Abstract
A five-year, prospective, randomized follow-up study was done to compare three methods for repair of a rupture of the anterior cruciate ligament of the knee: acute primary repair (Group 1), acute repair with a synthetic ligament-augmentation device (Group 2), and acute repair augmented with an autologous bone-patellar ligament-bone graft (Group 3). One hundred and fifty patients who had an acute rupture of the anterior cruciate ligament were randomized to one of the three repair groups, with fifty patients in each group. The patients were between sixteen and fifty years old (mean, twenty-nine years old). All patients had the operation within ten days after the injury. The rehabilitation protocol was identical for each group. The patients were evaluated prospectively at one, two, and five years with use of the Tegner scoring system for level of activity and the scoring system of Lysholm and Gillquist for function, and the stability of the knee was assessed with clinical examination and with use of the KT-1000 arthrometer. One hundred and thirty-one patients completed the study and ten other patients were known to have had a failure of the procedure, a 94 per cent rate of follow-up.All three groups had a lower level of activity at the five-year follow-up evaluation than they had had before the injury. The patients who had had augmentation with a patellar-ligament graft had a significantly higher mean level of activity at two years than those who had had non-augmented repair (p = 0.002) and those who had had repair with a synthetic ligament-augmentation device (p = 0.01). They also had a significantly higher mean level of function at two years than those who had had non-augmented repair (p = 0.0001) and those who had had repair with a synthetic ligament-augmentation device (p = 0.03) and a significantly higher mean level of function at five years than those who had had non-augmented repair (p = 0.004). The ability to attain full extension improved significantly in all three groups during the five-year follow-up period; the highest gains occurred in the group that had had augmentation with a patellar-ligament graft.Rotatory and anterior instability progressively increased during the follow-up period for all three groups. At one, two, and five years, the knees that had had repair with a patellar-ligament graft were significantly more stable than those that had had non-augmented repair and those that had had repair with a ligament-augmentation device (p < 0.0001 to p = 0.03).The findings of this study reinforce the conclusions of our two-year follow-up report that a non-augmented primary repair should not be performed, a repair with a ligament-augmentation device has an unacceptably high rate of failure (more than one-third of the patients), and a repair that is augmented with the patellar ligament has the best outcome.
The role of acute repair in the treatment of ruptures of the anterior cruciate ligament is controversial8. Acute primary repair was a common mode of treatment in the past16. However, in several recent long-term follow-up studies, the results have been reported to deteriorate with time6,7,15,23, leading to a common belief that this procedure is inadequate. In a continued attempt to preserve the blood supply and the neurosensory role of the anterior cruciate ligament9,18, some surgeons have augmented the primary repair of the anterior cruciate ligament. The short-term results of repair with augmentation have been encouraging1,3,20. However, we are aware of no long-term, prospective, randomized follow-up study in which the results after acute primary repair have been compared with those after repair with synthetic or biological augmentation. Therefore, we began a five-year, prospective, randomized study to compare the results after acute primary repair, acute repair with synthetic augmentation, and acute repair augmented with an autologous bone-patellar ligament-bone graft. We did not use a control group of non-operatively managed patients because instability of the knee is a frequent result in very active patients1,19. The study design was approved by the Norwegian Regional Board of Research Ethics, and written informed consent was obtained from each patient.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Department of Orthopaedic Surgery, Trondheim University Hospital, 7006 Trondheim, Norway. Please address requests for reprints to Dr. Grøntvedt.
‡Department of Orthopaedic Surgery, University of Minnesota, 420 Delaware Street, S.E., Minneapolis, Minnesota 55455.
§Department of Orthopaedic Surgery, Aker Hospital, University of Oslo, 0514 Oslo, Norway.
¶Department of Orthopaedic Surgery, Haukeland Hospital, University of Bergen, N-5021 Bergen, Norway.
Subjects
The original study group included 150 consecutive patients who were between sixteen and fifty years old (mean, twenty-nine years old) and who had an acute rupture of the anterior cruciate ligament, but three patients dropped out during the first year of the study. The patients were managed at three hospitals in Norway from May 1986 to April 1988. One hundred and twenty-five (83 per cent) of the patients sustained the rupture during sports activities, with 116 (77 per cent) of the 150 patients being injured while snow-skiing or playing soccer or team handball. Patients who had a concomitant injury of the meniscus or of the medial collateral ligament were included; those who had a rupture of the posterior cruciate or lateral collateral ligament were excluded. There were nineteen meniscal tears in the patients who had non-augmented primary repair, sixteen in the patients who had repair with a synthetic ligament-augmentation device, and seventeen in the patients who had augmentation with a patellar-ligament graft. All were treated with partial arthroscopic resection or open repair. There were eight injuries of the medial collateral ligament in the patients who had non-augmented repair, eight in the patients who had repair with a ligament-augmentation device, and twelve in the patients who had repair with a patellar-ligament graft. All twenty-eight injuries were treated non-operatively. With the numbers available, we did not find a significant difference among the three groups with regard to the number of concomitant injuries of a meniscus or collateral ligament.
After the injury of the anterior cruciate ligament had been confirmed by Lachman and pivot-shift testing with the patient under general anesthesia and by arthroscopy, sealed envelopes were used to assign the patients randomly to one of three operative procedures: non-augmented repair (Group 1), repair with a synthetic ligament-augmentation device (Group 2), and repair with bone-patellar ligament-bone augmentation (Group 3). There were fifty patients in each group at the end of the inclusion period. There was no significant difference among the groups with regard to the pre-injury level of activity or the age or sex of the patients. All patients were examined and had an operation within ten days after the injury. Meniscal repairs or partial meniscectomies were performed when indicated. The location of the rupture in the anterior cruciate ligament was noted for each group (Table I).
One hundred and thirty-one of the 150 patients returned for evaluation at a minimum of five years (mean, six years; maximum, seven years) after the operation, and ten others were known to have had a reconstruction after the failure of the index operation; the rate of follow-up therefore was 94 per cent. Of the nine remaining patients, three moved abroad during the first year of the study, one had died, and five chose not to participate. All of the patients who are included in the present study were evaluated preoperatively and at one and two years after the operation, as we reported previously5.
Operative Procedures
All repair procedures were performed through a medial arthrotomy. In 1986, at the start of the present study, arthroscopy-assisted or endoscopic techniques were not used in Norway for a repair of a cruciate ligament. The repairs were carried out by general orthopaedic surgeons.
Group 1 (non-augmented repair): The primary repair was performed according to the method that was reported by Palmer. Multiple loop sutures are placed in the remnants of the anterior cruciate ligament in an anteromedial-to-posterolateral direction. Two holes are drilled through the anatomical insertion sites of the anterior cruciate ligament in the femur and tibia with a two-millimeter-diameter Steinmann pin. The sutures are tied over the bone-bridge while tension is manually applied with the knee in 30 degrees of flexion. For this procedure as well as for the two augmented repairs, sufficient tension is applied to prevent more than five millimeters of anterior tibial translation during a Lachman test.
Group 2 (repair with a ligament-augmentation device): The primary repair was performed with the addition of a braided polypropylene augmentation segment as a ligament-augmentation device (LAD; 3M, St. Paul, Minnesota). A Steinmann pin, two millimeters in diameter, is placed just anterior and medial to the anatomical insertion site of the anterior cruciate ligament in the tibia. The hole is overdrilled with a four-millimeter-diameter cannulated reamer. The posterior edge of bone of the tibial tunnel is smoothed with a rasp. A small notchplasty is performed. The over-the-top position is identified and is chamfered with a curved rasp to produce a small groove. The ligament-augmentation device is passed through the tibial tunnel and is sutured to the periosteum of the anteromedial aspect of the tibia with six, seven, or eight interrupted non-absorbable sutures. The remnants of the anterior cruciate ligament are sutured to the ligament-augmentation device with 2—0 Vicryl suture (Ethicon, Somerville, New Jersey). The ligament-augmentation device with the attached anterior cruciate ligament is then routed over the top of the femoral condyle. With the knee in 30 degrees of flexion and while manual tension is applied, the construct is fixed to the lateral femoral condyle with the belt-buckle staple technique. The ligament-augmentation device is skeletally fixed at only one end and is not used as a prosthesis.
Group 3 (repair with a patellar-ligament graft): A bone-patellar ligament-bone graft is obtained from the central third of the ligament. A two-millimeter-diameter Steinmann pin is drilled from the medial aspect of the proximal and anterior part of the tibia, emerging five millimeters anterior and medial to the center of the tibial attachment of the anterior cruciate ligament. This hole is overdrilled with a ten-millimeter-diameter cannulated reamer. With use of the inside-out technique at maximum knee flexion, another Steinmann pin is placed posterior and superior to the center of the femoral attachment of the anterior cruciate ligament. This femoral hole is also overdrilled to ten millimeters. A notchplasty is performed. Tibial fixation is obtained with an interference screw and two sets of non-absorbable sutures passed through holes in the bone block and tied around a cortical-bone screw and washer outside the tunnel. The remnants of the anterior cruciate ligament are sutured to the graft. With the knee in 30 degrees of flexion and while manual tension is applied, the construct is fixed in the femoral tunnel with an interference screw and reinforcing sutures.
Rehabilitation
At the end of the operation, a cast was applied from the proximal part of the thigh to the ankle, with the knee in 30 degrees of flexion. This remained in place for two weeks, after which a brace was used for six more weeks. No weight-bearing on the extremity was permitted during the first eight weeks after the operation. The brace allowed motion from 30 to 60 degrees of flexion during the third and fourth weeks and from 30 to 90 degrees thereafter. No dynamic strengthening exercises were permitted for the last 30 degrees of extension during the first twelve weeks. No participation in contact sports was allowed during the first year after the operation. The rehabilitation was controlled by the same group of physical therapists following the same standard rehabilitation protocol at each of the three hospitals.
Follow-up
The patients were evaluated prospectively by two of us (T. G. and L. E.). The level of activity was evaluated with use of the Tegner activity score. The functional status was graded according to the scoring system designed by Lysholm and Gillquist, which emphasizes the importance of pain, instability, and locking. The physical examination included an assessment of the range of motion (restriction in flexion and extension compared with the uninjured side) and an evaluation of the stability of the knee with manual testing and use of an arthrometer. Anterolateral rotatory instability was evaluated with use of the MacIntosh, Slocum, or flexion-rotation drawer test and was graded as negative (0), trace-positive (1+), moderate (2+), or severe (3+). Anterior instability in 20 degrees of flexion was evaluated with the Lachman test and was graded as negative (0), slight (1+) when there was less than five millimeters of displacement, moderate (2+) when there was five to ten millimeters, or severe (3+) when there was more than ten millimeters, compared with that of the normal knee. Instrumented testing of anterior laxity was performed with a KT-1000 arthrometer (MedMetric, San Diego, California) at 20 degrees of flexion and a load of eighty-nine newtons (twenty pounds).
Isokinetic strength of the quadriceps and hamstrings was also evaluated with use of an isokinetic testing machine (Biodex, Shirley, New York) in fifty-one of the patients at one of the participating hospitals. These patients included sixteen from Group 1, seventeen from Group 2, and eighteen from Group 3. The strength on the injured side was compared with that on the normal side at the five-year follow-up evaluation only. The tests were performed in extension and flexion at 60 and 240 degrees per second. Peak torque and total work were measured. The agonist-antagonist values of the uninvolved and involved sides were calculated at both 60 and 240 degrees per second.
Statistical Methods
The goals of the statistical analysis were to compare the differences in the measurements of outcome among the three groups at each of the follow-up time-points and to compare the changes in the outcomes during the five-year evaluation period. A non-parametric analysis of variance (the Kruskal-Wallis test) was used to determine the significance of the over-all difference in the outcome measurements among the three groups at a particular point in time. Pairwise comparisons among the three groups at each time-point were made with use of the Mann-Whitney U test. The Wilcoxon test for paired data was used to test for the significance of the changes in the outcome measurements for a particular group between any two time-points in the follow-up period. Comparisons were considered significant at a p value of 0.05 or less. We previously reported measurements collected preoperatively and at one and two years postoperatively5. These earlier data were reanalyzed along with the five-year data for the present report.
Activity Level
The mean Tegner activity score for the group that had had non-augmented repair (Group 1) significantly decreased (p < 0.05) from before the injury (6.7 points) to the one-year follow-up examination (4.8 points), remained at that same level at two years, and slightly increased at five years (5.0 points) (Fig. 1). For the patients who had had repair with a ligament-augmentation device (Group 2), the mean score significantly decreased (p < 0.05) from before the injury (6.2 points) to one year after the operation (4.9 points), slightly increased at two years (5.0 points), and remained at that same level at five years. The group that had had repair with a patellar-ligament graft (Group 3) also had a significant decrease (p < 0.01) from before the injury (6.5 points) to one year after the operation (5.1 points), a significant increase to near the level before the injury at two years (6.0 points) (p = 0.008), and then a significant decrease at five years (5.3 points) (p = 0.008).
There were no significant differences in the mean activity scores among the three group at the follow-up evaluations at one and five years, but there was a significant difference at two years (p = 0.004). The patients in Group 3 had a significantly higher mean score at two years than those in Group 1 (p = 0.002) and Group 2 (p = 0.01).
The percentage of patients who had an activity score of 7 to 10 points, which was representative of a high-performance athlete, decreased from before the injury (range, 49 per cent [twenty-three] of forty-seven to 62 per cent [thirty-one] of fifty) to the one-year follow-up visit (range, 22 per cent [eleven] of fifty to 32 per cent [sixteen] of fifty), increased at two years (range, 29 per cent [fourteen] of forty-eight to 39 per cent [eighteen] of forty-six), and then decreased at the five-year evaluation (range, 21 per cent [nine] of forty-two to 27 per cent [thirteen] of forty-eight) (Table II).
Functional Level
The mean Lysholm and Gillquist functional score for Group 1 significantly decreased between the one-year (90.0 points) and two-year follow-up evaluations (86.7 points) (p = 0.001) and slightly increased to 88.3 points at five years (Fig. 2). At five years, eleven (27 per cent) of the remaining forty-one patients in Group 1 (non-augmented repair) had functional scores within the categories of poor and fair (0 to 83 points) (Table II). The mean functional score for the patients in Group 2 (repair with a ligament-augmentation device) remained at nearly the same level at one year (88.9 points) and two years (88.5 points) and slightly increased at five years (91.4 points). At the five-year evaluation, six (14 per cent) of the remaining forty-two patients in Group 2 had functional scores within the categories of poor and fair. The mean functional score for the patients in Group 3 (repair with a patellar-ligament graft) significantly increased between one and two years (from 91.9 to 94.7 points) (p = 0.02) and slightly decreased at five years (93.3 points). Only two (4 per cent) of the remaining forty-eight patients in Group 3 were rated as having poor or fair function at five years.
There were significant differences in the mean functional scores among the three groups at two and five years (p < 0.0001 and p = 0.01, respectively), but with the numbers available not at one year. At the two-year follow-up examination, the patients in Group 3 had a significantly higher mean functional score than those in Groups 1 and 2 (p = 0.0001 and 0.03, respectively). At the five-year follow-up, the patients in Group 3 had a significantly higher mean functional score than the patients in Group 1 (p = 0.004). The main factors responsible for the relative reduction in the functional scores for the patients in Group 1 were pain and the feeling of instability. The patients in Groups 2 and 3 did not have these same levels of pain and instability.
Range of Motion
Extension Deficit
In the group that had had non-augmented repair (Group 1), 24 per cent (twelve) of the patients had an extension deficit of more than 5 degrees at the one-year follow-up examination (Table III). This prevalence decreased to 23 per cent (eleven) of forty-eight patients at two years and to 5 per cent (two) of forty-one patients at five years. Thirty-eight per cent (eighteen) of the forty-seven patients who had had repair with a ligament-augmentation device (Group 2) had an extension deficit of more than 5 degrees at one year; this prevalence decreased to 27 per cent (twelve) of forty-four patients at two years and to 14 per cent (six) of forty-two patients at five years. Sixty-two per cent (thirty-one) of the fifty patients who had had repair with a patellar-ligament graft (Group 3) had an extension deficit of more than 5 degrees at the one-year evaluation, 52 per cent (twenty-four) of forty-six patients had one at two years, and 17 per cent (eight) of forty-eight patients had one at five years.
Postoperative limitation of motion was corrected by a second operation, between six and twelve months after the initial procedure, in four of the patients in Group 3. These four patients were subsequently found to have an extension deficit of less than 10 degrees and a flexion deficit of less than 20 degrees when motion of the involved knee was compared with that of the uninjured knee.
In general, the ability to extend the knee improved significantly in all three groups from the two to the five-year follow-up examination (p = 0.05 for Group 1, p = 0.04 for Group 2, and p = 0.004 for Group 3), with the highest gains occurring in Group 3. With the numbers available, we could detect no significant differences among the three groups at five years. Only one patient in Group 1 and one in Group 3 had an extension deficit of more than 10 degrees at five years.
Flexion Deficit
Thirty-two per cent (sixteen) of the fifty patients in Group 1 had a flexion deficit of more than 10 degrees at one year (Table III). This prevalence decreased to 10 per cent (four) of forty-one patients at five years. The prevalence of patients in Group 2 who had a flexion deficit of more than 10 degrees decreased from 19 per cent (nine) of forty-seven patients at the one-year follow-up evaluation to 12 per cent (five) of forty-two at five years. In Group 3, 26 per cent (thirteen) of the fifty patients had a flexion deficit of more than 10 degrees at the one-year evaluation, decreasing to 13 per cent (six) of forty-eight patients at five years.
In general, the patients in the three groups had an improvement in the ability to attain full flexion during the one to five-year follow-up period. However, four patients in Group 1, five in Group 2, and six in Group 3 still demonstrated a flexion deficit of between 10 and 30 degrees at five years.
Stability
The repair was considered to have failed when the results of the Lachman and pivot-shift tests were at least 2+ and the KT-1000 device demonstrated three millimeters or more of laxity on the tested side than on the contralateral side.
Anterolateral Rotatory Instability
Ten per cent (five) of the fifty patients in Group 1 had 2+ or 3+ (moderate or severe) anterolateral rotatory instability at the one-year follow-up evaluation (Table IV). This prevalence significantly increased to 42 per cent (twenty) of forty-eight patients at two years (p < 0.0001) and to 51 per cent (twenty-one) of forty-one patients at five years. Thirteen per cent (six) of forty-seven patients in Group 2 had 2+ or 3+ rotatory instability at one year. This prevalence remained unchanged (14 per cent [six] of forty-four patients) at two years and then significantly increased to 36 per cent (fifteen) of forty-two patients at five years (p = 0.005). In Group 3, there were no patients who had more than 1+ rotatory instability at the one and two-year follow-up examinations, but 8 per cent (four) of forty-eight patients had 2+ or 3+ instability at five years. On the basis of the MacIntosh, Slocum, or flexion-rotation drawer test, rotatory instability progressively increased during the five-year follow-up period for all three groups. The knees in Group 3 were significantly more stable than those in Groups 1 and 2 at one, two, and five years (p < 0.0001 to p = 0.03). The knees in Group 2 were significantly more stable than those in Group 1 at two and five years (p = 0.0005 and 0.04, respectively).
Lachman Test
Thirty-four per cent (seventeen) of the fifty patients in Group 1 had 2+ or 3+ (moderate or severe) anterior instability at the one-year follow-up examination (Table IV). This prevalence significantly increased to 54 per cent (twenty-six) of forty-eight patients at two years (p = 0.002) and then decreased to 44 per cent (eighteen) of forty-one patients at the five-year evaluation. Twenty-one per cent (ten) of the forty-seven patients in Group 2 had moderate or severe anterior instability at one year, increasing to 25 per cent (eleven) of forty-four patients at two years and to 29 per cent (twelve) of forty-two patients at five years. In Group 3, only 8 per cent (four) of the fifty patients had 2+ or 3+ anterior instability at one year, decreasing to 2 per cent (one) of forty-six patients at the two-year evaluation and increasing to 10 per cent (five) of forty-eight patients at five years. The knees in Group 3 were significantly more stable than those in Groups 1 and 2 at the one, two, and five-year follow-up evaluations (p < 0.0001 to p = 0.01). The knees in Group 2 were significantly more stable than those in Group 1 at two and five years (p = 0.02 and 0.05, respectively).
Measurements with the KT-1000 Arthrometer
The mean anterior laxity in Group 1 significantly increased from 2.7 millimeters of translation at one year to 3.7 millimeters at two years (p = 0.0002) and then decreased to 2.7 millimeters at five years (Fig. 3). The mean anterior laxity for Group 2 did not significantly change during the follow-up period (2.1 millimeters at one year, 2.2 millimeters at two years, and 2.3 millimeters at five years). The mean anterior laxity for Group 3 also did not significantly change during the follow-up period (1.7 millimeters at one year, 1.2 millimeters at two years, and 1.3 millimeters at five years).
There were significant differences in the mean anterior laxity among the three groups at the two and five-year follow-up evaluations (p < 0.0001 and p = 0.0001, respectively). At two and five years, Groups 1 and 2 had a mean anterior laxity that was significantly greater than that in Group 3 (p < 0.0001 to p = 0.009). Group 1 had significantly higher anterior laxity than Group 2 at two years (p = 0.045).
Forty-eight per cent (twenty-three) of forty-eight patients in Group 1, 42 per cent (eighteen) of forty-three patients in Group 2, and 32 per cent (fifteen) of forty-seven patients in Group 3 had a difference in anterior laxity compared with that of the contralateral side (a side-to-side difference) of three millimeters or more at one year (Table IV). At five years, the prevalence of patients who had a side-to-side difference in anterior laxity of three millimeters or more had increased to 66 per cent (twenty-seven) of forty-one patients in Group 1 and to 48 per cent (twenty) of forty-two patients in Group 2. It had decreased to 21 per cent (ten) of forty-eight patients in Group 3.
Isokinetic Testing
Fifty-one patients had testing of isokinetic muscle strength at the five-year follow-up evaluation. The values for peak torque in extension at 60 degrees per second were significantly higher for Group 1 (p = 0.01) and for Group 3 (p = 0.04), compared with the values for Group 2. With the numbers available, there were no significant differences among the three groups with regard to the results of the other isokinetic tests.
Reoperations
Four patients in Group 3 had a second operation between six and twelve months after the first procedure because of persistent loss of extension. Seven patients in Group 1 and three in Group 2 had reconstruction with an autologous bone-patellar ligament-bone graft for gross instability, between the one and five-year follow-up evaluations. The ligament-augmentation device was torn in two of the patients in Group 2. One of these patients sustained a new rotational injury while playing soccer. In the third patient, the ligament-augmentation device loosened from the belt-buckle fixation on the lateral femoral condyle.
Complications
Two superficial infections developed in Group 2. One superficial infection and one deep infection developed in Group 3. These infections resolved with the administration of antibiotics. One deep venous thrombosis was detected in Group 1. At the one-year follow-up examination, patellofemoral crepitation was found during active extension of the flexed knee in 34 per cent (seventeen) of the fifty patients in Group 1, in 43 per cent (twenty) of the forty-seven patients in Group 2, and in 56 per cent (twenty-eight) of the fifty patients in Group 3. With the numbers available, we could find no definite increase in crepitation at the two-year follow-up examination and no change between two and five years.
It is difficult to compare the outcomes of repair procedures for the treatment of rupture of an anterior cruciate ligament because most published studies have been retrospective and non-randomized and have involved different operative techniques, rehabilitation regimens, and follow-up protocols. The present study is prospective and randomized, but its weakness is that the operations were performed in three different hospitals by general orthopaedic surgeons. However, the similarity between the results for the group that had had repair with a patellar-ligament graft and those reported by Clancy et al. suggests that the operative technique was adequate. Sandberg et al. performed a prospective study comparing the results of non-operative treatment (immobilization in a cast) with those of direct repair of the anterior cruciate ligament. Although, with the numbers available, there was no significant difference between the functional outcomes of the two types of treatment, they found a higher percentage of patients who had a positive pivot-shift test in the group that had received non-operative treatment. In long-term retrospective follow-up studies, Fruensgaard et al. and Sommerlath et al. reported that non-operative treatment and simple suture of an acutely torn anterior cruciate ligament produced similar outcomes. In the present study, all three groups had approximately the same levels of activity and function at one year. However, the tests for stability demonstrated a high percentage of failures at one year in the group that had had non-augmented repair (Group 1). Instability then significantly increased at two years and somewhat leveled off at five years. Seven patients in Group 1 had a reoperation because of gross instability. The problems in Group 1 are also reflected by the finding that the levels of activity and function remained low from two to five years. Most failures of non-augmented repairs became apparent during the first two years after the operation, and nearly 50 per cent of the repairs without augmentation became unstable at two and five years. These results were inferior to the results in the two groups that had received augmentation, reinforcing the findings of previous studies1,15,23 on simple repair of a ruptured anterior cruciate ligament.
The ligament-augmentation device was developed by Kennedy et al.11,13 to augment different types of reconstruction of the anterior cruciate ligament until biological remodeling of the graft tissue has produced sufficient strength in the ligament. The ligament-augmentation device apparently was first used by Schabus in the repair of acute ruptures of the anterior cruciate ligament. He hypothesized that the synthetic material would initially share load with the repaired anterior cruciate ligament, protecting it from disruption or attenuation until sufficient tissue growth and remodeling of the ligament had occurred. This would be followed, hypothetically, by a gradual transfer of load from the ligament-augmentation device back to the repaired anterior cruciate ligament. Schabus reported promising clinical results with this procedure.
One of us (L. E.) and colleagues performed a study of cadavera4, analyzing the biomechanics of the same procedures considered in the present follow-up study. The ligament-augmentation device was found to carry approximately 75 per cent of the total load in the composite graft at extension and at 30 degrees, thus providing stress relief for the repair tissue. The ligament-augmentation device carried only 25 per cent of the composite load at 90 degrees.
The patients who had had repair with a ligament-augmentation device (Group 2) in the present study had little change in the levels of activity and function during the five-year follow-up period. At one year, the tests for stability revealed a rate of failure that was approximately half that of the group that had had a non-augmented repair (Group 1), with little change between one and two years. At the five-year follow-up examination, the prevalence of failure due to instability in Group 2 had significantly increased. However, the stability in Group 2 was still significantly better than that in Group 1. Three patients in Group 2 had a reoperation because of gross instability. Our results showed that some of the ligament-augmentation devices broke early after the operation. This may have been caused by damage to the ligament-augmentation device from abrasion by the edge of bone at the end of the tunnels. The later failures may have been caused by fatigue of the ligament-augmentation device. None of the late failures in Group 2 were related to major injuries.
Biological augmentation has been performed with use of the hamstrings21, iliotibial band1, prepatellar retinaculum10, and patellar ligament. Our choice of biological augmentation was based on animal studies by Cabaud et al. and clinical studies by Clancy et al. Cabaud et al. demonstrated that complete reruptures consistently occurred after simple repair of an acutely injured anterior cruciate ligament. They documented improved results when the anterior cruciate ligament was augmented with a patellar-ligament graft. Again, this augmentation procedure theoretically provides additional stability to the repair while growth and remodeling take place and allows earlier and more intensive rehabilitation.
In a prospective study of augmentation of repairs of the anterior cruciate ligament with a bone-patellar ligament-bone graft, Clancy et al. reported that fifty (96 per cent) of fifty-two patients had a good or excellent result. The mean time to the lastest follow-up evaluation in their study was forty-eight months. In the present study, the patients in Group 3 had a significant increase in the levels of both activity (p = 0.008) and function (p = 0.02) from one to two years. From two to five years, the level of activity decreased but was still at a higher level than that of the other two groups. None of the patients in Group 3 had more than a trace-positive pivot-shift sign at one or two years, although this sign was positive in four (8 per cent) of forty-eight of these patients at five years. In Group 3, only nine (20 per cent) of forty-four patients had a side-to-side difference of three millimeters or more as measured by the KT-1000 arthrometer at two years and only ten (21 per cent) of forty-eight patients had such a difference at five years. No patient in this group had a reoperation because of instability; four had a reoperation because of a decreased range of motion.
With the numbers available, we could not find a significant difference in the range of motion among the three groups at the five-year follow-up evaluation. Early problems of extension lag after reconstruction of the patellar ligament have been reported previously14,17,22. The prevalence of this condition has decreased with delayed operations and with arthroscopic and endoscopic techniques. The early problem of a limited range of motion in Group 3 in the present study was probably related to the acute open procedure, the anteromedial placement of the tibial tunnel, and the conservative rehabilitation protocol.
With regard to all of the outcome measurements other than an early limitation in extension, which decreased with time, patellar-ligament augmentation was superior to direct repair and to the use of the ligament-augmentation device. On the basis of these data, we conclude that a non-augmented repair should no longer be performed. In addition, our results show an unacceptably high rate of failure of repairs augmented with the ligament-augmentation device.
Excellent results were obtained with the repairs that included augmentation with the patellar-ligament graft, even though the procedures were performed with an arthrotomy by general orthopaedists and the rehabilitative protocol was extremely conservative. It is conceivable that the more modern endoscopic technique and intensive rehabilitation would produce a similar outcome with less morbidity. A two-year follow-up study of 105 patients who had had a reconstruction of the anterior cruciate ligament arthroscopically, used no brace, and began immediate weight-bearing has so far demonstrated similar stability and less morbidity.
Andersson, C.; Odensten, M.; Good, L.; and |and |Gillquist, J.: Surgical or non-surgical treatment of acute rupture of the anterior cruciate ligament. A randomized study with long-term follow-up. J. Bone and Joint Surg.,71-A: 965-974, Aug. 1989.71-A965
1989
Cabaud, H. E.; Feagin, J. A.; and |and |Rodkey, W. G.: Acute anterior cruciate ligament injury and augmented repair. Experimental studies. Am. J. Sports Med.,8: 395-401, 1980.8395
1980
[PubMed][CrossRef]
Clancy, W. G., Jr.; Ray, J. M.; and |and |Zoltan, D. J.: Acute tears of the anterior cruciate ligament. Surgical versus conservative treatment. J. Bone and Joint Surg.,70-A: 1483-1488, Dec. 1988.70-A1483
1988
Engebretsen, L.; Lew, W. D.; Lewis, J. L.; and |and |Hunter, R. E.: Knee mechanics after repair of the anterior cruciate ligament. A cadaver study of ligament augmentation. Acta Orthop. Scandinavica,60: 703-709, 1989.60703
1989
[CrossRef]
Engebretsen, L.; Benum, P.; Fasting, O.; Mølster, A.; and |and |Strand, T.: A prospective, randomized study of three surgical techniques for treatment of acute ruptures of the anterior cruciate ligament. Am. J. Sports Med.,18: 585-590, 1990.18585
1990
[PubMed][CrossRef]
Feagin, J. A., Jr., and |and |Curl, W. W.: Isolated tear of the anterior cruciate ligament: 5-year follow-up study. Am. J. Sports Med.,4: 95-100, 1976.495
1976
[PubMed][CrossRef]
Fruensgaard, S.; Krøner, K.; and |and |Riis, J.: Suture of the torn anterior cruciate ligament. 5-year follow-up of 60 cases using an instrumental stability test. Acta Orthop. Scandinavica,63: 323-325, 1992.63323
1992
[CrossRef]
Gillquist, J.: Repair and reconstruction of the anterior cruciate ligament: is it good enough?. J. Arthroscopy,9: 68-71, 1993.968
1993
[CrossRef]
Johansson, H.; Sjölander, P.; and |and |Sojka, P.: A sensory role for the cruciate ligaments. Clin. Orthop.,268: 161-178, 1991.268161
1991
[PubMed]
Jonsson, T.; Peterson, L.; and |and |Renstrøm, P.: Anterior cruciate ligament repair with and without augmentation. A prospective 7-year study of 51 patients. Acta Orthop. Scandinavica,61: 562-566, 1990.61562
1990
[CrossRef]
Kennedy, J. C.; Roth, J. H.; Mendenhall, H. V.; and |and |Sanford, J. B.: Presidential address. Intraarticular replacement in the anterior cruciate ligament-deficient knee. Am. J. Sports Med.,8: 1-8, 1980.81
1980
[PubMed][CrossRef]
Lysholm, J., and |and |Gillquist, J.: Evaluation of knee ligament surgery results with special emphasis on use of a scoring scale. Am. J. Sports Med.,10: 150-154, 1982.10150
1982
[PubMed][CrossRef]
McPherson, G. K.; Mendenhall, H. V.; Gibbons, D. F.; Plenk, H.; Rottmann, W.; Sanford, J. B.; Kennedy, J. C.; and |and |Roth, J. H.: Experimental mechanical and histologic evaluation of the Kennedy ligament augmentation device. Clin. Orthop.,196: 186-195, 1985.196186
1985
[PubMed]
Mohtadi, N. G.; Webster-Bogaert, S.; and |and |Fowler, P. J.: Limitation of motion following anterior cruciate ligament reconstruction. A case-control study. Am. J. Sports Med.,19: 620-625, 1991.19620
1991
[PubMed][CrossRef]
Odensten, M.; Lysholm, J.; and |and |Gillquist, J.: Suture of fresh ruptures of the anterior cruciate ligament. A 5-year follow-up. Acta Orthop. Scandinavica,55: 270-272, 1984.55270
1984
[CrossRef]
Palmer, I.: On the injuries to the ligaments of the knee joint. A clinical study. Acta Chir. Scandinavica, Supplementum 53, 1938.
Paulos, L. E.; Rosenberg, T. D.; Drawbert, J.; Manning, J.; and |and |Abbott, P.: Infrapatellar contraction syndrome. An unrecognized cause of knee stiffness with patella entrapment and patella infera. Am. J. Sports Med.,15: 331-341, 1987.15331
1987
[PubMed][CrossRef]
Pitman, M. I.; Nainzadeh, N.; Menche, D.; Gasalberti, R.; and |and |Song, E. K.: The intraoperative evaluation of the neurosensory function of the anterior cruciate ligament in humans using somatosensory evoked potentials. Arthroscopy,8: 442-447, 1992.8442
1992
[PubMed][CrossRef]
Sandberg, R.; Balkfors, B.; Nilsson, B.; and |and |Westlin, N.: Operative versus non-operative treatment of recent injuries to the ligaments of the knee. A prospective randomized study. J. Bone and Joint Surg.,69-A: 1120-1126, Oct. 1987.69-A1120
1987
Schabus, R.: Die Bedeutung der Augmentation für die Rekonstruktion des Vorderen Kreutzbandes. Acta Chir. Austriaca,76 (Supplementum): 1-48, 1988.76 (Supplementum)1
1988
Sgaglione, N. A.; Warren, R. F.; Wickiewicz, T. L.; Gold, D. A.; and |and |Panariello, R. A.: Primary repair with semitendinosus tendon augmentation of acute anterior cruciate ligament injuries. Am. J. Sports Med.,18: 64-73, 1990.1864
1990
[PubMed][CrossRef]
Shelbourne, K. D., and |and |Wilckens, J. H.: Arthrofibrosis in the acute anterior cruciate ligament reconstruction: the effect of timing of reconstruction and rehabilitation protocol. Orthop. Trans.,15: 88-89, 1991.1588
1991
Sommerlath, K.; Lysholm, J.; and |and |Gillquist, J.: The long-term course after treatment of acute anterior cruciate ligament ruptures. A 9 to 16 year followup. Am. J. Sports Med.,19: 156-162, 1991.19156
1991
[PubMed][CrossRef]
Tegner, Y., and |and |Lysholm, J.: Rating systems in the evaluation of knee ligament injuries. Clin. Orthop.,198: 43-49, 1985.19843
1985
[PubMed]