Biomechanical studies have demonstrated that single-bundle anterior cruciate ligament reconstruction is insufficient for controlling both translation and rotation in extension1,2. For patients with generalized ligamentous laxity, some authors have stated that the risk of instability is greater with a conventionally reconstructed anterior cruciate ligament3. Double-bundle anterior cruciate ligament reconstruction has been shown to provide more improvement in knee stability, theoretically because it more closely resembles the anatomic structure of the anterior cruciate ligament1,2. Until now, however, little has been reported regarding the clinical outcome of double-bundle anterior cruciate ligament reconstruction in patients with generalized ligamentous laxity. In 2004, a double-bundle anterior cruciate ligament reconstruction procedure with autogenous quadriceps tendon with use of two femoral sockets and one tibial tunnel was developed by the senior author (S.-J.K.)4.
The purpose of the present study was to compare the clinical outcome of double-bundle reconstruction with use of the quadriceps tendon-bone autograft and that of single-bundle reconstruction with use of a bone-patellar tendon-bone autograft in patients with generalized ligamentous laxity. We hypothesized that the clinical outcome of anterior cruciate ligament reconstruction would be different on the basis of the reconstruction technique used.
Patients
The records of 305 patients who underwent anterior cruciate ligament reconstruction between June 2002 and October 2005 were evaluated. All patients were treated by the senior author (S.-J.K.). The patients were selected if they met the following criteria: (1) they had a unilateral, isolated anterior cruciate ligament injury without an injury of the contralateral knee; (2) they had no history of surgery involving the lower extremity; (3) they had no articular cartilage erosion of more than grade II (fissuring and fragmentation <0.5 in [13 mm] in diameter), according to the Outerbridge classification, at the time of surgery; and (4) a meniscectomy, when performed, had involved less than one-third of the entire meniscus. On the basis of one or more of these criteria, 115 patients were excluded from this study. Single-bundle anterior cruciate ligament reconstruction with a bone-patellar tendon-bone autograft was performed in 118 patients, and a double-bundle anterior cruciate ligament reconstruction with a quadriceps tendon-bone autograft was done in seventy-two patients. The type of reconstruction selected was dictated merely by the need to obtain a sufficient size of graft for a double-bundle anterior cruciate ligament reconstruction. The selection of a quadriceps tendon-bone graft was dependent on the thickness of the tendon as measured on magnetic resonance imaging: a quadriceps tendon-bone graft was selected if the thickness of the tendon was >7 mm. Otherwise, a bone-patellar tendon-bone autograft was used. Following the Beighton and Horan criteria5 (Table I), which have gained international acceptance and have been widely used to diagnose generalized ligamentous laxity6, we excluded patients with fewer than four positive findings of ligamentous laxity. After taking all of these factors into account, sixty-one patients were eligible for this study; thirty-two had an anterior cruciate ligament reconstruction with a bone-patellar tendon-bone autograft (group 1), and twenty-nine had an anterior cruciate ligament reconstruction with a quadriceps tendon-bone autograft (group 2).
Surgical Technique
Single-Bundle Reconstruction with Use of a Bone-Patellar Tendon-Bone Autograft
The bone-patellar tendon-bone graft was harvested with a width of 10 mm. The patellar and tibial bone blocks were trapezoidal in shape and were 25 mm long and 8 mm deep. The patellar paratenon was sutured. A tibial tunnel was drilled with use of a 10-mm-diameter cannulated reamer. A femoral guide pin was positioned at 10:30 o'clock on the right knee or 1:30 o'clock on the left knee with the knee flexed 70° to 90°, and the femoral socket was reamed to a depth of 30 mm. The previously prepared bone-patellar tendon-bone graft tendon was passed through the tibial tunnel, across the joint, and into the femoral socket. The graft was secured within the femoral socket with use of a bioabsorbable interference screw with the knee held at 90° of flexion. The graft was pretensioned by pulling it tightly and moving the knee through a full range of motion ten times. The graft was then fixed within the tibial tunnel with a bioabsorbable interference screw at 10° to 15° of knee flexion.
Double-Bundle Reconstruction with Use of a Quadriceps Tendon-Bone Autograft
We harvested a piece of quadriceps tendon, including the full thickness of all three tendon layers, with use of a ParaSmillie graft harvester (Linvatec, Largo, Florida). The patellar bone block was harvested in a rectangular shape that was 20 mm long, 8 mm deep, and 11 mm wide with use of an oscillating saw. The quadriceps tendon was then split coronally into two bundles at a ratio of about 2:3. Each tendon bundle was sutured with use of a baseball stitch over a length of 30 mm. The 7-mm and 9-mm-diameter EndoPearl devices (Linvatec) were fixed to each tendon end with use of number-2 Ethibond sutures (Ethicon, Somerville, New Jersey). Two drill-holes were made in the patellar bone plug. The number-2 Ethibond sutures were then passed through the holes in the bone plug.
High anterolateral and low anteromedial arthroscopy portals were established7. An 11-mm-diameter tibial tunnel was made with use of the same method as in the single-bundle technique. A 6-mm over-the-top guide was placed on the proximal cortex of the intercondylar notch at about the 11 o'clock (right knee) or 1 o'clock (left knee) position. A guide pin was inserted, and the femoral socket for the anteromedial bundle was drilled over the guide pin with a 9-mm-diameter reamer to a depth of 45 mm. The posterolateral femoral socket was then drilled through an accessory anteromedial portal located 1.5 cm medial to the anteromedial portal just above the medial meniscus and 5 mm anterior to the medial femoral condyle. A 7-mm-diameter reamer was placed in the center of the posterolateral bundle footprint, which was located approximately at the crossing point of the long axis line of the anterior cruciate ligament attachment and a vertical line drawn through the contact point between the femoral condyle and the tibial plateau at 90° of knee flexion8. The socket was placed at the 9 o'clock position (right knee) or the 3 o'clock position (left knee). With use of an accessory anteromedial portal, and once the guide pin was in an acceptable position, the socket was drilled with a 7-mm-diameter reamer in a counterclockwise direction to a depth of approximately 35 mm. The anteromedial portal was extended vertically to 2.5 cm with a knife to facilitate passage of the graft. A leading suture was passed through the tibial tunnel and was pulled out with a grasper through the anteromedial portal. The distal leading sutures attached to the end of the bone plug were pulled out through the tibial tunnel. Two additional leading sutures were passed through each femoral socket and were pulled out through the anteromedial portal. The two leading sutures were then tied to the sutures in each limb of the tendon graft, and each was pulled into one of the two sockets in the lateral femoral condyle. The distal bone plug was passed through the anteromedial portal and placed in the tibial tunnel by pulling the leading sutures distally. Next, the proximal leading sutures were used to guide the graft into the closed end of the femoral sockets.
Bioabsorbable interference screws were used for both femoral and tibial fixation of the grafts. To fix the anteromedial and posterolateral bundles at different angles, we followed a sequence of graft-tensioning and fixation. First, femoral fixation of the posterolateral bundle graft was obtained with a bioabsorbable interference screw inserted through the low anteromedial portal with the knee at 90° of flexion. Next, tension was applied by pulling on the Ethibond suture in the bone plug in the tibial tunnel while moving the knee ten times through a full range of motion. The knee was then brought to a position of 10° of flexion, and the distal bone plug was tensioned and secured by a bioabsorbable interference screw, achieving posterolateral bundle fixation at 10° of knee flexion. Then, tension was applied to the graft in the femoral socket for the anteromedial bundle while the knee was moved ten times through a full range of motion. Femoral fixation of the anteromedial graft bundle was then obtained with use of a bioabsorbable interference screw tightened with a flexible screwdriver, which was placed through the low anteromedial portal with the knee initially at 90° of flexion (because of easy engagement), followed by 70° of flexion to achieve the final fixation.
Postoperative Rehabilitation
The rehabilitation protocol was identical for both groups. All patients were permitted immediate partial weight-bearing using crutches. Patients were allowed to bear their full weight approximately four weeks after surgery. By the twelfth week, jogging, swimming, and cycling were permitted. A return to sports involving jumping, pivoting, or sidestepping was allowed after six months.
Clinical Assessments
We obtained clinical outcomes from data obtained before surgery and at the twenty-four-month follow-up visit. Manual examinations were performed by the senior author (S.-J.K.). Ligament stability was assessed with use of the Lachman and pivot-shift tests. The Lachman test was graded on a scale of 0 (<3 mm), 1+ (3 to 5 mm), 2+ (6 to 10 mm), or 3+ (>10 mm). The pivot-shift test was performed with the hip in abduction and the tibia in internal rotation. The pivot-shift phenomenon was graded on a scale of 0 (absent), 1+ (subluxation), 2+ (jump), or 3+ (transient lock). The side-to-side difference in anterior tibial translation was measured with a KT-2000 arthrometer (MEDmetric, San Diego, California) at 30° of knee flexion. The Hospital for Special Surgery knee ligament questionnaire and the Lysholm knee scoring scale were used to evaluate functional outcome.
Statistical Analysis
Groups 1 and 2 were compared with use of the Mann-Whitney test for the Hospital for Special Surgery and Lysholm scores, and the unpaired Student t test was used to compare laxity assessments. Laxity assessment comparison was confirmed with use of the Kolmogorov-Smirnov goodness-of-fit test to satisfy the assumptions for the Student t test. The differences in the Lachman test and pivot-shift test were analyzed with the chi-square test. The level of significance was set at p < 0.05. The results are given as the mean and the standard deviation.
Source of Funding
There was no external funding source for this investigation.
There were fourteen male and eighteen female patients in group 1 and eleven male and eighteen female patients in group 2. The average age at the time of surgery was 28.9 years (range, eighteen to forty-two years) in group 1 and 25.3 years (range, twenty to thirty-nine years) in group 2. The interval from injury to surgical treatment ranged from three to twelve months. With the numbers studied, no significant difference was found between groups 1 and 2 with regard to sex, age at the time of surgery, and time to surgery (Table II).
Physical Examination
Preoperative Status
Twenty-three of the thirty-two patients in group 1 were scored on the Lachman test as grade 2+, and nine patients were grade 3+. In group 2, twenty-two of the twenty-nine patients were scored as grade 2+ and seven patients were grade 3+. No patient had grade-1+ laxity in either group.
With regard to the pivot-shift test, three patients had a grade-1+ pivot shift in each group. In group 1, twenty-seven patients had a grade 2+ and two patients had a grade 3+. In group 2, twenty-five patients had a grade 2+ and one patient had a grade 3+. With these numbers, no significant difference between the two groups was found in the preoperative Lachman grade or pivot-shift test (Table III).
Postoperative Status
On the basis of the Lachman test (Table IV), the patients in group 1 had significantly more laxity than did the patients in group 2 (p = 0.032). In the thirty-two patients in group 1, eighteen patients had grade-1+ laxity and two patients had grade 2+. In the twenty-nine patients in group 2, nine patients were classified as grade 1+, and no patient was grade 2+. No patient had a grade-3+ laxity in either group. With regard to the pivot-shift test, three patients had a grade-1+ pivot shift in group 1, while no patient had an abnormal pivot shift in group 2 (p = 0.091) (Table IV).
Anterior Tibial Translation
The results from the KT-2000 arthrometer measurement of anterior tibial translation are presented in Table V. The postoperative mean side-to-side difference was greater in group 1 (3.37 ± 1.76 mm; range, 1.00 to 8.00 mm) than in group 2 (2.03 ± 1.11 mm; range, 0.00 to 3.50 mm) (p = 0.02). Two patients had a side-to-side difference of >5 mm in group 1, while no patient had a side-to-side difference of >5 mm in group 2.
Subjective Data
Substantial improvements in both the Hospital for Special Surgery and Lysholm scores between the preoperative and the follow-up examination were observed in both groups (Table VI), but the values were not significantly different between the groups. The postoperative mean Hospital for Special Surgery score was 90.8 points (range, 75 to 100 points) and the Lysholm score was 89.4 points (range, 72 to 99 points) in group 1. In group 2, the postoperative mean Hospital for Special Surgery score was 92.1 points (range, 75 to 100 points) and the Lysholm score was 91.1 points (range, 73 to 100 points).
Complications
In group 1, six patients had anterior knee pain with kneeling. Another patient had a limitation of about 5° of full knee extension, which may have been caused by four weeks of immobilization that was needed to treat a postoperative soft-tissue infection. At the time of the last follow-up, he reported moderate knee pain after walking for more than thirty minutes and he did not participate in sports activities. In group 2, the posterior femoral cortex was perforated during fixation of the anteromedial bundle in one patient. In that case, instead of screw fixation, the leading sutures of the anteromedial bundle were fixed to the anterolateral femoral cortex with use of a suture washer. After two years, the anterior tibial translation in this patient, measured with a KT-2000 arthrometer, was 4.5 mm. He had minimal knee pain during strenuous activities and had mild discomfort with twisting motions, although we could not detect any rotatory instability on physical examination. In a thirty-nine-year-old patient, the fixed patellar bone plug in the tibial tunnel migrated proximally during pretensioning of the femoral graft. This was attributed to weak fixation because of poor bone quality in the proximal part of the tibia. For more secure fixation of the patellar bone plug, two bioabsorbable screws were inserted after repositioning the bone plug. After two years, the patient had 3.5 mm of anterior drawer laxity. He did not have any discomfort or pain around the knee and participated in weekly sports activities.
This study was conducted to investigate ligament stability and functional outcomes after anterior cruciate ligament reconstruction in patients with generalized ligamentous laxity. We compared single-bundle reconstruction with use of a bone-patellar tendon-bone autograft and double-bundle reconstruction with use of a quadriceps tendon-bone autograft. In previous studies, the overall outcomes of anterior cruciate ligament reconstruction with a bone-patellar tendon-bone autograft have been satisfactory in terms of anterior-posterior translation9-12. The average amount of anterior tibial translation in patients who undergo anterior cruciate ligament reconstruction with a bone-patellar tendon-bone graft has been reported to range from 1.51 to 2.7 mm9-12. This is better than the average result of 3.37 mm that we found in the current study. Over time, bone-patellar tendon-bone autografts may develop increased stiffness and strength13, as well as revascularization14. Despite these advantages, patients with generalized ligamentous laxity may experience inferior results because of the inherent connective tissue extensibility of the autograft as well as laxity of the secondary knee restraints caused by the composition of the connective tissue and the orientation of the various soft-tissue structures3,15.
The quadriceps tendon-bone graft is less often used for anterior cruciate ligament reconstruction, but Stäubli et al.16 showed its initial mechanical properties to be adequate for anterior cruciate ligament reconstruction. Lee et al.17 suggested that quadriceps tendon-bone grafts offer a reliable, low-morbidity autograft alternative for anterior cruciate ligament reconstruction on the basis of promising outcomes, including knee stability after more than two years of follow-up. In patients with generalized ligamentous laxity, however, the risk of increased anterior tibial translation with a quadriceps tendon-bone autograft may not be lower than that with a bone-patellar tendon-bone autograft with use of conventional single-bundle reconstruction18. In this study, we found that patients who received a double-bundle anterior cruciate ligament reconstruction with use of a quadriceps tendon autograft showed less anterior-posterior laxity than the patients who received a single-bundle reconstruction with use of a bone-patellar tendon-bone autograft. We believe that the primary reason for better stability with the double-bundle reconstruction is closely related to the biomechanics of double-bundle reconstruction. Mae et al.19 reported that, biomechanically, the double-bundle anterior cruciate ligament reconstruction provided better stability compared with the single-bundle reconstruction under an anterior tibial load of 100 N at smaller flexion angles. They also noted that the posterolateral bundle acted dominantly in extension, while the anteromedial bundle mainly resisted anterior tibial load in flexion. Yasuda et al.20 further stated that, with the double-bundle technique, overloading of one bundle could be avoided early after anterior cruciate ligament reconstruction, allowing good maturation of both bundles.
Although we observed a trend toward increased rotational instability (on the basis of the pivot-shift test results) in patients with single-bundle reconstruction, no significant difference was found between the groups. The pivot-shift test is less reproducible and less reliable because of its subjectivity. While Yagi et al.21, with use of three-dimensional electromagnetic sensors, demonstrated that the double-bundle reconstruction resulted in better pivot-shift control than did a single-bundle reconstruction, better methods are needed to identify and quantify rotational instability more clearly and the results should be correlated with patient discomfort during twisting of the knee. We used the EndoPearl device to secure the quadriceps tendon-bone autografts and to prevent graft slippage in the early phase because Arneja et al.22 noted that the application of this device in conjunction with a bioabsorbable screw results in a substantially decreased laxity.
No significant difference was observed between the groups with regard to the Lysholm and Hospital for Special Surgery scores, a finding which is in agreement with several other studies21,22. Kneeling pain occurred more often and was greater in patients who received the bone-patellar tendon-bone grafts, which is also similar to other reports23,24. Although the cause of anterior knee pain after anterior cruciate ligament reconstruction remains unknown, several authors have reported a low prevalence of kneeling pain with the use of quadriceps tendon-bone autografts17,18,25. One reason for kneeling pain in patients undergoing anterior cruciate ligament reconstruction with a bone-patellar tendon-bone autograft might be damage of the infrapatellar branch of the saphenous nerve and neuroma formation. In this study, three of six patients with kneeling pain had anterior or lateral knee paresthesias26.
There are some inherent limitations to this study that warrant review before definite conclusions can be drawn. First, we had a relatively small number of patients who had positive responses on the pivot-shift test. Consequently, we were unable to determine whether rotational laxity influenced patient satisfaction with regard to surgical outcome. Second, we did not fully evaluate the activity scale. In daily life without strenuous activity, a certain degree of anterior instability may be tolerated. Therefore, it is possible that limitations in knee function could be masked by an involuntarily low activity level. Third, the patients were not assigned to the treatment groups in a randomized manner. Although the selection of quadriceps tendon-bone autografts depended on the thickness of the quadriceps tendon, we are not completely convinced of the accuracy of magnetic resonance imaging measurement because its reliability was not assessed. In addition, the surgeon's preference may have entered into the graft selection.
In conclusion, on the basis of the evaluation of ligamentous laxity measured with the KT-2000 arthrometer, double-bundle anterior cruciate ligament reconstruction with use of a quadriceps tendon-bone autograft provided less anterior translation in comparison with single-bundle anterior cruciate ligament reconstruction with use of a bone-patellar tendon-bone autograft. However, with the numbers studied, no significant difference in functional outcome could be identified between the groups. 