Surgical techniques for reconstruction of the posterior cruciate ligament have improved as the anatomy and biomechanics of this ligament have become better understood. Conventionally, the transtibial tunnel technique has been used. This technique has a "killer turn" at the posterior orifice of the tibial tunnel, which may lead to friction and stretching, followed by failure of the graft1. This method of posterior cruciate ligament reconstruction often produces unsatisfactory results with residual knee laxity2-4.
The tibial inlay method for reconstruction of the posterior cruciate ligament was introduced in 19955 and has evolved into an arthroscopic procedure6. The inlay technique can theoretically avoid the "killer turn" by direct fixation of the bone plug on the posterior aspect of the tibia.
The focus of conventional posterior cruciate ligament reconstruction has been on reconstructing the anterolateral bundle. However, recent biomechanical studies have shown that the double-bundle technique restores normal knee kinematics more effectively than do single-bundle techniques7,8.
Biomechanical studies have revealed that the tibial inlay technique has substantial advantages1,5 and that the double-bundle technique has a theoretical advantage as well4,5. However, the preferred surgical procedure has not been determined in a clinical setting.
The purpose of this study was to compare the clinical results of arthroscopic tibial inlay single-bundle and double-bundle techniques with those of the conventional transtibial single-bundle technique for posterior cruciate ligament reconstruction.
The general inclusion criteria for the study were (1) isolated posterior knee instability of greater than grade 2 (>10 mm), (2) an uninjured contralateral knee, (3) no previous surgery on the affected knee, and (4) no fracture around the knee.
Twenty-nine patients who had undergone an isolated posterior cruciate ligament reconstruction, met the above inclusion criteria, and had been followed for more than twenty-four months were the subjects of the study. All patients had pain or a feeling of instability in the affected knee before the surgery. Revision surgical cases and patients with generalized laxity were excluded.
Our retrospective study included eight patients treated with the transtibial single-bundle technique (Group T) from February 1998 to March 2004 and twenty-one patients who underwent the arthroscopic tibial inlay technique from August 2001 to March 2004, with the latter group including eleven single-bundle procedures (Group I1) and ten double-bundle procedures (Group I2). The average period of follow-up was 46.4 months in Group T, 36.3 months in Group I1, and 29.4 months in Group I2. All of the arthroscopic procedures were performed by one surgeon (S.-J.K.), and an Achilles tendon allograft was used in all twenty-nine patients.
The study population comprised five men and three women in Group T, eight men and three women in Group I1, and seven men and three women in Group I2. At the time of the operation, the average age was 32.4 years in Group T, 31.9 years in Group I1, and 33.6 years in Group I2.
The causes of the injuries included sixteen motor-vehicle accidents, nine sports injuries, and four accidental falls. The mean time from the injury to the operation was 9.4 months. Associated injuries included a lateral meniscal tear (five cases), a medial meniscal tear (two cases), and a small chondral defect of the lateral femoral condyle (one case) (Table I).
We evaluated Lysholm scores9, the range of motion, and Telos stress radiographs of all patients at the time of follow-up.
A positive posterior drawer test performed with the knee in 90° of flexion was our major criterion for posterior instability. With grade-2 instability, the tibial plateau ends up being displaced flush with the medial femoral condyle, which corresponds to 5 to 10 mm of posterior tibial translation compared with that of the contralateral knee. With grade-3 instability, the medial tibial plateau is displaced posterior to the medial femoral condyle, which corresponds to >10 mm of posterior tibial translation compared with that of the contralateral knee. All twenty-nine patients had a grade-3 posterior drawer test and a positive posterior sag test. Valgus and varus stress radiographs were made to evaluate collateral ligament laxity. Posterolateral instability was checked with use of the reverse shift test and the external rotation recurvatum test.
Radiographic Evaluation
We obtained posterior stress radiographs with use of a Telos device (Telos, Marburg, Germany) with a 35-lb (15.9-kg) posterior load applied to the proximal part of the tibia at 90° of knee flexion10. The stress radiographs were analyzed with use of the posterior aspects of the femoral and tibial condyles as peripheral osseous landmarks, as described by Stäubli and Jakob11. After a line was drawn parallel to the tibial plateau, a vertical line at the midpoint of two lines originating from the most posterior portions of the medial and lateral tibial condyles was drawn. Another vertical line was then drawn in the same manner, at the midpoint of two lines originating from the posterior portions of both femoral condyles, to avoid a rotational error due to tibial motion10,11. Posterior displacement was measured between the femoral and tibial vertical midpoint lines with use of a computerized radiographic system (Centricity Enterprise Web, version 2.0; GE Healthcare, Milwaukee, Wisconsin).
To assess the reliability of the postoperative Telos stress radiographs, each radiograph was measured three times by three different orthopaedic surgeons. The average of the three individual mean values was used to determine the final side-to-side difference in posterior tibial translation.
Operative Technique
The transtibial and arthroscopic tibial inlay methods used in this study were previously described by one of us (S.-J.K.) and colleagues6,12,13.
Transtibial Single-Bundle Technique
A tibial tunnel, 11 mm in diameter, was made at the anterolateral cortex of the proximal part of the tibia. The knee was held in 110° of flexion while the proximal part of the tibia was forced backward, and a plastic sheath was placed in contact with the lateral femoral condyle. Then, the femoral socket was created at the one o'clock position in a right knee (and at the eleven o'clock position in a left knee), 2 to 3 mm proximal to the articular junction. A 9-mm EndoPearl device (Linvatec, Largo, Florida) was attached to the whipstitched tendon side of the prepared Achilles tendon allograft. The EndoPearl side of the graft was passed through the tibial tunnel and into the femoral socket and was then secured by an absorbable interference screw. Subsequent tibial fixation was accomplished with an absorbable interference screw (Figs. 1-A and 1-B).
Arthroscopic Tibial Inlay Technique (Single-Bundle and Double-Bundle)
In the single-bundle procedure (Figs. 2-A and 2-B), the tibial tunnel was made as described for the transtibial method. The Achilles tendon bone plug for the tibial tunnel inlay was customized into a cylindrical shape, 11 mm in diameter and 15 mm in height, perpendicular to the fibers of the Achilles tendon with use of a cannulated coring reamer (Arthrex, Naples, Florida). The femoral socket was made in the same manner as used with the transtibial technique. The skin incision for the parapatellar anteromedial portal was extended to 2.5 cm to allow easy passage of the graft. The bone plug was pulled distally to pass through the extended parapatellar anteromedial portal and the intercondylar notch and was engaged into the tibial tunnel. The graft was then fixed with an absorbable interference screw and an additional suture washer on the proximal tibial cortex. Femoral fixation was then obtained with a second absorbable interference screw while the knee was held in 90° of flexion.
With the double-bundle technique (Figs. 3-A and 3-B), a 9-mm headed reamer was used for the anterolateral femoral socket and a 7-mm reamer was used for the posteromedial femoral socket, which was located 4 to 5 mm posterior to the articular junction at the three o'clock position in the right knee (and at the nine o'clock position in the left knee). The proximal width of the Achilles tendon allograft was divided at a ratio of 2 to 3 (from medial to lateral) and split distally along its fibers into superficial and deep layers. A 9-mm EndoPearl device was attached to the deep-layer bundle, and a 7-mm EndoPearl device was attached to the superficial-layer bundle. The deep-layer bundle was fixed in the anterolateral socket with an absorbable interference screw placed through the far anterolateral portal while the knee was in 90° of flexion, and the proximal part of the tibia was drawn forward. The superficial-layer bundle was fixed in the posteromedial femoral tunnel with an absorbable interference screw inserted from the far anterolateral portal with use of a flexible screw driver with the knee in 90° of flexion. When the screw became well engaged, the rest of it was advanced with the knee in 30° or 45° of flexion to obtain final fixation.
Postoperative Rehabilitation
A leg splint in full extension was worn by all patients for two weeks. Quadriceps-strengthening and straight-leg-raising exercises were initiated on the first postoperative day, and partial weight-bearing was permitted with crutches. Motion to 90° was performed as tolerated two weeks after the operation. Closed-chain exercises were allowed at six weeks postoperatively. A posterior cruciate ligament brace was worn for six weeks after removal of the leg splint. A low-impact sports program was initiated at six months after the operation.
Statistical Methods
Two-sample nonparametric Mann-Whitney tests were used to compare Lysholm scores, side-to-side differences in range of motion, and side-to-side differences in posterior translation between Groups I2 and T and between Groups I1 and T. Wilcoxon signed-rank tests were used to compare the preoperative and final Lysholm scores in each group. All tests were performed with a 95% confidence level; p < 0.05 was considered significant.
Source of Funding
There was no external funding source for this study.
Preoperative Evaluation
The mean preoperative Lysholm scores (and standard deviation) were 51.3 ± 6.7 points in Group T, 49.3 ± 6.0 points in Group I1, and 51.3 ± 7.3 points in Group I2. There was no significant difference between Groups I2 and T (p = 0.929) or between Groups I1 and T (p = 0.591).
Follow-up Evaluation
Mean Side-to-Side Difference in Posterior Tibial Translation
The mean side-to-side differences in posterior tibial translation as measured with Telos stress radiography were 5.6 ± 2.00 mm in Group T, 4.7 ± 1.62 mm in Group I1, and 3.6 ± 1.43 mm in Group I2 (Table II). There was a significant difference between Groups I2 and T (p = 0.023). However, there was no significant difference between Groups I1 and T (p = 0.374) (Table III).
When Groups I1 and I2 were merged into one arthroscopic inlay group and compared with Group T (the transtibial group), no significant difference in the side-to-side difference in posterior translation could be identified (p = 0.076).
Range of Motion
Final examination with a goniometer showed the mean side-to-side difference in knee flexion to be 2.8° ± 0.70° in Group T, 4.1° ± 2.59° in Group I1, and 3.4° ± 0.84° in Group I2. With the numbers studied, there was no significant difference between Groups I2 and T (p = 0.102) or between Groups I1 and T (p = 0.343) with regard to this parameter. We also did not find a significant difference between the combined arthroscopic inlay group and the transtibial group (p = 0.154). None of the patients had an extension deficit at their final follow-up examination.
Lysholm Knee Score
At the final postoperative evaluation, the mean Lysholm scores were 86.8 ± 7.53 points in Group T, 79.7 ± 11.67 points in Group I1, and 84.3 ± 9.74 points in Group I2. No significant difference in this score was identified between Groups I2 and T (p = 0.755) or between Groups I1 and T (p = 0.137), and no significant difference was identified between the combined arthroscopic inlay group and the transtibial group (p = 0.294). However, the mean postoperative clinical Lysholm score was significantly improved compared with the preoperative score in all three groups (p = 0.012 for Group T, p = 0.003 for Group I1, and p = 0.005 for Group I2).
Intraoperative Complications
There were two intraoperative complications in the arthroscopic inlay group. In one patient, the cylindrical bone plug broke during screw fixation in the tibial tunnel. The remaining portion of the bone plug was stabilized by fixation with extra interference screws. The other complication was a rupture of the suture securing the bone plug during the tensioning procedure. In this case, immobilization was used postoperatively for an additional four weeks.
Many operative techniques, varying with regard to graft selection and fixation pattern, have been described for posterior cruciate ligament reconstruction. Unlike anterior cruciate ligament surgery, posterior cruciate ligament reconstruction is still associated with problems such as residual posterior laxity and inconsistent functional results1-4. The major concerns regarding posterior cruciate ligament reconstruction are the biomechanical advantages of transtibial over tibial inlay techniques14-17 and of single-bundle over double-bundle techniques7,8,18,19.
With the transtibial technique, excessive angulation of the tendon graft in the posterior orifice of the tibial tunnel can cause failure of the graft20. Several techniques such as creating an anterolateral tibial tunnel, chamfering the tibial bone tunnel exit, and aperture fixation of the soft-tissue graft have been advocated to minimize this problem of the "killer turn."20-23 In this study, we used a one-incision transtibial technique, which creates a femoral socket from the inside-out. Dunlop et al.21 and Handy et al.24 discussed concerns regarding the sharper graft/femoral tunnel angle used with the inside-out technique in their cadaver study and recommended an outside-in technique for femoral tunnel placement. One of us (S.-J.K.) and Min12 solved this problem with high flexion of the knee, drawing the tibia back more than 10 mm and placing the reamer as close as possible to the lateral femoral condyle. Because the surgical indication for posterior cruciate ligament reconstruction is instability with greater than grade-2 posterior displacement of the tibia, this approach is practical and a satisfactory femoral tunnel can be created.
Berg5 reported on posterior cruciate ligament reconstruction with an open tibial inlay technique with use of a popliteal fossa exposure to avoid the problem of ligament graft-tunnel margin abrasion. Use of this method was supported by biomechanical studies in 20007 and 20011, although more contemporary studies have not shown that inlay methods have a biomechanical advantage14,16,17. Also, Seon and Song25 and MacGillivray et al.26 noted no significant difference in the two-year clinical results between the open tibial inlay single-bundle technique and the arthroscopic transtibial single-bundle technique.
The open tibial inlay technique carries with it the morbidity of a posterior capsulotomy27, and one of us (S.-J.K.) and colleagues presented the all-arthroscopic inlay method in 20046 to overcome the disadvantages of creating an arthrotomy in the open inlay technique. Mariani and Margheritini28 also reported on the arthroscopic inlay technique. They used a rectangular bone-block graft and fixed it in the tibial groove by tying a suture over the anterior tibial cortex. Campbell et al.29 found that the suture fixation technique for a tibial inlay posterior cruciate ligament reconstruction appears to approximate the early strength of inlay screw fixation. They concluded that the arthroscopic inlay technique, even when done with use of suture fixation, may offer initial fixation stability similar to that of the open inlay technique without the need for a posterior capsulotomy. We used a cylindrical bone-block graft fixed with an interference screw in the tibial tunnel combined with suture washer fixation. We found that the mean side-to-side difference in posterior tibial translation did not differ significantly between the arthroscopic inlay single-bundle group and the transtibial single-bundle (conventional) group.
Race and Amis8 and Harner et al.7 demonstrated the biomechanical superiority of double-bundle posterior cruciate ligament reconstruction over single-bundle reconstruction in their cadaver studies. In 2001, Stähelin et al.30 reported on double-bundle posterior cruciate ligament reconstruction with use of hamstring tendons. Since then, there have been many comparisons of the clinical results between double-bundle and single-bundle transtibial reconstructions31-33, with no significant differences found. Bergfeld et al. performed a cadaver study of the tibial inlay technique and found no biomechanical advantages of a double-bundle graft over a single-bundle graft18. Their findings differed from those of Valdevit et al.19, who showed that a double-bundle graft had mechanical advantages. In 2005, one of us (S.-J.K.) and Park described an arthroscopic tibial inlay double-bundle technique13. Campbell et al.34 reported a similar technique, in which they used a bifid patellar tendon allograft to obtain bone-to-bone fixation in outside-in femoral tunnels and suture fixation of rectangular tibial bone blocks. In the current study, we used a bifid Achilles tendon allograft and an inside-out femoral tunnel with interference screw aperture fixation.
Zehms et al.35 presented the results of a cadaver study showing that the arthroscopic inlay double-bundle technique provides stability comparable with that provided by the open inlay technique with a potential for less operative morbidity. While, to our knowledge, no one has reported the results of a clinical comparison of arthroscopic inlay double-bundle and single-bundle reconstructions, the arthroscopic inlay double-bundle reconstruction used in our study produced better posterior stability than the conventional (transtibial single-bundle) method.
For the purpose of focusing on posterior stability, we analyzed patients with isolated posterior cruciate ligament instability without generalized laxity. To minimize treatment variables, one surgeon used the same graft (an Achilles tendon allograft) and the same fixation method (with interference screws) in all three groups. We compared the methods only with regard to the restoration of posterior stability. A limitation of our study was the small number of cases that were analyzed because of the rarity of isolated posterior cruciate ligament injury without combined instability. The retrospective study design was also a limitation of our study.
In conclusion, we prefer to use the arthroscopic tibial inlay double-bundle technique for posterior cruciate ligament reconstruction because better stability can be achieved with that method. 