Autologous Chondrocyte Transplantation - SAGE Journals

0363-5465/102/3030-0002$02.00/0 THE AMERICAN JOURNAL OF SPORTS MEDICINE, Vol. 30, No. 1 © 2002 American Orthopaedic Society for Sports Medicine

Autologous Chondrocyte Transplantation Biomechanics and Long-Term Durability Lars Peterson,*† MD, PhD, Mats Brittberg,‡ MD, PhD, Illka Kiviranta,§ MD, PhD, Evy Lundgren Åkerlund,储 PhD, and Anders Lindahl,a MD, PhD From *Gothenburg Medical Center and ‡Kungsbacka Hospital, Gothenburg University, Gothenburg, Sweden, §Department of Surgery, Jyva¨skyla¨ Central Hospital, Jyva¨skyla¨, Finland, 储 Department of Cell and Molecular Biology, University of Lund, Lund, Sweden, and a Institute of Laboratory Medicine, Sahlgrenska University Hospital, Gothenburg, Sweden cartilage is limited. Full-thickness defects larger than a few millimeters in diameter fail to heal, even after nonoperative management such as the use of continuous passive motion31 or after arthroscopic debridement and lavage.14 The persistence of untreated focal chondral lesions is primarily caused by the lack of chondroprogenitor cells within the defect. Treatment options are therefore directed to the recruitment of bone marrow cells because the underlying marrow provides an easily accessed source of potential cartilage precursors. A variety of methods have been developed to violate the subchondral bone plate to allow fibrin clot formation within the defect. Unfortunately, these marrow-stimulation treatments, such as abrasion,8 drilling,7 and microfracture,3 produce a fibrous repair tissue that lacks the mechanical characteristics necessary for durability. However, more hyaline-like repair tissue based on increased type II collagen has been reported in horses 12 months after microfracturing.9 Clinically, these treatment options may have a limited duration and a relatively poor long-term prognosis. Patients may require additional surgical interventions or may have to adjust their activity level to that of their knee function. In one study of patients who had undergone abrasion arthroplasty, Nehrer and colleagues23 frequently found fibrous, soft, spongiform tissue combined with central degeneration (Outerbridge27 grade IV crater). Clinical failure occurred at a mean time of 21 months after the treatment (range, 0.5 to 60 months). The use of autologous cultured chondrocyte transplantation was initiated during the early 1980s. Animal studies demonstrated the production of hyaline-like repair tissue when cultured chondrocytes were implanted beneath a periosteal patch.6, 11 The clinical use of autologous chondrocyte transplantation was pioneered in Sweden in 1987, and the technique has been used extensively during the last 5 years in both the United States and in Europe. The first report of clinical use of autologous chondrocyte transplantation in 1994 demonstrated that 14 of 16 pa-

ABSTRACT We evaluated the durability of autologous chondrocyte transplantation grafts in 61 patients treated for isolated cartilage defects on the femoral condyle or the patella and followed up for a mean of 7.4 years (range, 5 to 11). Durability was determined by comparing the clinical status at the long-term follow-up with that found 2 years after the transplantation. After 2 years, 50 of the 61 patients had good or excellent clinical results, and 51 of 61 had good or excellent results at 5 to 11 years later. Grafted areas from 11 of the patients were evaluated with an electromechanical indentation probe during a second-look arthroscopy procedure (mean follow-up, 54.3 months; range, 33 to 84); stiffness measurements were 90% or more of those of normal cartilage in eight patients. Eight of twelve 2-mm biopsy samples taken from these patients showed hyaline characteristics with safranin O staining and a homogeneous appearance in polarized light. Three fibrous and eight hyaline biopsy specimens stained positive to aggrecan and to cartilage oligomeric matrix protein. Hyaline-like specimens stained positive for type II collagen, and fibrous, for type I collagen. Autologous chondrocyte transplantation for the treatment of articular cartilage injuries has a durable outcome for as long as 11 years.

The presence of isolated injuries to the cartilage of the knee is considered a risk factor for more extensive joint damage because the intrinsic repair capacity of articular

† Address correspondence and reprint requests to Lars Peterson, MD, PhD, Gothenburg Medical Center, Gruvgatan 6, SE-421 30 Va¨stra Fro¨lunda, Gothenburg, Sweden. No author or related institution has received any financial benefit from research in this study. See “Acknowledgments” for funding information.

2

Vol. 30, No. 1, 2002

tients with injuries to the femoral condyles had good or excellent results and that 11 of 15 biopsy samples from these patients showed hyaline-like cartilage.5 However, only two of the seven patients with patellar injuries had good function after treatment, and only one of seven had cartilage with a hyaline-like appearance.5 Many surgeons worldwide have now replicated these encouraging early clinical results in the femoral condyle.1, 10, 12, 15, 21 These studies have generally demonstrated improvement in 85% of femoral lesions with short- to medium-term follow-up. To date, however, no study has examined the mechanical characteristics of this repair tissue or correlated shortterm clinical improvement with long-term durability results. More than 700 patients had been treated with autologous chondrocyte transplantation by our group by the end of 1999. More than 400 of these patients had been observed for 2 years or more, and the group of patients whose treatment outcome had been evaluated over a long term (5 to 10 years) numbered more than 60. We therefore undertook this study of the first 61 consecutive patients treated with autologous chondrocyte transplantation to achieve the following four objectives: 1) to correlate observed clinical success in the short term (2 years) with long-term clinical success out to 11 years, 2) to examine the histologic characteristics of the repair tissue and to correlate these findings with clinical results, 3) to assess the biomechanics of the repair tissue with use of a standardized measurement device (indentation probe) and correlate these findings with patient outcome, and 4) to correlate the biomechanical findings with the histologic nature of the repair tissue.

MATERIALS AND METHODS Patients Of 72 examined patients in whom biopsies were taken for transplantation, transplantations were completed in 61 patients between October 1987 and August 1994. The patients were divided among the following treatment groups by type and location of the defect: group 1, isolated femoral condyle lesions (19); group 2, osteochondritis dissecans lesions (14); group 3, patellar lesions (17); and group 4, femoral condyle lesions with ACL reconstruction (11). Four patients in our initial study who were treated with carbon fiber implants after graft failure were not examined (two femoral condyles and two patellae) but were included as treatment failures in the analysis.5 Fiftysix of the remaining 57 patients agreed to return either for clinical evaluation and documentation of adverse events or to complete questionnaires with regard to their clinical course. One patient in group 1 was examined at 2 years but later died of unrelated causes and is thus not included in the final follow-up. The minimum follow-up duration from the time of implantation was 5 years. All of the patients had had moderate-to-large (1.3 to 12.0 cm2) full-thickness (Outberbridge27 grade III or IV) chondral defects of the knee, had severe symptoms, and were categorized as having poor results by the clinical

Autologous Chondrocyte Transplantation

3

grading system of Brittberg et al.5 Twenty-seven of the 56 patients had undergone previous orthopaedic procedures (Table 1). Chondrocyte Transplantation Technique The patients were treated in three hospitals by three different surgeons, and the treatments were performed in two separate operations, as described previously.22 Briefly, an arthroscopic surgical procedure was performed initially, with the patient under general or spinal anesthesia and in a tourniquet-controlled bloodless field. The defect was examined, and slivers of cartilage (300 to 500 mg) were obtained from the upper minor load-bearing area of the medial femoral condyle in the injured knee for culture and later implantation. In the operating room, the samples of cartilage were placed in a sterile glass tube containing 0.9% NaCl at ambient temperature and subsequently transferred to the cell-culture laboratory. Blood was collected preoperatively from the patient into 10 separate 9-ml vacuum serum sample tubes (Venoject II, Terumo Europe, Leuven, Belgium). The blood was allowed to coagulate for a minimum of 30 minutes at room temperature and centrifuged within 1 hour (15 minutes at 1000 to 1300 relative centrifugation force). The serum was sterile filtered within 24 hours from sampling (Millex GS, 0.22 ␮m, Millipore, Bedford, Massachusetts) and kept refrigerated (⫺4° to ⫺8°C). Cell isolation was initiated not later than 6 hours after the operation. The cartilage pieces were minced and washed twice in Ham’s F-12 medium (Gibco/BRL, Paisley, Scotland) supplemented with gentamicin sulfate (50 ␮g/ml), amphotericin B (2 ␮g/ml), and L-ascorbic acid (50 ␮g/ml). The minced cartilage was digested overnight (16 to 20 hours) in a 25-cm2 culture bottle (Costar, Cambridge, Massachusetts) containing 5 ml of Ham’s F-12 medium supplemented as described previously, clostridial collagenase (0. 8 mg/ml, No. C-9407, ⬎1200 IU/mg, Sigma, Freehold, New Jersey), and deoxyribonuclease I (0. 1 mg/ml, No. D-5025, Sigma). The isolated cells were washed in Ham’s F-12 medium once and resus-

TABLE 1 Previous Surgical Procedures Group

Procedures

1

2 3 2 1

tibial osteotomies lateral meniscectomies medial meniscectomies removal of a Baker’s cyst

2

4 2 2 1 1 1

fixations of loose fragments lateral meniscectomies medial meniscectomies ACL reconstruction drilling tibial osteotomy

3

1 extensor mechanism reconstruction 1 medial meniscectomy 1 patellar groove reconstruction

4

3 ACL reconstructions 2 medial meniscectomies 2 lateral meniscectomies

4

Peterson et al.

pended in 5 ml of culture medium containing Dulbecco’s Modified Eagle Medium/F12 1:1 (Gibco/BRL) with the addition of 10% of the patient’s own serum and supplemented with gentamicin sulfate (50 ␮g/ml), amphotericin B (2 ␮g/ml), L-ascorbic acid (50 ␮g/ml), and L-glutamine (Gibco/BRL). The cells were incubated in 25-cm2 tissueculture flasks (Falcon Primaria, Becton-Dickinson, Oxford, United Kingdom) in 7% CO2 in air at 37°C. After 1 week, the cells were suspended by means of treatment with trypsin-ethylene diametetraacetic acid (0.125%) and transferred to a 75-cm2 culture flask (Costar) with the addition of the culture medium. Implantation was performed 14 to 21 days after obtaining the slivers of cartilage that were cultured. Three days before transplantation, the cell cultures were subjected to evaluation with thorough quality-control procedures, consisting of testing for bacterial and fungal growth and photographic recording of the morphologic characteristics of the cells. The chondrocytes were trypsinized and washed twice in Ham’s F-12 medium containing 20% autologous serum (implantation medium). The cells were released for implantation if the cell viability was more than 85%, as determined with trypan blue staining. Finally, the cells were suspended in 0.3 to 0.4 ml of implantation medium in a tuberculin syringe and transferred to the surgical department in a sterile package. The average number of cells implanted was 4.5 ⫻ 106. At the time of chondrocyte implantation, prophylactic antibiotics were given to the patient for 24 hours, beginning at the initiation of the procedure. With the patient under general or spinal anesthesia, a medial or lateral parapatellar arthrotomy was performed in a tourniquetcontrolled bloodless field. The chondral lesion was debrided back to the best cartilage available, while still maintaining a contained lesion, if possible. The surrounding cartilage had minimal or no fissuring, and a nerve hook probe was used to confirm that it was not undermined. At all times, care was taken not to penetrate the subchondral bone plate or to provoke bleeding from the wound bed. A periosteal flap was harvested from the proximal medial subcutaneous border of the tibia. The flap was fitted and sutured to the surrounding rim of the debrided cartilage with interrupted 5– 0 or 6 – 0 Vicryl (Ethicon Inc., New Brunswick, New Jersey) or Dexon (Sherwood, Davis & Geck, St. Louis, Missouri) sutures with the deep cambium layer facing toward the subchondral bone plate. The periosteal rim was sealed with fibrin glue (Tisseel, Immuno AG, Vienna, Austria) except for one corner, where the implanted chondrocytes were injected into the defect. After cell injection, this corner was closed with a final suture and application of fibrin sealant. The joint capsule, the retinaculum, and the skin were sutured in separate layers, and the knee was covered with an elastic bandage. Continuous passive motion was administered for 48 hours after the operation. Rehabilitation on crutches began with gradual weightbearing for 8 weeks, progressing to full weightbearing by 10 to 12 weeks. The emphasis during rehabilitation was on functional use of the limb and on active muscle recruitment. If required, reconstruction of the ACL was performed

American Journal of Sports Medicine

before implantation during the same operative procedure. After a medial parapatellar arthrotomy, a vascularized distally based medial-third patellar tendon and retinaculum graft complex was rerouted through standard tibial and femoral bone tunnels. Femoral fixation was achieved through a separate lateral incision with the suture attachment over a post. Autologous chondrocyte transplantation was then performed as described. Clinical Evaluation The patient’s clinical status was evaluated on an annual basis using five scoring systems: the method described by Lysholm and Gillquist,16 the modified Cincinnati (Noyes) knee score,24 the overall Cincinnati knee-rating score,25 the Wallgren-Tegner activity score,32 and the overall Brittberg clinical grading score.5 In addition, the Brittberg visual analog score and a questionnaire with regard to the patient’s perception of the surgical outcome were also used.28 Treatment durability was defined as the percentage of patients whose outcome was graded good or excellent 2 years after treatment and who maintained this status with a minimum follow-up of 5 years. Baseline clinical scores were compared with follow-up data by paired Student’s t-tests for the Tegner-Wallgren activity score, the modified Cincinnati knee score, the Lysholm score, and the Brittberg visual analog score. Statistical analysis was performed using the Stata software package (version 6.0, Stata Corp., College Station, Texas). All reported P values are two-tailed. For all statistical tests, a two-tailed P ⬍ 0.05 was considered significant. Safety assessment was performed by an independent evaluator (Chiltern International, Gothenburg, Sweden), who retrospectively reviewed inpatient and outpatient records for adverse events. Mechanical Assessment of Repair Tissue We asked 18 randomly chosen patients in treatment groups 1, 2, and 4 whether they were willing to undergo a second-look arthroscopic procedure and a biopsy for evaluation of the biomechanical and histologic characteristics of the repair tissue. None of the patients in treatment group 3 were asked to participate in this part of the study because the indentation instrument could be used transarthroscopically only on the femoral condyles and not on the patella because of the measurement angles. Eleven of the 18 patients agreed to the follow-up operation at a mean of 54.3 months after surgery. The grafted area in the patient’s knee was classified according to the integration of the graft into the surrounding cartilage (0 to 4 points), the degree of defect fill (0 to 4 points), and the macroscopic appearance (0 to 4 points), giving a total possible defect-repair score of 12 points. This scoring system has been described previously (Table 2 in reference 28). The mechanical characteristics of the repair tissue were assessed through the use of an electromechanical indention probe (Artscan Inc., Kuopio, Finland) that allowed for quantification of the cartilage stiffness under arthroscopic

Vol. 30, No. 1, 2002


5

Figure 1. The indentation instrument (left) and the indentation surface with the probe (right). control (Fig. 1). Measurements were taken from both the repaired area and an adjacent (control) location. Detailed descriptions of the instrument have been previously published.17–20 Briefly, the probe consists of a measurement rod (5 mm in diameter and 150 mm in length) that can be sterilized and introduced arthroscopically into the knee. The end of the rod has a cylindrical 1.3 mm long ⫻ 300 ␮m diameter indenter located in the center of an inclined (20°) reference plate. The indenter is joined to a bending beam that measures the force applied by a pair of transducers. A second pair of transducers measures the force by which the surgeon presses the reference plate against the cartilage surface. These measurements are transmitted to a personal computer and displayed on a video monitor. An accurate measurement is obtained by pressing the reference plate against the cartilage surface (under arthroscopic visualization) until a constant 10-N force is displayed on the monitor, at which time the calculated cartilage stiffness is displayed. A previous study has demonstrated more than 95% reproducibility when this technique is used.18 Histologic Evaluation of Repair Tissue In conjunction with the indentation probing, biopsy samples from 11 patients were obtained from the center of the grafted area with a 2-mm core biopsy instrument. The specimens were fixed in 4% formaldehyde for 24 hours, were embedded in paraffin, and were sectioned in 8-␮m sections. Staining was performed with Alcian blue van Gieson, hematoxylin and eosin, and Alcian blue or safranin O or both. For immunohistochemical analysis, paraffin sections of cartilage tissue were deparaffinized in xylene for 5 minutes, hydrated in ethanol at decreasing concentrations (100%, 90%, 70%) for 5 minutes each, and rinsed in distilled water for 5 minutes. Sections used for collagen type II staining were then exposed to high temperature in 10-mM citrate-buffer in a microwave oven (2 minutes at maximum effect, 10 minutes at medium effect, and 20 minutes at low effect) to unmask the epitopes. All sections were washed with phosphate-buffered saline so-

lution containing 0.5% bovine serum albumin for 10 minutes at room temperature and digested with 2 mg/ml of hyaluronidase (Sigma) in phosphate-buffered saline solution, pH 5.0, for 15 minutes at 37° C. After washing with phosphate-buffered saline solution containing 0.5% bovine serum albumin for 15 minutes, the sections were incubated with 1% H2O2 in phosphate-buffered saline solution for 15 minutes at room temperature to remove endogenous peroxidase activity. This step was followed by incubation with phosphate-buffered saline solution containing 0.5% bovine serum albumin for 5 minutes. The sections were then incubated with normal goat serum for 15 minutes at room temperature. Incubation followed overnight at 4°C with polyclonal antibodies against either the cartilage oligomeric matrix protein serum diluted 1:4000 in phosphate-buffered saline solution containing 0.5% bovine serum albumin, or aggrecan (serum diluted 1:2000 in phosphate-buffered saline solution containing 0.5% bovine serum albumin). Sections were also incubated with monoclonal antibodies against collagen type I (10 ␮g/ml) or collagen type II (10 ␮g/ml) diluted in phosphate-buffered saline solution containing 0.5% bovine serum albumin. After being washed in phosphate-buffered saline solution, sections were incubated with biotinylated secondary goat antirabbit antibodies (Vector Laboratories Inc.; diluted 1:200 in blocking buffer) or goat antimouse antibodies (Vector Laboratories Inc; diluted 1:200 in blocking buffer) at room temperature for 30 minutes. Washed sections were then incubated with Vectastain ABC reagent (Vector Laboratories, Inc.) for 1 hour at room temperature and washed. The color was developed using 1 mg/ml of diamino benzidine in 0.1M Tris-HCl, pH 7.2, and 0.02% H2O2. Sections were rinsed in water for 5 minutes followed by 75% ethanol, 95% ethanol, and 99.5% ethanol for 5 minutes each, and then rinsed three times in xylene for 3 minutes each time at room temperature. Samples were mounted in Pertex (Histolab Products AB, Gothenburg, Sweden). Two independent pathologists who were blinded to the treatment outcomes each described the histologic appear-

6

Peterson et al.


ance of the repair tissue on the basis of the criteria of cellularity (chondrocytes in lacunae), matrix appearance (metachromatic staining with Alcian blue), columnar structure of the cartilage, clonal formation, and the appearance under polarized light. Under polarized light, the fibrous repair tissue was clearly indicated by the presence of collagen bundles, compared with the opaque appearance of hyaline cartilage because of the organized collagen. The repair tissue was classified as hyaline-like, fibrous, or mixed tissue. The staining of collagen type I and type II cartilage oligomeric matrix protein, and aggrecan was classified as moderately stained to intensely stained.

RESULTS Group 1: Femoral Condyle Lesions Nineteen patients in this group were included in the follow-up, and their results are shown in Table 2. The average defect size was 3.4 cm2 (range, 1.6 to 7.5). The patients had undergone an average of 1.5 operations (range, 0 to 4) and had a mean duration of symptoms of 3.4 years (range, 0.3 to 15.7) before implantation. Eighteen of the 19 patients were evaluated 5 years or more after the implantation. One patient in this group was evaluated 2 years after implantation but later died from unrelated causes. The mean follow-up duration at the latest evaluation was 7.4 years (range, 5 to 11). The clinical grade was good or excellent in 89% of patients at 2 years after treatment and this overall rate of improvement was maintained at the latest follow-up. Tegner-Wallgren activity scores at 2 years and at latest follow-up increased significantly compared with the preoperative score (P ⬍ 0.001). The mean modified Cincinnati grading also improved significantly at 2 years (P ⬍ 0.001) and at the final follow-up (P ⬍ 0.001). The Lysholm score and the Brittberg visual analog scale score showed similar results, with statistically significant improvement at 2 years (P ⬍ 0.001) and a continued trend toward additional improvement at final follow-up (not significant). The two treatment failures in this patient group ocTABLE 2 Clinical Results at the Different Points Preoperatively and after Autologous Chondrocyte Transplantation for Patients with Femoral Condyle Injuries Variables

No. of patients Mean age (years) Clinical grading (No.) Good Excellent Graft failures (new/total) Cincinnati score Brittberg VASa score Tegner-Wallgren activity score a

Visual analog scale.

Preoperative

2-year follow-up

Final follow-up Mean, 7.4 (range, 5 to 11)

19 30.4

19 32.4

18 40.8

0 0

5 12 2/2

5 12 0/2

1.2 84.5 7

8.9 14.2 9.8

9.2 12.2 10

curred at 11 and 14 months after implantation. No patients experienced graft failure or had a decline in clinical status between the evaluation at 2 years and the follow-up at 5 years or more (mean, 7.4), resulting in a durability of 100% for this group. One patient had an arthroscopic debridement of an unrelated pathologic lesion and osteotomy performed 2 years after the autologous chondrocyte transplantation but ended up with an excellent result after 5 years. Group 2: Osteochondritis Dissecans Lesions In this group, 14 patients were followed up; summary data are shown in Table 3. The average defect size was 4.8 cm2 (range, 1.5 to 12.0). The patients had undergone an average of 3.2 operations (range, 0 to 8) and had a mean duration of symptoms of 7.3 years (range, 0.9 to 14.8) before treatment. All of the patients were evaluated at 5 years or more after the implantation; none of the patients were lost to follow-up. Overall, the clinical score was good or excellent in 86% of patients by 2 years after implantation. Twelve patients with good or excellent ratings at 2 years maintained their condition at the longer-term follow-up; one patient with a good result at 2 years improved to an excellent rating at the latest follow-up. The Tegner-Wallgren activity scores increased significantly from preoperatively to 2 years (P ⬍ 0.001) and at final follow-up (P ⬍ 0.001). The modified Cincinnati grading also increased significantly at 2 years (P ⬍ 0.001) and at final follow-up (P ⬍ 0.001). The Lysholm score and Brittberg visual analog scale score showed similar results, with the largest improvement in score occurring from preoperatively to 2 years (both changes P ⬍ 0.001) and with slight additional improvement occurring between 2 years and final follow-up. The two treatment failures in this group both occurred before the 2-year evaluation. One patient received a carbon fiber implant after failure of the autologous chondrocyte transplantation due to an early return to sport (tennis). The second patient, who was treated again with autologous chondrocyte transplantation, was doing well at

TABLE 3 Clinical Results at the Different Points Preoperatively and after Autologous Chondrocyte Transplantation for Patients with Osteochondritis Dissecans Lesions of the Femoral Condyle Variables

Preoperative

2-year follow-up


No. of patients Mean age (years) Clinical grading (No.) Good Excellent Graft failures (new/total) Cincinnati score Brittberg VASa score Tegner-Wallgren activity score

14 26.9

14 29.0

14 34.8

0 0

3 9 2/2

2 10 0/2

1.7 80.1 6

8.2 21.8 8.8

9.0 19.4 9.8

a

Visual analog scale.

Vol. 30, No. 1, 2002


TABLE 4 Clinical Results at the Different Points Preoperatively and after Autologous Chondrocyte Transplantation for Patients with Patellar Lesions Variables

Preoperative

2-year follow-up


No. of patients Mean age (years) Clinical grading (No.) Good Excellent Graft failures (new/ total) Cincinnati score Brittberg VASc score Tegner-Wallgren activity score

17 27.3

17 26.9

17 35.4

0 0

4 7 2/4

6a 7 2b/4

1.6 68.1 5.5

6.8 27.3 9.2

6.6 27.8 8.8

TABLE 5 Clinical Results at the Different Points Preoperatively and after Autologous Chondrocyte Transplantation for Patients with Femoral Condyle Lesions and ACL Reconstructions Variables

Preoperative

2-year follow-up


No. of patients Mean age (years) Clinical grading (No.) Good Excellent Graft failures (new/ total) Cincinnati score Brittberg VASb score Tegner-Wallgren activity score

11 28.5

11 30.4

11 38.8

0 0

3 6a 2/2

3 6a 0/2

2.3 83.2 7

7.5 25.9 9.6

7.6 24.2 9.8

a

a

b

b

Two patients graded fair improved after the 2-year follow-up. Two patients graded fair were reoperated. c Visual analog scale.

the most recent evaluation; the treatment failure was likely due to an insufficiently radical debridement at the time of the initial autologous chondrocyte transplantation. No additional patients experienced graft failure or had a decline in clinical status between the 2-year evaluation and the follow-up evaluation at 5 or more years (mean, 6.5), resulting in durability of 100% for this group. None of the patients underwent arthroscopy after the autologous chondrocyte transplantation because of graft hypertrophy or any other implantation problem. Group 3: Patellar Lesions Table 4 shows the data with regard to the 17 patients who had repair of patellar defects. The mean defect size was 4.4 cm2 (range, 1.0 to 12.0). The patients had undergone an average of 2.7 previous operations (range, 1 to 6) and had been symptomatic for a mean of 7.2 years (range, 0.6 to 13.2) before the implantation. All of the patients were evaluated 5 years or more after the implantation, and none were lost to follow-up. The overall clinical score was good or excellent in 65% of patients at 2 years after transplantation. All patients with good and excellent ratings at 2 years maintained their ratings at the final follow-up; two patients had improvements from fair to good between the 2-year and final follow-ups. The Tegner-Wallgren activity score increased significantly from preoperatively to 2 years (P ⬍ 0.001) and from preoperatively to final follow-up (P ⬍ 0.001). The modified Cincinnati grading increased significantly at 2 years (P ⬍ 0.001) and at final follow-up (P ⬍ 0.001). The Lysholm and Brittberg visual analog scale scores also demonstrated statistically significant improvements from preoperatively to 2-year follow-up (P ⬍ 0.001), and improvement was maintained at the final follow-up. Two patients whose results were previously graded as fair underwent further treatment and were considered to have failure of treatment at the longer-term evaluation. Four treatment failures occurred in this group, two before the 2-year evaluation (16 and 24 months) and two shortly

7

One patient underwent ACL reoperation at 12 months. Visual analog scale.

after this time. None of the patients who had good or excellent ratings experienced graft failure or had a decline in clinical status, resulting in durability of 100% for this group. Additionally, two patients whose results were previously rated as only fair improved to good during this time. Two patients had a partial lateral meniscectomy 2 and 5 years after the autologous chondrocyte transplantation. No other patients were operated on for graft hypertrophy or for any other implantation problem. Group 4: Femoral Condyles and ACL Reconstruction The data with regard to the 11 patients in group 4 are shown in Table 5. The mean defect size was 4.0 cm2 (range, 1.5 to 10.5). The patients had undergone an average of 1.7 previous operations (range, 1 to 3) and had a mean duration of symptoms of 6.3 years (range, 0.4 to 21.3) before the treatment with autologous chondrocyte transplantation. All of the patients were evaluated at 5 years or more after implantation, and no patients were lost to follow-up. At 2 years after autologous chondrocyte transplantation, the overall score was good or excellent in 91% of patients. Patients with good or excellent ratings at 2 years maintained this rating at the longer-term follow-up. The mean Tegner-Wallgren activity score increased significantly from preoperatively to 2 years (P ⬍ 0.001), and from preoperatively to final follow-up (P ⬍ 0.001). The modified Cincinnati grading had increased significantly at 2 years (P ⬍ 0.001) and at the final follow-up (P ⬍ 0.001). The Lysholm score and Brittberg visual analog scale score tracked the results seen with other scores, demonstrating a statistically significant improvement at 2 years (both scales, P ⬍ 0.001) that was maintained at the final follow-up. Two autologous chondrocyte transplantation treatment failures occurred in this group, one before the 2-year follow-up and the second, in a patient who had been graded fair at 2 years, at 32 months after treatment. One additional patient with a fair result at 1 year was reoperated on because of ligament rupture and ended with an excel-

8

Peterson et al.

American Journal of Sports Medicine TABLE 6 Characteristics of the 11 Patients Who Underwent Second-Look Arthroscopy

Patient no.

9 12 13 19 27 35 43 56 63 63 65 70 Mean SEM a b c

Sex

Age (years)

Defect size (cm2)

Defect locationa

Time between transplant and biopsy (months)

M M F F M M M M F F M F

40 46 22 32 26 18 26 28 32 32 46 46

2.2 2.0 4.0 4.5 3.3 7.5 4.0 5.6 3.3 1.0 3.2 3.0 3.6 0.5

MFC MFC MFC MFC MFC MFC MFC LFC MFCb MFCc MFC MFC

84 78 77 56 61 56 47 38 33 33 34 33 54.3 5.3

Clinical grading

Excellent Excellent Excellent Excellent Excellent Excellent Excellent Excellent Excellent Excellent Good Excellent

MFC, medial femoral condyle; LFC, lateral femoral condyle. Weightbearing. Nonweightbearing. TABLE 7 Mechanical Assessment and Immunohistologic Evaluation Scores

Patient no.

9 12 13 19 27 35 43 56 63 63 65 70 Mean SEM a b

Histologic grading

Hyaline Hyaline Hyaline Hyaline Hyaline Fibrous Hyaline Hyaline Hyaline Fibrous Fibrous Fibrous

Integration to surrounding cartilage (0 to 4)

Defect fill score (0 to 12)

4 4 4 4 4 4 3 4 4 4 1 4 3.7 0.2

11 11 12 10 12 12 9 12 12 10 4 12 10.6 0.6

Indentation value (N) Graft

Normal cartilage

1.4 1.7 2.6 4.3 2.7 1.1 3.7 3.4 4.2 1.9 1.6 1.3 2.4 0.3

0.8 3.5 2.5 2.4 2.4 1.2 3.8 3.7 3.5 4.6 5.0 4.0 3.2 0.3

Immunohistologic evaluation resultsa Collagen type I

⫺ ⫺ ⫺ ⫺ ⫺ ⫹ ⫺ ⫺ ⫺ ⫹ ⫹ ⫹

Collagen type II

⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫺ ⫹⫹ ⫹⫹ ⫹⫹ ⫺ ⫺ ⫺

Aggrecan

⫹ ⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹

COMPb

⫹⫹ ⫹⫹ ⫹⫹⫹ ⫹⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹⫹ ⫹⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹

⫺, no staining; ⫹, lightly stained; ⫹⫹, moderately stained; ⫹⫹⫹, intensely stained. Cartilage oligomeric matrix protein.

lent result at 2 years. No other patient underwent an operation because of graft hypertrophy or for any other implantation problem. No patients with good or excellent ratings experienced graft failure or had a decline in clinical status between the 2-year evaluation and the follow-up evaluation at 5 years (mean, 7.8) or more, resulting in durability of 100% for this group. It should be noted in assessing the change in clinical status of the patients in this group that the individual contributions to the results of the two major reconstructive procedures performed in this patient group cannot be separated.

Mechanical Assessment of Repair Tissue The characteristics of the 11 patients who volunteered to undergo second-look arthroscopy and mechanical and histologic evaluation are given in Table 6. Nine of the patients belonged to treatment group 1, one to group 2 (No.

19), and one to group 4 (No. 27). The integration into the surrounding cartilage was incomplete in two patients (Nos. 43 and 65) (Table 7). The mean defect repair score for the 11 patients was 10.6 ⫾ 0.6 points of the 12 possible points. A slightly fibrillated surface was seen in 6 of the 11 defects, probably because of the remaining periosteum. One graft had a biologically unacceptable appearance with fibrous tissue (No. 65). Indentation measurements were performed at the central part of the graft and on the surrounding normal cartilage (control) on the same condyle. Twelve grafts in the 11 patients who agreed to undergo second-look arthroscopy were evaluated. Considerable variation was seen in both the grafted and the normal (control) indentation values (Table 7). Overall, the stiffness in the grafted areas was 2.4 ⫾ 0.3 N (mean ⫾ SD), compared with 3.2 ⫾ 0.3 N in the adjacent control location (P ⫽ 0.21). The mean stiffness measurement in the grafted areas with hyaline

Vol. 30, No. 1, 2002


9

Figure 3. Biopsy tissue sample taken from the central portion of the grafted area 77 months after implantation in patient number 13 (indentation value, 2.6 N; staining, Alcian blue van Gieson; magnification, ⫻4). The arrow indicates the articular surface.

Figure 2. A, a biopsy specimen taken from the central portion of the grafted area of patient number 12, 78 months after autologous chondrocyte implantation (indentation value, 1.7 N; staining, Alcian blue van Gieson; magnification, ⫻4). B, a biopsy tissue sample taken from the central portion of the grafted area of patient number 63, 33 months after autologous chondrocyte implantation (indentation value, 4.2 N; staining, Alcian blue van Gieson; magnification, ⫻4). The arrow indicates the articular surface. tissue was 3.0 ⫾ 1.1, compared with a stiffness of 1.5 ⫾ 0.35 observed in the fibrous tissue repairs (P ⫽ 0.02). In 8 of 12 stiffness measurements, the indentation measurement was 90% or more of the value observed in the control group. No adverse events were noted from the use of the probe.

Histologic Evaluation of Repair Tissue At the time of indentation probing, full-thickness osteochondral biopsy specimens were taken from the central part of the graft and from the surrounding normal cartilage (control). Eight of the 12 biopsy specimens were consistent with hyaline tissue, while 4 were fibrous in nature. Three of four patients with fibrous graft tissue also had control biopsy specimens that were fibrous. In the patient with two treated defects (No. 63, Table 7), the control biopsy specimen and that taken from a defect in the weightbearing area (Fig. 2) were consistent with hyaline tissue, whereas the defect in the nonweightbearing area was filled with fibrous cartilage (data not shown). Biopsy

Figure 4. Immunohistochemical staining of a biopsy specimen taken from patient number 13. Staining was performed with antibodies against collagen type II (A), aggrecan (B), and cartilage oligomeric matrix protein (C) (magnification, ⫻125). Staining for collagen type I was negative (not shown). D, control, omission of first antibody (magnification, ⫻4).

specimens designated as hyaline cartilage had a positive staining for safranin O and an organized structure in polarized light. Areas with cell clusters were seen, and in all specimens a fibrous structure, probably representing the remaining periosteum, could be seen facing the joint. (Figs. 2 and 3).

10

Peterson et al.

Immunohistochemistry All specimens taken from the second-look arthroscopy procedures were tested for immunoreactivity against cartilage oligomeric matrix protein, collagen types I and II, and aggrecan. All of the specimens stained positive for cartilage oligomeric matrix protein and aggrecan. Hyaline-like cartilage stained positive for type II collagen in the predominant part of the tissue (⬎50%). The fibrous area stained positive for type I collagen (Table 7) (Fig. 4).

DISCUSSION In the present study, we have investigated the long-term clinical durability of treatment with autologous chondrocyte transplantation in the first 61 consecutive patients treated. Good-to-excellent results were seen in 85% of patients (29 of 32) who had had defects on the femoral condyles, including osteochondritis dissecans lesions. Patients with lesions in this location who required concomitant ACL reconstruction had only a slightly reduced rate of success (81%; 9 of 11), and it is impossible to attribute the cause of the clinical failures to either the autologous chondrocyte transplantation or the ligament reconstruction. Although the results for patellar lesions were somewhat less consistent, we were encouraged by the number of good and excellent ratings (13 of 17) recorded in this challenging patient population. Overall, there were 10 treatment failures among the 61 patients (16%), with failure rates lowest among the patients who had isolated femoral condylar lesions (11%) and osteochondritis dissecans lesions (14%) and somewhat higher in those who had autologous chondrocyte transplantation combined with ACL repair (18%) and highest among the patients treated for patellar lesions (24%). A comparison of the outcome variables for each patient at 2 years after the implantation and at the longer-term follow-up evaluation underscores the importance of the early clinical status of each patient as a prognostic factor in the eventual long-term outcome. In this series, 7 of the 10 failures occurred within 2 years of the autologous chondrocyte transplantation. All of the patients whose treatment eventually failed were rated fair or poor at 2 years; none of the patients whose condition was initially rated as good or excellent experienced any deterioration. In nearly all cases, the treatment failures resulted from initial insufficient fill or integration of the defect (perhaps due to insufficient radical debridement) or early exposure to shear forces caused by noncompliance with the prescribed rehabilitation regimen. We hypothesize that the excellent durability of the treatment was a result of the hyaline nature of the repair tissue, which closely approximated the physical characteristics (stiffness) of the native cartilage in the majority of cases. The durability of the repair from 2 years to the time of the long-term follow-up evaluation contrasts sharply with the limited reported data for arthroscopic debridement and fibrocartilage repairs (such as abrasion, drilling, or microfracture). In a randomized, controlled trial of debridement compared with lavage alone, Hubbard14 re-


ported initial success with debridement at 1 year in 80% of patients who had not undergone previous operations; however, this result decreased to 59% at the 5-year evaluation. Ogilvie-Harris and Fitsialos26 examined the outcome of 441 of 551 patients treated with arthroscopic methods: debridement for grade I and II lesions and abrasion for grade III and IV lesions. They found that 68% of the patients had symptomatic relief for at least 2 years, whereas only 53% maintained this status at 4 years after treatment. Similarly, Bert2 reported that fibrocartilage repair tissue resulted from arthroscopic abrasion arthroplasty in some patients, but that this tissue lasted only up to 4 years. A recent study by Nehrer et al.23 that examined failed abrasion treatment reported a mean time to failure of 21 months. Examination of the failed repair revealed soft, fibrillated tissue, frequently with central degeneration. In addition to providing substantial clinical outcome data, the present study also contributes information on the mechanical and immunohistochemical nature of the grafted tissue—an area that remains in need of substantial study. In our earlier study, we found that the biopsy specimens taken from the grafted area of 11 of 15 patients with condylar lesions were characterized as hyaline-like repair tissue.5 In animal studies of cartilage repair in rabbits and dogs, the repair tissue also had a hyaline character.4, 6 The technique used in this study was to obtain a 2-mm cylindrical biopsy specimen, which enabled us to have a thicker sample of the repair tissue compared with the shaving biopsy specimen acquired in the earlier study. Of the 12 biopsy specimens tested, 8 were consistent with hyaline cartilage and 4 were fibrous in appearance. The control specimens of nonrepaired tissue from the four patients who had fibrous repair tissue were consistent with fibrous cartilage in three, indicating that the repair tissue formation could be dependent on the nature of the surrounding cartilage. In the three patients who had fibrous cartilage in the nongrafted area, the biopsy specimen had a fibrous appearance but with a more regular organization of the collagen bundles in polarized light than the fibrous cartilage in the nongrafted area. In one patient, we also found that the location of the grafted area was critical with regard to the formation of the repair tissue. Weightbearing seemed to promote the formation of hyaline repair tissue compared with tissue formation in a nonweightbearing area. We are aware that the 11 patients studied represented only a subset of the total population because only a limited number of patients agreed to undergo a second-look arthroscopic procedure. Immunohistochemical analysis was also performed on 12 biopsy specimens. The biopsy specimens characterized as hyaline cartilage stained positive for type II collagen, aggrecan, and cartilage oligomeric matrix protein. In contrast, fibrous biopsy specimens did not stain positive for type II collagen but stained positive for the others. Cartilage oligomeric matrix protein was first isolated in bovine cartilage, and the immunologic staining pattern is localized to the territorial matrix surrounding the chondrocytes.13 In our specimens, a staining pattern similar to that of normal cartilage was found, and a weak staining was found in the fibrous cartilage. Aggrecan is one of the

Vol. 30, No. 1, 2002

most abundant proteoglycans in cartilage and is seen even in patients with advanced osteoarthritis.29 The presence of aggrecan and cartilage oligomeric matrix protein in our biopsy samples suggests a chondrogenic phenotype of the repair tissue. However it is important to note that type II staining was present only in the eight biopsy samples that were hyaline-like cartilage and not in the four fibrous cartilage specimens, in which only type I staining was found. Recently, there has been some confusion with regard to the interpretation of the results of histologic examination of repair tissue biopsy specimens. The results of the present study demonstrated both that the grafted area contains a hyaline cartilage indistinguishable from the surrounding cartilage and that it is well anchored to the subchondral bone. However, most biopsy specimens have a superficial fibrous surface area resulting from the incorporation of the periosteal remnant. This remnant appears to have no bearing on either the immediate clinical effectiveness of the autologous chondrocyte transplantation or its long-term durability. Similar structural characteristics were recently demonstrated by a group assessing the outcome of two patients with repair tissue produced from commercially available autologous cultured chondrocytes (Genzyme Tissue Repair, Cambridge, Massachusetts).30 These findings, together with our observation that three of four of the fibrous repair tissue specimens were from patients who also had control samples of fibrous cartilage, underscore the importance of careful attention to biopsy technique. Data from shaving or pinch biopsies that do not include a full-core sample (including subchondral bone) and biopsy specimens taken without assessment of a similar sample of adjacent nongrafted tissue are insufficient for drawing conclusions with regard to the nature of the repair. The present study represents the first use of a mechanical probe to quantitatively assess the mechanical characteristics of the repair tissue. Early results seem to indicate that the probe has sufficient sensitivity to detect mechanical differences in hyaline cartilage compared with fibrous cartilage. In our small number of patients, stiffness measurements in hyaline cartilage were twice that of those in fibrous cartilage samples. Unfortunately, our sample of patients was too small to correlate clinical results with probe measurements. Early work has been initiated to validate this probe in a healthy patient population19; however, substantial additional work is required to analyze both normal knees and those undergoing repair to fully understand the value of this device. Nevertheless, we are encouraged by our early findings and believe the instrument may have a role as an early prognostic tool in the evaluation of articular cartilage repairs.

CONCLUSIONS Since the first report on autologous chondrocyte transplantation in 1994,5 we have considered the potential durability of the repair to be an advantage. The long-term results presented in this article confirm our initial belief that treatment with autologous chondrocyte transplanta-


11

tion results in a durable repair for the majority of patients. Our results also indicate the importance of patient status at 2 years as an indicator of future outcome. Patients who have returned to the normal activities of daily living and sport by 2 years after autologous chondrocyte transplantation are able to continue these activities in the long term. In most patients with graft failure, the grafts fail within the first 2 years after implantation. After that period, graft survival is close to 100% between 3 and 8 years later. This outcome contrasts sharply with that of fibrocartilage repairs, which often have poor durability, even in the intermediate term.

ACKNOWLEDGMENTS This project was supported by grants from the Swedish Medical Research Council; Gothenburg University; Genzyme Tissue Repair, Cambridge, Massachusetts; and The Lundberg Foundation, Gothenburg, Sweden. We are indebted to Eva Sjo¨ gren-Jansson and Helena Sterner for excellent cell cultures performed at the Cartilage Laboratory at Sahlgrenska University Hospital. We thank Professor Dick Heinegård for kindly providing the cartilage oligomeric protein and aggrecan antisera and Dr. John Mo for the collagen type II antisera.

REFERENCES 1. Bahuaud J, Maitrot R, Bouvet R, et al: Implantation of autologous chondrocytes for cartilaginous lesions in young patients. A study of 24 cases [in French]. Chirurgie 123: 568 –571, 1998 2. Bert JM: Abrasion arthroplasty. Oper Tech Orthop 7: 294 –299, 1997 3. Blevins FT, Steadman JR, Rodrigo JJ, et al: Treatment of articular cartilage defects in athletes: An analysis of functional outcome and lesion appearance. Orthopedics 21: 761–768, 1998 4. Breinan HA, Minas T, Barone L, et al: Histological evaluation of the course of healing of canine articular cartilage defects treated with cultured autologous chondrocytes. Tissue Eng 4: 101–114, 1998 5. Brittberg M, Lindahl A, Nilsson A, et al: Treatment of deep cartilage defects in the knee with autologous chondrocyte transplantation. N Engl J Med 331: 889 – 895, 1994 6. Brittberg M, Nilsson A, Lindahl A, et al: Rabbit articular cartilage defects treated with autologous cultured chondrocytes. Clin Orthop 326: 270 –283, 1996 7. Dzioba RB: The classification and treatment of acute articular cartilage lesions. Arthroscopy 4: 72– 80, 1988 8. Friedman MJ, Berasi CC, Fox JM, et al: Preliminary results with abrasion arthroplasty in the osteoarthritic knee. Clin Orthop 182: 200 –205, 1984 9. Frisbie DD, Trotter GW, Powers BE, et al: Arthroscopic subchondral bone plate microfracture technique augments healing of large chondral defects in the radial carpal bone and medial femoral condyle of horses. Vet Surg 28: 242–255, 1999 10. Gillogly SD, Voight M, Blackburn T: Treatment of articular cartilage defects of the knee with autologous chondrocyte implantation. J Orthop Sports Phys Ther 28: 241–251, 1998 11. Grande DA, Pitman MI, Peterson L, et al: The repair of experimentally produced defects in rabbit articular cartilage by autologous chondrocyte transplantation. J Orthop Res 7: 208 –218, 1989 12. Haugegaard M, Lundsgaard C, Vibe-Hansen H: Treatment of cartilage defects with autologous chondrocyte implantation [abstract]. Dos Bulletin (Danica Societas Orthopaedica) 27(3): 53, 1998 13. Hedbom E, Antonsson P, Hjerpe A, et al: Cartilage matrix proteins. An acidic oligomeric protein (COMP) detected only in cartilage. J Biol Chem 267: 6132– 6136, 1992 14. Hubbard MJS: Articular debridement versus washout for degeneration of the medial femoral condyle: A five year study. J Bone Joint Surg 78B: 217–219, 1996

12

Peterson et al.

15. Lohnert J, Ruhnau K, Gossen A, et al: Autologous chondrocyte transplantation (ACT) in the knee joint: First clinical results. Arthroskopie 12: 34 – 42, 1999 16. Lysholm J, Gillquist J: Evaluation of knee ligament surgery results with special emphasis on use of a scoring scale. Am J Sports Med 10: 150 –154, 1982 17. Lyyra T, Arokoski JP, Oksala N, et al: Experimental validation of arthroscopic cartilage stiffness measurement using enzymatically degraded cartilage samples. Phys Med Biol 44: 525–535, 1999 18. Lyyra T, Jurvelin J, Pitkanen P, et al: Indentation instrument for the measurement of cartilage stiffness under arthroscopic control. Med Eng Phys 17: 395–399, 1995 19. Lyyra T, Kiviranta I, Va¨ a¨ ta¨ inen U, et al: In vivo characterization of indentation stiffness of articular cartilage in the normal human knee. J Biomed Mater Res 48: 482– 487, 1999 20. Lyyra-Laitinen T, Niinimaki M, Toyras J, et al: Optimization of the arthroscopic indentation instrument for the measurement of thin cartilage stiffness. Phys Med Biol 44: 2511–2524, 1999 21. Minas T: The role of cartilage repair techniques, including chondrocyte transplantation, in focal chondral knee damage. Instr Course Lect 48: 629 – 643, 1999 22. Minas T, Peterson L: Chondrocyte transplantation. Oper Tech Orthop 7: 323–333, 1997 23. Nehrer S, Spector M, Minas T: Histologic analysis of tissue after failed cartilage repair procedures. Clin Orthop 365: 149 –162, 1999

American Journal of Sports Medicine 24. Noyes FR, Barber SD, Mooar LA: A rationale for assessing sports activity levels and limitations in knee disorders. Clin Orthop 246: 238 –249, 1989 25. Noyes FR, Barber-Westin SD: Arthroscopic-assisted allograft anterior cruciate ligament reconstruction in patients with symptomatic arthrosis. Arthroscopy 13: 24 –32, 1997 26. Ogilvie-Harris DJ, Fitsialos DP: Arthroscopic management of the degenerative knee. Arthroscopy 7: 151–157, 1991 27. Outerbridge RE: The etiology of chondromalacia patellae. J Bone Joint Surg 43B: 752–757, 1961 28. Peterson L, Minas T, Brittberg M, et al: Two- to nine-year outcome after autologous chondrocyte transplantation of the knee. Clin Orthop 374: 212–234, 2000 29. Poole A, Rizkallan G, Ionescu M, et al: Osteoarthritis in the human knee: A dynamic process of cartilage matrix degradation, synthesis and reorganization. Agents Actions Suppl 39: 3–13, 1993 30. Richardson JB, Caterson B, Evans EH, et al: Repair of human articular cartilage after implantation of autologous chondrocytes. J Bone Joint Surg 81B: 1064 –1068, 1999 31. Salter RB, Simmonds DF, Malcolm BW, et al: The biological effect of continuous passive motion on the healing of full-thickness defects in articular cartilage. An experimental investigation in the rabbit. J Bone Joint Surg 62A: 1232–1251, 1980 32. Wallgren K, Norlin R, Gillquist J: Activity score for the evaluation of orthopedic patients [abstract]. Acta Orthop Scand 58: 453, 1987