Most orthopaedic surgeons treat isolated medial collateral ligament injuries nonoperatively1,2. However, there is no agreement on how to treat a medial collateral ligament rupture that is combined with an anterior cruciate ligament tear3-15. An increasing body of evidence suggests that the medial collateral ligament can be treated nonoperatively in these cases as well4,7,10-12,14,16,17. Repair of all torn ligaments has led to a functionally stable knee, but there has been a high prevalence of postoperative complications affecting knee motion3,5,14,18,19.
Complications affecting knee motion are more common when operative treatment is performed early20-22, and it has been claimed that concurrent medial collateral ligament repair adds to the risk19. However, early anterior cruciate ligament reconstruction may provide better results than delayed anterior cruciate ligament reconstruction, with no greater problems related to loss of motion23.
In 1996, we started a prospective randomized study on the treatment of combined anterior cruciate ligament-medial collateral ligament ruptures. Nonoperative treatment of the medial collateral ligament led to results that were similar to those obtained with medial collateral ligament repair in terms of subjective function of the knee, postoperative laxity, range of motion, muscle power, return to activities, the Lysholm score, and overall performance at the time of the two-year follow-up16.
In the present study, we focused on postoperative recovery. The purpose of the study was to evaluate the effect of early repair of the concomitant medial collateral ligament injury (as compared with the effect of nonoperative treatment of the medial collateral ligament injury) on the range of motion of the knee and the strength of the quadriceps muscle in patients with combined anterior cruciate ligament-medial collateral ligament injuries who were managed with early reconstruction of the torn anterior cruciate ligament. The main outcome measures were (1) range of motion of the knee (knee extension deficit and knee flexion) and (2) quadriceps muscle power deficit. We also wanted to determine the ability of the rehabilitation program to restore normal knee motion after early operative intervention.
The present study was approved by our institutional review board, and all patients provided informed consent. The present trial was registered in the Helsinki University Trial Registry (no. T102020Z04). Between December 1996 and November 2001, we operated on forty-seven consecutive patients who had a combined anterior cruciate ligament and grade-III medial collateral ligament injury. The sample size was selected to have at least 80% power to identify a 10-point difference in Lysholm scores. The inclusion criteria were an age of at least eighteen years, an acute complete rupture of the anterior cruciate ligament, a rupture of most or all of the medial ligamentous structures (a grade-III medial collateral ligament rupture), and no other ligamentous deficiency. No patient who met the inclusion criteria was dropped from the study. The primary diagnosis was made on the basis of clinical examination. All patients had a positive Lachman test and/or pivot shift test, >10 mm of medial joint space opening on manual valgus stress testing at 25°, and at least a subtle pathologic opening of the medial space in almost full extension (=5° of flexion). The diagnosis was confirmed and concomitant injuries were detected with use of magnetic resonance imaging. Magnetic resonance imaging (available for forty-four patients) was performed with either a 1.5-T scanner (Sigma LX 1.5T; GE Medical Systems, Milwaukee, Wisconsin) (eleven patients, managed between 2000 and 2001) or a 0.23-T open scanner (Outlook GP; Picker Nordstar, Helsinki, Finland) (thirty-three patients, managed between 1997 and 2000), with use of routine clinical sequences and dedicated knee coils. Patients with posterior cruciate or lateral collateral ligament injuries or only grade-I or II medial collateral ligament injuries were excluded.
Anterior cruciate ligament reconstruction was performed in all patients within twenty-three days (range, four to twenty-three days) after the injury. The patients were randomized into two groups according to the treatment method for medial collateral ligament rupture. In Group I the medial collateral ligament was repaired, and in Group II it was treated nonoperatively. Blocking was used to balance the number in each group. After blocking of every ten patients, five were allocated to each group. On the day of surgery, a sealed and serially numbered opaque randomization envelope was opened in the operating room by the attending surgeon (J.H., J.L., or E.H.). The only variable between the groups was the mode of medial collateral ligament treatment.
Group I (reconstruction of the anterior cruciate ligament and repair of the medial collateral ligament) comprised twenty-three patients (eight men and fifteen women) with a mean age of forty years (range, twenty-two to sixty-four years). The patients in Group I underwent surgery at a mean of 9.2 days (range, four to seventeen days) after the original injury. Group II (reconstruction of the anterior cruciate ligament and nonoperative treatment of the medial collateral ligament) comprised twenty-four patients (twelve men and twelve women) with a mean age of thirty-eight years (range, twenty-one to fifty-seven years). The patients in Group II underwent surgery at a mean of 10.5 days (range, five to twenty-three days) after the original injury.
Operative Technique
An arthroscopic anterior cruciate ligament reconstruction technique was used. All patients were examined under anesthesia. Routine diagnostic arthroscopy and meniscal surgery were performed, followed by anterior cruciate ligament reconstruction and, in Group I, medial collateral ligament repair. In both groups, anterior cruciate ligament reconstruction was performed with use of the transtibial technique and a bone-patellar tendon-bone autograft24. The graft was fixed with interference screws (ConMed Linvatec, Largo, Florida). In Group I, the medial ligamentous structures were repaired with suture anchors (Mitek Surgical Products, Norwood, Massachusetts) (fifteen patients) or were sutured through small bone tunnels (six patients) and the repair was completed with direct suturing (twenty-three patients); in two patients, the repair was performed with direct suturing only. We did not use any augmentations. Achieving an anatomic repair of the medial collateral ligament and testing the effect of range of motion on the suture line before closing the wound were also important steps. In cases of femoral avulsion, we tried to avoid too-anterior attachment. We also tried to avoid over-tightening of the anterior part of the superficial medial collateral ligament. Before suturing the medial structures, we flexed the knee to see how the torn structures behaved. By doing so, we avoided disruption of the suture line. For anterior cruciate ligament reconstruction the incision was 5 to 6 cm, and for medial collateral ligament repair an additional medial incision of 10 to 12 cm was needed.
The anterior cruciate ligament was ruptured proximally in almost all cases (91%). The operative findings in Group I showed that the medial collateral ligament was ruptured proximally only in eleven cases (48%), proximally and in the midsubstance in three cases (13%), in the midsubstance only in six cases (26%), in the midsubstance and distally in two cases (9%), and distally only in one case (4%). During the operation, all twenty-three patients were found to have a superficial medial collateral ligament rupture, twenty-two patients were found to have a posterior oblique ligament rupture with partial disruption of the posteromedial capsule, and eight patients were found to have a total rupture of the posteromedial capsule. Because the preoperative findings in Group II were similar, we believe that the extent of medial collateral ligament injury was similar in those patients, although this was not confirmed by means of direct visual inspection at the time of surgery.
Postoperative Rehabilitation and Follow-up Protocol
The rehabilitation program was the same in both groups. Full weight-bearing and full range of motion were permitted immediately with a hinged knee brace. Crutches were recommended for two to three weeks to allow the patient to regain a normal gait. The brace was used continuously for six weeks and thereafter during the daytime for two weeks. The primary goal in the early rehabilitation period was to normalize the gait and the range of motion of the knee. Aggressive muscle rehabilitation was delayed and began with closed kinetic chain exercises. Sports activities were allowed gradually. High-level sports activities were allowed nine to twelve months after the operation if 90% of the muscle strength (in comparison with the uninvolved knee) had been achieved. Low-level activities were allowed earlier despite a lack of muscle strength. Most of our patients had not participated in aggressive sports before the injury.
Postoperatively, the patients were monitored by a physiotherapist. If a patient exhibited any signs of mobilization problems, monitoring was carried out weekly during the first few weeks. The brace did not allow full hyperextension; it only allowed extension to 0°. It did not limit flexion. In Group I (combined repair), patients visited the physiotherapist a mean of thirteen times (range, seven to twenty-two times). In Group II (reconstruction of anterior cruciate ligament and nonoperative treatment of the medial collateral ligament), the patients visited the physiotherapist a mean of nine times (range, four to fifteen times). The physiotherapists supervised range of motion and stretching of the knee in both flexion and extension and also taught the patients how to normalize gait. One of the authors (J.H.) clinically monitored the patients postoperatively at six, twelve, thirty-six, fifty-two, and 104 weeks. At fifty-two and 104 weeks of follow-up, isokinetic extension-flexion muscle torque of the knee joint was measured with a dynamometer (System 3 Quick-Set; Biodex, Shirley, New York), and the one-leg-hop test was performed.
In the present report, full extension means 0° of flexion-extension of the knee. For the presentation of knee flexion deficit values, the uninvolved knee is used as a reference.
Statistical Analysis
The material was tested for kurtosis and skewness with a one-sample Kolmogorov-Smirnov test of kurtosis. The nonparametric Mann-Whitney test was used for calculations. The level of significance was set at p = 0.05.
The study was conducted in conformity with the principles of the Declaration of Helsinki and was approved by our institutional committee of ethics. The subjects gave informed consent to participate in the study.
Source of Funding
No funds were received in support of the present study.
All forty-seven patients were available at the time of the latest follow-up (mean, 116 weeks; range, eighty-nine to 160 weeks), and most of the patients were available at six weeks (forty-one patients), twelve weeks (forty patients), thirty-six weeks (thirty-seven patients), and fifty-two weeks (thirty-five patients) as well. Physiotherapy records were available from one, two, and three weeks after the operation. We could not identify any significant differences between the two groups with respect to age, sex, the timing of the operation, preoperative stability, or meniscal or chondral abnormality.
The mean extension deficit of the knee at one, two, and three weeks postoperatively was 5°, 5°, and 3°, respectively, in Group I (combined repair) and 3°, 2°, and 2°, respectively, in Group II (reconstruction of the anterior cruciate ligament and nonoperative treatment of the medial collateral ligament). Thereafter, the deficits were <1°. At twelve weeks, only one patient in each group did not exhibit normal knee extension; the absolute deficit values for these two patients were 4° and 5°. At fifty-two weeks, no patient exhibited an extension deficit. Two years after the operation, the mean knee extension was -1.0° in Group I (n = 23), and -0.8° in Group II (n = 24). Postoperative knee extension deficits are presented in Figure 1.
Flexion of the knee was poorer in Group I (combined repair) than in Group II (reconstruction of the anterior cruciate ligament and nonoperative treatment of the medial collateral ligament) at all time intervals (Fig. 2), and the difference between the groups was significant at six weeks (100° compared with 112°; p = 0.009), twelve weeks (119° compared with 128°; p = 0.043), and thirty-six weeks (130° compared with 136°; p = 0.011) after the operation. At fifty-two and 104 weeks, the values in Group I were 132° and 134°, respectively, and the values in Group II were 137° and 137°, respectively; the difference between the groups was not significant at either interval (p = 0.07 and p = 0.23). Only two patients did not exhibit normal range of motion at 104 weeks. In Group I, a fifty-nine-year-old woman in whom moderate arthrosis was noted at the time of surgery had a 13° flexion deficit. In Group II, a forty-six-year-old man who had undergone a previous medial meniscectomy exhibited slight medial arthrosis and had an 8° flexion deficit. Thus, 96% (forty-five) of the forty-seven patients exhibited normal flexion according to the system of the International Knee Documentation Committee25. Six patients in Group I and three patients in Group II did not achieve as much flexion in the involved knee as was present in the uninvolved knee; the absolute values of flexion deficit were 3°, 4°, 5°, 5°, 5°, and 13° for the six patients in the former group and 4°, 5°, and 8° for the three patients in the latter group. No operative interventions were performed because of loss-of-motion complications.
The quadriceps muscle power deficit at a speed of 60°/sec at fifty-two and 104 weeks is presented in Figure 3. At fifty-two weeks, the quadriceps muscle power deficit (in comparison with the uninvolved side) was 30.7% in Group I (combined repair) and 20.5% in Group II (reconstruction of the anterior cruciate ligament and nonoperative treatment of the medial collateral ligament) (p = 0.015). At 104 weeks, a difference persisted, but it was not significant (14.4% compared with 9.7%; p = 0.2). There were similar findings at a speed of 180°/sec, with deficits of 25% and 15% (p = 0.049) at fifty-two weeks and of 6% and 6.6% (p > 0.05) at 104 weeks (Fig. 4). At fifty-two and 104 weeks, the values on the one-leg-hop test in Group I were 83.1% and 90.2%, respectively, and the values in Group II were 86.1% and 93.4%, respectively. The differences in the results of the one-leg-hop test were not significant at either follow-up interval.
Complications affecting knee motion following anterior cruciate ligament reconstruction are more common when the operation is performed early20-22, and it has been claimed that concomitant medial collateral ligament repair adds to the risk19. In the present study, despite early operative intervention, forty-five of the forty-seven patients regained a full range of motion without the need for any manipulative or supplementary operative measures. Both of the patients who did not regain full flexion of the knee had medial compartment arthritis at the time of the injury. Regaining the full range of motion and quadriceps power occurred more slowly in Group I (combined repair) than in Group II (reconstruction of the anterior cruciate ligament and nonoperative treatment of the medial collateral ligament). The differences diminished over time, and, by two years, no differences existed in terms of range of motion or quadriceps muscle power. The average age of our patients undergoing anterior cruciate ligament reconstruction was almost forty years, and that factor, as well as the severity of the injury, might have affected the time scale of the recovery period.
The present study had some limitations. The operations were performed by three different surgeons (J.H., J.L., E.H.), and most of the postoperative measurements were performed by one of the operating surgeons (J.H.). In addition, we were unable to follow every patient at every time interval. Our study did, however, have some advantages. The study groups were homogeneous. Every patient had grade-III medial collateral ligament and anterior cruciate ligament ruptures with no other ligamentous injury, and the method of repairing the anterior cruciate ligament was the same. Although some data were missing during the study, there were no dropouts and all of the patients were available for follow-up at the end of the study.
Several previous studies have shown the same tendency of increased motion problems and slower quadriceps muscle power gains following the operative treatment of both the anterior cruciate ligament and the medial collateral ligament5,11,14,18,19, and proximal or middle-third medial collateral ligament tears have been associated with a higher complication rate than distal tears have18,26. In our series, almost all tears were proximal or in the middle third of the ligament. Achieving an anatomic repair of the medial collateral ligament and checking the knee range of motion and its effect on the suture line before closing the wound were considered to be important steps in the surgical procedure. However, the medial side may have been more painful in the patients undergoing medial collateral ligament repair, thus leading to restricted motion in the early postoperative phase.
Arthrofibrosis can result from either injury or surgical intervention, or both. The likelihood that arthrofibrosis will develop increases with the severity of a knee joint injury, the extensiveness of the related surgery, and the length of time that the knee is subsequently immobilized. In an effort to minimize the development of arthrofibrosis, we allowed immediate full weight-bearing as tolerated and full range of motion with a hinged knee brace. Postoperatively, the patients were monitored by a physiotherapist. If the patient exhibited any signs of mobilization problems, monitoring was arranged weekly during the first weeks after surgery. With use of this protocol, none of our patients required operative intervention because of loss of motion and only two patients failed to exhibit normal range of motion at the two-year follow-up. We were most concerned about the development of an extension deficit, which is far more difficult to eliminate later. The patients were encouraged early to fully extend the knees, and the physiotherapists concentrated on preventing the development of an extension deficit. The results indicated success as only one patient in each group failed to exhibit normal knee extension at twelve weeks and no patient exhibited an extension deficit by one year.
There was a remarkable deficit in knee extension power in both groups at one year, although it was significantly less in Group II (reconstruction of the anterior cruciate ligament and nonoperative treatment of the medial collateral ligament). The patients were encouraged to perform muscle-strengthening exercises but were not systematically supervised while doing so. The majority of the patients informed us of poor compliance in performing the exercises, and the poor muscle-testing results subsequently motivated many of the patients to perform the exercises. By two years, the overall results were acceptable.
Closed kinetic chain exercises were strongly recommended for rehabilitation after the anterior cruciate ligament surgery because they are considered safer than other exercises27,28. This view is not, however, accepted by all authors29, as these exercises may be insufficient for strengthening the quadriceps muscle30, and remaining weakness of the quadriceps muscle is commonly found after knee ligament surgery30,31. The current trend is toward accelerated protocols with immediate range-of-motion training and weight-bearing and a return to sports within four to six months32. Our protocol was more conservative with regard to the return to sports, which was permitted after nine to twelve months. Currently, we believe that rehabilitation for a patient with a knee ligament injury should be supervised by a physiotherapist and should focus on reducing pain and swelling; on restoring the range of motion, strength, and endurance; and on enhancing proprioception and dynamic stability in order to restore function and to minimize disability33.
In conclusion, although previous studies16 that have compared the outcomes of operative and nonoperative treatment of complete and isolated medial collateral ligament injuries have revealed equally good outcomes at the time of the two-year follow-up, we found some significant differences when we compared operative and nonoperative approaches to combined anterior cruciate ligament-medial collateral ligament injuries. Nonoperative treatment of the medial collateral ligament component allows faster restoration of flexion and quadriceps muscle power. All of our patients achieved full knee extension. At one year, muscle power was better in Group II (reconstruction of the anterior cruciate ligament and nonoperative treatment of the medial collateral ligament), and, at two years the muscle power and range of motion of the knee were equal for both groups. These results favor nonoperative treatment of the torn medial collateral ligament in patients with combined anterior cruciate ligament and medial collateral ligament injuries. 