Knee Surg Sports Traumatol Arthrosc (2005) 13: 451–457
KNEE
DOI 10.1007/s00167-004-0535-3
Stefan Marlovits Gabriele Striessnig Florian Kutscha-Lissberg Christoph Resinger Silke M. Aldrian Vilmos Ve´csei Siegfried Trattnig
Received: 2 October 2003 Accepted: 5 April 2004 Published online: 16 October 2004 Springer-Verlag 2004
S. Marlovits (&) Æ G. Striessnig F. Kutscha-Lissberg Æ C. Resinger S. M. Aldrian Æ V. Ve´csei Department of Traumatology, Medical University of Vienna, Waehringer Guertel 18–20, 1090 Vienna, Austria E-mail:
[email protected] Tel.: +43-1-40400 5964 Fax: +43-1-40400 5947 S. Marlovits Æ V. Ve´csei Ludwig Boltzmann Institute for Biomechanics and Cell Biology, Waehringer Guertel 18–20, 1090 Vienna, Austria S. Trattnig Department of Radiology, Medical University of Vienna, Waehringer Guertel 18–20, 1090 Vienna, Austria
Early postoperative adherence of matrix-induced autologous chondrocyte implantation for the treatment of full-thickness cartilage defects of the femoral condyle
Abstract Matrix-induced autologous chondrocyte implantation (MACI) is a tissue-engineering technique for the treatment of fullthickness articular cartilage defects and requires the use of a threedimensional collagen type I–III membrane seeded with cultured autologous chondrocytes. The cellscaffold construct is implanted in the debrided cartilage defect and fixed only with fibrin glue, with no periosteal cover or further surgical fixation. In a clinical pilot study, the MACI technique was used for the treatment of full-thickness, weightbearing chondral defects of the femoral condyle in 16 patients. All patients were followed prospectively and the early postoperative attachment rate, 34.7 days (range: 22–47) after the scaffold implantation, was determined. With the use of highresolution magnetic resonance imaging (MRI), the transplant was graded as completely attached, partially attached, or detached. In 14 of 16 patients (87.5%), a
Introduction Autologous chondrocyte implantation (ACI) is a biological approach for the treatment of large full-thickness chondral defects of the knee, based on the implantation of a suspension of cultured autologous chondrocytes beneath a tightly-sealed periosteal flap [3]. The clinical
completely-attached graft was found, and the cartilage defect site was totally covered by the implanted scaffold and repair tissue. In one patient (6.25%), a partial attachment occurred with partial filling of the chondral defect. A complete detachment of the graft was found in one patient (6.25%), which resulted in an empty defect site with exposure of the subchondral bone. Interobserver variability for the MRI grading of the transplants showed substantial agreement (j=0.775) and perfect agreement (jw=0.99). In conclusion, the implantation and fixation of a cell-scaffold construct in a deep cartilage defect of the femoral condyle with fibrin glue and with no further surgical fixation leads to a high attachment rate 34.7 days after the implantation, as determined with high resolution MRI. Keywords Cartilage defect Æ Matrixinduced autologous chondrocyte implantation Æ High resolution magnetic resonance imaging
experience with this technique now exceeds 10 years, with subjective and objective improvement in the majority of the treated patients [3, 23, 24]. However, despite the promising clinical results obtained, the use of ACI carries a number of limitations, essentially correlated with the complexity of the surgical procedure and biological response of the periosteum [4,
452
20, 23]. High detachment rates for sutured periosteal and fascial flaps were found in an experimental study on goats [7]. Periosteal hypertrophy after ACI, resulting in clinical symptoms, was found in 10–26% of cases [23]. In addition, as cells are grown in a liquid suspension, a lack of chondrocyte differentiation and limited in vitro matrix production are a concern [19]. Recently, the use of three-dimensional scaffolds has been shown to support the maintenance of a chondrocyte-differentiated phenotype [12, 13, 27]. In consequence, efforts are now focused on a tissue-engineered approach, which combines laboratory-grown cells with appropriate three-dimensional biocompatible scaffolds for the purpose of generating new tissues or tissue equivalents [10, 14, 28]. Matrix-induced autologous chondrocyte implantation (MACI) is a tissue-engineering technique that requires the use of a three-dimensional collagen type I–III membrane seeded with cultured autologous chondrocytes [2, 5]. The scaffold is implanted in the debrided cartilage defect and fixed only with fibrin glue, with no periosteal cover or fixing stitches. A sufficient attachment is essential for the successful use of this technique, and the use of a fibrin glue alone for the fixation of the graft may facilitate the detachment of the transplanted cell carrier and may lead to treatment failure. To determine the early postoperative adherence rate, invasive procedures are not possible. Magnetic resonance imaging (MRI), as a noninvasive technique, is very useful for the evaluation of the morphologic status of cartilage defects and the repair tissue throughout the postoperative period [16, 22, 25, 26]. Further improvement in accuracy and sensitivity for the assessment of cartilage repair procedures can be obtained with high resolution MRI [16]. In this clinical pilot study, MACI was used for the treatment of full-thickness, weight-bearing chondral defects of the femoral condyle of the knee joint. All patients were followed prospectively, and the early postoperative adherence rate 4–5 weeks after scaffold implantation was determined using high resolution MRI.
Materials and methods
and IV according to International Cartilage Repair Society (ICRS) criteria. Patient demographics are summarized in Table 1. Operation The MACI procedure was carried out in two stages. During the initial arthroscopy, the cartilage defect was graded, and 200–300 mg of full-thickness cartilage was harvested from the non-weight-bearing surface of the intercondylar notch. The cartilage biopsy was then placed into the transport medium, enzymatically digested, and cultured in monolayer. Culture of the cells was carried out in serum taken from the patient’s blood at the time of surgery under good manufacturing practice (GMP) conditions (Verigen, Leverkusen, Germany). After 3–4 weeks, a cell number of 15–20 million was achieved, and the cells were seeded on a type I–III collagen bilayer membrane. The viability of the cells was determined using the trypan blue dye exclusion assay [11]. During the second operation, a miniarthrotomy allowed access to the cartilage defect in the specified joint compartment. The chondral defect was carefully prepared, and all fissured and undermined cartilage was removed using small closed curettes. The debridement was performed on the subchondral bone plate, avoiding any perforations or subchondral bleeding. A template of the defect was obtained, and the measurements were transferred to the cell-seeded collagen membrane. The membrane was carefully cut out, and implanted into the defect in such a way that the cells were facing the subchondral bone. A commercially-available fibrin glue (Tissucol; Baxter, Vienna, Austria) was used for the graft fixation (Fig. 1). The stability of the transplant was tested by intraoperative movement of the joint. After wound closure, the joint was covered with a compressive elastic bandage. Continuous passive motion (CPM) was started on the first postoperative day, with daily increases in range of motion. CPM was used for 6–8 h daily for up to 6 weeks postoperatively. Isometric strengthening of the Table 1 Patient characteristics Variables
Patients In a prospective pilot study, 16 patients (one female and 15 males) with a mean age of 33.1 years (range: 20.1– 44.3) were included. All patients gave informed consent, and treatment was approved by the local ethics board. The sites of the defects were the following: medial femoral condyle (n=10) and lateral femoral condyle (n=6). The average size of the defects was 4.7 cm2 (range: 2.6– 10.9 cm2) and the defects were classified as Grade III
Demographic data Age (years) Sex (M/F) Baseline characteristics Knee (right/left) Localization (MFC/LFC) Defect size
33.1±7.1 (20.1–44.3) 15/1 8/8 13/3 4.7±2.3 (2.6–10.9)
Mean±SEM (min-max) SEM standard error of mean, min minimum value, max maximum value, MFC medial femoral condyle, LFC lateral femoral condyle
453
rectangular field of view: 80%; matrix: 256·256; scan percentage: 80%; slice thickness: 4 mm; gap: 0.4 mm; number of acquisitions: 3; number of slices: 19); and 3Dgradient echo sequence with fat-suppression (TR: shortest; TE: 4.6 ms; flip angle: 45; FOV: 180 mm; rectangular field of view: 90%; matrix: 256·256; scan percentage: 75%; slices: 50; slice thickness: 2 mm; number of acquisitions: 1). For high-resolution imaging, a surface-phased array coil was placed over the knee compartment of interest, including the cartilage repair site. With this surface coil, the following sequence was applied: sagittal dual TSE (TR: 2,400 ms; TE: first, shortest; second, 120 ms; TSE factor: 12; FOV: 120 mm; rectangular field of view: 75%; matrix: 512·512; scan percentage: 80%; slice thickness: 2 mm; gap: 0.2 mm; number of acquisitions: 4; number of slices: 23). Scoring
Fig. 1 Matrix-induced autologous chondrocyte transplantation on the right medial femoral condyle. Deep cartilage defect of the weight-bearing articular surface of the femoral condyle before (a) and after defect preparation down to the subchondral bone plate (b). Preparation of a template of the defect size (c) and completelyfilled defect with autologous chondrocytes seeded on collagen scaffold fixed only with fibrin glue (d)
quadriceps began in the first week, while partial weightbearing was allowed after 6 weeks and full weightbearing after 10 weeks. High resolution MRI All MR examinations were performed on a 1.0 T MR unit (Philips, Gyroscan Intera, Best, The Netherlands). Initially, we used a standard knee MRI protocol that incorporates proton density and fast-spin echo (FSE) acquisitions. Using the circular polarized knee coil, the following sequences were applied: sagittal T1-SE (TR: shortest; TE:13.8 ms; FOV: 150 mm; rectangular field of view: 95%; matrix: 256·512; scan percentage: 100%; slice thickness: 3 mm; gap: 0.3 mm; number of acquisitions: 2; number of slices: 23); sagittal dual TSE (TR: 2,500 ms; TE: first, shortest, second, 120 ms; TSE factor: 12; FOV: 180 mm; rectangular field of view: 100%; matrix: 256·512; scan percentage: 100%; slice thickness: 3 mm; gap: 0.3 mm; number of acquisitions: 3; number of slices: 23); coronal STIR-TSE (TR:1,200 ms; TE: 13 ms; TI: 130 ms; TSE factor: 5; FOV: 200 mm;
For the scoring, we used a simple evaluation system and graded the transplant as completely attached, partially attached, or detached (Fig. 2). Attachment was scored when the prepared defect site was completely filled with the graft and no gaps were visible. In partial attachment, the graft covered only a part of the defect. If the cartilage defect zone was completely empty, this was rated as detachment. If detachment was present, partial or complete separation of the entire repair tissue with displacement from the defect site or partial or complete dissolution was possible. To evaluate the validity and reliability of the scoring, the MRI scans were read by two independent radiologists, and the interobserver variability was determined using kappa statistics (kappa [j] and weighted kappa [jw]; [6, 9]). The j statistic provides a measure of interobserver agreement that adjusts for agreement based on chance. j estimates the amount of agreement that is independent of chance, and varies from )1.0 (perfect disagreement) through 0.0 (change agreement) to 1.0 (perfect agreement). According to Landis and Koch, a j-value of 0.81 indicates ‘‘almost perfect’’ agreement, while those values between 0.61 and 0.80 are ‘‘substantial,’’ those between 0.41 and 0.60 are ‘‘moderate,’’ and those between 0.21 and 0.40 are ‘‘fair’’ [17, 18].
Results The cartilage biopsy and the in vitro cell processing were successful in all patients. The viability of the cells at the time of seeding on the collagen carrier reached a mean value of 96% (standard deviation: 0.024). After defect preparation and scaffold implantation, the stability and adherence of the transplant was checked intraoperatively by moving the knee joint. During the
454
Fig. 2 Schematic drawing of the scoring system for the evaluation of the transplants graded as completely attached, partially attached, or detached. Attachment was scored when the prepared defect site was completely filled with the graft and no gaps were visible (a). In partial attachment, the graft covered only a part of the defect (a and b). If the cartilage defect zone was completely empty, this was rated as detachment (c)
knee movement, in all cases, a stable, non-displaced graft was present. Joint closure was only performed after verification of graft adherence (Fig. 1). In the early postoperative period, no infections or further complications were observed, and all patients followed the rehabilitation protocol. According to the study protocol, all patients were evaluated 4–5 weeks after the scaffold implantation with high resolution MRI (mean: 34.7 days after the operation; range: 22–47 days). The examination time was 35 min, with 5 min for the preparation of the patient and positioning of the coil. The signal intensity of the graft and the repair tissue in high resolution MRI using fast spin echo (dual T2FSE) sequences was always markedly hyperintense, and appeared brighter compared to the adjacent native cartilage (Fig. 2). The defect size and the position of the transplant were always clearly distinguishable from the adjacent cartilage.
Fig. 3 High resolution MRI of femoral condyles with the prepared cartilage defect and the implanted scaffold 4–5 weeks after implantation (sagittal T2-FSE). Complete attachment and filling of the defect is shown in a and b. Partial attachment and incomplete filling was observed in one patient (c), where the anterior part of the defect was covered and the posterior part was empty. A complete detachment of the transplant with exposure of the subchondral bone is shown in d (arrows indicate the transplanted areas)
A completely-attached graft and repair tissue was found in 14 of 16 patients (87.5%). In these patients, the cartilage defect site was totally covered by the implanted scaffold (Fig. 2a and b). A partial attachment was observed in one patient (6.25%, Fig. 3), and a detachment in one patient. In the patient with the partial attachment, the graft was partially dissolved and covered only the anterior parts of the prepared femoral defect, whereas the posterior part was empty (Fig. 2c). In the patient with a detachment of the graft, the cartilage defect site was completely empty down to the border of the subchondral bone (Fig. 2d). In this case, a displaced graft, forming a loose body in the knee joint, was not observed. Fig. 4 shows the adherence characteristics (attachment and detachment) of matrix-induced autologous chondrocyte implantation for the 16 patients. The j-statistics for the evaluation of the interobserver variability of the scoring system showed substantial agreement, with a j-value of 0.775, and almost perfect agreement, with a jw-value of 0.99.
455
Fig. 4 Adherence (attachment and detachment) of matrix-induced autologous chondrocyte implantation in 16 patients
Discussion The classic autologous chondrocyte transplantation was described by Britberg as the first generation of a celltransplantation technique for cartilage repair, which is characterized by the combination of two chondrogenic factors: the implanted suspension of chondrocytes and the cambium cells of the periosteum [3, 4]. The steps in open ACT implantation include arthrotomy, defect preparation, periosteum procurement, periosteum fixation, periosteum water-tight integrity testing, autologous fibrin glue sealant, chondrocyte implantation, wound closure, and rehabilitation [20]. This procedure has certain disadvantages, including the potential leakage of cells from defects, the dedifferentiation of a cellular phenotype as the cells are grown in monolayer before implantation, the uneven distribution of cells, and the substantial risk of periosteal complications [4]. Further improvement of the technology of autologous chondrocyte transplantation includes the use of biomaterials as cell carriers, and scaffolds for cell growth. These biomaterials secure the cells in the defect area and enhance their proliferation and differentiation. Many biomaterials have been tested for cartilage repair in in vitro and in vivo studies [15]. Resorbable collagen membranes, such as Chondro-Gide, a type-I/ type-III collagen scaffold (Geistlich Biomaterials, Pharma AG, Wolhausen, Switzerland), are a possible alternative to the periosteum [27]. The first clinical results with that membrane seem promising, and show a success rate comparable to classical ACT [2, 5]. Another biomaterial is the hyaluronan-based biodegradable polymer scaffold, Hyaff-11 (Fidia Advanced Biopolymers, Abano Terme, Italy), where good experimental and initial clinical data have been published [12, 13, 21]. Other
biomaterials that are in experimental and clinical use are polymers of polylactin and polyglactin (Ethicon, Norderstedt, Germany), which may offer new possibilities for cartilage repair [8]. In general, these biomaterials are effective on the cell biology of seeded chondrocytes in the phase of cell proliferation and cell differentiation [13, 27, 28]. In the clinical application for the treatment of full-thickness cartilage defects, biomaterials are used as cell carriers and three-dimensional, tissue-engineered cell-scaffold constructs. After debridement of the defect, the biomaterials with seeded cells are trimmed to exactly match the defect size and implanted without the use of a periosteal cover or fixing stitches. In most techniques, only a fibrin glue is used for the fixation of the graft [2, 21]. Because of the plasticity of the tissue-engineered constructs and no requirement for periosteal grafting, minor incisions are used and initial attempts at arthroscopic implantation have been reported [8]. For the scientific evaluation of these matrix-associated techniques, a systematic approach is necessary to compare the results with the published results after autologous chondrocyte transplantation using a periosteal membrane. One major concern with the use of biomaterials for cell-based repair of cartilage defects with no surgical graft fixation, using stitches or other fixation devices, is graft detachment, which can lead to graft failure. For the early postoperative evaluation of graft attachment, invasive methods are not possible. However, magnetic resonance imaging is a useful tool because it is noninvasive and characterizes the cartilage morphology adequately and accurately [1, 26]. In our pilot study for the evaluation of the MACI technique, we used high resolution imaging for the detection of subtle changes in the repair tissue and the adjacent structures. A surface coil, which substantially increases the signal-to-noise ratio, was placed over the knee compartment of interest, and the dual TSE sequence was adapted using a high matrix size and a small field of view to provide high spatial resolution [16]. No contrast medium is necessary to enhance image quality, thus improving patient compliance and cooperation [29]. We used this technique for the evaluation of the graft adherence of the MACI method 4–5 weeks after the implantation. The aim was the description of graft adherence and graft detachment. With a simple evaluation system, we scored the transplant as completely attached, partially attached, or detached (Fig. 2). To evaluate the validity of the scoring system, all MRI scans were read by two independent radiologists and the interobserver variability was determined using j-statistics [9]. The determined j-value of 0.775 showed a ‘‘substantial’’ interobserver agreement [17]. The usual calculation of j does not allow for a consideration of partial agreement between two different responses when there are more than two responses possible; one response
456
either agrees or disagrees with another. In other words, calculation of j does not permit a weighting of the degree of agreement or disagreement between possible responses when this would be appropriate. For example, in a study of a specific radiographic finding for which there are three possible responses (finding present, finding indeterminant, or finding absent) calculation of j does not recognize that there is less disagreement between the responses ‘‘finding present’’ and ‘‘finding indeterminant’’ than there is between ‘‘finding present’’ and ‘‘finding absent’’. The calculation of ‘‘weighted j’’, however, allows for a consideration of such differences [6]. In our study, in addition to the calculation of j, the weighted j-score for three possible responses was calculated. The jw-value of 0.99 showed an ‘‘almost perfect’’ interobserver agreement. Sixteen patients with singular full-thickness, weightbearing chondral lesions of the femoral condyle were analyzed prospectively. Overall, a complete attachment and adherence was found in 14 patients. Partial attachment and complete detachment were each found in one patient. The possible reasons for the early detachment of the implanted biomaterials include weak mechanical fixation and adherence and/or a possible faster biochemical degradation that led to partial or total dissolution. In classical ACT, most complications are directly related to the periosteal graft. Early problems include periosteal graft detachment and delamination, and late periosteal hypertrophy. Delamination is most commonly encountered during the early postoperative period (
