstructions, while others were classified on the basis of multiple lesions. Among the 116 total defects, the most common defect locations were the medial femoral condyle (n=51) and the trochlea (n=28) followed by the patella (17), lateral femoral condyle (15), and lateral (4) and medial (1) tibial plateaus.
	(c) Intraoperative appearance of lateral compartment right knee after radical debride- ment of degen-erated cartilage mar-gins, back to full thickness cartilage borders.
	(d) Intraoperative appearance of lateral compartment right knee after ACI and cadaveric cryopre-servered lateral meniscus allograft transplantation.

Results

SF-36 overall physical health scores improved at each follow- up visit, reaching significance at 24 months (p <0.001). Mental health overall scores improved at each visit, reaching significance at 36 months (p=0.024), with vitality and social functioning components demonstrating the greatest improve-ment.

The overall WOMAC score, as well as physical function and pain subscores, showed significant improvement at 24 months (p <0.01 in each case).

Similarly, outcomes at 24 months evaluated by the Knee Society Knee Score demonstrated improvement (p<0.001).Changes in the Knee Society Function Score did not reach sig-nificance (p = 0.16).

Results from the Cincinnati Knee score reflected significant improvement at 12 and 24 months (p<0.001).

Management of Complications

To date there have been no intra-articular joint infections following ACI in our patients. Minor superficial wound infections have occurred, as well as complications related to open arthrotomy.

The most common postoperative problem seen is periosteal hypertrophy. This usually manifests 3 to 7 months after surgery (at second look arthroscopy) as a proliferative hypertrophic periosteal healing response. Patients may present with new onset catching from a previously smooth track-ing knee with symptoms of pain and effusion. If this occurs, activity level should be decreased and arthroscopic shaving is recommended to resolve the prominent overgrowth. Periosteal problems may occur as often as 25% of the time. In most cases, the catching response settles and the patient remains asymptomatic.

Intra-articular adhesions are rare except in the case of large intra-articular periosteal patches taken from the femur, which can lead to intra-articular fibrosis. If this occurs, adhesions are best released with arthroscopic electrocautery, elec-troablation, or shaving of adhesions. After intra-articular adhesions are released, the grafts are visualized to ensure that there are no adhesions to the grafts. Gentle manipulation then confirms adequate release.

Discussion

Rationale for the treatment of cartilage damage in younger patients depends on a thorough understanding of the predisposing factors for the chondrosis and the stage of disease. Implantation with autologous cultured chondrocytes allows for resurfacing of larger defect areas, with good to excellent results in 90% of patients with isolated lesions of the femoral condyle. Patellar lesions may be successfully treated, but strict attention must be given to correction of malalignment. Results in patients with tibial and salvage lesions are encouraging, however these results should be viewed with caution due to the small number of patients with two year follow-up. At present, ACI involves an open technique with the inherent disadvantages of adhesions and prolonged recovery. However, these disadvantages must be weighed against the procedure's ability to produce a hyaline-like tissue with greater durability than the fibrocartilaginous tissue that is produced by traditional marrow-stimulation techniques. For this reason, it is recommended that the treating surgeon match the treatment procedure to patient demographics and expectations, and the location and size of the chondral lesion. Based on the available literature, algorithms have been published (7-10) recommending that ACI be reserved as first-line treatment for high-demand patients with large lesions (> 2cm squared ) and as revision therapy in patients with lesions of all sizes who have failed alternative techniques.