Effects of ACL Graft Placement on In Vivo Knee Function and Cartilage Thickness Distributions

Transcription

1 Effects of ACL Graft Placement on In Vivo Knee Function and Cartilage Thickness Distributions Louis E. DeFrate Departments of Orthopaedic Surgery, Mechanical Engineering and Materials Science and Biomedical Engineering, Duke University, Durham, North Carolina Received 15 December 2016; accepted 23 January 2017 Published online 24 March 2017 in Wiley Online Library (wileyonlinelibrary.com). DOI /jor ABSTRACT: Injuries to the anterior cruciate ligament (ACL) frequently lead to early-onset osteoarthritis. Despite advancement in surgical techniques, ACL reconstruction has a limited ability to prevent these degenerative changes. While previous studies have investigated knee function after ACL reconstruction, in vivo investigations of the effects of graft placement on in vivo joint function and cartilage health are limited. This review presents a series of studies that used novel imaging and 3D modeling techniques to determine the in vivo placement of the ACL graft on the femur using two different ACL reconstruction techniques. These techniques resulted in two distinct graft placement groups: one where the ACL was placed anatomically near the center of the native ACL footprint and another where the graft was placed anteroproximally on the femur, centered outside the ACL footprint. We quantified the effects of graft placement on graft deformation during in vivo loading and how these variables affected knee motion. Finally, we quantified whether femoral placement of the graft affected cartilage thickness. Our results demonstrate that achieving anatomic graft placement on the femur is critical to restoring native ACL function and normal knee kinematics. Knees with grafts that more closely restored normal ACL function, and thus knee motion, experienced less focal cartilage thinning than did those that experienced abnormal knee motion. These results suggest that achieving anatomic graft placement is a critical factor in restoring normal knee motion and potentially slowing the development of degenerative changes after ACL reconstruction. ß 2017 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 35: , Keywords: deformation; cartilage; kinematics; MR imaging; osteoarthritis The anterior cruciate ligament (ACL) is one of the most commonly injured knee ligaments, with more than 400,000 injuries occurring annually in the United States. 1 ACL injuries are associated with pain, altered knee motion, meniscus injury, and early-onset osteoarthritis (OA). 2 9 While ACL reconstruction has generally been successful in improving patient-reported outcomes and in returning athletes to sports, long-term studies have reported a high incidence of joint degeneration after ACL reconstruction. 3,13 18 For example, previous work has indicated that 18% of patients had radiographic evidence of OA after only 5 years. 19 Others have reported OA in more than 50% of patients years after surgery. 14,17 Despite advances in reconstructive techniques, early-onset OA after ACL reconstruction remains a problem. 3,13 18,20 24 Previous work suggests that altered knee motion plays an important role in early-onset OA after ACL injuries Several recent studies have indicated that ACL deficiency increases anterior translation, 26,30,31 medial translation, 26,30 and internal rotation 2,4,26,32 of the tibia relative to the femur under various in vivo loading conditions. Similarly, recent studies have hypothesized that reconstruction s This work was recognized with the 2016 Kappa Delta Young Investigator Award at the annual meetings of the American Academy of Orthopaedic Surgeons and the Orthopaedic Research Society (March 2016). Grant sponsor: National Football League Charities; Grant sponsor: Arthrex; Grant sponsor: National Institutes of Health; Grant numbers: AR055659, AR065527, AR Correspondence to: Louis E. DeFrate (T: ; F: ; lou.defrate@duke.edu) # 2017 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. inability to restore normal joint motion contributes to joint degeneration after surgery. 29,33 35 A factor that potentially influences the ability of reconstruction to restore normal knee biomechanics is placement of the ACL graft In particular, previous work suggests that non-anatomic placement of the graft on the femur may be a common problem in ACL reconstruction Cadaveric studies suggest that graft placement closer to the anatomical footprint more closely reproduces native knee motion However, there is little data quantifying the influence of graft placement on in vivo knee function in response to physiological loading conditions. The aim of this review is to present our results that directly address this lack of data. Our aims were: to determine the in vivo placement of the ACL graft on the femur, to determine how graft placement affected graft deformation and knee kinematics during in vivo weight bearing flexion, and to determine whether femoral graft placement affected cartilage thickness. Our results demonstrated that achieving anatomic graft placement on the femur is critical to restoring native ACL function and normal knee kinematics. Because patients in the anatomic placement group had less evidence of cartilage degeneration compared to the non-anatomic group, these results suggest that it may be possible to slow the progression of cartilage degeneration following ACL reconstruction. In Vivo Measurement of ACL Graft Placement on the Femur 49 A 2009 study estimated that 90% of ACL reconstructions performed in the U.S. are performed transtibially, where the femoral tunnel is placed through the tibial tunnel. 50 Moreover, controversy remains as 1160

2 IN VIVO ACL GRAFT PLACEMENT 1161 to the appropriate surgical technique to use in ACL reconstruction. 51 Recent studies suggest that some transtibial techniques have a limited ability to place the ACL graft anatomically on the femur due to constraints imposed by the tibial tunnel. 45,49,52 A cadaveric study from our laboratory indicated that the transtibial technique can result in grafts centered near the anteroproximal border of the native femoral ACL attachment site, while grafts placed independently of the tibial tunnel (Fig. 1) resulted in more anatomic placement. 45 Despite recent interest in achieving anatomic graft placement, there is limited in vivo data on graft placement relative to the native ACL footprint. 49,53 We recently developed a method to quantify graft placement in vivo using advanced MR imaging and 3D modeling techniques 49 so that graft placement could be determined relative to the ACL anatomy of the contralateral knee. Using this technique, we compared the femoral graft placement of a single-incision transtibial technique with that of a two-incision tibial tunnel-independent technique in patients. We hypothesized that the independent technique would allow for more anatomic placement of the femoral tunnel compared to the transtibial technique, as graft placement using the independent technique is not constrained by the position and orientation of the tibial tunnel. Eight patients who had undergone unilateral transtibial ACL reconstruction (five men, three women; mean age: 35 years; range: years) and eight patients with a tibial tunnel-independent reconstruction (five men, three women; mean age: 31 years; range: years) were recruited for this study. All subjects were between 6 and 36 months after surgery, had a healthy contralateral knee, and were recruited sequentially from the clinical practice of two experienced orthopaedic surgeons at our institution. All Figure 1. A two-incision technique was used to place the femoral guide pin transfemorally from the outside-in, independently of the tibial tunnel. In this tibial tunnel-independent technique, the guide was placed at the center of the anterior cruciate ligament (ACL), as judged visually by the surgeon. Adapted from Kaseta et al. 45 patients were doing well clinically, and had returned to sports without restriction. In the transtibial group, 49 the tibial tunnel was placed using a Concept Precision guide pin (ConMed Linvatec; Largo, FL). The guide was placed at approximately 57 and 65 in the sagittal and coronal planes, respectively. 49,54,55 The tibial tunnel passed through the anterior fibers of the MCL and was centered approximately 7 mm anterior to the PCL. A reamer equal in size to the graft diameter was used for each tibial tunnel. The tibial tunnel location was aimed to allow placement of a 7 mm offset guide at approximately the 1:30 position or the 10:30 position. A cannulated reamer was then passed through the tibial tunnel and over the guide pin to create the femoral socket. Notchplasty, when performed, was limited to the anterior portion of the notch and did not alter the femoral attachment site. In the grafts placed using the independent technique, the location and shape of the ACL footprint was identified arthroscopically through the anteromedial and anterolateral portals. A guide pin was placed through the center of the tibial footprint of the ACL and a graft-size-appropriate cannulated reamer was used to create the tibial tunnel. Using the anteromedial portal, the femoral tunnel was placed by positioning the Retro-Drill guide (Arthrex; Naples, FL) near the center of the ACL footprint, as estimated visually by the surgeon. 45,56 A guide-pin was placed from outside the joint through a small incision over the lateral femoral cortex just proximal to the lateral femoral condyle and anterior to the intermuscular septum. The guide pin was drilled through the femur to the tip of the aiming guide. 45,49 The pin was threaded to allow the placement of a cutter of the appropriate size on the guide pin as it entered the joint through the femoral ACL footprint. The cutter was used to create a socket into the femur to the desired depth. No notchplasty was necessary. At follow-up, each subject underwent 3T MR imaging (Siemens, Trio Tim; Malvern, PA) of both the injured and contralateral native knees (Fig. 2). Sagittal, coronal, and axial plane images were acquired using a double-echo steady state sequence (DESS) with a field of view of cm, a matrix of pixels, and slice thickness of 1 mm (Flip angle: 25, TR: 17 ms, TE: 6 ms). The contours of each subject s femurs, normal ACL attachment sites, and apertures of the graft tunnels were traced within each MR image. These outlines were used to generate 3D models of both knees using solid modeling software. 49 The position and shape of the ACL and tunnels were cross-referenced and confirmed using the axial, sagittal, and coronal image sets to generate the 3D model of the joint. Next, all right knee models were mirrored and converted to left knees for analysis. An iterative closest point technique 49,57 was used to align the two models to allow for the comparison of the tunnel placement relative to the native ACL. Previous

3 1162 DEFRATE Figure 2. High resolution MR images (left) of both the reconstructed and contralateral native knees were used to generate 3D models of both femurs for each patient. The two models were aligned using an iterative closest point technique, allowing for the comparison of the graft tunnel placement relative to the location of the contralateral native ACL attachment. As shown in two patients, the independent technique (middle) resulted in placement of the graft closer to the center of the ACL attachment site compared to the transtibial technique (right), which resulted in graft placement centered near the border of the native ACL. Reprinted with permission. 49 validation indicates that this technique has an accuracy of 0.3 mm in measuring the centers of the ACL footprint and tunnels. 49,58 The contralateral knee was used as a control because a high degree of bilateral symmetry in ACL geometry has been demonstrated in subjects with no history of knee injury. 53,59 From these models, we calculated the distance between each tunnel site and the native ACL of the non-surgical limb using a coordinate system with an origin at the center of the native ACL. The transtibial technique resulted in placement of the graft further from the center of the native ACL with greater variability compared to the tibial tunnelindependent technique. The tunnel center was an average of 9 mm from the center of the ACL attachment with the transtibial technique, whereas it was 3 mm with the tibial tunnel-independent technique (p < 0.05, Fig. 3). The transtibial technique resulted in a more anterior and superior placement of the tunnel (near the anteroproximal border of the ACL) than the tunnel-independent technique (p < 0.05). Consistent with our cadaver model of femoral placement, 45 these findings suggest that the transtibial technique may place the ACL graft non-anatomically due to constraints imposed by the tibial tunnel. The advantage of our MR imaging technique is that graft placement is directly measured relative to the native ACL anatomy. CT techniques cannot directly visualize the ACL attachment and must rely on bony Figure 3. The mean position (mm, mean standard deviation) of the center of the tunnels using the transtibial and tibial tunnel-independent techniques relative to the center of the ACL (shown in pink). The transtibial technique resulted in grafts that were placed anterior and superior to the center of the native ACL attachment. The tibial tunnel-independent technique resulted in placement near the center of the native ACL attachment (S, superior; I, inferior; A, anterior; P, posterior). Reprinted with permission. 49 landmarks to determine placement. Our results indicate that using the transtibial technique may lead to an anteroproximal placement of the graft on the femur, resulting in a more vertical graft that is unable to restore the oblique orientation of the native ACL. Effects of Femoral Graft Placement on In Vivo Graft Deformation 59 Our previous study identified two distinct graft placement groups: one with anatomical graft placement (using a tibial tunnel-independent technique), and one with anteroproximal placement on the femur (using a transtibial technique). 49 Having established these two different graft placements, we then investigated whether graft placement affected the graft s ability to restore orientation and length relative to the native ACL. 59 We hypothesized that anatomically placed grafts would more closely restore the function of the native ACL, while grafts placed anteroproximally on the femur would result in more vertical grafts. Such data is important because in order to reproduce the force vector of the native ACL, a graft should mimic both the orientation and length of the ACL during in vivo loading. Twenty-two patients (16 men and six women, mean age: 31 years, range: years) with unilateral ACL reconstruction were recruited from the clinics of two surgeons at the Duke University Sports Medicine Center. Patients were recruited retrospectively and were all within 6 and 36 months of surgery. Chart reviews were conducted to identify potential candidates meeting the recruitment criteria. Exclusion criteria included those with varus-valgus deformity, osteoarthritis, articular cartilage defects, meniscus injury, or any history of other trauma or surgery to either knee. Because isolated ACL tears are infrequent, 60 patients with minor tears of the meniscus (requiring removal of less than 10% of the meniscus) in the operative knee were included in the study. All patients completed the same 6-month rehabilitation protocol at the Duke University Sports Medicine Center, and had stable knees under Lachman and pivot shift examinations. At the time of the study, all patients were doing well and had returned to sports activity without restriction. All patients who met the recruitment criteria were sorted by operative date and invited in chronological order to participate. Twelve patients (nine men, three women; mean age: 32 years; mean follow-up: 20 months) received a

4 IN VIVO ACL GRAFT PLACEMENT 1163 transtibial reconstruction from a single surgeon while ten patients (seven men, three women; mean age: 30 years; mean follow-up: 18 months) received a tibial tunnel-independent reconstruction from another surgeon. 59 In the anteroproximal graft placement group, six patients received hamstring grafts (8 9 mm in diameter), while six had patellar tendon grafts (9 10 mm in diameter). Five patients had intact menisci and the remaining seven had tears requiring removal of less than 10% of the meniscus (five lateral and two medial). In the anatomic group, all patients received hamstrings grafts ranging from 7.5 to 9 mm in diameter. Four patients had intact menisci, and the remaining six had tears requiring removal of less than 10% of the meniscus (three lateral and three medial). Although graft types were not identical in the two groups, subgroup analyses did not detect any differences in the length and orientation of the patellar tendon and hamstrings grafts within the anteroproximal group. As described in Section In Vivo Measurement of ACL Graft Placement on the Femur, each subject underwent 3T MR imaging of both their injured and uninjured knees (Fig. 4) to make 3D models of both knees. Next, subjects were imaged using two orthogonally placed fluoroscopes (Pulsera, Philips; Amsterdam, The Netherlands) while performing a quasi-static lunge from 0 to 90 of knee flexion in increments of 15 (Fig. 4). 61,62 The biplanar fluoroscopic images and 3D knee models were used to reproduce the in vivo motion of each subject s graft and native ACL (Fig. 4) using a model-based matching technique with an accuracy of 0.1 mm and ,57 From these models, the elongation and coronal and sagittal plane orientation of the native ACL and graft were recorded (Fig. 5). In the sagittal plane, grafts placed anteroproximally on the femur were more vertically oriented than the native ACL from 0 to 60 of flexion (p < 0.01, Fig. 6). In contrast, there were no statistically significant differences detected between grafts placed anatomically and the native ACL at any angle (p > 0.10). At full extension, the grafts placed anteroproximally were an average of 12 more vertical than the native ACL in the sagittal plane (p ¼ 0.001), whereas those placed anatomically were 3 less vertical than the native ACL (p ¼ 0.6). In the coronal plane (Fig. 7), grafts placed anteroproximally were more vertical than the native ACL between 30 and 90 of flexion (p < 0.03). Consistent with the sagittal plane, no statistically significant differences in coronal plane angle were observed between the grafts placed anatomically and the native ACL between 0 and 90 of flexion (p > 0.3). At 60 of flexion, the grafts placed anteroproximally were a mean of 5.4 more vertical than the native ACL (p < 0.001), while those placed anatomically were 3.4 less vertical than the native ACL (p ¼ 0.3). Figure 4. MR imaging was used to create 3D models of both the femur and tibia, including the attachment sites of the ACL and graft (top left). Biplanar fluoroscopy was then used to record the motion of each subject s knees while performing a quasi-static lunge on each leg (top right). The 3D models were registered to the biplanar radiographic images to reproduce the motion of each subject s knees during activity (bottom left). From these models, both the length and orientation (in both the coronal and sagittal planes) of the ACL and graft were measured (bottom right). Reprinted with permission. 59

5 1164 DEFRATE These results indicate that grafts placed anteroproximally on the femur result in grafts that are longer and oriented more vertically in both the sagittal and coronal planes than the native ACL. The anatomic grafts (placed near the center of the femoral ACL footprint) more closely restored native ACL length and orientation. These findings are important because a vertically-oriented graft is unlikely to restore the force vector of the native ACL, as this orientation is inefficient in restraining motion in the transverse plane, and therefore is unlikely to constrain the increases in anterior translation, medial translation, and internal rotation that are observed with ACL deficiency. 26,27 Figure 5. A combination of 3D modeling, MR imaging, and biplanar fluoroscopy were used to quantify the length and orientation of the native ACL and grafts placed anatomically (top) and anteroproximally (bottom) on the femur, as demonstrated in two subjects during in vivo weight bearing. In the sagittal plane, anatomic placement resulted in a graft that more closely restored the orientation of the native ACL, while the anteroproximal grafts resulted in a graft oriented more vertically than the native ACL. Reprinted with permission. 63 Grafts placed anteroproximally were significantly longer than the native ACL at all flexion angles (p < 0.001, Fig. 8). Grafts placed anatomically were significantly longer than the native ligament for all flexion angles except at 15 and 45 (p < 0.01). Averaged across all flexion angles, the anteroproximally placed grafts were 5.6 mm longer than the native ACL, which was significantly greater than the difference of 2.1 mm for the anatomic group (p < 0.001). Effects of Femoral Graft Placement on In Vivo Knee Kinematics 63 Having demonstrated that the reconstruction affects graft placement, 49 subsequently altering graft length and orientation, 59 our next step was to determine if graft function would alter knee kinematics. Our hypothesis was that since anatomically placed grafts more closely mimicked native ACL deformation during flexion (compared to the longer and more vertical grafts placed anteroproximally on the femur), the anatomic reconstruction would more closely restore normal knee kinematics. We quantified in vivo knee kinematics of the same 22 patients from our study of in vivo graft deformation (Section Effects of Femoral Graft Placement on In Vivo Graft Deformation ). 59 Kinematics of both the native and reconstructed knee were measured using the models that reproduced the in vivo motion of each subject s knees during the lunge. Coordinate systems were created simultaneously on the injured and healthy knees by aligning the models using an iterative closest point technique. 26,49,57 Anterior-posterior translation, medial-lateral translation, and internalexternal rotation of the tibia relative to the femur were measured in these models. 26 To directly compare the ability of a reconstruction to restore each patient s normal knee function, the relative differences between the reconstructed and uninjured contralateral knees were calculated. Patients with anteroproximal graft placement on the femur had increased anterior tibial translation in the reconstructed knee relative to the contralateral Figure 6. In the sagittal plane, anteroproximal grafts were more vertically oriented than the native ACL (left), while the anatomic grafts (right) more closely restored the orientation of the native ACL (mean and 95% confidence intervals). ( p < 0.05.) Reprinted with permission. 59

6 IN VIVO ACL GRAFT PLACEMENT 1165 Figure 7. In the coronal plane, anteroproximal grafts were more vertically oriented than the native ACL (left), while the anatomic grafts (right) more closely restored the orientation of the native ACL (mean and 95% confidence intervals). ( p < 0.05.) Reprinted with permission. 59 side between 0 and 60 (p < 0.03, Fig. 9). The maximum increase in translation occurred at 30 of flexion, with a mean increase of 3.4 mm relative to the contralateral knee (p ¼ 0.003). Patients with anatomic placement, however, showed no differences in anterior tibial translation between the reconstructed and contralateral sides at any flexion angle (p > 0.32). Patients with anteroproximal graft placement on the femur also had increased medial tibial translation in the reconstructed knee relative to the uninjured knee between 0 and 75 of flexion (p < 0.05, Fig. 10). The maximum increase in translation occurred at 15 of flexion, with a mean increase of 1.1 mm relative to the contralateral knee (p ¼ 0.005). Patients with anatomic placement showed no differences in medial tibial translation between the reconstructed and contralateral knees at any flexion angle (p > 0.21). Additionally, patients with anteroproximal graft placement had increased internal tibial rotation in the reconstructed knee relative to the contralateral side between 0 and 60 (p < 0.04, Fig. 11). The maximum increase in rotation occurred at 30 of flexion, where a mean increase of 3.5 of internal rotation was detected (p ¼ 0.01). Patients with anatomic placement showed no differences between the reconstructed and contralateral sides at any flexion angle (p > 0.19). The altered kinematics observed in patients with non-anatomically placed grafts are consistent with our findings on graft deformation. 59 Grafts placed anteroproximally are oriented more vertically and are unlikely to provide sufficient restraint to motions in the transverse plane. Therefore, increased anterior and medial translation and internal rotation are likely to be observed with grafts placed anteroproximally. This is consistent with previous reports of increased anterior translation, medial translation, and internal rotation in patients with ACL deficiency. 4,26,27,35 In contrast, the more oblique grafts, placed anatomically, more closely reproduced normal knee motion. These results suggest that anatomic placement of the graft on the femur during ACL reconstruction is critical to restoring normal knee motion, which is important because abnormal knee motion has been thought to contribute to the degenerative changes observed after ACL injury. 4,26,27,64,65 Effects of Femoral Graft Placement on Cartilage Thickness 66 Our results have shown that grafts placed anatomically more closely reproduce the orientation and length of the native ACL during in vivo knee function 59 and as a result, better restore normal knee motion. 63 On the other hand, grafts placed anteroproximally are in a more vertical orientation and do not provide sufficient restraint to transverse plane motions. Such an orientation could explain the increased anterior Figure 8. With regard to length, anteroproximal grafts (top left) were longer than the native ACL at all flexion angles (mean and 95% confidence intervals). Anatomic grafts (top right) were also longer than the native ACL. However, when the difference in length from the native ACL was averaged across all flexion angles, anteroproximal grafts were longer compared to the anatomic grafts (bottom). ( p < 0.05.) Reprinted with permission. 59

7 1166 DEFRATE Figure 9. The increase in anterior tibial translation of the reconstructed knee relative to the contralateral native knee was measured as a function of flexion (mean and 95% confidence intervals). Zero denotes an anterior translation in the reconstructed knee that exactly mimicked the motion of the contralateral side. Patients with grafts placed anteroproximally on the femur had increased anterior tibial translation relative to the contralateral side between 0 and 60 of flexion, while the anatomically placed grafts more closely restored anterior tibial translation. ( p < 0.05.) Reprinted with permission. 63 translation, medial translation, and internal rotation seen in these patients and those with ACL deficiencies. 26,63 These results are important because altered joint kinematics potentially elicit changes in cartilage loading, and could predispose the knee to degenerative changes. 26,27,64,65,67 In ACL deficient knees, cartilage contact in the medial compartment is shifted towards the medial tibial spine, a region where degeneration is observed in patients. 5,26,27,67 70 Therefore, we analyzed cartilage thickness in the corresponding region on the femur to quantify the effects of graft placement on cartilage thickness in this region. The same twenty-two subjects from the previous studies (Sections Effects of Femoral Graft Placement on In Vivo Graft Deformation and Effects of Femoral Graft Placement on In Vivo Knee Kinematics ) were analyzed. In addition, the normal side-to-side variation of cartilage thickness in ten male control subjects (mean age: 30 years) without history of knee injury was evaluated. After a 30-minute period of non-weight bearing, 71 sagittal MR images were acquired using the same Figure 10. The increase in medial tibial translation of the reconstructed knee relative to the contralateral native knee was measured as a function of flexion (mean and 95% confidence intervals). Zero denotes a medial tibial translation in the reconstructed knee that exactly mimicked the motion of the native knee. Patients with grafts placed anteroproximally on the femur had significantly increased medial tibial translation relative to the native knee between 0 and 75 of flexion, while the anatomically placed grafts more closely restored medial tibial translation. ( p < 0.05.) Reprinted with permission. 63 Figure 11. The increase in internal tibial rotation of the reconstructed knee relative to the contralateral native knee was measured as a function of flexion (mean and 95% confidence intervals). Zero denotes an internal tibial rotation of the reconstructed knee that exactly mimicked the motion of the native knee. Patients with grafts placed anteroproximally on the femur had increased internal tibial rotation relative to the native knee between 0 and 60 of flexion, while the anatomically placed grafts more closely restored internal tibial rotation. ( p < 0.05.) Reprinted with permission. 63 sequences as previously described (Sections In Vivo Measurement of ACL Graft Placement on the Femur and Effects of Femoral Graft Placement on In Vivo Graft Deformation ). In addition to the bony geometry, these images were used to generate 3D models of the articular cartilage surfaces To measure changes in cartilage thickness relative to the same points in both knees, an iterative closest point technique 49,59,63 was used to align the 3D bone models of the native and reconstructed knees with each other. This MR imaging and 3D modeling technique has been previously validated, 67 with a coefficient of repeatability in measuring cartilage thickness of 0.03 mm. 75 In the present study, three sampling points along the lateral aspect of the medial femoral condyle were selected for analysis (Fig. 12). Thickness measurements were performed by averaging cartilage thickness at each vertex on the model within a 2.5 mm radius of the sampling point. The side-to-side differences in cartilage thickness in each patient group were compared. In the anteroproximal graft placement group, cartilage thickness along the lateral aspect of the medial condyle of the reconstructed knee decreased by 8% compared to that of the contralateral knee with a native ACL (p ¼ 0.02). In the anatomic group, no statistically significant side-to-side differences were observed between the reconstructed and native knees, with a mean difference of 1%. In the ten normal control subjects with healthy knees, no statistical differences were observed in the thickness of left versus right knees in the same regions, with a mean difference of 2%. In conclusion, the minimal variation in side-to-side cartilage thickness in the normal subjects suggests that the contralateral knee is an appropriate control for these measurements. Additionally, the anatomic graft placement group did not experience significant cartilage thinning, suggesting that restoring normal knee motion might be protective of cartilage health. Finally, abnormal motion resulting from anteroproximal graft placement on the femur potentially alters

8 IN VIVO ACL GRAFT PLACEMENT 1167 Figure 12. Thickness maps of the femoral cartilage were generated from the 3D knee models. Thickness measurements were made at three evenly-spaced points along the lateral aspect of the medial femoral condyle (within the red ovals). In subjects with anatomic graft placement (top), there were minimal side-toside differences in thickness in this region. In in subjects with anteroproximal graft placement, there was a decrease in the thickness of the cartilage (bottom). Reprinted with permission. 66 normal cartilage loading patterns. 26,27,70 Such altered cartilage loading could disrupt normal cartilage homeostasis, potentially explaining the resulting decreased cartilage thickness along the lateral aspect of the medial femoral condyle observed in patients with anteroproximal graft placement. This region is consistent with regions where joint degeneration has been observed clinically in patients with chronic ACL deficiency. 5,76 Future work is needed to quantify changes in cartilage thickness at multiple time points before and after ACL reconstruction. CONCLUSIONS ACL reconstruction is a commonly performed procedure that generally has excellent clinical outcomes. Numerous long term studies, however, have questioned its ability to prevent degenerative changes compared to non-operative treatments. 3,14,15,20 24 It has been suggested that abnormal kinematics after ACL reconstruction might be a factor predisposing the knee to these degenerative changes. 29,33 In these studies, we examined the effects of graft placement on the restoration of normal in vivo joint function. First, we used an in vivo 3D model to measure the ability of two different reconstruction techniques to place the graft anatomically on the femur relative to the native ACL footprint. 45,49 The transtibial technique used in this study placed the femoral tunnel near the anteroproximal border of the ACL attachment site, while the tibial tunnelindependent technique resulted in placement near the center of the ACL footprint. Next, we showed that anatomic placement is important for restoring the normal deformation and orientation of the native ACL under in vivo loading conditions. Grafts placed anteroproximally on the femur were longer and more vertically oriented relative to the native ACL. In contrast, grafts placed near the center of the ACL footprint, via the tunnel-independent technique, mimicked the orientation and length of the native ACL. A more anatomically-placed graft would be expected to more closely restore normal joint kinematics, since it reproduces the deformation of the native ACL. A vertically-oriented graft is likely to provide insufficient restraint to motions in the transverse plane because the oblique orientation of the ACL serves an important role in restraining anterior translation, medial translation, and internal rotation of the tibia. 26,27,67 Our follow-up study on the effects of graft placement on in vivo knee kinematics confirmed these findings. The more vertical grafts resulting from anteroproximal placement provided insufficient restraint to anterior translation, medial translation, and internal rotation while the more anatomic placement from the tibial tunnel-independent technique resulted in grafts that mimicked the native ACL and closely reproduced normal knee motion. These results demonstrate that achieving anatomic femoral tunnel placement more closely restores normal knee function, and provide surgeons with important clinical implications regarding joint degeneration. Recent in vivo studies of ACL deficiency have indicated that increased anterior translation, medial translation, and internal rotation elevate peak contact strains and alter contact regions within the joint, 26,27,67,70 shifting contact in the medial compartment towards the medial tibial spine and medial intercondylar notch a region where cartilage degeneration has been observed clinically in patients with ACL deficiency. 5,26,27,67 70 Our results suggest that a more anatomic placement of the ACL graft might reduce these elevated contact strains. Restoring normal joint contact is likely an important factor in preventing degenerative changes. 4,64,67,77,78 These data were further supported by our study on the changes in cartilage thickness between uninjured and reconstructed knees for the anatomic and anteroproximal placement groups. Patients with grafts placed anteroproximally on the femur had significantly decreased cartilage thickness along the lateral aspect of the medial femoral condyle, while anatomic graft placement maintained cartilage thickness in this same region. Overall, these findings suggest that achieving anatomic graft placement is crucial to reproducing normal knee kinematics and might slow the progression of joint degeneration following ACL reconstruction. Future studies should evaluate whether graft placement has an effect on the development of degenerative changes in the joint with long term follow-up. AUTHOR S CONTRIBUTIONS Dr. L.E.D. was responsible for all aspects of this review paper summarizing the work for which he was awarded the 2016 Kappa Delta Young Investigator Award.

9 1168 DEFRATE ACKNOWLEDGEMENTS This body of work was recognized with the 2016 Kappa Delta Young Investigator Award. This work was supported by grants from the National Football League Charities, Arthrex, and the National Institutes of Health (AR055659, AR065527, AR063325). I would like to thank Dr. William E. Garrett, M.D., Ph.D. and Dr. Charles E. Spritzer, M.D. for the countless invaluable discussions and the myriad contributions to these projects. Their contributions were essential to completing this work. I would also like to thank Dr. Ermias S. Abebe, M.D. and Dr. Eziamaka C. Okafor, M.D. for leading these projects during their research years during medical school. Thanks to Dr. C. T. Moorman, M.D. and Dr. Dean C. Taylor, M.D. from the James R. Urbaniak, MD, Sports Sciences Institute for their input and discussions regarding this work. I would also like to thank Duke University Medical Center colleagues Dr. Maria K. Kaseta, M.D., Dr. Brian L. Charnock, M.D., Dr. Robert T. Sullivan, M.D., Gangadhar M. Utturkar, M.S., Dr. Jong Pil Kim, M.D., Dr. T. Scott Dziedzic, M.D., Dr. R. Lee Cothran, M.D., Molly Widmyer, M.S., Dr. Kevin Taylor, M.D., Dr. Tripp Mostertz, M.D., Dr. Amber T. Collins, Ph.D., and Hattie C. Cutcliffe, M.S. for their valuable contributions to this work. I would like to thank Dr. Farshid Guilak, Ph.D. (Washington University at St. Louis) and Dr. Guoan Li, Ph.D. (Massachusetts General Hospital) for all of their helpful advice and mentorship. Finally, I acknowledge Donald T. Kirkendall, ELS, a contracted medical editor, for his help in preparing this manuscript for submission. REFERENCES 1. Junkin DM, Jr., Johnston DL, Fu FH, et al Knee ligament injuries. In: Kibler WB, editor. Orthopaedic knowledge update Sports medicine 4. Rosemont, IL: American Academy of Orthopaedic Surgeons. p Georgoulis AD, Papadonikolakis A, Papageorgiou CD, et al Three-dimensional tibiofemoral kinematics of the anterior cruciate ligament-deficient and reconstructed knee during walking. Am J Sports Med 31: Fithian DC, Paxton LW, Goltz DH Fate of the anterior cruciate ligament-injured knee. Orthop Clin North Am 33: Andriacchi TP, Briant PL, Bevill SL, et al Rotational changes at the knee after ACL injury cause cartilage thinning. Clin Orthop Relat Res 442: Fairclough JA, Graham GP, Dent CM Radiological sign of chronic anterior cruciate ligament deficiency. Injury 21: Hill CL, Seo GS, Gale D, et al Cruciate ligament integrity in osteoarthritis of the knee. Arthritis Rheum 52: Roos H, Adalberth T, Dahlberg L, et al Osteoarthritis of the knee after injury to the anterior cruciate ligament or meniscus: the influence of time and age. Osteoarthritis Cartilage 3: Beynnon BD, Fleming BC Anterior cruciate ligament strain in-vivo: a review of previous work. J Biomech 31: Griffin LY, Albohm MJ, Arendt EA, et al Understanding and preventing noncontact anterior cruciate ligament injuries: a review of the Hunt Valley II meeting, January Am J Sports Med 34: Freedman KB, D Amato MJ, Nedeff DD, et al Arthroscopic anterior cruciate ligament reconstruction: a metaanalysis comparing patellar tendon and hamstring tendon autografts. Am J Sports Med 31: Krych AJ, Jackson JD, Hoskin TL, et al A metaanalysis of patellar tendon autograft versus patellar tendon allograft in anterior cruciate ligament reconstruction. Arthroscopy 24: Lee DY, Karim SA, Chang HC Return to sports after anterior cruciate ligament reconstruction a review of patients with minimum 5-year follow-up. Ann Acad Med Singapore 37: Grossman MG, ElAttrache NS, Shields CL, et al Revision anterior cruciate ligament reconstruction: three- to nine-year follow-up. Arthroscopy 21: Lohmander LS, Ostenberg A, Englund M, et al High prevalence of knee osteoarthritis, pain, and functional limitations in female soccer players twelve years after anterior cruciate ligament injury. Arthritis Rheum 50: von Porat A, Roos EM, Roos H High prevalence of osteoarthritis 14 years after an anterior cruciate ligament tear in male soccer players: a study of radiographic and patient relevant outcomes. Ann Rheum Dis 63: Holm I, Oiestad BE, Risberg MA, et al No differences in prevalence of osteoarthritis or function after open versus endoscopic technique for anterior cruciate ligament reconstruction: 12-year follow-up report of a randomized controlled trial. Am J Sports Med 40: Salmon LJ, Russell VJ, Refshauge K, et al Long-term outcome of endoscopic anterior cruciate ligament reconstruction with patellar tendon autograft: minimum 13-year review. Am J Sports Med 34: Janssen RP, du Mee AW, van Valkenburg J, et al Anterior cruciate ligament reconstruction with 4-strand hamstring autograft and accelerated rehabilitation: a 10-year prospective study on clinical results, knee osteoarthritis and its predictors. Knee Surg Sports Traumatol Arthrosc 21: Pinczewski LA, Deehan DJ, Salmon LJ, et al A fiveyear comparison of patellar tendon versus four-strand hamstring tendon autograft for arthroscopic reconstruction of the anterior cruciate ligament. Am J Sports Med 30: Smith MV, Nepple JJ, Wright RW, et al Knee osteoarthritis is associated with previous meniscus and anterior cruciate ligament surgery among elite college American football athletes. Sports Health. nlm.nih.gov/pubmed/ [Epub ahead of print]. 21. Tsoukas D, Fotopoulos V, Basdekis G, et al No difference in osteoarthritis after surgical and non-surgical treatment of ACL-injured knees after 10 years. Knee Surg Sports Traumatol Arthrosc 24: Butler RJ, Minick KI, Ferber R, et al Gait mechanics after ACL reconstruction: implications for the early onset of knee osteoarthritis. Br J Sports Med 43: Lohmander LS, Englund PM, Dahl LL, et al The longterm consequence of anterior cruciate ligament and meniscus injuries: osteoarthritis. Am J Sports Med 35: Fink C, Hoser C, Hackl W, et al Long-term outcome of operative or nonoperative treatment of anterior cruciate ligament rupture-is sports activity a determining variable? Int J Sports Med 22: Andriacchi TP, Dyrby CO Interactions between kinematics and loading during walking for the normal and ACL deficient knee. J Biomech 38: Defrate LE, Papannagari R, Gill TJ, et al The 6 degrees of freedom kinematics of the knee after anterior cruciate ligament deficiency: an in vivo imaging analysis. Am J Sports Med 34: Li G, Moses JM, Papannagari R, et al Anterior cruciate ligament deficiency alters the in vivo motion of the

10 IN VIVO ACL GRAFT PLACEMENT 1169 tibiofemoral cartilage contact points in both the anteroposterior and mediolateral directions. J Bone Joint Surg Am 88: Logan M, Dunstan E, Robinson J, et al Tibiofemoral kinematics of the anterior cruciate ligament (ACL)-deficient weightbearing, living knee employing vertical access open interventional multiple resonance imaging. Am J Sports Med 32: Tashman S, Kolowich P, Collon D, et al Dynamic function of the ACL-reconstructed knee during running. Clin Orthop Relat Res 454: Li G, Papannagari R, DeFrate LE, et al The effects of ACL deficiency on mediolateral translation and varus-valgus rotation. Acta Orthopaedica 78: Zhang LQ, Shiavi RG, Limbird TJ, et al Six degreesof-freedom kinematics of ACL deficient knees during locomotion-compensatory mechanism. Gait Posture 17: Gao B, Zheng NN. Alterations in three-dimensional joint kinematics of anterior cruciate ligament-deficient and -reconstructed knees during walking. Clin Biomech (Bristol, Avon) 25: Papannagari R, Gill TJ, Defrate LE, et al In vivo kinematics of the knee after anterior cruciate ligament reconstruction: a clinical and functional evaluation. Am J Sports Med 34: , Epub 2006 Aug Scanlan SF, Chaudhari AM, Dyrby CO, et al Differences in tibial rotation during walking in ACL reconstructed and healthy contralateral knees. J Biomech 43: Gao B, Zheng NN Alterations in three-dimensional joint kinematics of anterior cruciate ligament-deficient and -reconstructed knees during walking. Clin Biomech (Bristol, Avon) 25: Giron F, Buzzi R, Aglietti P Femoral tunnel position in anterior cruciate ligament reconstruction using three techniques. A cadaver study. Arthroscopy 15: Giron F, Cuomo P, Aglietti P, et al Femoral attachment of the anterior cruciate ligament. Knee Surg Sports Traumatol Arthrosc 14: Farrow LD, Chen MR, Cooperman DR, et al Morphology of the femoral intercondylar notch. J Bone Joint Surg 89: Fu FH, Bennett CH, Ma CB, et al Current trends in anterior cruciate ligament reconstruction. Part II. Operative procedures and clinical correlations. Am J Sports Med 28: Kohn D, Busche T, Carls J Drill hole position in endoscopic anterior cruciate ligament reconstruction. Results of an advanced arthroscopy course. Knee Surg Sports Traumatol Arthrosc 6:S13 S Zavras TD, Race A, Amis AA The effect of femoral attachment location on anterior cruciate ligament reconstruction: graft tension patterns and restoration of normal anterior-posterior laxity patterns. Knee Surg Sports Traumatol Arthrosc 13: Arnold MP, Kooloos J, van Kampen A Single-incision technique misses the anatomical femoral anterior cruciate ligament insertion: a cadaver study. Knee Surg Sports Traumatol Arthrosc 9: Cain EL, Jr., Clancy WG, Jr Anatomic endoscopic anterior cruciate ligament reconstruction with patella tendon autograft. Orthop Clin North Am 33: Harner CD, Fu FH, Irrgang JJ, et al Anterior and posterior cruciate ligament reconstruction in the new millennium: a global perspective. Knee Surg Sports Traumatol Arthrosc 9: Kaseta MK, Defrate LE, Charnock BL, et al Reconstruction technique affects femoral tunnel placement in ACL reconstruction. Clin Orthop Relat Res 466: Loh JC, Fukuda Y, Tsuda E, et al Knee stability and graft function following anterior cruciate ligament reconstruction: comparison between 11 o clock and 10 o clock femoral tunnel placement. Arthroscopy 19: Scopp JM, Jasper LE, Belkoff SM, et al The effect of oblique femoral tunnel placement on rotational constraint of the knee reconstructed using patellar tendon autografts. Arthroscopy 20: Yamamoto Y, Hsu WH, Woo SL, et al Knee stability and graft function after anterior cruciate ligament reconstruction: a comparison of a lateral and an anatomical femoral tunnel placement. Am J Sports Med 32: Abebe ES, Moorman CT, 3rd, Dziedzic TS, et al Femoral tunnel placement during anterior cruciate ligament reconstruction: an in vivo imaging analysis comparing transtibial and 2-incision tibial tunnel-independent techniques. Am J Sports Med 37: Duquin TR, Wind WM, Fineberg MS, et al Current trends in anterior cruciate ligament reconstruction. J Knee Surg 22: Robin BN, Jani SS, Marvil SC, et al Advantages and disadvantages of transtibial, anteromedial portal, and outside-in femoral tunnel drilling in single-bundle anterior cruciate ligament reconstruction: a systematic review. Arthroscopy 31: Kopf S, Forsythe B, Wong AK, et al Nonanatomic tunnel position in traditional transtibial single-bundle anterior cruciate ligament reconstruction evaluated by threedimensional computed tomography. J Bone Joint Surg 92: Scanlan SF, Lai J, Donahue JP, et al Variations in the three-dimensional location and orientation of the ACL in healthy subjects relative to patients after transtibial ACL reconstruction. J Orthop Res 30: Howell SM, Gittins ME, Gottlieb JE, et al The relationship between the angle of the tibial tunnel in the coronal plane and loss of flexion and anterior laxity after anterior cruciate ligament reconstruction. Am J Sports Med 29: Simmons R, Howell SM, Hull ML Effect of the angle of the femoral and tibial tunnels in the coronal plane and incremental excision of the posterior cruciate ligament on tension of an anterior cruciate ligament graft: an in vitro study. J Bone Joint Surg 85-A: Wittstein JR, Garrett WE Prevalence of the remnant femoral attachment of the ruptured anterior cruciate ligament. Clin Orthop Relat Res 467: Caputo AM, Lee JY, Spritzer CE, et al In vivo kinematics of the tibiotalar joint after lateral ankle instability. Am J Sports Med 37: Taylor KA, Cutcliffe HC, Queen RM, et al In vivo measurement of ACL length and relative strain during walking. J Biomech 46: Abebe ES, Kim JP, Utturkar GM, et al The effect of femoral tunnel placement on ACL graft orientation and length during in vivo knee flexion. J Biomech 44: Olsson O, Isacsson A, Englund M, et al Epidemiology of intra- and peri-articular structural injuries in traumatic knee joint hemarthrosis data from 1145 consecutive knees with subacute MRI. Osteoarthritis Cartilage 24: Utturkar GM, Irribarra LA, Taylor KA, et al The effects of a valgus collapse knee position on in vivo ACL elongation. Ann Biomed Eng 41: Carter TE, Taylor KA, Spritzer CE, et al In vivo cartilage strain increases following medial meniscal tear and correlates with synovial fluid matrix metalloproteinase activity. J Biomech 48: