- Email: [email protected]
Evaluation of the isometry of different points of the patella and femur for medial patellofemoral ligament reconstruction
Clinical Biomechanics 38 (2016) 8–12
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
Evaluation of the isometry of different points of the patella and femur for medial patellofemoral ligament reconstruction☆ Riccardo Gomes Gobbi PhD a,⁎, César Augusto Martins Pereira a, David Sadigursky MD b, Marco Kawamura Demange PhD a, Luis Eduardo Passarelli Tírico PhD a, José Ricardo Pécora PhD a, Gilberto Luis Camanho PhD a a b
Hospital das Clínicas da Universidade de São Paulo, SP, Brazil Clínica ortopédica e Traumatológica, COT-Martagão, BA, Brazil
a r t i c l e
i n f o
Article history: Received 19 February 2016 Accepted 4 August 2016 Keywords: Patellar dislocation Patellofemoral joint Knee Biomechanical phenomena Range of motion Articular
a b s t r a c t Background: The location of patellar and femoral fixation of the graft in medial patellofemoral ligament reconstructions has been widely discussed. This study aimed to assess the distances between different patellar and femoral fixation points to identify the least anisometric pairs of points. Methods: Ten cadaver knees were attached to an apparatus that simulated an active range of motion of 120°, with three metallic markers fixed onto the medial side of the patella, and seven markings onto the medial epicondyle. The examined points included the proximal patella pole (1), the patellar center (3), the midpoint between points 1 and 3 (2), a point directly on the epicondyle (6), points 5 mm anterior (5) and posterior (7) to the epicondyle, points 5 mm anterior to point 5 (4) and 5 mm posterior to point 7 (8), and points 5 mm proximal (9) and distal (10) to the epicondyle. The distances between patella and femur points were measured by a photogrammetry system at 15° intervals. Findings: The pair of points that exhibited the lowest average variability in distance, and hence was the most isometric, was the patella center combined with the anterior to the medial femoral epicondyle. The pairs of points that exhibited the highest average variability in distance, and hence were the least isometric, were the ones located distal or posterior to the medial femoral epicondyle, with less influence by the patellar location. Interpretation: Surgeons should avoid positioning the graft distally or posterior to the epicondyle due to the increase in anisometry. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction The medial patellofemoral ligament (MPFL) is considered the most important patellar medial stabilizer and accounts for approximately 50–60% of patellar lateralization resistance (Conlan et al., 1993; Desio et al., 1998; Hautamaa et al., 1998). Despite this importance, valuation of its reconstruction in the treatment of patellar instability is relatively recent and has mainly occurred in the last 20 years. Despite the wide range of techniques described for its reconstruction, with different graft sources and fixation methods, evidence has been accumulating showing good clinical results for this surgery, with very low (b 10%) instability recurrence rates (Deie et al., 2005; Drez et al., 2001; Ellera Gomes et al., 2004; Fernandez et al., 2005; Mikashima et al., 2006; Nomura and Inoue, 2006). However, there are still controversies surrounding the surgical technique that hinder universal standardization of the procedure. An important and widely discussed aspect is the location of patellar and femoral fixation of the graft in reconstructions. The femoral ligament origin has been described in the adductor tubercle – AT (Conlan et al., 1993; Desio et al., 1998; Tuxoe et al., 2002), in the anterior portion of the medial
http://dx.doi.org/10.1016/j.clinbiomech.2016.08.002 0268-0033/© 2016 Elsevier Ltd. All rights reserved.
epicondyle – ME (Feller et al., 1993), on the anterior portion of the medial epicondyle distal to the adductor tubercle (Smirk and Morris, 2003), on the medial epicondyle itself (Hautamaa et al., 1998; Reider et al., 1981; Warren and Marshall, 1979) and proximal and posterior to the epicondyle, immediately distal to the adductor tubercle (Nomura et al., 2000). Variation in these points can cause errors with significant consequences to patellar biomechanics, causing increased joint pressure, pain, chondral degeneration and graft failure (Beck et al., 2007; Elias and Cosgarea, 2006; Sandmeier et al., 2000). Thus, despite MPFL behavior not being isometric, increasing tension with knee extension (Amis et al., 2003), studies are needed to analyze different patellar and femoral fixation points to identify those that offer more consistent performance. Although some studies have specifically addressed this issue, they are few in number (Smirk and Morris, 2003; Steensen et al., 2004; Stephen et al., 2012; Yoo et al., 2012), and all of them present technical limitations, such as excessive soft tissue removal, distance measurement method and number of points studied. If a ligament or stabilizing structure is removed, released or incised (such as medial retinaculum), it is not possible to adequately study isometry, because the patella may rest in a non-anatomical position, shifted
R.G. Gobbi et al. / Clinical Biomechanics 38 (2016) 8–12
laterally for example, throughout the study. Some studies used sliding sutures for obtaining distances, which add to the error of the measuring device its manipulation and positioning by humans. The present study aimed to assess the distances between different patellar and femoral MPFL fixation points along the range of motion of cadaver knees using a precise measurement method and preservation and proper tensioning of soft tissue to identify the least anisometric pairs of points. This knowledge will be useful to help surgeons predict the effect of malpositioning of the graft around the ideal points. 2. Material and methods A total of 10 cadaver knees, without throclear dysplasia and normal patellar height, with no previous injuries or procedures, were used. The knees were obtained from the Arthroscopy Laboratory of our institution after local Institutional Review Board approval (number 1115/08). The knees had the femur and tibia firmly fixed to an apparatus that applies traction through the quadriceps muscle (connected to a Kratos® K5002 universal mechanical testing machine, with a 5 tf load cell) and resistance against the tibia, simulating an active range of motion of 120° to 0° flexion (Fig. 1). Traction by the Kratos machine was progressive to bring the knee from full flexion to full extension and applied at the speed of 50 mm/min; it varied from 80 N in deep flexion to 600 N in final extension. The resistance applied to the tibia consisted of a 5,65 kg positioned in the pulley (letter U in Fig. 1). No dissection of soft tissue around the knee joint was performed, preserving all structures. On each knee, three metallic markers (nails) were fixed onto the medial side of the patella, and a cancellous screw was fixed onto the medial epicondyle (ME), onto which a thermoplastic marker with seven markings was attached. Minimum dissection was performed on these fixation points, enough only to keep the markers near the bone surface. The obtained points were the proximal pole patella (P1), the patellar midpoint (P3), the midpoint between the two previous measures on the patella (P2), a point directly on the ME (F6), points 5 mm anterior (F5) and posterior (F7) to the ME, points 5 mm anterior to F5 (F4) and 5 mm posterior to F7 (F8), and points 5 mm proximal (F9) and distal (F10) to the ME (Fig. 2). The markers positioning was based on
9
finding the medial epicondyle. It is very easy to locate it through palpation, even through intact skin. The same is also true when considering the location of the proximal and distal ends of the patella. These points were chosen because they could be easily located during surgery (all femoral points were based on the ME, which is easy to locate by palpation) to increase the clinical applicability of the findings, respecting the anatomical descriptions of the MPFL (Feller et al., 1993; Hautamaa et al., 1998; Reider et al., 1981; Smirk and Morris, 2003; Warren and Marshall, 1979). To measure the distances between different points of the patella and femur (analysis of isometry between pairs of points), a threedimensional photogrammetry system was used based on the model developed by Abdel-Aziz and Karara (2015). This system consisted of a computer software, two digital cameras and a calibrator with markers with known distances between them engineered by an industrial machine with a margin of error of 1.2 μm. The two cameras simultaneously photographed the medial portion of the knee at angles of 120°, 105°, 90°, 75°, 60°, 45°, 30°, 15° and 0° of flexion, with the aid of a digital goniometer connected to the cameras' remote controls. The system allows three-dimensional measurements between the points at different camera positions (Fig. 3). This system allowed us to obtain images to measure the distance between the patellar and femoral points at every 15° of flexion in a 0–120° range of motion. Each arc was performed three times, with 21 pairs of points (three patellar with seven femoral), with three photos for each flexion angle (27 images from each camera for each knee). The distance between points was measured with the software, and the value used for the statistical tests was the mean score of the three measures for each pair of points. An error analysis of the method was performed calculating the distances between the markers of the calibrator obtained by the software. Based on means and standard deviations of maximum distance variation through range of motion for each pair of points, analysis of variance (ANOVA) with repeated measures was applied, followed by Tukey's repeated measures test to identify which point pair had the lowest and highest distance variation compared to the other pairs.
Fig. 1. Schematic representation of the left knee positioned in the mechanical device coupled to the universal mechanical testing machine. Front and side view of the device components. A - load cell, B - steel wire, C - sheave, D - quadriceps grip guide, E - quadriceps grip, F - support table, G - femoral grip support, H - femoral rod, I - patella, J - femur, K - tibia, L - tibial grip, M spherical plain bearing, N - axial arm, O - radial arm, P - counterweight, Q - drive shaft, R - bearing housings, S - pulley, T - pulley steel wire, U - weight (5,65 kg), V - angular optical encoder (digital goniometer), W - femur markers, X - patella markers.
10
R.G. Gobbi et al. / Clinical Biomechanics 38 (2016) 8–12
mean error of 0.08 mm, with standard deviation of 0.05 mm ranging from −0.11 mm to 0.31 mm. 4. Discussion
Fig. 2. Positioning of the points used to measure P1 to P3 on the patella and F4 to F10 on the femur (femoral central point: F6).
3. Results The behavior of the mean of the three measurements obtained for each flexion angle between the point pairs studied is shown in Figs. 4–6. Analyzing the mean difference in distance variability along the knee flexion movement between all pairs of points revealed a significant difference, with p b 0.001 (Table 1). When considered individually, the point pair that exhibited the lowest average variability in distances in knee movement of 2.55 mm was 3–4 (patella center and 10 mm anterior to medial epicondyle), being the most isometric. The point pair that exhibited the highest average variability in distances of 8.22 mm in knee flexion angles was 1–10 (proximal patella and 10 mm posterior to medial epicondyle); however, it was not significantly different from point pairs 1–8, 2–8, 2–10 and 3–8 (p N 0.05). The distance was not increased between any pair of points with knee flexion, which is compatible with MPFL biomechanics. Patellar positioning had less influence on isometry. The femoral points that exhibited the highest average variability in distance, and hence were the least isometric, were the ones located distal or 10 mm posterior to the medial femoral epicondyle. The margin of error analysis compared the known distances between the calibrator markers and the same distances found by the software. After comparing 2708 obtained distances, the authors found a
The most important finding of the study was that the least isometric points for MPFL reconstruction were located distal and 10 mm posterior to the ME on the femur, with little influence based on patellar location; while the most isometric points were located at patellar center and 10 mm anterior to the ME. Correct positioning of the graft is critical because small errors can lead to significant increases in patellar joint contact pressure (Beck et al., 2007; Elias and Cosgarea, 2006). Due to the high variability of descriptions of the femoral MPFL origin, (Conlan et al., 1993; Desio et al., 1998; Feller et al., 1993; Hautamaa et al., 1998; Nomura et al., 2000; Reider et al., 1981; Smirk and Morris, 2003; Tuxoe et al., 2002; Warren and Marshall, 1979) several studies have been conducted evaluating the “isometry” of different points. Smirk and Morris (2003) conducted an isometric study of four knees using various reference points on the patella and around the medial epicondyle. Although no combination of points presented isometry, the points with the least variation in knee movement distance were on the patella, the superior third and at its equator; on the femur, one point 10 mm distal and 5 mm posterior to the adductor tubercle, and one point 5 mm posterior and another 5 mm distal to the former. Steensen et al. (2004) after dissecting the MPFLs of 11 knees, reported that the most isometric points were the most distal part of the MPFL on the patella and the most proximal of the MPFL on the femur. Stephen et al. (2012) conducted another study on MPFL isometry in eight knees and reported that the point between the medial epicondyle and the adductor tubercle was the most isometric. Positioning errors on the proximal-distal axis generated more anisometry than errors on the anterior–posterior axis. Changes in patellar insertion had little effect on MPFL isometry. A similar result was reported by Yoo et al. (2012). They measured these distances using three-dimensional knee reconstruction of 10 patients who underwent computed tomography (CT) scans at various flexion angles. A summary of the findings from these studies is presented in Table 2. In our study, we found the patellar midpoint and the point 10 mm anterior to the ME (points 3–4) to be the most isometric points (absolute measures). These findings are consistent with the points identified by Smirk and Morris (2003) and Steensen et al. (2004) in relation to the patella location. The least isometric points (absolute measures) were the most proximal to the patella and the most distal on the femur (points 1–10). Also statistically equivalent were the points posterior to the ME (points 1–8, 2–8, 2–10, and 3–8). Therefore, our findings suggest that placing the graft on the femur directly posterior or distal to the ME is risky. It should be noted that no point pair under study generated an
Fig. 3. Right knee positioned in the device coupled to the universal mechanical testing machine, with cameras in position.
R.G. Gobbi et al. / Clinical Biomechanics 38 (2016) 8–12
11
Fig. 4. Mean profiles in knee flexion angles of the absolute distances between points, with patellar point 1 as the reference.
Fig. 6. Mean profiles in knee flexion angles of the absolute distances between points, with patellar point 3 as the reference.
increase in distance between the points (tensioning) with knee flexion, compatible with MPFL biomechanics. The differences of findings among studies described in Table 2 can occur for several reasons. The studies used different measurement methods, different set of points in patella and femur, different set of flexion angles and different soft tissue tensioning. Although the present study offers a precise measurement system, a large number of insertion points and flexion angles and adequate soft tissue tensioning, the results should be critically interpreted as a whole by the readers. The present study has four main limitations. First, we did not study one of the points often cited as a femoral MPFL insertion point, located between the ME and the AT and described by Nomura et al. (2000). We did not include this point because we chose to study only points that are easily locatable during surgery and used only the ME as a parameter. We believe that to consistently place this point, we would need fluoroscopy, unavailable in our methodology. Second, despite point F9 vicinity, we know that our findings do not exactly match the most frequently recommended location for femoral insertion, described radiographically by Schottle et al. (2007). However, it is important to separate the concepts of anatomical and isometric reconstruction, which still lead to important debates in anterior cruciate ligament surgery. The option of using anatomical landmarks does not diminish the
value of knowing the behavior of points in the surrounding area. This fact is especially true when considering a structure such as the MPFL, with insertion described at various points around the ME. If a surgeon places the femoral attachment of the graft in a less than ideal position for any given reason, it is important to know the isometric behavior of these points around the medial epicondyle. Our findings suggest that if the mistake occurs towards posterior or distal, it should be repositioned. It's important to emphasize that our aim is not to replace the points described by Nomura and Schöttle as it is well known that the MPFL is not isometric. We intend to offer surgeons more information about behavior of possible points around the ideal one. The third limitation is that our method, being a linear measure, cannot account for the wrapping that the MPFL suffers around the femur at deeper flexion angles. This wrapping would cause an increase in length in deep flexion, possibly causing the results to show a smaller decrease in length than what was seen in the results. Considering that no method used in previous studies can account for all the possible patterns of ligament's wrapping, this limitation has been accepted. And lastly, as occurs in most
Fig. 5. Mean profiles in knee flexion angles of the absolute distances between points, with patellar point 2 as the reference.
Table 1 Description of maximum distance variation between each pair of points in knee flexion movement. Distances
Mean
SD
N
p
Points 1–4 Points 1–5 Points 1–6 Points 1–7 Points 1–8a Points 1–9 Points 1–10a Points 2–4 Points 2–5 Points 2–6 Points 2–7 Points 2–8a Points 2–9 Points 2–10a Points 3–4b Points 3–5 Points 3–6 Points 3–7 Points 3–8a Points 3–9 Points 3–10
4.80 5.44 6.24 7.11 7.95 4.53 8.22 3.68 4.62 5.80 7.04 8.21 4.17 7.73 2.55 3.60 5.10 6.68 8.17 3.68 6.84
2.34 2.52 2.49 2.42 2.33 2.12 2.59 1.94 2.17 2.16 2.09 2.02 1.80 2.33 1.33 1.69 1.74 1.69 1.63 1.39 2.00
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
b0.001
a b
Least isometric points (not significantly different between them). Most isometric point.
12
R.G. Gobbi et al. / Clinical Biomechanics 38 (2016) 8–12
Table 2 Summary of the findings of MPFL isometry studies. Smirk and Morris
Stephen et al.
Yoo et al.
Steensen et al.
Gobbi et al. (current study)
Patella points Femur points Measurement method
3 7 Sliding suture
3 5 Sliding suture
0.03 to 0.2 mm 7 (0–120°) No, only 1.25 kg load 4 Not described Proximal and central patella and femur between ME and AT
0.01 mm accuracy 12 (0–110°) Yes 8 1.1 mm Inferior patella and central femur
3 3 Direct measurement with compass Not noted 5 (0–120°) No 11 1.1 mm Inferior patella–superior femur
3 7 Photogrammetry
Method error Number of measured angles Adequate soft tissue tensioning? Knees Least variation Most isometric points
2 4 3-D tomography, virtual ligament Not noted 5 (0–120°) No - no contraction 10 6.8 mm Central patella and nomura's
cadaveric studies, the knees had normal anatomy, and presumably no lateral maltracking, what limits the clinical application of the results in patients with lateral subluxation. The pathological tracking probably change the patterns described here and allow more wrapping around the femur. We believe that our work provides important data for the study of insertion points that knee surgeons can come across when positioning femoral graft fixation. Comparing our study with the small number of articles that explore this issue, the present study used an adequate number of knees, made use of an accurate method for measuring distances, did not violate the anatomical structures around the knee and used a muscle tensioning method that generated movement through muscle contraction. Especially soft tissue preservation and tensioning are considered of utmost importance, once its laxity and excessive dissection cause the patella to be non-anatomically positioned, causing bias to the measures. 5. Conclusion The most isometric pair of points for MPFL reconstruction was located at the patella center and anterior to the medial epicondyle, while the least isometric were located distally and posterior to the medial epicondyle, with little influence based on patellar location. Conflicts of interest The authors declare that they have no conflict of interest. References Abdel-Aziz, Y., Karara, H., 2015. Direct linear transformation from comparator coordinates into object space coordinates in close-range photogrammetry. Photogramm. Eng. Rem. S. 81, 103–107. Amis, A.A., Firer, P., Mountney, J., Senavongse, W., Thomas, N.P., 2003. Anatomy and biomechanics of the medial patellofemoral ligament. Knee 10, 215–220. http://dx.doi. org/10.1016/s0968-0160(03)00006-1. Beck, P., Brown, N.A., Greis, P.E., Burks, R.T., 2007. Patellofemoral contact pressures and lateral patellar translation after medial patellofemoral ligament reconstruction. Am. J. Sports Med. 35, 1557–1563. http://dx.doi.org/10.1177/0363546507300872. Conlan, T., Garth Jr., W.P., Lemons, J.E., 1993. Evaluation of the medial soft-tissue restraints of the extensor mechanism of the knee. J. Bone Joint Surg. Am. 75, 682–693. Deie, M., Ochi, M., Sumen, Y., Adachi, N., Kobayashi, K., Yasumoto, M., 2005. A long-term follow-up study after medial patellofemoral ligament reconstruction using the transferred semitendinosus tendon for patellar dislocation. Knee Surg. Sports Traumatol. Arthrosc. 13, 522–528. http://dx.doi.org/10.1007/s00167-005-0641-x. Desio, S.M., Burks, R.T., Bachus, K.N., 1998. Soft tissue restraints to lateral patellar translation in the human knee. Am. J. Sports Med. 26, 59–65.
0.08 mm 9 (0–120°) Yes 10 2.55 mm Central patella–1 cm anterior ME
Drez Jr., D., Edwards, T.B., Williams, C.S., 2001. Results of medial patellofemoral ligament reconstruction in the treatment of patellar dislocation. Arthroscopy 17, 298–306. http://dx.doi.org/10.1053/jars.2001.21490. Elias, J.J., Cosgarea, A.J., 2006. Technical errors during medial patellofemoral ligament reconstruction could overload medial patellofemoral cartilage: a computational analysis. Am. J. Sports Med. 34, 1478–1485. http://dx.doi.org/10.1177/0363546506287486. Ellera Gomes, J.L., Stigler Marczyk, L.R., Cesar de Cesar, P., Jungblut, C.F., 2004. Medial patellofemoral ligament reconstruction with semitendinosus autograft for chronic patellar instability: a follow-up study. Arthroscopy 20, 147–151. http://dx.doi.org/ 10.1016/j.arthro.2003.11.006. Feller, J.A., Feagin Jr., J.A., Garrett Jr., W.E., 1993. The medial patellofemoral ligament revisited: an anatomical study. Knee Surg. Sports Traumatol. Arthrosc. 1, 184–186. Fernandez, E., Sala, D., Castejon, M., 2005. Reconstruction of the medial patellofemoral ligament for patellar instability using a semitendinosus autograft. Acta Orthop. Belg. 71, 303–308. Hautamaa, P.V., Fithian, D.C., Kaufman, K.R., Daniel, D.M., Pohlmeyer, A.M., 1998. Medial soft tissue restraints in lateral patellar instability and repair. Clin. Orthop. Relat. Res. 174–182. Mikashima, Y., Kimura, M., Kobayashi, Y., Miyawaki, M., Tomatsu, T., 2006. Clinical results of isolated reconstruction of the medial patellofemoral ligament for recurrent dislocation and subluxation of the patella. Acta Orthop. Belg. 72, 65–71. Nomura, E., Inoue, M., 2006. Hybrid medial patellofemoral ligament reconstruction using the semitendinous tendon for recurrent patellar dislocation: minimum 3 years' follow-up. Arthroscopy 22, 787–793. http://dx.doi.org/10.1016/j.arthro.2006.04.078. Nomura, E., Horiuchi, Y., Kihara, M., 2000. Medial patellofemoral ligament restraint in lateral patellar translation and reconstruction. Knee 7, 121–127. Reider, B., Marshall, J.L., Koslin, B., Ring, B., Girgis, F.G., 1981. The anterior aspect of the knee joint. J. Bone Joint Surg. Am. 63, 351–356. Sandmeier, R.H., Burks, R.T., Bachus, K.N., Billings, A., 2000. The effect of reconstruction of the medial patellofemoral ligament on patellar tracking. Am. J. Sports Med. 28, 345–349. Schottle, P.B., Schmeling, A., Rosenstiel, N., Weiler, A., 2007. Radiographic landmarks for femoral tunnel placement in medial patellofemoral ligament reconstruction. Am. J. Sports Med. 35, 801–804. http://dx.doi.org/10.1177/0363546506296415. Smirk, C., Morris, H., 2003. The anatomy and reconstruction of the medial patellofemoral ligament. Knee 10, 221–227. Steensen, R.N., Dopirak, R.M., McDonald 3rd, W.G., 2004. The anatomy and isometry of the medial patellofemoral ligament: implications for reconstruction. Am. J. Sports Med. 32, 1509–1513. http://dx.doi.org/10.1177/0363546503261505. Stephen, J.M., Lumpaopong, P., Deehan, D.J., Kader, D., Amis, A.A., 2012. The medial patellofemoral ligament: location of femoral attachment and length change patterns resulting from anatomic and nonanatomic attachments. Am. J. Sports Med. 40, 1871–1879. http://dx.doi.org/10.1177/0363546512449998. Tuxoe, J.I., Teir, M., Winge, S., Nielsen, P.L., 2002. The medial patellofemoral ligament: a dissection study. Knee Surg. Sports Traumatol. Arthrosc. 10, 138–140. http://dx.doi. org/10.1007/s00167-001-0261-z. Warren, L.F., Marshall, J.L., 1979. The supporting structures and layers on the medial side of the knee: an anatomical analysis. J. Bone Joint Surg. Am. 61, 56–62. Yoo, Y.S., Chang, H.G., Seo, Y.J., Byun, J.C., Lee, G.K., Im, H., Song, S.Y., 2012. Changes in the length of the medial patellofemoral ligament: an in vivo analysis using 3-dimensional computed tomography. Am. J. Sports Med. 40, 2142–2148. http://dx.doi.org/10.1177/ 0363546512453301.
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
Evaluation of the isometry of different points of the patella and femur for medial patellofemoral ligament reconstruction☆ Riccardo Gomes Gobbi PhD a,⁎, César Augusto Martins Pereira a, David Sadigursky MD b, Marco Kawamura Demange PhD a, Luis Eduardo Passarelli Tírico PhD a, José Ricardo Pécora PhD a, Gilberto Luis Camanho PhD a a b
Hospital das Clínicas da Universidade de São Paulo, SP, Brazil Clínica ortopédica e Traumatológica, COT-Martagão, BA, Brazil
a r t i c l e
i n f o
Article history: Received 19 February 2016 Accepted 4 August 2016 Keywords: Patellar dislocation Patellofemoral joint Knee Biomechanical phenomena Range of motion Articular
a b s t r a c t Background: The location of patellar and femoral fixation of the graft in medial patellofemoral ligament reconstructions has been widely discussed. This study aimed to assess the distances between different patellar and femoral fixation points to identify the least anisometric pairs of points. Methods: Ten cadaver knees were attached to an apparatus that simulated an active range of motion of 120°, with three metallic markers fixed onto the medial side of the patella, and seven markings onto the medial epicondyle. The examined points included the proximal patella pole (1), the patellar center (3), the midpoint between points 1 and 3 (2), a point directly on the epicondyle (6), points 5 mm anterior (5) and posterior (7) to the epicondyle, points 5 mm anterior to point 5 (4) and 5 mm posterior to point 7 (8), and points 5 mm proximal (9) and distal (10) to the epicondyle. The distances between patella and femur points were measured by a photogrammetry system at 15° intervals. Findings: The pair of points that exhibited the lowest average variability in distance, and hence was the most isometric, was the patella center combined with the anterior to the medial femoral epicondyle. The pairs of points that exhibited the highest average variability in distance, and hence were the least isometric, were the ones located distal or posterior to the medial femoral epicondyle, with less influence by the patellar location. Interpretation: Surgeons should avoid positioning the graft distally or posterior to the epicondyle due to the increase in anisometry. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction The medial patellofemoral ligament (MPFL) is considered the most important patellar medial stabilizer and accounts for approximately 50–60% of patellar lateralization resistance (Conlan et al., 1993; Desio et al., 1998; Hautamaa et al., 1998). Despite this importance, valuation of its reconstruction in the treatment of patellar instability is relatively recent and has mainly occurred in the last 20 years. Despite the wide range of techniques described for its reconstruction, with different graft sources and fixation methods, evidence has been accumulating showing good clinical results for this surgery, with very low (b 10%) instability recurrence rates (Deie et al., 2005; Drez et al., 2001; Ellera Gomes et al., 2004; Fernandez et al., 2005; Mikashima et al., 2006; Nomura and Inoue, 2006). However, there are still controversies surrounding the surgical technique that hinder universal standardization of the procedure. An important and widely discussed aspect is the location of patellar and femoral fixation of the graft in reconstructions. The femoral ligament origin has been described in the adductor tubercle – AT (Conlan et al., 1993; Desio et al., 1998; Tuxoe et al., 2002), in the anterior portion of the medial
http://dx.doi.org/10.1016/j.clinbiomech.2016.08.002 0268-0033/© 2016 Elsevier Ltd. All rights reserved.
epicondyle – ME (Feller et al., 1993), on the anterior portion of the medial epicondyle distal to the adductor tubercle (Smirk and Morris, 2003), on the medial epicondyle itself (Hautamaa et al., 1998; Reider et al., 1981; Warren and Marshall, 1979) and proximal and posterior to the epicondyle, immediately distal to the adductor tubercle (Nomura et al., 2000). Variation in these points can cause errors with significant consequences to patellar biomechanics, causing increased joint pressure, pain, chondral degeneration and graft failure (Beck et al., 2007; Elias and Cosgarea, 2006; Sandmeier et al., 2000). Thus, despite MPFL behavior not being isometric, increasing tension with knee extension (Amis et al., 2003), studies are needed to analyze different patellar and femoral fixation points to identify those that offer more consistent performance. Although some studies have specifically addressed this issue, they are few in number (Smirk and Morris, 2003; Steensen et al., 2004; Stephen et al., 2012; Yoo et al., 2012), and all of them present technical limitations, such as excessive soft tissue removal, distance measurement method and number of points studied. If a ligament or stabilizing structure is removed, released or incised (such as medial retinaculum), it is not possible to adequately study isometry, because the patella may rest in a non-anatomical position, shifted
R.G. Gobbi et al. / Clinical Biomechanics 38 (2016) 8–12
laterally for example, throughout the study. Some studies used sliding sutures for obtaining distances, which add to the error of the measuring device its manipulation and positioning by humans. The present study aimed to assess the distances between different patellar and femoral MPFL fixation points along the range of motion of cadaver knees using a precise measurement method and preservation and proper tensioning of soft tissue to identify the least anisometric pairs of points. This knowledge will be useful to help surgeons predict the effect of malpositioning of the graft around the ideal points. 2. Material and methods A total of 10 cadaver knees, without throclear dysplasia and normal patellar height, with no previous injuries or procedures, were used. The knees were obtained from the Arthroscopy Laboratory of our institution after local Institutional Review Board approval (number 1115/08). The knees had the femur and tibia firmly fixed to an apparatus that applies traction through the quadriceps muscle (connected to a Kratos® K5002 universal mechanical testing machine, with a 5 tf load cell) and resistance against the tibia, simulating an active range of motion of 120° to 0° flexion (Fig. 1). Traction by the Kratos machine was progressive to bring the knee from full flexion to full extension and applied at the speed of 50 mm/min; it varied from 80 N in deep flexion to 600 N in final extension. The resistance applied to the tibia consisted of a 5,65 kg positioned in the pulley (letter U in Fig. 1). No dissection of soft tissue around the knee joint was performed, preserving all structures. On each knee, three metallic markers (nails) were fixed onto the medial side of the patella, and a cancellous screw was fixed onto the medial epicondyle (ME), onto which a thermoplastic marker with seven markings was attached. Minimum dissection was performed on these fixation points, enough only to keep the markers near the bone surface. The obtained points were the proximal pole patella (P1), the patellar midpoint (P3), the midpoint between the two previous measures on the patella (P2), a point directly on the ME (F6), points 5 mm anterior (F5) and posterior (F7) to the ME, points 5 mm anterior to F5 (F4) and 5 mm posterior to F7 (F8), and points 5 mm proximal (F9) and distal (F10) to the ME (Fig. 2). The markers positioning was based on
9
finding the medial epicondyle. It is very easy to locate it through palpation, even through intact skin. The same is also true when considering the location of the proximal and distal ends of the patella. These points were chosen because they could be easily located during surgery (all femoral points were based on the ME, which is easy to locate by palpation) to increase the clinical applicability of the findings, respecting the anatomical descriptions of the MPFL (Feller et al., 1993; Hautamaa et al., 1998; Reider et al., 1981; Smirk and Morris, 2003; Warren and Marshall, 1979). To measure the distances between different points of the patella and femur (analysis of isometry between pairs of points), a threedimensional photogrammetry system was used based on the model developed by Abdel-Aziz and Karara (2015). This system consisted of a computer software, two digital cameras and a calibrator with markers with known distances between them engineered by an industrial machine with a margin of error of 1.2 μm. The two cameras simultaneously photographed the medial portion of the knee at angles of 120°, 105°, 90°, 75°, 60°, 45°, 30°, 15° and 0° of flexion, with the aid of a digital goniometer connected to the cameras' remote controls. The system allows three-dimensional measurements between the points at different camera positions (Fig. 3). This system allowed us to obtain images to measure the distance between the patellar and femoral points at every 15° of flexion in a 0–120° range of motion. Each arc was performed three times, with 21 pairs of points (three patellar with seven femoral), with three photos for each flexion angle (27 images from each camera for each knee). The distance between points was measured with the software, and the value used for the statistical tests was the mean score of the three measures for each pair of points. An error analysis of the method was performed calculating the distances between the markers of the calibrator obtained by the software. Based on means and standard deviations of maximum distance variation through range of motion for each pair of points, analysis of variance (ANOVA) with repeated measures was applied, followed by Tukey's repeated measures test to identify which point pair had the lowest and highest distance variation compared to the other pairs.
Fig. 1. Schematic representation of the left knee positioned in the mechanical device coupled to the universal mechanical testing machine. Front and side view of the device components. A - load cell, B - steel wire, C - sheave, D - quadriceps grip guide, E - quadriceps grip, F - support table, G - femoral grip support, H - femoral rod, I - patella, J - femur, K - tibia, L - tibial grip, M spherical plain bearing, N - axial arm, O - radial arm, P - counterweight, Q - drive shaft, R - bearing housings, S - pulley, T - pulley steel wire, U - weight (5,65 kg), V - angular optical encoder (digital goniometer), W - femur markers, X - patella markers.
10
R.G. Gobbi et al. / Clinical Biomechanics 38 (2016) 8–12
mean error of 0.08 mm, with standard deviation of 0.05 mm ranging from −0.11 mm to 0.31 mm. 4. Discussion
Fig. 2. Positioning of the points used to measure P1 to P3 on the patella and F4 to F10 on the femur (femoral central point: F6).
3. Results The behavior of the mean of the three measurements obtained for each flexion angle between the point pairs studied is shown in Figs. 4–6. Analyzing the mean difference in distance variability along the knee flexion movement between all pairs of points revealed a significant difference, with p b 0.001 (Table 1). When considered individually, the point pair that exhibited the lowest average variability in distances in knee movement of 2.55 mm was 3–4 (patella center and 10 mm anterior to medial epicondyle), being the most isometric. The point pair that exhibited the highest average variability in distances of 8.22 mm in knee flexion angles was 1–10 (proximal patella and 10 mm posterior to medial epicondyle); however, it was not significantly different from point pairs 1–8, 2–8, 2–10 and 3–8 (p N 0.05). The distance was not increased between any pair of points with knee flexion, which is compatible with MPFL biomechanics. Patellar positioning had less influence on isometry. The femoral points that exhibited the highest average variability in distance, and hence were the least isometric, were the ones located distal or 10 mm posterior to the medial femoral epicondyle. The margin of error analysis compared the known distances between the calibrator markers and the same distances found by the software. After comparing 2708 obtained distances, the authors found a
The most important finding of the study was that the least isometric points for MPFL reconstruction were located distal and 10 mm posterior to the ME on the femur, with little influence based on patellar location; while the most isometric points were located at patellar center and 10 mm anterior to the ME. Correct positioning of the graft is critical because small errors can lead to significant increases in patellar joint contact pressure (Beck et al., 2007; Elias and Cosgarea, 2006). Due to the high variability of descriptions of the femoral MPFL origin, (Conlan et al., 1993; Desio et al., 1998; Feller et al., 1993; Hautamaa et al., 1998; Nomura et al., 2000; Reider et al., 1981; Smirk and Morris, 2003; Tuxoe et al., 2002; Warren and Marshall, 1979) several studies have been conducted evaluating the “isometry” of different points. Smirk and Morris (2003) conducted an isometric study of four knees using various reference points on the patella and around the medial epicondyle. Although no combination of points presented isometry, the points with the least variation in knee movement distance were on the patella, the superior third and at its equator; on the femur, one point 10 mm distal and 5 mm posterior to the adductor tubercle, and one point 5 mm posterior and another 5 mm distal to the former. Steensen et al. (2004) after dissecting the MPFLs of 11 knees, reported that the most isometric points were the most distal part of the MPFL on the patella and the most proximal of the MPFL on the femur. Stephen et al. (2012) conducted another study on MPFL isometry in eight knees and reported that the point between the medial epicondyle and the adductor tubercle was the most isometric. Positioning errors on the proximal-distal axis generated more anisometry than errors on the anterior–posterior axis. Changes in patellar insertion had little effect on MPFL isometry. A similar result was reported by Yoo et al. (2012). They measured these distances using three-dimensional knee reconstruction of 10 patients who underwent computed tomography (CT) scans at various flexion angles. A summary of the findings from these studies is presented in Table 2. In our study, we found the patellar midpoint and the point 10 mm anterior to the ME (points 3–4) to be the most isometric points (absolute measures). These findings are consistent with the points identified by Smirk and Morris (2003) and Steensen et al. (2004) in relation to the patella location. The least isometric points (absolute measures) were the most proximal to the patella and the most distal on the femur (points 1–10). Also statistically equivalent were the points posterior to the ME (points 1–8, 2–8, 2–10, and 3–8). Therefore, our findings suggest that placing the graft on the femur directly posterior or distal to the ME is risky. It should be noted that no point pair under study generated an
Fig. 3. Right knee positioned in the device coupled to the universal mechanical testing machine, with cameras in position.
R.G. Gobbi et al. / Clinical Biomechanics 38 (2016) 8–12
11
Fig. 4. Mean profiles in knee flexion angles of the absolute distances between points, with patellar point 1 as the reference.
Fig. 6. Mean profiles in knee flexion angles of the absolute distances between points, with patellar point 3 as the reference.
increase in distance between the points (tensioning) with knee flexion, compatible with MPFL biomechanics. The differences of findings among studies described in Table 2 can occur for several reasons. The studies used different measurement methods, different set of points in patella and femur, different set of flexion angles and different soft tissue tensioning. Although the present study offers a precise measurement system, a large number of insertion points and flexion angles and adequate soft tissue tensioning, the results should be critically interpreted as a whole by the readers. The present study has four main limitations. First, we did not study one of the points often cited as a femoral MPFL insertion point, located between the ME and the AT and described by Nomura et al. (2000). We did not include this point because we chose to study only points that are easily locatable during surgery and used only the ME as a parameter. We believe that to consistently place this point, we would need fluoroscopy, unavailable in our methodology. Second, despite point F9 vicinity, we know that our findings do not exactly match the most frequently recommended location for femoral insertion, described radiographically by Schottle et al. (2007). However, it is important to separate the concepts of anatomical and isometric reconstruction, which still lead to important debates in anterior cruciate ligament surgery. The option of using anatomical landmarks does not diminish the
value of knowing the behavior of points in the surrounding area. This fact is especially true when considering a structure such as the MPFL, with insertion described at various points around the ME. If a surgeon places the femoral attachment of the graft in a less than ideal position for any given reason, it is important to know the isometric behavior of these points around the medial epicondyle. Our findings suggest that if the mistake occurs towards posterior or distal, it should be repositioned. It's important to emphasize that our aim is not to replace the points described by Nomura and Schöttle as it is well known that the MPFL is not isometric. We intend to offer surgeons more information about behavior of possible points around the ideal one. The third limitation is that our method, being a linear measure, cannot account for the wrapping that the MPFL suffers around the femur at deeper flexion angles. This wrapping would cause an increase in length in deep flexion, possibly causing the results to show a smaller decrease in length than what was seen in the results. Considering that no method used in previous studies can account for all the possible patterns of ligament's wrapping, this limitation has been accepted. And lastly, as occurs in most
Fig. 5. Mean profiles in knee flexion angles of the absolute distances between points, with patellar point 2 as the reference.
Table 1 Description of maximum distance variation between each pair of points in knee flexion movement. Distances
Mean
SD
N
p
Points 1–4 Points 1–5 Points 1–6 Points 1–7 Points 1–8a Points 1–9 Points 1–10a Points 2–4 Points 2–5 Points 2–6 Points 2–7 Points 2–8a Points 2–9 Points 2–10a Points 3–4b Points 3–5 Points 3–6 Points 3–7 Points 3–8a Points 3–9 Points 3–10
4.80 5.44 6.24 7.11 7.95 4.53 8.22 3.68 4.62 5.80 7.04 8.21 4.17 7.73 2.55 3.60 5.10 6.68 8.17 3.68 6.84
2.34 2.52 2.49 2.42 2.33 2.12 2.59 1.94 2.17 2.16 2.09 2.02 1.80 2.33 1.33 1.69 1.74 1.69 1.63 1.39 2.00
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
b0.001
a b
Least isometric points (not significantly different between them). Most isometric point.
12
R.G. Gobbi et al. / Clinical Biomechanics 38 (2016) 8–12
Table 2 Summary of the findings of MPFL isometry studies. Smirk and Morris
Stephen et al.
Yoo et al.
Steensen et al.
Gobbi et al. (current study)
Patella points Femur points Measurement method
3 7 Sliding suture
3 5 Sliding suture
0.03 to 0.2 mm 7 (0–120°) No, only 1.25 kg load 4 Not described Proximal and central patella and femur between ME and AT
0.01 mm accuracy 12 (0–110°) Yes 8 1.1 mm Inferior patella and central femur
3 3 Direct measurement with compass Not noted 5 (0–120°) No 11 1.1 mm Inferior patella–superior femur
3 7 Photogrammetry
Method error Number of measured angles Adequate soft tissue tensioning? Knees Least variation Most isometric points
2 4 3-D tomography, virtual ligament Not noted 5 (0–120°) No - no contraction 10 6.8 mm Central patella and nomura's
cadaveric studies, the knees had normal anatomy, and presumably no lateral maltracking, what limits the clinical application of the results in patients with lateral subluxation. The pathological tracking probably change the patterns described here and allow more wrapping around the femur. We believe that our work provides important data for the study of insertion points that knee surgeons can come across when positioning femoral graft fixation. Comparing our study with the small number of articles that explore this issue, the present study used an adequate number of knees, made use of an accurate method for measuring distances, did not violate the anatomical structures around the knee and used a muscle tensioning method that generated movement through muscle contraction. Especially soft tissue preservation and tensioning are considered of utmost importance, once its laxity and excessive dissection cause the patella to be non-anatomically positioned, causing bias to the measures. 5. Conclusion The most isometric pair of points for MPFL reconstruction was located at the patella center and anterior to the medial epicondyle, while the least isometric were located distally and posterior to the medial epicondyle, with little influence based on patellar location. Conflicts of interest The authors declare that they have no conflict of interest. References Abdel-Aziz, Y., Karara, H., 2015. Direct linear transformation from comparator coordinates into object space coordinates in close-range photogrammetry. Photogramm. Eng. Rem. S. 81, 103–107. Amis, A.A., Firer, P., Mountney, J., Senavongse, W., Thomas, N.P., 2003. Anatomy and biomechanics of the medial patellofemoral ligament. Knee 10, 215–220. http://dx.doi. org/10.1016/s0968-0160(03)00006-1. Beck, P., Brown, N.A., Greis, P.E., Burks, R.T., 2007. Patellofemoral contact pressures and lateral patellar translation after medial patellofemoral ligament reconstruction. Am. J. Sports Med. 35, 1557–1563. http://dx.doi.org/10.1177/0363546507300872. Conlan, T., Garth Jr., W.P., Lemons, J.E., 1993. Evaluation of the medial soft-tissue restraints of the extensor mechanism of the knee. J. Bone Joint Surg. Am. 75, 682–693. Deie, M., Ochi, M., Sumen, Y., Adachi, N., Kobayashi, K., Yasumoto, M., 2005. A long-term follow-up study after medial patellofemoral ligament reconstruction using the transferred semitendinosus tendon for patellar dislocation. Knee Surg. Sports Traumatol. Arthrosc. 13, 522–528. http://dx.doi.org/10.1007/s00167-005-0641-x. Desio, S.M., Burks, R.T., Bachus, K.N., 1998. Soft tissue restraints to lateral patellar translation in the human knee. Am. J. Sports Med. 26, 59–65.
0.08 mm 9 (0–120°) Yes 10 2.55 mm Central patella–1 cm anterior ME
Drez Jr., D., Edwards, T.B., Williams, C.S., 2001. Results of medial patellofemoral ligament reconstruction in the treatment of patellar dislocation. Arthroscopy 17, 298–306. http://dx.doi.org/10.1053/jars.2001.21490. Elias, J.J., Cosgarea, A.J., 2006. Technical errors during medial patellofemoral ligament reconstruction could overload medial patellofemoral cartilage: a computational analysis. Am. J. Sports Med. 34, 1478–1485. http://dx.doi.org/10.1177/0363546506287486. Ellera Gomes, J.L., Stigler Marczyk, L.R., Cesar de Cesar, P., Jungblut, C.F., 2004. Medial patellofemoral ligament reconstruction with semitendinosus autograft for chronic patellar instability: a follow-up study. Arthroscopy 20, 147–151. http://dx.doi.org/ 10.1016/j.arthro.2003.11.006. Feller, J.A., Feagin Jr., J.A., Garrett Jr., W.E., 1993. The medial patellofemoral ligament revisited: an anatomical study. Knee Surg. Sports Traumatol. Arthrosc. 1, 184–186. Fernandez, E., Sala, D., Castejon, M., 2005. Reconstruction of the medial patellofemoral ligament for patellar instability using a semitendinosus autograft. Acta Orthop. Belg. 71, 303–308. Hautamaa, P.V., Fithian, D.C., Kaufman, K.R., Daniel, D.M., Pohlmeyer, A.M., 1998. Medial soft tissue restraints in lateral patellar instability and repair. Clin. Orthop. Relat. Res. 174–182. Mikashima, Y., Kimura, M., Kobayashi, Y., Miyawaki, M., Tomatsu, T., 2006. Clinical results of isolated reconstruction of the medial patellofemoral ligament for recurrent dislocation and subluxation of the patella. Acta Orthop. Belg. 72, 65–71. Nomura, E., Inoue, M., 2006. Hybrid medial patellofemoral ligament reconstruction using the semitendinous tendon for recurrent patellar dislocation: minimum 3 years' follow-up. Arthroscopy 22, 787–793. http://dx.doi.org/10.1016/j.arthro.2006.04.078. Nomura, E., Horiuchi, Y., Kihara, M., 2000. Medial patellofemoral ligament restraint in lateral patellar translation and reconstruction. Knee 7, 121–127. Reider, B., Marshall, J.L., Koslin, B., Ring, B., Girgis, F.G., 1981. The anterior aspect of the knee joint. J. Bone Joint Surg. Am. 63, 351–356. Sandmeier, R.H., Burks, R.T., Bachus, K.N., Billings, A., 2000. The effect of reconstruction of the medial patellofemoral ligament on patellar tracking. Am. J. Sports Med. 28, 345–349. Schottle, P.B., Schmeling, A., Rosenstiel, N., Weiler, A., 2007. Radiographic landmarks for femoral tunnel placement in medial patellofemoral ligament reconstruction. Am. J. Sports Med. 35, 801–804. http://dx.doi.org/10.1177/0363546506296415. Smirk, C., Morris, H., 2003. The anatomy and reconstruction of the medial patellofemoral ligament. Knee 10, 221–227. Steensen, R.N., Dopirak, R.M., McDonald 3rd, W.G., 2004. The anatomy and isometry of the medial patellofemoral ligament: implications for reconstruction. Am. J. Sports Med. 32, 1509–1513. http://dx.doi.org/10.1177/0363546503261505. Stephen, J.M., Lumpaopong, P., Deehan, D.J., Kader, D., Amis, A.A., 2012. The medial patellofemoral ligament: location of femoral attachment and length change patterns resulting from anatomic and nonanatomic attachments. Am. J. Sports Med. 40, 1871–1879. http://dx.doi.org/10.1177/0363546512449998. Tuxoe, J.I., Teir, M., Winge, S., Nielsen, P.L., 2002. The medial patellofemoral ligament: a dissection study. Knee Surg. Sports Traumatol. Arthrosc. 10, 138–140. http://dx.doi. org/10.1007/s00167-001-0261-z. Warren, L.F., Marshall, J.L., 1979. The supporting structures and layers on the medial side of the knee: an anatomical analysis. J. Bone Joint Surg. Am. 61, 56–62. Yoo, Y.S., Chang, H.G., Seo, Y.J., Byun, J.C., Lee, G.K., Im, H., Song, S.Y., 2012. Changes in the length of the medial patellofemoral ligament: an in vivo analysis using 3-dimensional computed tomography. Am. J. Sports Med. 40, 2142–2148. http://dx.doi.org/10.1177/ 0363546512453301.