- Email: [email protected]
Histologic and Biomechanical Analysis of Anterior Cruciate Ligament Graft to Bone Healing in Skeletally Immature Sheep
Histologic and Biomechanical Analysis of Anterior Cruciate Ligament Graft to Bone Healing in Skeletally Immature Sheep Rupert Meller, M.D., Elmar Willbold, M.D., Eric Hesse, M.D., Beatrix Dreymann, M.D., Michael Fehr, M.D., Carl Haasper, M.D., Christof Hurschler, M.D., Christian Krettek, M.D., F.R.A.C.S., F.R.C.S.Ed., and Frank Witte, M.D.
Purpose: It was our aim to establish an animal model and to investigate the tendon graft–to– bone and physis healing process in skeletally immature sheep after reconstruction of the anterior cruciate ligament (ACL). Methods: Thirty-two immature sheep aged 4 months underwent a fully transphyseal ACL reconstruction by use of a soft-tissue graft. The animals were subsequently killed after 3, 6, 12, and 24 weeks and analyzed histologically and biomechanically. Results: There was a transient hypertrophy of the physis tissue at the passing site of the graft. Anchoring Sharpey-like fibers evolved as early as 3 weeks after surgery. A strong expression of collagen III messenger ribonucleic acid within the first 6 weeks preceded this anchoring process. The maximum load to failure of the tendon graft in the reconstructed knees initially decreased to 37.8 ⫾ 17.8 N after 3 weeks and was restored to 522.9 ⫾ 113 N after 24 weeks. Tendon graft stiffness was restored to 86% when compared with the control knees. Conclusions: The early anchoring by Sharpey fibers was found at 3 weeks with continued maturation to 24 weeks. This development of anchoring fibers corresponded to that of biomechanical strength, starting with 5% of the normal knee at 3 weeks and then 15.2% at 6 weeks, 41.2% at 12 weeks, and 69% at 24 weeks. Tendon graft–to– bone and physis healing in skeletally immature sheep is further characterized by a transient hypertrophy of the physis cartilage. The physis recovers well from the trauma of drilling and placement of a soft-tissue graft. The early development of Sharpey-like fibers results in a solid integration of the graft into bone in a timely manner. Clinical Relevance: ACL reconstruction in skeletally immature individuals is still controversial. This study describes in detail the histologic and biomechanical stages of tendon graft healing to the bone and physis. These data enrich the existing knowledge of previous studies in adult sheep and may provide a basis for further research in the controversial field of ACL reconstruction during growth. Key Words: Anterior cruciate ligament reconstruction—Tendon graft—Growth—Skeletally immature—In situ hybridization—Sheep model.
A
lthough there exists a broad scope of studies and experiences regarding most aspects of anterior cruciate ligament (ACL) reconstruction in adults,
From the Trauma Department (R.M., E.H., C. Haasper, C.K.) and Orthopaedic Department (E.W., B.D., C. Hurschler, F.W.), Hannover Medical School, and Small Animal Clinic, University of Veterinary Medicine Hannover (M.F.), Hannover, Germany. Supported by the Research Commission of Hannover Medical School, Hannover, Germany. The authors report no conflict of interest. Received January 27, 2008; accepted June 27, 2008. Address correspondence and reprint requests to Rupert Meller, M.D., Trauma Department, Hannover Medical School (MHH), Carl-Neuberg-Strasse 1, 30625 Hannover, Germany. E-mail: meller. [email protected] © 2008 by the Arthroscopy Association of North America 0749-8063/08/2411-0848$34.00/0 doi:10.1016/j.arthro.2008.06.021
only limited data are available regarding the tendon graft–to– bone healing process in skeletally immature subjects, although intraligamentous ruptures of the ACL are diagnosed in this group with increasing frequency.1-5 To gain more insight into the healing process after ACL reconstruction in skeletally immature subjects, a corresponding sheep model was established. We hypothesized that the graft incorporation into the bone tunnel and physis tissue follows distinct mechanisms and that the incorporation is reflected by a rapid gain in biomechanical strength. The purpose of the study was to investigate the basic mechanisms of the graft tissue incorporation into the bone tunnel and the physis tissue, as well as to determine the biomechanical behavior of the graft during the first 6 months after surgery.
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 24, No 11 (November), 2008: pp 1221-1231
1221
1222
R. MELLER ET AL.
FIGURE 1. (A) Schema of a right sheep knee showing physis (white triangle) and position and orientation of graft within tibial and femoral tunnels. (B) A general view of the explanted and in plastic-embedded implantation site (black rectangle in A) shows the femoral tunnel (the walls of which are marked by black dots) crossing the physis (white triangles). The tunnel is filled with the well-incorporated ligamentous graft (bright white fibers). The articular tunnel entrance is marked by a black arrow (1, fixation site; 2, tunnel midportion [physis]; 3, articular tunnel entrance; bar, 4,500 m.).
METHODS Study Design All procedures were performed with permission of all responsible authorities and in accordance with the National Institutes of Health guidelines for the use of laboratory animals. Thirty-two black-headed sheep aged 4 months were acquired from the Niedersächsischer Schafzüchterverband (Hannover, Germany). Before surgery, a veterinarian confirmed the health status of the sheep. All animals underwent a fully transphyseal reconstruction of the right ACL. The left knees served as controls. Four groups of eight animals each were killed after 3, 6, 12, and 24 weeks. Two animals from each group were used for the histologic analysis and six for the biomechanical analysis. Operative Technique The right knee joint was exposed through an anteromedial incision with release of the medial parapatellar retinaculum. The patella was displaced laterally, the anterior fat pad was sharply separated, and the ACL was exposed and removed. A split of the ipsilateral superficial flexor digitorum tendon and the
gastrocnemius tendon was harvested and used as graft. The tibial and femoral tunnels were created by use of a fully transphyseal technique (Fig 1A). The tibial tunnel placement was centered to the attachment area of the anteromedial bundle of the ACL. The starting point was 15 mm distal and 10 mm medial to the most prominent point of the tuberosity. By use of a K-wire drill guide adjusted at 40°, a 2.4-mm K-wire was first placed and then overdrilled by a 4.5-mm EndoButton hand drill (Smith & Nephew, Andover, MA) to avoid thermal damage to the physis. The femoral tunnel entrance was located in the origin area of the ACL. The K-wire was aimed toward the lateral aspect of the distal third of the thigh, ensuring extra-articular positioning of the fixation device. This desired position was best achieved in 30° of knee flexion. The graft was fixed by use of the EndoButton proximally and a suture washer distally (Smith & Nephew Endoscopy). Gross Inspection and Tissue Sampling Knee joints used for histologic analysis were exposed by an anteromedial arthrotomy. The graft sites were scrutinized for their integrity and then sharply dissected. Knee joints dedicated for biomechanical
TENDON GRAFT TO BONE HEALING DURING GROWTH analysis were disarticulated in the hip joint and frozen in full extension at ⫺20°C. Twelve hours before biomechanical testing, samples were thawed to room temperature. The capsule was sharply dissected, and the graft was checked for its integrity. General Histology After 3, 6, 12, and 24 weeks, the dissected femoral transplantation sites were fixed in 3.7% commercial formalin for 7 days at 4°C. After embedding and polymerization in methyl methacrylate (Technovit 9100 New; Heraeus-Kulzer, Hanau, Germany) according to established protocols (Fig 1B), the tissue blocks were first trimmed by use of an Exakt 310 Diamond Band Saw (EXAKT, Norderstedt, Germany). By use of an RM 2155 microtome (Leica, Bensheim, Germany), 5-m thin sections were cut and placed onto poly-L-lysine– coated glass slides. Before the different staining procedures, the sections were first dried for 2 days at 37°C and then deacrylated in xylol (15 minutes) and 2-methoxyethyl acetate (10 minutes), cleared through a decreasing ethanol series (isopropyl alcohol, 96% ethanol, and 70% ethanol, for 2 minutes each), and rehydrated by use of distilled water. Toluidine Blue Staining For toluidine blue staining, sections were incubated in 0.1% toluidine blue O (Sigma, Taufkirchen, Germany) for 20 seconds, washed in distilled water, dehydrated in ethanol, and mounted in Eukitt (Labonord, Mönchengladbach, Germany). In Situ Hybridization for Collagen III Messenger RNA For messenger RNA (mRNA) in situ hybridization, ZytoFast RNA in situ Hybridisierungs System Biotin (ZytoVision, Bremerhaven, Germany) was used. The sections were first decalcified for 60 minutes in 10% ethylenediaminetetraacetic acid and were then washed and gently digested for 15 minutes with pepsin. Digestion was stopped by washing for 10 seconds with ethanol. Slides were air dried, and the sections were hybridized with the collagen III–specific biotin-labeled ribonucleic acid probe overnight at 55.5°C. Slides were washed with buffer, and a biotin-specific antibody conjugated with alkaline phosphatase was applied overnight at 4°C. The slides were washed again with buffer, and the probes were visualized by use of nitro blue tetrazolium chloride/5-bromo-4chloro-3-indolyl phosphate (NBT/BCIP) under micro-
1223
scopic control. To assess the specificity of our protocol, we carried out both a positive control consisting of a poly-dT oligonucleotide detecting the poly-A tails of mRNA, as well as the integrity of the mRNA,andanegativecontrolconsistingofa“randomsequence” oligonucleotide to detect nonspecific binding. In addition, controls to exclude nonspecific binding of the antibody as well as controls to exclude nonspecific chromogenic reaction were carried out. Finally, the sections were washed and mounted with Kaiser’s glycerol gelatin (Merck, Darmstadt, Germany). The probe sequence was 5=-biotinTggCTACTTCTCgCTCTgCTTCATCCCACT-3=, and it was synthesized by TIB MOLBIOL (Berlin, Germany). Microscopy and Photography Photomicrographs were taken with a Zeiss Axioskop 40 microscope in combination with a Zeiss AxioCam MRc digital camera and Zeiss AxioVision software (Zeiss, Oberkochen, Germany). Specimen Preparation and Cross-Sectional Area for Biomechanical Evaluation The knee joints were dissected of all remaining soft tissue. The femur and tibia were cut 10 cm above and below the joint line, leaving a femur-ACL-tibia complex (FATC) of the control left knee and a femurgraft-tibia complex (FGTC) of the operated right knee for uniaxial tensile testing. The bone ends were then embedded in metal cylinders by use of polyurethane casting resin and mounted to a custom-made adjusting device. The cross-sectional area of each ACL (intact left knee) and tendon graft (right knee) was measured 3 times with a laser micrometer (Takikawa Engineering, Tokyo, Japan). The means of 3 measurements of each knee were used for statistical analysis. The crosssectional area was used to determine the strain and Young’s modulus. Failure Testing For failure testing, a materials testing machine (Zwick 1445; Zwick, Ulm, Germany) was used. The FATC and FGTC, mounted in the testing fixture, were carefully aligned parallel to the axis of the applied load in 60° of knee flexion. For the preconditioning of the tendon graft, a preload of 5 N was applied with a load displacement rate of 6 mm/min, and 20 loading cycles were performed (12 mm/min; with upper and lower limits of 5 N and 0 N, respectively). The surface of the ACLs and tendon grafts were marked with
1224
R. MELLER ET AL.
high-contrast grease that allowed the determination of the surface strain. This marker motion was recorded by a digital video camera (Sony DCR-HC17E; Sony, Tokyo, Japan). Knees were then loaded until failure with a displacement rate of 6 mm/min. Maximum load to failure and stiffness were calculated from the load displacement curve. The Young’s modulus was calculated from the stress-strain curve. Failure modes (femoral or tibial graft pullout, graft rupture at the femoral or tibial insertion site, midsubstance failure, or bony avulsion) were recorded. Data Analysis All data were analyzed with SPSS software (version 14.0; SPSS, Chicago, IL). The biomechanical data were analyzed with a 2-way analysis of variance. Independent variables were the time of sacrifice and the left or right knee. Paired t tests were used to look for side-to-side differences. The differences were considered to be significant at a probability level of P ⱕ .0125 after Bonferroni correction for multiple comparisons.
RESULTS Postoperative Course Two animals died of pneumonia postoperatively (3and 6-week groups) and were excluded from the study, leaving 30 animals for the final evaluation. All other animals were healthy and showed a quick return to full mobilization. The macroscopic appearance of the hind limbs did not show obvious angular or rotational deformities. All grafts were in place. Intratunnel Graft Remodeling Similar to Intra-Articular Portion Remodeling Three weeks after surgery, the tendon graft showed acellular areas. However, other regions of the graft already showed a significant cell repopulation. After 6 weeks, the graft was highly vascularized and many fibroblasts were present between the collagen fibers. During the next 6 weeks, the collagen fibers became more organized and the vascularization and cell density decreased. After 6 months, the collagen fibers were densely packed and aligned in parallel with fibroblasts and capillaries in between. The histologic structure of the tendon graft appeared close to that of the contralateral control knees.
Tunnel Wall Remodeling Starts at Fixation Sites and Proceeds Toward Joint Because of the drilling procedure, the tunnel walls showed an initial destruction of the cancellous bone. After 3 weeks, the trabeculae appeared rugged. Numerous gaps allowed a direct contact between the graft and the bone marrow (Fig 2A). After 6 weeks, new trabeculae were formed and the gaps in the tunnel walls were mostly closed by newly formed woven bone. A corresponding osteoid formation was also present (Fig 2B). During the next 6 months, the remodeling process further proceeded and the thickness of the wall increased significantly (Figs 2C, 2D). Early Development of Sharpey-Like Fibers In the early stages of the graft healing process, only scattered fibers passing from the fibrovascular interface into the tunnel walls could be identified. These fibers were predominantly seen adjacent to the fixation sites (Fig 3A and B). After 6 weeks, these Sharpeylike fibers increased noticeably in density and were then also present in the tunnel midportion and at the articular tunnel entrance. The fibers were aligned oblique to the supposed tensile stresses within the tendon graft and seemed to anchor the fibrovascular interface tissue into the bone (Fig 3C and D). After 12 weeks (Fig 3E and F) and 24 weeks (Fig 3G and H), the number and size of these Sharpey-like fibers further increased and they also reached much deeper into the surrounding bone. This architecture resembled an indirect insertion anatomy. Transient Hypertrophy of Physis Cartilage Three weeks after surgery, we found no signs of necrosis or other damage to the physis that might be correlated with tunnel preparation. After 3 weeks, the physis cartilage expanded considerably into the fibrovascular interface tissue but without its typical architecture. The chondrocytes were smaller in size, and their nuclei stained stronger. A distinct layer of smallsized cells was present between the graft and the physis (Fig 4A). During the next 3 weeks, this hypertrophic physis tissue showed an orientation toward the tensile forces along the graft (Fig 4B). After 12 weeks, the physis tissue within the tunnel retracted (Fig 4C), and after 24 weeks, it was in line with the remodeled tunnel wall (Fig 4D). At that point, the chondrocytes again showed the typical architecture and were arranged in columns.
TENDON GRAFT TO BONE HEALING DURING GROWTH
1225
FIGURE 2. Remodeling of tunnel wall: Toluidine blue staining of tunnel wall after 3 weeks (A), 6 weeks (B), 12 weeks (C), and 24 weeks (D). (A) After 3 weeks, the tunnel wall still shows the destruction of the cancellous bone (stars). Numerous gaps (arrows) perforate the wall and allow a direct contact between the graft and the bone marrow. (B) After 6 weeks, nearly all of the gaps are closed and the surface of the wall has become smoother. Only small bulges (arrows) are still present. After 12 weeks (C) and 24 weeks (D), the thickness of the wall had significantly increased.
Early Expression of Collagen III mRNA To investigate the formation of anchoring fibers, in situ hybridization for collagen III, a major component of Sharpey-like fibers, was used. After 3 weeks (Fig 5A and C) and 6 weeks (Fig 5B), positive cells were predominantly found in the fibrovascular interface tissue in close vicinity to the bony parts of the tunnel wall, especially in the elongation of the initially destroyed trabeculae (Fig 5C). Only a few positive cells were scattered in the graft and the bone. After 12 weeks (Fig 5D) and 24 weeks (Fig 5E), positive cells were more likely allocated within the tendon graft and the bone. The number of collagen III–positive cells and their staining intensity significantly decreased from 12 to 24 weeks. Cross-Sectional Area The mean cross-sectional area of the intact ACLs was 23.39 ⫾ 3.5 mm2. There was a significant in-
crease in the cross-sectional area of the tendon grafts throughout the remodeling process, from 25.1 ⫾ 7.3 mm2 after 3 weeks to 51.1 ⫾ 14.9 mm2 after 24 weeks (P ⬍ .005). Failure Testing The 2-way analysis of variance showed a significant effect of time and side on the maximum load to failure and the stiffness. In ACL-reconstructed knees the load to failure raised steadily, indicating an improvement of the function of the tendon graft (Fig 6A). The intact contralateral knees (FATC) had a mean load to failure of 759.2 ⫾ 114.1 N. The operated knees (FGTC), after 3 weeks, reached only 37.8 ⫾ 17.8 N (5% of intact FATC, P ⬍ .0005). There was a nonsignificant increase in the load to failure of the FGTC during the next 3 weeks (115.9 ⫾ 59.6 N, 15.2% of intact ACLs; P ⫽ .02). After 12
1226
R. MELLER ET AL.
FIGURE 3. Remodeling of tendon-to-bone interface and increasing anchoring of graft by Sharpey-like fibers: Toluidine blue staining (A, C, E, G) and corresponding polarization optics (B, D, F, H) after 3 weeks (A, B), 6 weeks (C, D), 12 weeks (E, F), and 24 weeks (G, H). (A, B) Three weeks after transplantation, only very few fibers (arrows) passing from the tendon to bone are present. (C, D) After 6 weeks, the number of these fibers had significantly increased and they reach deeper into the bone. After 12 weeks (E, F) and 24 weeks (G, H), the number and size of these fibers further increased. Stars indicate the tunnel wall.
TENDON GRAFT TO BONE HEALING DURING GROWTH
1227
FIGURE 4. Development of tendon-to-physis interface: Toluidine blue staining of physis after 3 weeks (A), 6 weeks (B), 12 weeks (C), and 24 weeks (D). (A) After 3 weeks, the physis (diamonds) had protruded into the graft, indicating that the physis was not damaged during the operation. (B) After 6 weeks, the physis had retracted, showing a clear orientation toward the tensile forces. After 12 weeks (C) and 24 weeks (D), the physis fully retracted to be in line with the wall of the tunnel. Stars indicate the tunnel wall.
weeks, the FGTC failed at 316.4 ⫾ 134.9 N (41.7% of intact ACLs, P ⬍ .001). Finally, after 24 weeks, the FGTC gained further stability, resulting in a load to failure of 522.9 ⫾ 113.9 N (69% of intact ACLs). No statistically significant side-to-side difference could be observed at this time point (P ⫽ .05). The corresponding stiffness value was 136.3 ⫾ 28.5 N/mm for the intact contralateral knees (FATC). After 3 weeks, the operated knees (FGTC) reached 19.6 ⫾ 8.2 N/mm and then steadily increased to 56.3 ⫾ 22.5 N/mm after 6 weeks and 84.6 ⫾ 21.7 N/mm after 12 weeks (P ⬍ .001 for 3, 6, and 12 weeks). After 24 weeks, the stiffness reached 113.7 ⫾ 5.7 N/mm (P ⫽ .05), implying a restoration to 86.5% when compared with intact knees (Fig 6B). No statistically significant sideto-side difference could be observed at this time point. The Young’s modulus of the FATC was 217.3 ⫾ 57.9 N/mm2. The Young’s modulus of the FGTC decreased to 35.5 ⫾ 17.6 N/mm2 after 3 weeks and
increased to 74.6 ⫾ 37.6 N/mm2 after 24 weeks. This was significantly different from the control at 24 weeks (P ⬍ .0005). Different Modes of Failure The modes of failure for the intact knees were either fracture of the distal femoral physis (11 knees) or a tibial bony avulsion of the ACLs (11 knees) (the posterolateral bundle constantly failed first). After 3 weeks, the mode of failure of the FGTC was predominantly by graft pullout from the femoral tunnel. One graft after 3 weeks failed at its midportion. Similar modes of failure could be found after 6 weeks, with 3 femoral pullouts and 2 midportion ruptures. After 12 and 24 weeks, grafts failed by midportion ruptures in 2 knees and by tibial bony avulsions in 4 knees, respectively.
1228
R. MELLER ET AL.
FIGURE 5. Spatiotemporal expression of collagen III mRNA after 3 weeks (A, E), 6 weeks (B), 12 weeks (C), and 24 weeks (D). (A) After 3 weeks, collagen III– expressing cells are predominantly located in the tendon-to-bone interface, especially in the elongation of the ruptured trabeculae (black arrows in A, E), which is indicative of the formation of Sharpey-like fibers and the remodeling of the bony tunnel wall (stars). After 6 weeks (B) and 12 weeks (C), this process proceeds further on. After 24 weeks (D) and the completion of the tunnel remodeling, only low collagen III mRNA expression can be detected.
DISCUSSION To gain more insight into the healing processes after ACL reconstruction during growth, we used skeletally
immature sheep as a large animal model. A nonanatomic fixation far away from the joint line was used to protect the physis tissue. The tendon graft–to– bone
TENDON GRAFT TO BONE HEALING DURING GROWTH
FIGURE 6. (A) Biomechanical evaluation of maximum load to failure of intact knee and ACL-reconstructed knee. The horizontal line shows the mean value for the load-to-failure data of the intact contralateral knees, and the dotted lines show the corresponding SD. Asterisks indicate statistically significant side-to-side differences at 3, 6, and 12 weeks. (B) Biomechanical evaluation of stiffness (in Newtons per millimeter) at failure testing of the native, ACL-intact knee and the ACL-reconstructed knee. Asterisks indicate statistically significant side-to-side differences at 3, 6, and 12 weeks (P ⬍ .0125). The means of the measured values are connected by lines to visualize their change over time in the different time groups.
healing process in these cases progresses via a fibrous interzone, resulting in an indirect insertion anatomy.6 We investigated the spatiotemporal patterns of tendon graft–to– bone and physis healing as well as the recovery of the biomechanical properties. We hypothesized that the graft incorporation into the bone tunnel and physis tissue follows distinct mechanisms and that the incorporation is reflected by a rapid gain in biomechanical strength. The hypothesis was affirmed, because we have seen an early development of Sharpey-like fibers as soon as 3 weeks after surgery. This strong incorporation re-
1229
sulted in equivalent biomechanical properties. We found a quick remodeling of the tendon graft within the tunnel and an early formation of Sharpey-like fibers as soon as 3 weeks after surgery. We also found a complex response of the physis cartilage at the passing site of the tendon graft. The physis recovered well from the trauma of drilling and placement of a soft-tissue graft. The biomechanical data fully correspond with the morphologic data and showed an increasing recovery of the failure load to almost 70% of the intact knee and a recovery of the stiffness to 86.5%. The predominant mode of failure after 3 weeks was by pullout of the graft; at later stages, the mode changed either to an intraligamentous rupture or to a tibial osseous avulsion. These data are in line with those from similar biomechanical studies.7 The crosssectional area of the tendon graft significantly increased in the postoperative course, which is again in line with findings from previous studies in adult sheep.7 Using the same fixation device as in our study and comparable testing protocols, Goradia et al.8 identified Sharpey fibers in adult sheep not before 8 weeks after surgery, and the load to failure after 24 weeks was only 27.8% of that of the intact ACLs. Walsh et al.9 identified Sharpey fibers in their untreated controls not before 12 to 26 weeks, and they found a maximum load to failure of about 200 N after 26 weeks. In a recent report, Chudik et al.1 performed 3 different femoral techniques of ACL reconstruction using a skeletally immature canine model and looked for angular and rotational deformities. They concluded that an ACL reconstruction in the skeletally immature individual is complicated by the presence of active physeal and epiphyseal cartilage surrounding the growing knee. They indicated that the growth abnormalities might be related to the persistent tension of the graft (80 N at the time of fixation). In addition, the high rate of graft elongation and failure is, in part, considered to be growth related. In contrast, by use of pre-tensioning of only 20 N, the grafts in our study remained in place and appeared vital at the time of sacrifice. Although this might also be related to different growth rates of different animal models, the excessively high pre-tensioning of 80 N in the model of Chudik et al. might be one reason for the growth disturbances and graft failures. This is in line with a previous study by Edwards et al.10 Despite the difference in the timing of the development of the Sharpey-like fibers and the absence of the physis cartilage in adult sheep, the underlying mechanisms of graft incorporation appeared the same. Im-
1230
R. MELLER ET AL.
mediately after surgery, the space between the grafted tendon and the original cancellous bone was filled by a fibrovascular interface tissue.11 Within this fibrovascular interface tissue, healing occurred both by the establishment of collagen fibers and by bone ingrowth. In this study both processes were very well reflected by the spatial pattern of cells expressing collagen III mRNA. Collagen III is widely spread throughout mesodermal derivates, and it often precedes collagen I expression temporally during development; Sharpey-like fibers contain especially high amounts of collagen III.12,13 The dynamic of the healing process was further accomplished by the presence of numerous cells of the fibrovascular interface and by scattered osteoblasts and osteoclasts within the tunnel wall.11 The graft incorporation typically started at the fixation sites and proceeded toward the articular tunnel entrance, where graft motion may impair an early graft incorporation.14 In previous studies sheep models have been used to analyze tendon graft–to– bone healing after ACL reconstruction in adult animals.6,7 However, our approach was aimed at investigating the tendon graft– to– bone and physis healing in skeletally immature sheep. Histologic analyses showed a sequential healing pattern with an early development of Sharpey-like fibers and a transient physis hypertrophy followed by a strong incorporation of the ACL graft into the bone tunnel. After 6 months, biomechanical properties were comparable to the intact contralateral ACLs, with no significant difference remaining. This study may help to better understand the pathophysiology of ACL reconstruction during skeletal growth. This study is not without limitations. One such limitation is the transference of animal models to humans. Cummings and Grood,15 in their studies, could show that an ACL reconstruction in ovine and caprine models always results in inferior biomechanical results when compared with human reconstructions. This has to be kept in mind when trying to transfer animal studies to clinical practice. However, the ovine stifle joint is known to approximate the human knee very closely.16 Another drawback is the short observation time of 6 months. Weiler et al.,17 in their long-term study (2 years) of ACL reconstruction in adult sheep, showed that after 6 months, the cell count stays the same and that there are no major histologic changes to expect. Several other studies used 24 weeks and even shorter periods to study graft remodeling.15,18,19 Finally, we only used 2 animals for the histologic analysis per time point. The general principles and mechanisms of the graft remodeling
process have been extensively studied in the past. In addition, recent studies on graft remodeling also used only 2 animals for the histologic analysis and 6 for the biomechanical evaluation.19 CONCLUSIONS The early anchoring by Sharpey fibers was found at 3 weeks, with continued maturation to 24 weeks. This development of anchoring fibers corresponded with that of biomechanical strength, starting with 5% of the normal knee at 3 weeks and then 15.2% at 6 weeks, 41.2% at 12 weeks, and 69% at 24 weeks. Tendon graft–to– bone and physis healing in skeletally immature sheep is further characterized by a transient hypertrophy of the physis cartilage. The physis recovers well from the trauma of drilling and placement of a soft-tissue graft. The early development of Sharpeylike fibers results in a solid integration of the graft into bone in a timely manner. Acknowledgment: The authors gratefully acknowledge the help of Friederike Fritz, Alexandra Neddermann, Frederike Schiborra, Sabine Thoben, Sophie Müller, and Heike Ulrich. They also thank Klaus Otto and Karl Napierski for their excellent animal care.
REFERENCES 1. Chudik S, Beasley L, Potter H, Wickiewicz T, Warren R, Rodeo S. The influence of femoral technique for graft placement on anterior cruciate ligament reconstruction using a skeletally immature canine model with a rapidly growing physis. Arthroscopy 2007;23:1309-1319.e1. Available online at www. arthroscopyjournal.org. 2. Bales CP, Guettler JH, Moorman CT III. Anterior cruciate ligament injuries in children with open physes: Evolving strategies of treatment. Am J Sports Med 2004;32:1978-1985. 3. Simonian PT, Metcalf MH, Larson RV. Anterior cruciate ligament injuries in the skeletally immature patient. Am J Orthop 1999;28:624-628. 4. Nottage WM, Matsuura PA. Management of complete traumatic anterior cruciate ligament tears in the skeletally immature patient: Current concepts and review of the literature. Arthroscopy 1994;10:569-573. 5. Behr CT, Potter HG, Paletta GA Jr. The relationship of the femoral origin of the anterior cruciate ligament and the distal femoral physeal plate in the skeletally immature knee. An anatomic study. Am J Sports Med 2001;29:781-787. 6. Weiler A, Hoffmann RF, Bail HJ, Rehm O, Sudkamp NP. Tendon healing in a bone tunnel. Part II: Histologic analysis after biodegradable interference fit fixation in a model of anterior cruciate ligament reconstruction in sheep. Arthroscopy 2002;18:124-135. 7. Weiler A, Peine R, Pashmineh-Azar A, Abel C, Sudkamp NP, Hoffmann RF. Tendon healing in a bone tunnel. Part I: Biomechanical results after biodegradable interference fit fixation in a model of anterior cruciate ligament reconstruction in sheep. Arthroscopy 2002;18:113-123.
TENDON GRAFT TO BONE HEALING DURING GROWTH 8. Goradia VK, Rochat MC, Grana WA, Rohrer MD, Prasad HS. Tendon-to-bone healing of a semitendinosus tendon autograft used for ACL reconstruction in a sheep model. Am J Knee Surg 2000;13:143-151. 9. Walsh WR, Stephens P, Vizesi F, Bruce W, Huckle J, Yu Y. Effects of low-intensity pulsed ultrasound on tendon-bone healing in an intra-articular sheep knee model. Arthroscopy 2007;23:197-204. 10. Edwards TB, Greene CC, Baratta RV, Zieske A, Willis RB. The effect of placing a tensioned graft across open growth plates. A gross and histologic analysis. J Bone Joint Surg Am 2001;83:725-734. 11. Kawamura S, Ying L, Kim HJ, Dynybil C, Rodeo SA. Macrophages accumulate in the early phase of tendon-bone healing. J Orthop Res 2005;23:1425-1432. 12. Birk DE, Mayne R. Localization of collagen types I, III and V during tendon development. Changes in collagen types I and III are correlated with changes in fibril diameter. Eur J Cell Biol 1997;72:352-361. 13. Ingraham JM, Hauck RM, Ehrlich HP. Is the tendon embryogenesis process resurrected during tendon healing? Plast Reconstr Surg 2003;112:844-854.
1231
14. Rodeo SA, Kawamura S, Kim HJ, Dynybil C, Ying L. Tendon healing in a bone tunnel differs at the tunnel entrance versus the tunnel exit: An effect of graft-tunnel motion? Am J Sports Med 2006;34:1790-1800. 15. Cummings JF, Grood ES. The progression of anterior translation after anterior cruciate ligament reconstruction in a caprine model. J Orthop Res 2002;20:1003-1008. 16. Tapper JE, Fukushima S, Azuma H, et al. Dynamic in vivo kinematics of the intact ovine stifle joint. J Orthop Res 2006; 24:782-792. 17. Weiler A, Unterhauser FN, Bail HJ, Huning M, Haas NP. Alpha-smooth muscle actin is expressed by fibroblastic cells of the ovine anterior cruciate ligament and its free tendon graft during remodeling. J Orthop Res 2002;20:310-317. 18. Weiler A, Forster C, Hunt P, et al. The influence of locally applied platelet-derived growth factor-BB on free tendon graft remodeling after anterior cruciate ligament reconstruction. Am J Sports Med 2004;32:881-891. 19. Yoshikawa T, Tohyama H, Katsura T, et al. Effects of local administration of vascular endothelial growth factor on mechanical characteristics of the semitendinosus tendon graft after anterior cruciate ligament reconstruction in sheep. Am J Sports Med 2006;34:1918-1925.
Purpose: It was our aim to establish an animal model and to investigate the tendon graft–to– bone and physis healing process in skeletally immature sheep after reconstruction of the anterior cruciate ligament (ACL). Methods: Thirty-two immature sheep aged 4 months underwent a fully transphyseal ACL reconstruction by use of a soft-tissue graft. The animals were subsequently killed after 3, 6, 12, and 24 weeks and analyzed histologically and biomechanically. Results: There was a transient hypertrophy of the physis tissue at the passing site of the graft. Anchoring Sharpey-like fibers evolved as early as 3 weeks after surgery. A strong expression of collagen III messenger ribonucleic acid within the first 6 weeks preceded this anchoring process. The maximum load to failure of the tendon graft in the reconstructed knees initially decreased to 37.8 ⫾ 17.8 N after 3 weeks and was restored to 522.9 ⫾ 113 N after 24 weeks. Tendon graft stiffness was restored to 86% when compared with the control knees. Conclusions: The early anchoring by Sharpey fibers was found at 3 weeks with continued maturation to 24 weeks. This development of anchoring fibers corresponded to that of biomechanical strength, starting with 5% of the normal knee at 3 weeks and then 15.2% at 6 weeks, 41.2% at 12 weeks, and 69% at 24 weeks. Tendon graft–to– bone and physis healing in skeletally immature sheep is further characterized by a transient hypertrophy of the physis cartilage. The physis recovers well from the trauma of drilling and placement of a soft-tissue graft. The early development of Sharpey-like fibers results in a solid integration of the graft into bone in a timely manner. Clinical Relevance: ACL reconstruction in skeletally immature individuals is still controversial. This study describes in detail the histologic and biomechanical stages of tendon graft healing to the bone and physis. These data enrich the existing knowledge of previous studies in adult sheep and may provide a basis for further research in the controversial field of ACL reconstruction during growth. Key Words: Anterior cruciate ligament reconstruction—Tendon graft—Growth—Skeletally immature—In situ hybridization—Sheep model.
A
lthough there exists a broad scope of studies and experiences regarding most aspects of anterior cruciate ligament (ACL) reconstruction in adults,
From the Trauma Department (R.M., E.H., C. Haasper, C.K.) and Orthopaedic Department (E.W., B.D., C. Hurschler, F.W.), Hannover Medical School, and Small Animal Clinic, University of Veterinary Medicine Hannover (M.F.), Hannover, Germany. Supported by the Research Commission of Hannover Medical School, Hannover, Germany. The authors report no conflict of interest. Received January 27, 2008; accepted June 27, 2008. Address correspondence and reprint requests to Rupert Meller, M.D., Trauma Department, Hannover Medical School (MHH), Carl-Neuberg-Strasse 1, 30625 Hannover, Germany. E-mail: meller. [email protected] © 2008 by the Arthroscopy Association of North America 0749-8063/08/2411-0848$34.00/0 doi:10.1016/j.arthro.2008.06.021
only limited data are available regarding the tendon graft–to– bone healing process in skeletally immature subjects, although intraligamentous ruptures of the ACL are diagnosed in this group with increasing frequency.1-5 To gain more insight into the healing process after ACL reconstruction in skeletally immature subjects, a corresponding sheep model was established. We hypothesized that the graft incorporation into the bone tunnel and physis tissue follows distinct mechanisms and that the incorporation is reflected by a rapid gain in biomechanical strength. The purpose of the study was to investigate the basic mechanisms of the graft tissue incorporation into the bone tunnel and the physis tissue, as well as to determine the biomechanical behavior of the graft during the first 6 months after surgery.
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 24, No 11 (November), 2008: pp 1221-1231
1221
1222
R. MELLER ET AL.
FIGURE 1. (A) Schema of a right sheep knee showing physis (white triangle) and position and orientation of graft within tibial and femoral tunnels. (B) A general view of the explanted and in plastic-embedded implantation site (black rectangle in A) shows the femoral tunnel (the walls of which are marked by black dots) crossing the physis (white triangles). The tunnel is filled with the well-incorporated ligamentous graft (bright white fibers). The articular tunnel entrance is marked by a black arrow (1, fixation site; 2, tunnel midportion [physis]; 3, articular tunnel entrance; bar, 4,500 m.).
METHODS Study Design All procedures were performed with permission of all responsible authorities and in accordance with the National Institutes of Health guidelines for the use of laboratory animals. Thirty-two black-headed sheep aged 4 months were acquired from the Niedersächsischer Schafzüchterverband (Hannover, Germany). Before surgery, a veterinarian confirmed the health status of the sheep. All animals underwent a fully transphyseal reconstruction of the right ACL. The left knees served as controls. Four groups of eight animals each were killed after 3, 6, 12, and 24 weeks. Two animals from each group were used for the histologic analysis and six for the biomechanical analysis. Operative Technique The right knee joint was exposed through an anteromedial incision with release of the medial parapatellar retinaculum. The patella was displaced laterally, the anterior fat pad was sharply separated, and the ACL was exposed and removed. A split of the ipsilateral superficial flexor digitorum tendon and the
gastrocnemius tendon was harvested and used as graft. The tibial and femoral tunnels were created by use of a fully transphyseal technique (Fig 1A). The tibial tunnel placement was centered to the attachment area of the anteromedial bundle of the ACL. The starting point was 15 mm distal and 10 mm medial to the most prominent point of the tuberosity. By use of a K-wire drill guide adjusted at 40°, a 2.4-mm K-wire was first placed and then overdrilled by a 4.5-mm EndoButton hand drill (Smith & Nephew, Andover, MA) to avoid thermal damage to the physis. The femoral tunnel entrance was located in the origin area of the ACL. The K-wire was aimed toward the lateral aspect of the distal third of the thigh, ensuring extra-articular positioning of the fixation device. This desired position was best achieved in 30° of knee flexion. The graft was fixed by use of the EndoButton proximally and a suture washer distally (Smith & Nephew Endoscopy). Gross Inspection and Tissue Sampling Knee joints used for histologic analysis were exposed by an anteromedial arthrotomy. The graft sites were scrutinized for their integrity and then sharply dissected. Knee joints dedicated for biomechanical
TENDON GRAFT TO BONE HEALING DURING GROWTH analysis were disarticulated in the hip joint and frozen in full extension at ⫺20°C. Twelve hours before biomechanical testing, samples were thawed to room temperature. The capsule was sharply dissected, and the graft was checked for its integrity. General Histology After 3, 6, 12, and 24 weeks, the dissected femoral transplantation sites were fixed in 3.7% commercial formalin for 7 days at 4°C. After embedding and polymerization in methyl methacrylate (Technovit 9100 New; Heraeus-Kulzer, Hanau, Germany) according to established protocols (Fig 1B), the tissue blocks were first trimmed by use of an Exakt 310 Diamond Band Saw (EXAKT, Norderstedt, Germany). By use of an RM 2155 microtome (Leica, Bensheim, Germany), 5-m thin sections were cut and placed onto poly-L-lysine– coated glass slides. Before the different staining procedures, the sections were first dried for 2 days at 37°C and then deacrylated in xylol (15 minutes) and 2-methoxyethyl acetate (10 minutes), cleared through a decreasing ethanol series (isopropyl alcohol, 96% ethanol, and 70% ethanol, for 2 minutes each), and rehydrated by use of distilled water. Toluidine Blue Staining For toluidine blue staining, sections were incubated in 0.1% toluidine blue O (Sigma, Taufkirchen, Germany) for 20 seconds, washed in distilled water, dehydrated in ethanol, and mounted in Eukitt (Labonord, Mönchengladbach, Germany). In Situ Hybridization for Collagen III Messenger RNA For messenger RNA (mRNA) in situ hybridization, ZytoFast RNA in situ Hybridisierungs System Biotin (ZytoVision, Bremerhaven, Germany) was used. The sections were first decalcified for 60 minutes in 10% ethylenediaminetetraacetic acid and were then washed and gently digested for 15 minutes with pepsin. Digestion was stopped by washing for 10 seconds with ethanol. Slides were air dried, and the sections were hybridized with the collagen III–specific biotin-labeled ribonucleic acid probe overnight at 55.5°C. Slides were washed with buffer, and a biotin-specific antibody conjugated with alkaline phosphatase was applied overnight at 4°C. The slides were washed again with buffer, and the probes were visualized by use of nitro blue tetrazolium chloride/5-bromo-4chloro-3-indolyl phosphate (NBT/BCIP) under micro-
1223
scopic control. To assess the specificity of our protocol, we carried out both a positive control consisting of a poly-dT oligonucleotide detecting the poly-A tails of mRNA, as well as the integrity of the mRNA,andanegativecontrolconsistingofa“randomsequence” oligonucleotide to detect nonspecific binding. In addition, controls to exclude nonspecific binding of the antibody as well as controls to exclude nonspecific chromogenic reaction were carried out. Finally, the sections were washed and mounted with Kaiser’s glycerol gelatin (Merck, Darmstadt, Germany). The probe sequence was 5=-biotinTggCTACTTCTCgCTCTgCTTCATCCCACT-3=, and it was synthesized by TIB MOLBIOL (Berlin, Germany). Microscopy and Photography Photomicrographs were taken with a Zeiss Axioskop 40 microscope in combination with a Zeiss AxioCam MRc digital camera and Zeiss AxioVision software (Zeiss, Oberkochen, Germany). Specimen Preparation and Cross-Sectional Area for Biomechanical Evaluation The knee joints were dissected of all remaining soft tissue. The femur and tibia were cut 10 cm above and below the joint line, leaving a femur-ACL-tibia complex (FATC) of the control left knee and a femurgraft-tibia complex (FGTC) of the operated right knee for uniaxial tensile testing. The bone ends were then embedded in metal cylinders by use of polyurethane casting resin and mounted to a custom-made adjusting device. The cross-sectional area of each ACL (intact left knee) and tendon graft (right knee) was measured 3 times with a laser micrometer (Takikawa Engineering, Tokyo, Japan). The means of 3 measurements of each knee were used for statistical analysis. The crosssectional area was used to determine the strain and Young’s modulus. Failure Testing For failure testing, a materials testing machine (Zwick 1445; Zwick, Ulm, Germany) was used. The FATC and FGTC, mounted in the testing fixture, were carefully aligned parallel to the axis of the applied load in 60° of knee flexion. For the preconditioning of the tendon graft, a preload of 5 N was applied with a load displacement rate of 6 mm/min, and 20 loading cycles were performed (12 mm/min; with upper and lower limits of 5 N and 0 N, respectively). The surface of the ACLs and tendon grafts were marked with
1224
R. MELLER ET AL.
high-contrast grease that allowed the determination of the surface strain. This marker motion was recorded by a digital video camera (Sony DCR-HC17E; Sony, Tokyo, Japan). Knees were then loaded until failure with a displacement rate of 6 mm/min. Maximum load to failure and stiffness were calculated from the load displacement curve. The Young’s modulus was calculated from the stress-strain curve. Failure modes (femoral or tibial graft pullout, graft rupture at the femoral or tibial insertion site, midsubstance failure, or bony avulsion) were recorded. Data Analysis All data were analyzed with SPSS software (version 14.0; SPSS, Chicago, IL). The biomechanical data were analyzed with a 2-way analysis of variance. Independent variables were the time of sacrifice and the left or right knee. Paired t tests were used to look for side-to-side differences. The differences were considered to be significant at a probability level of P ⱕ .0125 after Bonferroni correction for multiple comparisons.
RESULTS Postoperative Course Two animals died of pneumonia postoperatively (3and 6-week groups) and were excluded from the study, leaving 30 animals for the final evaluation. All other animals were healthy and showed a quick return to full mobilization. The macroscopic appearance of the hind limbs did not show obvious angular or rotational deformities. All grafts were in place. Intratunnel Graft Remodeling Similar to Intra-Articular Portion Remodeling Three weeks after surgery, the tendon graft showed acellular areas. However, other regions of the graft already showed a significant cell repopulation. After 6 weeks, the graft was highly vascularized and many fibroblasts were present between the collagen fibers. During the next 6 weeks, the collagen fibers became more organized and the vascularization and cell density decreased. After 6 months, the collagen fibers were densely packed and aligned in parallel with fibroblasts and capillaries in between. The histologic structure of the tendon graft appeared close to that of the contralateral control knees.
Tunnel Wall Remodeling Starts at Fixation Sites and Proceeds Toward Joint Because of the drilling procedure, the tunnel walls showed an initial destruction of the cancellous bone. After 3 weeks, the trabeculae appeared rugged. Numerous gaps allowed a direct contact between the graft and the bone marrow (Fig 2A). After 6 weeks, new trabeculae were formed and the gaps in the tunnel walls were mostly closed by newly formed woven bone. A corresponding osteoid formation was also present (Fig 2B). During the next 6 months, the remodeling process further proceeded and the thickness of the wall increased significantly (Figs 2C, 2D). Early Development of Sharpey-Like Fibers In the early stages of the graft healing process, only scattered fibers passing from the fibrovascular interface into the tunnel walls could be identified. These fibers were predominantly seen adjacent to the fixation sites (Fig 3A and B). After 6 weeks, these Sharpeylike fibers increased noticeably in density and were then also present in the tunnel midportion and at the articular tunnel entrance. The fibers were aligned oblique to the supposed tensile stresses within the tendon graft and seemed to anchor the fibrovascular interface tissue into the bone (Fig 3C and D). After 12 weeks (Fig 3E and F) and 24 weeks (Fig 3G and H), the number and size of these Sharpey-like fibers further increased and they also reached much deeper into the surrounding bone. This architecture resembled an indirect insertion anatomy. Transient Hypertrophy of Physis Cartilage Three weeks after surgery, we found no signs of necrosis or other damage to the physis that might be correlated with tunnel preparation. After 3 weeks, the physis cartilage expanded considerably into the fibrovascular interface tissue but without its typical architecture. The chondrocytes were smaller in size, and their nuclei stained stronger. A distinct layer of smallsized cells was present between the graft and the physis (Fig 4A). During the next 3 weeks, this hypertrophic physis tissue showed an orientation toward the tensile forces along the graft (Fig 4B). After 12 weeks, the physis tissue within the tunnel retracted (Fig 4C), and after 24 weeks, it was in line with the remodeled tunnel wall (Fig 4D). At that point, the chondrocytes again showed the typical architecture and were arranged in columns.
TENDON GRAFT TO BONE HEALING DURING GROWTH
1225
FIGURE 2. Remodeling of tunnel wall: Toluidine blue staining of tunnel wall after 3 weeks (A), 6 weeks (B), 12 weeks (C), and 24 weeks (D). (A) After 3 weeks, the tunnel wall still shows the destruction of the cancellous bone (stars). Numerous gaps (arrows) perforate the wall and allow a direct contact between the graft and the bone marrow. (B) After 6 weeks, nearly all of the gaps are closed and the surface of the wall has become smoother. Only small bulges (arrows) are still present. After 12 weeks (C) and 24 weeks (D), the thickness of the wall had significantly increased.
Early Expression of Collagen III mRNA To investigate the formation of anchoring fibers, in situ hybridization for collagen III, a major component of Sharpey-like fibers, was used. After 3 weeks (Fig 5A and C) and 6 weeks (Fig 5B), positive cells were predominantly found in the fibrovascular interface tissue in close vicinity to the bony parts of the tunnel wall, especially in the elongation of the initially destroyed trabeculae (Fig 5C). Only a few positive cells were scattered in the graft and the bone. After 12 weeks (Fig 5D) and 24 weeks (Fig 5E), positive cells were more likely allocated within the tendon graft and the bone. The number of collagen III–positive cells and their staining intensity significantly decreased from 12 to 24 weeks. Cross-Sectional Area The mean cross-sectional area of the intact ACLs was 23.39 ⫾ 3.5 mm2. There was a significant in-
crease in the cross-sectional area of the tendon grafts throughout the remodeling process, from 25.1 ⫾ 7.3 mm2 after 3 weeks to 51.1 ⫾ 14.9 mm2 after 24 weeks (P ⬍ .005). Failure Testing The 2-way analysis of variance showed a significant effect of time and side on the maximum load to failure and the stiffness. In ACL-reconstructed knees the load to failure raised steadily, indicating an improvement of the function of the tendon graft (Fig 6A). The intact contralateral knees (FATC) had a mean load to failure of 759.2 ⫾ 114.1 N. The operated knees (FGTC), after 3 weeks, reached only 37.8 ⫾ 17.8 N (5% of intact FATC, P ⬍ .0005). There was a nonsignificant increase in the load to failure of the FGTC during the next 3 weeks (115.9 ⫾ 59.6 N, 15.2% of intact ACLs; P ⫽ .02). After 12
1226
R. MELLER ET AL.
FIGURE 3. Remodeling of tendon-to-bone interface and increasing anchoring of graft by Sharpey-like fibers: Toluidine blue staining (A, C, E, G) and corresponding polarization optics (B, D, F, H) after 3 weeks (A, B), 6 weeks (C, D), 12 weeks (E, F), and 24 weeks (G, H). (A, B) Three weeks after transplantation, only very few fibers (arrows) passing from the tendon to bone are present. (C, D) After 6 weeks, the number of these fibers had significantly increased and they reach deeper into the bone. After 12 weeks (E, F) and 24 weeks (G, H), the number and size of these fibers further increased. Stars indicate the tunnel wall.
TENDON GRAFT TO BONE HEALING DURING GROWTH
1227
FIGURE 4. Development of tendon-to-physis interface: Toluidine blue staining of physis after 3 weeks (A), 6 weeks (B), 12 weeks (C), and 24 weeks (D). (A) After 3 weeks, the physis (diamonds) had protruded into the graft, indicating that the physis was not damaged during the operation. (B) After 6 weeks, the physis had retracted, showing a clear orientation toward the tensile forces. After 12 weeks (C) and 24 weeks (D), the physis fully retracted to be in line with the wall of the tunnel. Stars indicate the tunnel wall.
weeks, the FGTC failed at 316.4 ⫾ 134.9 N (41.7% of intact ACLs, P ⬍ .001). Finally, after 24 weeks, the FGTC gained further stability, resulting in a load to failure of 522.9 ⫾ 113.9 N (69% of intact ACLs). No statistically significant side-to-side difference could be observed at this time point (P ⫽ .05). The corresponding stiffness value was 136.3 ⫾ 28.5 N/mm for the intact contralateral knees (FATC). After 3 weeks, the operated knees (FGTC) reached 19.6 ⫾ 8.2 N/mm and then steadily increased to 56.3 ⫾ 22.5 N/mm after 6 weeks and 84.6 ⫾ 21.7 N/mm after 12 weeks (P ⬍ .001 for 3, 6, and 12 weeks). After 24 weeks, the stiffness reached 113.7 ⫾ 5.7 N/mm (P ⫽ .05), implying a restoration to 86.5% when compared with intact knees (Fig 6B). No statistically significant sideto-side difference could be observed at this time point. The Young’s modulus of the FATC was 217.3 ⫾ 57.9 N/mm2. The Young’s modulus of the FGTC decreased to 35.5 ⫾ 17.6 N/mm2 after 3 weeks and
increased to 74.6 ⫾ 37.6 N/mm2 after 24 weeks. This was significantly different from the control at 24 weeks (P ⬍ .0005). Different Modes of Failure The modes of failure for the intact knees were either fracture of the distal femoral physis (11 knees) or a tibial bony avulsion of the ACLs (11 knees) (the posterolateral bundle constantly failed first). After 3 weeks, the mode of failure of the FGTC was predominantly by graft pullout from the femoral tunnel. One graft after 3 weeks failed at its midportion. Similar modes of failure could be found after 6 weeks, with 3 femoral pullouts and 2 midportion ruptures. After 12 and 24 weeks, grafts failed by midportion ruptures in 2 knees and by tibial bony avulsions in 4 knees, respectively.
1228
R. MELLER ET AL.
FIGURE 5. Spatiotemporal expression of collagen III mRNA after 3 weeks (A, E), 6 weeks (B), 12 weeks (C), and 24 weeks (D). (A) After 3 weeks, collagen III– expressing cells are predominantly located in the tendon-to-bone interface, especially in the elongation of the ruptured trabeculae (black arrows in A, E), which is indicative of the formation of Sharpey-like fibers and the remodeling of the bony tunnel wall (stars). After 6 weeks (B) and 12 weeks (C), this process proceeds further on. After 24 weeks (D) and the completion of the tunnel remodeling, only low collagen III mRNA expression can be detected.
DISCUSSION To gain more insight into the healing processes after ACL reconstruction during growth, we used skeletally
immature sheep as a large animal model. A nonanatomic fixation far away from the joint line was used to protect the physis tissue. The tendon graft–to– bone
TENDON GRAFT TO BONE HEALING DURING GROWTH
FIGURE 6. (A) Biomechanical evaluation of maximum load to failure of intact knee and ACL-reconstructed knee. The horizontal line shows the mean value for the load-to-failure data of the intact contralateral knees, and the dotted lines show the corresponding SD. Asterisks indicate statistically significant side-to-side differences at 3, 6, and 12 weeks. (B) Biomechanical evaluation of stiffness (in Newtons per millimeter) at failure testing of the native, ACL-intact knee and the ACL-reconstructed knee. Asterisks indicate statistically significant side-to-side differences at 3, 6, and 12 weeks (P ⬍ .0125). The means of the measured values are connected by lines to visualize their change over time in the different time groups.
healing process in these cases progresses via a fibrous interzone, resulting in an indirect insertion anatomy.6 We investigated the spatiotemporal patterns of tendon graft–to– bone and physis healing as well as the recovery of the biomechanical properties. We hypothesized that the graft incorporation into the bone tunnel and physis tissue follows distinct mechanisms and that the incorporation is reflected by a rapid gain in biomechanical strength. The hypothesis was affirmed, because we have seen an early development of Sharpey-like fibers as soon as 3 weeks after surgery. This strong incorporation re-
1229
sulted in equivalent biomechanical properties. We found a quick remodeling of the tendon graft within the tunnel and an early formation of Sharpey-like fibers as soon as 3 weeks after surgery. We also found a complex response of the physis cartilage at the passing site of the tendon graft. The physis recovered well from the trauma of drilling and placement of a soft-tissue graft. The biomechanical data fully correspond with the morphologic data and showed an increasing recovery of the failure load to almost 70% of the intact knee and a recovery of the stiffness to 86.5%. The predominant mode of failure after 3 weeks was by pullout of the graft; at later stages, the mode changed either to an intraligamentous rupture or to a tibial osseous avulsion. These data are in line with those from similar biomechanical studies.7 The crosssectional area of the tendon graft significantly increased in the postoperative course, which is again in line with findings from previous studies in adult sheep.7 Using the same fixation device as in our study and comparable testing protocols, Goradia et al.8 identified Sharpey fibers in adult sheep not before 8 weeks after surgery, and the load to failure after 24 weeks was only 27.8% of that of the intact ACLs. Walsh et al.9 identified Sharpey fibers in their untreated controls not before 12 to 26 weeks, and they found a maximum load to failure of about 200 N after 26 weeks. In a recent report, Chudik et al.1 performed 3 different femoral techniques of ACL reconstruction using a skeletally immature canine model and looked for angular and rotational deformities. They concluded that an ACL reconstruction in the skeletally immature individual is complicated by the presence of active physeal and epiphyseal cartilage surrounding the growing knee. They indicated that the growth abnormalities might be related to the persistent tension of the graft (80 N at the time of fixation). In addition, the high rate of graft elongation and failure is, in part, considered to be growth related. In contrast, by use of pre-tensioning of only 20 N, the grafts in our study remained in place and appeared vital at the time of sacrifice. Although this might also be related to different growth rates of different animal models, the excessively high pre-tensioning of 80 N in the model of Chudik et al. might be one reason for the growth disturbances and graft failures. This is in line with a previous study by Edwards et al.10 Despite the difference in the timing of the development of the Sharpey-like fibers and the absence of the physis cartilage in adult sheep, the underlying mechanisms of graft incorporation appeared the same. Im-
1230
R. MELLER ET AL.
mediately after surgery, the space between the grafted tendon and the original cancellous bone was filled by a fibrovascular interface tissue.11 Within this fibrovascular interface tissue, healing occurred both by the establishment of collagen fibers and by bone ingrowth. In this study both processes were very well reflected by the spatial pattern of cells expressing collagen III mRNA. Collagen III is widely spread throughout mesodermal derivates, and it often precedes collagen I expression temporally during development; Sharpey-like fibers contain especially high amounts of collagen III.12,13 The dynamic of the healing process was further accomplished by the presence of numerous cells of the fibrovascular interface and by scattered osteoblasts and osteoclasts within the tunnel wall.11 The graft incorporation typically started at the fixation sites and proceeded toward the articular tunnel entrance, where graft motion may impair an early graft incorporation.14 In previous studies sheep models have been used to analyze tendon graft–to– bone healing after ACL reconstruction in adult animals.6,7 However, our approach was aimed at investigating the tendon graft– to– bone and physis healing in skeletally immature sheep. Histologic analyses showed a sequential healing pattern with an early development of Sharpey-like fibers and a transient physis hypertrophy followed by a strong incorporation of the ACL graft into the bone tunnel. After 6 months, biomechanical properties were comparable to the intact contralateral ACLs, with no significant difference remaining. This study may help to better understand the pathophysiology of ACL reconstruction during skeletal growth. This study is not without limitations. One such limitation is the transference of animal models to humans. Cummings and Grood,15 in their studies, could show that an ACL reconstruction in ovine and caprine models always results in inferior biomechanical results when compared with human reconstructions. This has to be kept in mind when trying to transfer animal studies to clinical practice. However, the ovine stifle joint is known to approximate the human knee very closely.16 Another drawback is the short observation time of 6 months. Weiler et al.,17 in their long-term study (2 years) of ACL reconstruction in adult sheep, showed that after 6 months, the cell count stays the same and that there are no major histologic changes to expect. Several other studies used 24 weeks and even shorter periods to study graft remodeling.15,18,19 Finally, we only used 2 animals for the histologic analysis per time point. The general principles and mechanisms of the graft remodeling
process have been extensively studied in the past. In addition, recent studies on graft remodeling also used only 2 animals for the histologic analysis and 6 for the biomechanical evaluation.19 CONCLUSIONS The early anchoring by Sharpey fibers was found at 3 weeks, with continued maturation to 24 weeks. This development of anchoring fibers corresponded with that of biomechanical strength, starting with 5% of the normal knee at 3 weeks and then 15.2% at 6 weeks, 41.2% at 12 weeks, and 69% at 24 weeks. Tendon graft–to– bone and physis healing in skeletally immature sheep is further characterized by a transient hypertrophy of the physis cartilage. The physis recovers well from the trauma of drilling and placement of a soft-tissue graft. The early development of Sharpeylike fibers results in a solid integration of the graft into bone in a timely manner. Acknowledgment: The authors gratefully acknowledge the help of Friederike Fritz, Alexandra Neddermann, Frederike Schiborra, Sabine Thoben, Sophie Müller, and Heike Ulrich. They also thank Klaus Otto and Karl Napierski for their excellent animal care.
REFERENCES 1. Chudik S, Beasley L, Potter H, Wickiewicz T, Warren R, Rodeo S. The influence of femoral technique for graft placement on anterior cruciate ligament reconstruction using a skeletally immature canine model with a rapidly growing physis. Arthroscopy 2007;23:1309-1319.e1. Available online at www. arthroscopyjournal.org. 2. Bales CP, Guettler JH, Moorman CT III. Anterior cruciate ligament injuries in children with open physes: Evolving strategies of treatment. Am J Sports Med 2004;32:1978-1985. 3. Simonian PT, Metcalf MH, Larson RV. Anterior cruciate ligament injuries in the skeletally immature patient. Am J Orthop 1999;28:624-628. 4. Nottage WM, Matsuura PA. Management of complete traumatic anterior cruciate ligament tears in the skeletally immature patient: Current concepts and review of the literature. Arthroscopy 1994;10:569-573. 5. Behr CT, Potter HG, Paletta GA Jr. The relationship of the femoral origin of the anterior cruciate ligament and the distal femoral physeal plate in the skeletally immature knee. An anatomic study. Am J Sports Med 2001;29:781-787. 6. Weiler A, Hoffmann RF, Bail HJ, Rehm O, Sudkamp NP. Tendon healing in a bone tunnel. Part II: Histologic analysis after biodegradable interference fit fixation in a model of anterior cruciate ligament reconstruction in sheep. Arthroscopy 2002;18:124-135. 7. Weiler A, Peine R, Pashmineh-Azar A, Abel C, Sudkamp NP, Hoffmann RF. Tendon healing in a bone tunnel. Part I: Biomechanical results after biodegradable interference fit fixation in a model of anterior cruciate ligament reconstruction in sheep. Arthroscopy 2002;18:113-123.
TENDON GRAFT TO BONE HEALING DURING GROWTH 8. Goradia VK, Rochat MC, Grana WA, Rohrer MD, Prasad HS. Tendon-to-bone healing of a semitendinosus tendon autograft used for ACL reconstruction in a sheep model. Am J Knee Surg 2000;13:143-151. 9. Walsh WR, Stephens P, Vizesi F, Bruce W, Huckle J, Yu Y. Effects of low-intensity pulsed ultrasound on tendon-bone healing in an intra-articular sheep knee model. Arthroscopy 2007;23:197-204. 10. Edwards TB, Greene CC, Baratta RV, Zieske A, Willis RB. The effect of placing a tensioned graft across open growth plates. A gross and histologic analysis. J Bone Joint Surg Am 2001;83:725-734. 11. Kawamura S, Ying L, Kim HJ, Dynybil C, Rodeo SA. Macrophages accumulate in the early phase of tendon-bone healing. J Orthop Res 2005;23:1425-1432. 12. Birk DE, Mayne R. Localization of collagen types I, III and V during tendon development. Changes in collagen types I and III are correlated with changes in fibril diameter. Eur J Cell Biol 1997;72:352-361. 13. Ingraham JM, Hauck RM, Ehrlich HP. Is the tendon embryogenesis process resurrected during tendon healing? Plast Reconstr Surg 2003;112:844-854.
1231
14. Rodeo SA, Kawamura S, Kim HJ, Dynybil C, Ying L. Tendon healing in a bone tunnel differs at the tunnel entrance versus the tunnel exit: An effect of graft-tunnel motion? Am J Sports Med 2006;34:1790-1800. 15. Cummings JF, Grood ES. The progression of anterior translation after anterior cruciate ligament reconstruction in a caprine model. J Orthop Res 2002;20:1003-1008. 16. Tapper JE, Fukushima S, Azuma H, et al. Dynamic in vivo kinematics of the intact ovine stifle joint. J Orthop Res 2006; 24:782-792. 17. Weiler A, Unterhauser FN, Bail HJ, Huning M, Haas NP. Alpha-smooth muscle actin is expressed by fibroblastic cells of the ovine anterior cruciate ligament and its free tendon graft during remodeling. J Orthop Res 2002;20:310-317. 18. Weiler A, Forster C, Hunt P, et al. The influence of locally applied platelet-derived growth factor-BB on free tendon graft remodeling after anterior cruciate ligament reconstruction. Am J Sports Med 2004;32:881-891. 19. Yoshikawa T, Tohyama H, Katsura T, et al. Effects of local administration of vascular endothelial growth factor on mechanical characteristics of the semitendinosus tendon graft after anterior cruciate ligament reconstruction in sheep. Am J Sports Med 2006;34:1918-1925.