A retrospective, comparative cohort study was conducted on patients who underwent 152 primary PCL reconstruction surgeries performed by a single surgeon between April 2004 and October 2008. Grade II PCL injuries combined with other ligamentous or grade III PCL injuries were considered for surgical treatment when the patients had persistent pain or functional disability (discomfort when going down stairs). In this situation, the status of the remnant PCL tissue (determined using MRI) was considered before surgery in order to select the most appropriate reconstruction option. The remnant PCL-preserving stent procedure using the transtibial tunnel technique was chosen if there was partial continuity or the remnant tissue was sufficient in the acute (< 3 months) or subacute stage (3-6 months). Remnant PCL tensioning and ALB reconstruction with the modified inlay technique was selected if there was abundant remnant tissue with continuity for tensioning in the chronic stage (> 6 months). A double-bundle reconstruction was performed if there was no remnant PCL or a very weak PCL remnant according to the MRI and arthroscopic findings in the subacute or chronic stage.

Thirty-three patients were excluded because they met one of the following exclusion criteria: 1) concomitant anterior cruciate ligament (ACL) or medial collateral ligament reconstruction (12 patients); 2) an associated fracture in the lower extremities that could affect knee function (2 patients); 3) combined severe life-threatening medical disease (1 patient); and 4) revision surgery (8 patients). Moreover, ten patients were also excluded because of loss to outpatient follow-up before 24 months.

The grading of PLC injuries has been described elsewhere (Table 1).15,19) To minimize the effect of combined PLC injuries, only patients with grade II PLC injury treated by PLC sling (modified Larson technique) simultaneously with a PCL reconstruction were enrolled in the present study. Twenty patients, assessed as having grade I or III PLC injuries, were excluded (Table 2). Therefore, the records of 89 patients, who had been followed up for a minimum of 24 months after PCL reconstruction were included in the analysis.

The study was approved by the Institutional Review Board at Chung-Ang University Hospital, and all patients provided informed consent. Out of 89 patients, 34 were treated with a PCL remnant-preserving ALB reconstruction using the transtibial tunnel technique in the acute and subacute stage of injury (group 1); 40 patients were treated with remnant tensioning and ALB reconstruction using the modified inlay technique in the chronic stage of injury (group 2); and 15 patients were treated with double-bundle reconstruction using the modified inlay technique (group 3). The mean age of patients in group 1 (32 males and 2 females), group 2 (36 males and 4 females), and group 3 (14 males and 1 female) was 31.1 years (range, 18 to 51 years), 31.2 years (range, 16 to 59 years), and 32.1 years (range, 19 to 24 years), respectively. An Achilles tendon allograft was used in the PCL double-bundle reconstruction and an autogenous hamstring tendon was used in the other two groups. The PLC reconstruction was performed with a contralateral autogenous hamstring tendon through the fibular head tunnel in all patients.

The surgical technique has been described elsewhere6,20) and we describe only the specific technical aspects of arthroscopically-assisted PCL reconstruction here. In order to prepare the femoral tunnel without removing the remains of the PCL, the surgeon placed the tip of the guide pin on the PCL at the 12:30 to 1 o'clock position for the right knee while looking through the anterolateral portal. After placing the femoral guide (Acufex Microsurgical, Mansfield, MA, USA) 5 to 6 mm proximal to the margin of the articular cartilage of the medial femoral condyle at the pin, a tunnel was made using the outside-in technique. A posteromedial portal was made to ensure proper preparation of the tibial tunnel. A switching stick was then inserted through the posteromedial portal via an arthroscope sleeve to reach the septum. Under direct visualization, the stick was gently pushed into the posterolateral compartment by piercing the posterior septum just posterior to the PCL remnant and then it was pushed until it reached the capsule at the posterolateral portal site. A skin incision was made at this point, which was approximately 1 cm posterior to the lateral femoral condyle and 1 cm proximal to the joint line, to create the posterolateral portal (modified transeptal portal). A motorized shaver was inserted through the posteromedial portal to reach the posteromedial compartment under direct visualization of the arthroscope which had been inserted earlier through the posterolateral portal. The posterior capsule was carefully separated from the PCL with the shaver and electric VAPR (DePuy Mitek, Raynham, MA, USA), starting from the articular surface level to the tibial attachment, without disrupting the remnant PCL bundle. It was important to keep the shaver's cutting surface in the anteroinferior direction to avoid damaging any posterior neurovascular structure. While looking arthroscopically through the posteromedial or posterolateral portal using a C-arm, the surgeon introduced the tip of the PCL tibial guide (Acufex Microsurgical) through the anteromedial portal and advanced it to the back side of the PCL. The surgeon positioned the guide 1.5 cm below the articular surface just distal to the tibial insertion of the PCL in order to avoid damaging the remnant PCL attachment. The drill guide angle of the tibia was oriented at 55° to 60°, and the tunnel diameter was determined as the same size as the graft diameter. The autogenous hamstring quadruple graft was passed through the femoral tunnel first, and then passed through the tibial tunnel with secured about the soft tissue impingement. For the tibial side fixation, a Rigidfix guide (DePuy Mitek) insert and two cross pin guides were placed into the lateral side of the tibial plateau. The graft tendon was fixed at the tibial tunnel, first with Rigidfix. After the tension and accurate positioning of the graft were assessed arthroscopically, the graft was cyclic loaded in tension with 15 to 20 lb and fixed with a biodegradable interference screw at the femoral tunnel additionally, and a post and tie were made with a screw and washer in both the tibial and femoral sides with the knee joint flexed at 90°.

The surgical technique has been described elsewhere,19,21-24) and we only describe the specific technical aspects of the arthroscopically-assisted PCL reconstruction here. The femoral tunnel was made in the same manner as the transtibial tunnel (stent operation). For the posterior approach, the knee is flexed 70° to 90° to provide easier access to the popliteal area and the operating table is tilted over 30° so that the operated knee is lower than the contralateral knee. After the posteromedial approach,25) the PCL tibial attachment was demarcated as a 1.5 × 2 cm area using an osteotome, and a 7-mm-thick bone block was detached from the distal to the proximal area using a 1.2 to 1.5 cm wide curved osteotome. A bony trough was made at the medial side of the PCL tibial insertion just distal to the portion of the detached bone block site to which the detached bone block and graft were to be fixed. The autogenous hamstring quadruple graft was passed through the knee joint to the femoral tunnel and fixed to the cortical bone of the distal and medial side of the tibial insertion of the PCL with a 10-mm staple. The remnant PCL was tensioned by pulling distally and the step-off as an anatomic reduction was confirmed at the anteromedial aspect of the tibiofemoral condyle. The remnant PCL was fixed through the tibia using a 5- or 6.5-mm cannulated screw with a spiked washer. The graft was cyclically loaded in tension with 15 to 20 lb of tension and fixed with a biodegradable interference screw at the femoral tunnel with the knee joint flexed 70° to 90° and additionally fixed with a 10-mm staple or post-tie on the femoral side.

For PCL double-bundle reconstruction surgery, the double femoral tunnel and modified tibial inlay technique with an allograft tendon was employed in all 15 patients. For the allograft, a tibial bone block with two femoral bundles was prepared. The Achilles tendon graft was transected sagittally into two strands, each sutured with No. 5 non-absorbable sutures using a locking whipstitch to to be 6 to 8 mm in diameter (Fig. 1). The ALB was made larger than the posteromedial bundle (PMB). The Achilles tibial bone block was 20 mm long, 15 mm wide, and 7 to 8 mm thick.

The femoral tunnel of the ALB was made in the same manner as the other two techniques. The femoral tunnel of the PMB was made from an accessory anterolateral portal using an inside-out technique, where the desired femoral tunnel was placed 8 mm proximal to the margin of the articular cartilage of the medial femoral condyle at the 3 (right knee) or 9 o'clock position (left knee) (Fig. 2). The original PCL remnant bundle was preserved where possible. Tibial fixation was achieved in the same manner with a modified inlay operation. After the posteromedial approach with a tilting operation table and lower down affected side which was flexion, abduction and external rotation of the hip and 90° flexion of the knee position on the ancillary table, the tibial attachment point of the PCL was demarcated and a bony trough was made just distal to the PCL tibial insertion where the detached bone block was to be fixed. The two femoral Achilles allograft bundles were passed through the knee joint to the femoral tunnel; the PMB was passed first, followed by the ALB. The Achilles bone block was fixed to the tibia using a 5- or 6.5-mm cannulated screw with a spiked washer.

The postoperative rehabilitation regimen has been described elsewhere,6,19) and it was customized for each patient depending on the rigidity of graft fixation, degree of isometricity, and results of the final intraoperative stress test after the reconstruction. Since the PLC reconstructions were performed at the same time as the PCL reconstructions, passive range of motion exercises usually started on the 3rd to 5th postoperative day. During the first six weeks, the range of motion was increased from 0° to 90° and the full range of motion was regained by the 12th to 24th postoperative week. The knee was immobilized with a long leg removable splint for two weeks after surgery. The PCL brace,26) which is made up with a tibial supporter including an elastic spring at the posterior aspect of the proximal tibia, was used until the 6th postoperative week. Full weight bearing without the use of crutches was generally achieved within the 8th postoperative week.

To compare the clinical and stability outcomes of the PCL reconstructions, serial postoperative evaluations were performed at 6 weeks, 3, 6, and 12 months, and every 12 months thereafter. Knee stability was evaluated by posterior stress radiography (push view) performed using a Telos stress device (Austin & Associates, Fallston, MD, USA) and by a maximum manual displacement test using a KT-1000 arthrometer (Instrumented Drawer Testing, KT-1000, MEDmetric, San Diego, CA, USA). The tests were performed preoperatively and at every follow-up after the 3rd postoperative month. The subjective and objective International Knee Documentation Committee (IKDC) score and Orthopädische Arbeitsgruppe Knie (OAK) score were used for the clinical evaluation. All evaluations were made by a single observer not involved in the operation. To test posetrolateral rotator instability, the posterolateral drawer test, dial test, and varus stress test were performed at every follow-up, and the results were compared with the results from the contralateral side, and classified as overconstrained, constrained and lax.19)

A power analysis was performed. If the subjective scores differed by more than 10 points and the posterior stress radiograph results differed by more than 2 mm, they were considered clinically different results. To achieve a power of 80% under a significance level of 5%, a sample size of 14 cases was needed for each study group. The Wilcoxon signed rank was used to analyze the pre- and postsurgery data. The statistical comparison between the 3 groups was made using a Kruskal-Wallis test. All statistical analyses were performed using SPSS ver. 10.0 (SPSS Inc., Chicago, IL, USA), and SAS ver. 9.1 (SAS Inc., Cary, NC, USA). Statistical significance was set at p < 0.05.

For the 34 patients in group 1, the mean side-to-side difference in posterior translation, as measured by posterior stress radiography, improved from 10.1 ± 2.5 mm preoperatively to 2.3 ± 1.4 mm at the last follow-up (p < 0.001). At the last evaluation, 23 patients (67.1%) exhibited a displacement of less than 3 mm, 8 patients (23.5%) had between a 3 and 5 mm displacement, and 3 patients (8.8%) showed displacement exceeding 5 mm (Fig. 3). The mean side-to-side difference as measured by the maximal manual test using the KT-1000 arthrometer also improved from 6.8 ± 2.0 mm preoperatively to 2.2 ± 2.2 mm at the last follow-up (p < 0.001).

For 40 patients in group 2, the mean side-to-side difference in posterior translation, as measured by posterior stress radiography, improved from 10.6 ± 2.4 mm preoperatively to 2.3 ± 1.5 mm at the last follow-up (p < 0.001). At the last evaluation, 23 patients (57.5%) exhibited a displacement of less than 3 mm, 15 patients (37.5%) had between a 3 and 5 mm displacement, and 2 patients (5.0%) showed displacement exceeding 5 mm. The mean side-to-side difference, as measured by the maximal manual test using the KT-1000 arthrometer, also improved from 8.4 ± 2.2 mm preoperatively to 2.0 ± 1.4 mm at the last follow-up evaluation (p < 0.001).

For the 15 patients in group 3, the mean side-to-side difference in posterior translation, as measured by posterior stress radiography, improved from 12.8 ± 3.2 mm preoperatively to 4.0 ± 2.5 mm at the last follow-up (p < 0.001). At the last evaluation, 3 patients (20%) exhibited a displacement of less than 3 mm, 9 patients (60%) had between a 3 and 5 mm displacement, and 3 patients (20%) showed displacement exceeding 5 mm. The mean side-to-side difference, as measured by the maximal manual test using the KT-1000 arthrometer, also improved, from 7.7 ± 2.2 mm preoperatively to 3.6 ± 1.9 mm at the last follow-up (p < 0.001).

One case in group 1 and one in group 3 showed a displacement greater than 10 mm compared to the normal side. Statistical analyses revealed both group 1 and 2 showed similar posterior stress radiography results or KT-1000 arthrometer stability. However, the results of group 3 were inferior to groups 1 and 2 at the last follow-up evaluation (p = 0.022).

At the last evaluation, rotational stability was assessed according to different knee flexion values (30° and 90° of flexion): 4 knees were overconstrained, 27 were constrained, and 3 were lax as compared with the normal side in group 1. In group 2, 6 knees were overconstrained, 31 were constrained, and 3 were lax. In group 3, 3 knees were overconstrained, 9 were constrained and 3 were lax. There were no statistical differences among the three groups in terms of rotational stability (p = 0.214).

The mean OAK score improved significantly for the 34 patients in group 1 from 71.7 ± 9.3 preoperatively to 85.0 ± 6.7 postoperatively (p < 0.001). The mean IKDC subjective score improved significantly from 46.7 ± 16.6 preoperatively to 65.1 ± 18.4 postoperatively (p < 0.001). With regard to the IKDC objective evaluation, 11 patients (32.4%) and 23 patients (67.6%) were rated as C (abnormal) and D (severely abnormal), respectively, at the preoperative evaluation. At the last evaluation, 21 patients (61.8%), 9 patients (26.5%), 3 patients (8.8%), and 1 patient (2.9%) were rated as A (normal), B (nearly normal), C and D, respectively (Fig. 4). Therefore, 88.3% of the patients had a rating of either A or B at the last evaluation.

For the 40 patients in group 2, the mean OAK score improved significantly from 63.5 ± 10.4 preoperatively to 88.9 ± 7.6 postoperatively (p < 0.001). The mean IKDC subjective score improved significantly from 63.5 ± 10.4 preoperatively to 79.7 ± 13.3 postoperatively (p < 0.001). With regard to the final IKDC evaluation, 13 patients (32.5%) and 27 patients (67.5%) were rated as C and D, respectively, at the preoperative evaluation. At the last evaluation, 17 patients (42.5%), 19 patients (47.5%), and 4 patients (10%) were rated as A, B, and C, respectively. Hence, 90% of the patients had a rating of either A or B at the last evaluation.

The mean OAK score for the 15 patients in group 3 improved significantly from 71.3 ± 12.9 preoperatively to 83.0 ± 5.9 postoperatively (p < 0.001). The mean IKDC subjective score improved significantly from 50.7 ± 17.6 preoperatively to 61.5 ± 12.9 postoperatively (p < 0.001). With regard to the final IKDC evaluation, 6 patients (40%) and 9 patients (60%) were rated as C and D, respectively, at the preoperative evaluation. At the last evaluation, 3 patients (20%), 9 patients (60%), 2 patients (13.3%), and 1 patient (6.7%) were rated as A, B, C, and D, respectively. Hence, 80% of the patients had a rating of either A or B at the last evaluation (Fig. 4). There was no a significant difference among the three groups in terms of the clinical results, namely the OAK score and IKDC objective and subjective scores.