Patient Selection and Study Design
Thirty-four patients were referred to our center between February 2004 and May 2008 for treatment following an unsuccessful PCL reconstruction, defined as recurrent pathologic knee laxity that was painful during daily or athletic activities. Laxity was considered to be pathologic if posterior stress radiography utilizing a TELOS stress device (Telos Medical/Austin & Associates, Fallston, Maryland) demonstrated a side-to-side difference of >5 mm in posterior translation.
Revision PCL reconstruction was not performed in six of these patients because of advanced arthritis or because the patient declined further treatment. The remaining twenty-eight patients underwent revision PCL reconstruction performed by a single surgeon. Conservative treatment consisting of physiotherapy and quadriceps strengthening exercises for at least three months had been unsuccessful in these patients, and they continued to have persistent knee pain, functional disability, and knee instability that prevented them from returning to activities such as running and descending stairs.
Two of these patients underwent arthroscopic PCL reconstruction involving tensioning of the PCL remnant, and another two patients underwent double-bundle PCL reconstruction involving the transtibial technique and graft material other than Achilles tendon allograft; these four patients and another two patients who were lost to follow-up were excluded from the study. Thus, twenty-two patients who had a minimum of twenty-four months of follow-up and met the inclusion criteria involving the surgical procedure were included in the study.
All twenty-two patients were treated with use of the same arthroscopic operative technique, which involved two femoral PCL tunnels, a modified tibial-inlay technique, and Achilles tendon allograft. Twenty of the twenty-two patients were men, and the mean age at the time of revision was 37.4 years (range, twenty-two to sixty-four years). The mean time from the primary PCL reconstruction to the revision reconstruction was 36.3 months (range, four to 120 months), and the mean duration of follow-up was 39.6 months (range, twenty-four to seventy-two months). This retrospective study was approved by the institutional review board of our hospital, and all patients provided informed consent.
Nineteen (86%) of the primary PCL reconstructions had been performed by surgeons at other hospitals using fourteen Achilles tendon allografts, two double-looped semitendinosus and gracilis tendon autografts, two tibialis posterior tendon allografts, and one bone-patellar tendon-bone autograft. The transtibial technique had been used for the tibial tunnel in twelve of these nineteen patients, and the tibial-inlay technique had been used in seven. Three patients had undergone primary reconstruction at our own hospital with use of either the transtibial technique with a tibialis posterior tendon allograft (two patients) or the modified tibial-inlay technique with a double-looped semitendinosus and gracilis tendon autograft (one patient). Fifteen (68%) of the twenty-two patients had also undergone concomitant procedures to reconstruct other knee ligaments; the lateral collateral ligament had been reconstructed in eight knees to correct posterolateral rotatory instability, the anterior cruciate ligament had been reconstructed in five, and the medial collateral ligament had been reconstructed in two.
Surgical Technique and Rehabilitation
The revision PCL reconstruction in all twenty-two patients involved a double femoral tunnel, a modified tibial-inlay technique, and Achilles tendon allograft. Nineteen (86%) of the patients also underwent at least one concomitant surgical procedure in the ipsilateral knee (Table I). A grade-2 or 3 posterolateral corner injury was present in seventeen (77%) of the patients, and reconstruction of the posterolateral corner was performed at the same time as the revision PCL reconstruction in these patients. A modified fibular-head tunnel method was used in the thirteen patients with a grade-2 posterolateral corner injury, and anatomical posterolateral corner reconstruction with a combination of the fibular-head and tibial-tunnel methods was performed in the four patients with a grade-3 posterolateral corner injury. The arthroscopically assisted PCL reconstruction technique involving use of a modified tibial inlay has been described previously8. Specific technical aspects of this method are described below.
The Achilles tendon allograft consisted of a tibial bone block and two femoral bundles (Fig. 1). The Achilles tendon graft was transected sagittally into two strands, and each strand was sutured to a diameter of 6 to 8 mm with a locking whipstitch using number-5 nonabsorbable sutures. The size of the anterolateral bundle was greater than that of the posteromedial bundle. The accompanying tibial bone block was 20 mm long, 15 mm wide, and 8 mm thick. An outside-in technique was used to create the femoral tunnel for the anterolateral bundle 5 to 6 mm proximal to the margin of the distal articular cartilage of the medial femoral condyle at the 12:30 to 1:00 o'clock position (right knee) or the 11:00 to 11:30 position (left knee). An inside-out technique utilizing an accessory anterolateral portal was used to create the femoral tunnel for the posteromedial bundle 8 to 9 mm proximal to the margin of the distal articular cartilage of the medial femoral condyle at the 3:00 o'clock position (right knee) or the 9:00 position (left knee). Thus, the tunnel for the posteromedial bundle was located more posteriorly and proximally than the tunnel for the anterolateral bundle.
The previous implants were then routinely removed, although the previous PCL graft or the remnant of the original PCL bundle was preserved if possible. The primary tunnel was reused if it was correctly situated. If the tunnel placement was grossly incorrect, a new tunnel was created at the correct position. However, the most difficult situation involved a tunnel position that was not entirely incorrect. In such a case, the tunnel was enlarged with use of a dilator to avoid generating a large bone defect and to preserve the original graft; on occasion, autogenous corticocancellous bone from the anterior iliac crest was impacted between the graft and the tunnel after the tunnel was widened. Once the femoral tunnels had been created, a 21-gauge wire loop was passed through each tunnel, directed toward the medial side of the tibial insertion site of the PCL, and later used to pass the graft from this site into the femoral tunnel.
The operating table was tilted 30° down on the affected side and the knee was flexed 70° to 90° to provide easier access to the popliteal area, and a posteromedial approach to the knee was made. The tibial attachment point of the PCL was demarcated, and an osseous trough was created just distal to the tibial insertion of the PCL for fixation of the detached bone block. The posteromedial femoral bundle of the Achilles tendon allograft was passed through the knee joint and into the femoral tunnel, followed by the anterolateral bundle. The Achilles bone block was fixed to the tibia with use of a 5-mm or 6.5-mm cannulated screw with a spiked washer. The operating table was then returned to the neutral position, the graft was cyclically loaded to a tension of 7 to 9 kg (15 to 20 lb), and the articular step-off was checked to confirm anatomical reduction. The knee joint was flexed 70° to 90°, and each graft strand was fixed at the femoral tunnel with a biodegradable interference screw. In addition, each graft strand was fixed on the femoral side with a 10-mm staple or a post-tie. A final assessment of the knee joint was performed to verify that full extension and flexion were possible and that there was no resistance.
Postoperative rehabilitation varied according to the rigidity of the graft fixation, the degree of isometricity, and the results of the final intraoperative stress test after the reconstruction. Because a posterolateral corner reconstruction was typically performed at the same time as the PCL reconstruction, passive knee range-of-motion exercises were usually initiated on the third to fifth postoperative day. Passive range of motion was attempted with the surgeon or the patient holding the proximal aspect of the tibia to prevent gravity-induced posterior displacement.
Knee motion was increased from 0° to 90° during the first six weeks. The knee was immobilized with a long removable splint for two to three weeks postoperatively and then with a PCL brace until the sixth postoperative week. The goal was to regain full knee movement by the twelfth to the twenty-fourth postoperative week.
Patient Assessment
The clinical outcome and knee stability were evaluated at six weeks after the revision PCL reconstruction, at three, six, and twelve months, and every twelve months thereafter. Knee stability was evaluated with use of posterior stress radiography (lateral radiography utilizing a TELOS stress device) and also with use of a maximum manual displacement test utilizing a KT1000 arthrometer (MEDmetric, San Diego, California). These tests were performed preoperatively and at each follow-up visit after the third postoperative month. Clinical evaluation was performed with use of the subjective International Knee Documentation Committee (IKDC) and objective Orthopädische Arbeitsgruppe Knie (OAK) scoring systems. Knee range of motion was measured preoperatively and postoperatively with a goniometer. The patient's activity level was evaluated at the time of the latest follow-up with use of the Tegner activity scale9. All evaluations were made by a single observer who had not been involved in the surgery. Testing for posterolateral rotatory instability was performed at each follow-up visit with use of the posterolateral drawer test, the dial test, and the varus stress test, and the results for the operatively treated and contralateral sides were compared.
Analysis of Failure
The probable causes of failure of the initial PCL reconstruction were assessed with use of previously published criteria10. For instance, tunnel placement was defined as abnormal if ≥75% of the tunnel width was outside the normal anatomical attachment site of the PCL on either the femur or the tibia11,12. The widths of the tibial and femoral tunnels were also measured to assess excessive osteolysis and widening, with tunnel widening defined as an increase of >50% in the area of the tunnel13. All measurements were performed on the monitor of a picture archiving and communications system (PACS; GE Healthcare, Waukesha, Wisconsin) with use of a mouse-controlled cursor and automated calculations performed by the computer. Two orthopaedic surgeons carried out the evaluations retrospectively with no knowledge of the arthroscopic findings, clinical history, or magnetic resonance imaging interpretations.
Statistical Analysis
A power analysis was performed prior to the study. A difference of >10 points in the subjective IKDC score and a difference of >2 mm on the posterior stress radiograph were deemed clinically important. A sample size of fourteen knees was found to be required to achieve a power of 80% at a significance level of 5%.
The Wilcoxon rank-sum test and the chi-square test were used to compare the preoperative and postoperative data as well as the outcomes of the primary and revision reconstructions. All statistical analyses were performed with use of SPSS for Windows (version 10.0; SPSS, Chicago, Illinois) and SAS software (version 9.1; SAS Institute, Cary, North Carolina). A p value of p ≤0.05 was considered significant.
Source of Funding
No external sources funded this research.
Causes of Failure of the Primary PCL Reconstructions
The factors that most likely explain the failure of the primary PCL reconstructions in the twenty-two patients are listed in Table II. Nine (41%) of the reconstructions appeared to have failed because of a single factor and the remaining thirteen (59%) appeared to have failed because of multiple factors. The most common probable causes of failure were posterolateral rotatory instability (seventeen knees, 77%) and improper graft tunnel placement (nine knees, 41%).
Stability
The side-to-side difference in posterior translation during posterior stress radiography improved from a mean (and standard deviation) of 9.9 ± 2.8 mm before the revision PCL reconstruction to 2.8 ± 1.8 mm at the time of the latest follow-up (p < 0.001). At the latest evaluation, the difference was <3 mm in nine (41%) of the twenty-two patients, between 3 and 5 mm in eleven (50%), and >5 mm in two (9%) (Figs. 2-A and 2-B). The mean side-to-side difference on the maximal manual displacement test performed with use of the KT1000 arthrometer also improved from 7.5 ± 2.4 mm preoperatively to 2.3 ± 1.3 mm at the latest follow-up evaluation (p < 0.001).
Clinical Results
The mean OAK score improved significantly from 69.0 ± 7.9 (range, 53 to 82) preoperatively to 78.1 ± 7.0 (range, 61 to 91) at the time of the latest follow-up (p < 0.001). At the preoperative evaluation, the score was classified as fair in twelve (55%) of the patients and as poor in ten (45%). At the latest evaluation, the score was classified as excellent in one patient (5%), as good in eleven (50%), as fair in seven (32%), and as poor in three (14%) Thus, twelve (55%) of the patients had a good-to-excellent rating at the latest evaluation, compared with none on the preoperative evaluation.
The mean subjective IKDC score improved significantly from 39.1 ± 11.3 (range, 23 to 63) preoperatively to 60.4 ± 13.5 (range, 41 to 84) at the time of the latest follow-up (p < 0.001). At the preoperative evaluation, the score was classified as C (abnormal) in six (27%) of the patients and as D (severely abnormal) in the other sixteen (73%). At the latest evaluation, however, the score was classified as A (normal) in two (9%) of the patients, as B (nearly normal) in fourteen (64%), as C in four (18%), and as D in two (9%). Thus, 73% of the patients had a rating of normal or nearly normal at the latest evaluation.
At the preoperative evaluation, knee extension averaged 0.8° (range, 0° to 5°) and flexion averaged 133.5° (range, 120° to 145°). At the time of the latest follow-up, extension averaged 1.2° (range, 0° to 5°) and flexion averaged 129.6° (range, 120° to 145°). The mean Tegner activity score at the latest follow-up evaluation was 4.2 (range, 2 to 7). The Tegner score was 5 (capable of doing heavy work and exercise such as cycling) in seven (32%) of the patients, 4 in ten (45%), and ≤3 in five (23%). Thus, seventeen (77%) of the patients were able to resume their normal activities of daily living.
Complications
There were no intraoperative complications. However, physical examination revealed that two patients who had undergone simultaneous PCL and posterolateral corner revision surgery with use of a modified fibular-head tunnel method developed recurrent grade-3 posterolateral instability during the follow-up period. This instability was treated with anatomical posterolateral corner reconstruction involving simultaneous grafting through the fibular head and the tibial tunnel, performed at eight months postoperatively in one patient and at thirty months in the other. Both of these patients had an IKDC objective score of B at the time of the latest follow-up, and the OAK score was 73 in the first patient and 82 in the second.
Seven patients could not perform a full squat at the time of the latest follow-up, and terminal knee flexion beyond 130° was not possible in these patients. Only one of these patients received additional treatment to regain a more normal range of knee motion. This patient underwent arthroscopic lysis of adhesions at six months postoperatively and was able to flex the knee up to 130° but not to squat fully at the time of the latest follow-up.
This study included twenty-two knees that underwent revision PCL reconstruction involving the same technique and graft material. The most common probable causes of failure of the primary PCL reconstruction were posterolateral rotatory instability (seventeen knees, 77%) and improper graft tunnel placement (nine knees, 41%). The revision reconstruction involved a double femoral tunnel, a modified tibial-inlay technique, and an Achilles tendon allograft. Although this was a small case series, the results were encouraging, with most patients showing improvement in both subjective criteria and objective measurements after the revision procedure.
Several studies have shown that the majority of knees that require revision PCL reconstruction had failed because of multiple factors7,10,14-16. These factors may involve the surgical technique, the integrity of the secondary restraints, lower limb malalignment, articular cartilage damage, prior meniscectomy, postoperative rehabilitation, and the motivations and expectations of the patients7,10,14-16. The reason that a PCL reconstruction fails may therefore be difficult to determine, and multiple factors may contribute. However, the authors of one of the previous studies reported that although multiple factors could cause failure of a primary PCL reconstruction, the two most common causes were associated posterolateral ligament deficiency and improper graft tunnel placement10. Our results were consistent with those findings.
Two clinical studies comparing double-bundle and single-bundle PCL reconstruction failed to detect any significant differences between the clinical outcome17,18. However, several studies on cadavers have shown that double-bundle PCL reconstruction is biomechanically superior to single-bundle reconstruction19,20. Furthermore, Kim et al.3 reported that arthroscopic tibial-inlay double-bundle PCL reconstruction resulted in better posterior stability compared with two single-bundle methods in twenty-nine patients with at least two years of follow-up. Thus, because most PCL revisions are associated with posterolateral corner injuries, revision with use of double-bundle PCL reconstruction could help to restore rotational stability. Although the double-bundle PCL reconstruction technique is technically demanding, we believe that it restores normal knee kinematics more effectively than single-bundle techniques do. Arthroscopic examination during a revision PCL reconstruction usually demonstrates preserved continuity and contour of the PCL, although the PCL might be attenuated. Therefore, we preserved the previous PCL graft or the remnants of the original PCL bundle when possible.
Many surgeons have suggested that a sharp bend (the so-called “killer turn” or “killer curve”) in the PCL reconstruction graft at the posterior opening of the tibial tunnel may result in stretching or failure of the graft21-23. Since 1995, we have advocated a modified tibial-inlay technique in which the posteromedial approach is facilitated by tilting the operating table 30° down on the affected side8,24. This technique is particularly useful for revision PCL reconstructions that are complicated by enlargement of the tibial graft tunnel. It is also useful for cases in which the previously placed tibial tunnel is located immediately adjacent to the anatomical attachment site of the PCL, making it impossible to correctly locate the new graft tunnel without overlapping and breaking into the old tunnel. The tibial-inlay method offers the advantages of superior fixation and graft healing at the tibial attachment site even when misplacement of the prior tibial tunnel interferes with the placement of a new tibial tunnel or when the prior tibial graft tunnel has become enlarged.
The importance of the posterolateral corner structures in maintaining knee stability has become better understood in recent years, as has the fact that these structures interact functionally with the cruciate ligaments25,26. Injuries to the posterolateral corner structures are often associated with injuries of the cruciate ligaments rather than occurring in isolation14,15. Biomechanical studies have demonstrated that cutting the posterolateral structures increases the in situ forces on the PCL16,19, and recurrent laxity after PCL reconstruction is most commonly related to untreated posterolateral corner injuries10,27,28. Thus, great efforts have been made to appropriately diagnose and treat concomitant posterolateral corner injuries. However, given the difficulty in diagnosing combined injuries of the PCL and the posterolateral corner, guidelines based on commonly used clinical evaluation methods would be helpful29,30. If posterolateral rotatory instability accompanying a PCL injury is grade 2 or less, we recommend reconstruction of the posterolateral corner with use of a fibular tunnel, as this is a simpler and biomechanically superior method that causes less surgical morbidity compared with the tibial tunnel method24,31. However, if the instability is grade 3, it may be preferable to reconstruct the posterolateral corner injuries anatomically by simultaneous grafting through the fibular head and the tibial tunnel.
This study has a number of limitations. First, the number of cases evaluated was relatively small. However, this limitation is mitigated by the advantage that all of the revision reconstructions were performed by a single surgeon with use of a standardized surgical technique, thus minimizing the effect of treatment variables. Moreover, we used multiple knee scores as well as stress radiography to evaluate the outcome of the reconstruction method. Other major limitations of this study are the retrospective study design, the absence of a control group, and the inclusion of patients with additional ligament injuries. The latter limitation, however, resulted from the fact that treatment of concomitant ligament injuries is required in most knees undergoing revision PCL reconstruction.
In summary, arthroscopic revision PCL reconstruction with use of the modified tibial-inlay double-bundle technique improved knee stability, as measured with posterior stress radiography and clinically, and outcomes. Associated posterolateral rotatory instability should be surgically corrected during PCL reconstruction to prevent graft failure resulting from abnormal opening of the lateral tibiofemoral joint.
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, no author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.