ARTHROSCOPIC RECONSTRUCTION OF THE ANTERIOR CRUCIATE LIGAMENT USING BONE-PATELLAR TENDON-BONE GRAFTS WITH AND WITHOUT AUGMENTATION

Transcription

1 ARTHROSCOPIC RECONSTRUCTION OF THE ANTERIOR CRUCIATE LIGAMENT USING BONE-PATELLAR TENDON-BONE GRAFTS WITH AND WITHOUT AUGMENTATION A PROSPECTIVE RANDOMISED STUDY T. Grøntvedt, L. Engebretsen, T. Bredland From Trondheim University Hospital, Norway I n 100 consecutive patients with chronic deficiency of the anterior cruciate ligament we reconstructed the ligament using a bone-patellar tendon-bone autograft either with or without a Kennedy ligament augmentation device. The patients had an aggressive rehabilitation programme supervised by two physiotherapists. They were followed prospectively for at least two years by one surgeon, and assessed after six months and at one and two years. No significant functional or clinical difference was found between the two groups and the stability of the knees did not deteriorate with time in either group. The addition of a Kennedy ligament augmentation device gave no better results than the bone-patellar tendon-bone technique alone. J Bone Joint Surg [Br] 1996;78-B: Received 16 October 1995; Accepted after revision 11 January 1996 The use of bone-patellar tendon-bone autologous grafts to reconstruct the anterior cruciate ligament generally provides satisfactory results (Daniel, Stone and Riehl 1990). It is the strongest of the biological grafts used for this purpose (Noyes et al 1984). Animal studies have shown, however, an initial decrease in strength in these grafts followed by a period of revascularisation and remodelling (Clancy et al 1981; Butler et al 1989). In this early period, the mechanical properties of the graft are reduced, and, theoretically, it may stretch or rupture if subjected to high loads. In man the duration of this period of diminished strength is not known, but most surgeons do not recommend a full return to twisting sports until six to 12 months after operation. T. Grøntvedt, MD, Orthopaedic Surgeon L. Engebretsen, MD, PhD, Orthopaedic Surgeon T. Bredland, MD, Orthopaedic Surgeon Trondheim University Hospital, 7006 Trondheim, Norway. Correspondence should be sent to Dr T. Grøntvedt British Editorial Society of Bone and Joint Surgery X/96/51203 $2.00 Kennedy et al 1980 proposed the use of a ligament augmentation device (3M, St Paul, Minnesota) and suggested that load-sharing between the device and the biological graft would protect the latter during the period of degeneration and weakening and allow it to regain mechanical strength without risk of elongation or rupture. This view was supported by Van Kampen, Mendenhall and McPherson 1987 who postulated that the load would be transferred from the synthetic device to the autogenous graft gradually as the biological graft remodelled and became stronger. Because of the strength of the device, earlier functional recovery should be possible. McPherson et al 1985, in animal studies, found no specific synovial reaction or signs of rejection when using the device, and this has been confirmed in clinical reports (Moyen et al 1992; Noyes and Barber 1992; Muren, Dahlstedt and Dalen 1995). Our prospective, randomised study was designed to compare the results of arthroscopically-assisted reconstructions of chronic rupture of the anterior cruciate ligament using the middle third of the patellar ligament with bone blocks from the patella and the tibial tuberosity, with or without reinforcement with the Kennedy ligament augmentation device, and using an aggressive regime of rehabilitation. Any difference in outcome should be detected during the remodelling and growth of the graft and be apparent by two years. Patients and methods Between May 1991 and January 1993, we included 100 consecutive patients of mean age 26 years (16 to 48) in the study. There were 45 men and 55 women. Ruptures of the anterior cruciate ligament had been confirmed by both clinical and arthroscopic examination. The mean time from injury to operation was 3.5 years (1 month to 15 years). Before operation the patients had agreed to participate in a protocol approved by the Norwegian Research Ethical Committee. They were randomised using the sealed-envelope method to receive arthroscopic reconstruction either with a bone-patellar tendon-bone ligament graft alone or combined with the Kennedy ligament augmentation device. The indications for operation were clinical instability with pain and/or swelling during sports or other physical activities. Sporting activities led to 95% of the ruptures, with VOL. 78-B, No. 5, SEPTEMBER

2 818 T. GRØNTVEDT, L. ENGEBRETSEN, T. BREDLAND Table I. Causes of injury in 100 patients Number Team handball 42 Soccer 38 Ski-ing 12 Other sports 3 Traffic accidents 4 Work accident 1 team handball and soccer responsible for most of them (Table I). There were 49 patients in the augmented and 51 in the non-augmented group. Of the augmented patients, 19 had had meniscal resections before the operation while 20 of the non-augmented patients had been so treated. At the time of reconstruction, 24 patients (49%) in the augmented and 20 (39%) in the non-augmented group had meniscal resections or re-resections. There were no statistically significant differences between the two groups with regard to the level of preinjury activity, age, gender or previous surgery. Patients with clinical signs of knee injuries other than rupture of the anterior cruciate ligament and meniscal damage were not included in the study. Ninety-two patients completed the two-year follow-up. One patient required removal of the ligament with the augmentation because of a deep infection postoperatively. This was registered as a complication, and the patient was not included in the follow-up examinations. Two patients, one in each group, sustained rupture of the anterior cruciate ligament on the contralateral side during the follow-up period and were also excluded from the study. Five patients, two in the augmented group and three in the other, had graft failure during follow-up; all had a further reconstruction before the final review. At the two-year examination, there were 45 patients in the augmented and 47 in the non-augmented group. No patient was lost to follow-up; the five patients who required reoperation are included as failures. Operative technique. The procedure was carried out under epidural anaesthesia with the use of a tourniquet. After a diagnostic arthroscopy with assessment and, if necessary, treatment of the menisci, a notch plasty was done. The graft was then harvested subcutaneously from the middle third of the ligamentum patellae, with bone blocks of 1 X 1 X 2.5 cm from the patella and the tibial tubercle, through two short transverse incisions, one over the patella and the other over the tibial tubercle. In the augmentation group, a Kennedy ligament augmentation device 8 mm wide and 18 cm long was sutured to the graft using a tension board (3M, St Paul, Minnesota) with a load of approximately 2 kg on both the device and the ligament. The bone blocks were trimmed to fit snugly into a tunnel of 10 mm in diameter. The anteromedial aspect of the tibia just medial to the tuberosity was then dissected subperiosteally, and, using the 3M drill guide (3M, St Paul, Minnesota), a Kirschner (K)- wire was brought in from the outside to the centre of the original insertion of the anterior cruciate ligament on the tibial plateau. The K-wire was overdrilled with a 10 mm cannulated reamer, and the intra-articular entrance of the tunnel chamfered to avoid sharp bone edges. A skin incision 4 to 5 cm long was then made parallel to and directly over the iliotibial band at the lateral femoral condyle. The iliotibial band was split along the length of its fibres and the vastus lateralis muscle lifted anteriorly, giving access to the posterolateral part of the lateral femoral condyle. Again by the use of the drill guide, a K-wire was brought in to exit at the posterior part of the anatomical origin of the anterior cruciate ligament on the femoral condyle. It was overdrilled with a 10 mm cannulated reamer and the intra-articular bone edges chamfered. The graft was brought into place by pulling it through the tibial tunnel into the knee, and then into the femoral tunnel. The bone block in the femoral tunnel was fixed from outside in with a cannulated interference screw of 7 or 9 X 25 mm (Linvatec Corporation, Largo, Florida) located on one of the cancellous sides of the block to avoid cutting the threads or damaging the augmentation device, which was then fixed to the lateral femoral condyle with two staples in a belt-buckle fashion. The composite graft or the single ligament was pretensioned with approximately 2 kg, and the bone block in the tibial tunnel fixed with an interference screw with the knee in 10 of flexion. When used, the tibial end of the augmentation device was fixed to the tibia with staples giving double-end fixation. With the arthroscope in place, the knee was then brought through a full range of motion to assess the graft tension and to check that there was no impingement of the graft in the notch with the knee in full extension. Rehabilitation. The rehabilitation programme was the same for both groups and supervised by the same two physiotherapists. Immediately after operation, the patients started knee movement, and full passive extension of the knee was carried out several times each day. No brace was used. Full weight-bearing was allowed as soon as tolerated. Closed-chain exercises with full extension and flexion of the knee were allowed as early as possible, but extension against resistance from 0 to 30 on a quadriceps bench was not allowed during the first six weeks. Jogging began after 10 to 12 weeks. After six months the patients were allowed to return to any sporting activity, provided that the strength of the thigh muscles on the operated side was at least 85% of that on the contralateral side, and that controlled functional training had been carried out without difficulty. Evaluation. The patients were evaluated preoperatively, and reviewed by one of the authors (TG) at six months, and at one and two years after surgery. Their subjective assessment of their knee function was recorded and graded as excellent, good, fair or poor. The level of activity was measured by the Tegner activity score (Tegner and Lysholm 1985). The functional status was graded according to the scoring system of Lysholm and Gillquist 1982, which emphasises the importance of pain, instability and swelling. THE JOURNAL OF BONE AND JOINT SURGERY

3 ARTHROSCOPIC RECONSTRUCTION OF THE ACL WITH AND WITHOUT AUGMENTATION 819 The physical examination included an assessment of the range of movement and of knee stability by manual and arthrometer testing. Anterior instability in 20 of flexion was evaluated by the Lachman test and was graded as negative, slight (1+, <5 mm), moderate (2+, 5 to 10 mm) or severe (3+, 10 mm or more), compared with the normal knee. Anterolateral rotatory instability was determined by the McIntosh, Slocum or flexion-rotation-drawer tests, and was graded as negative, just positive (1+), moderate (2+) or severe (3+). Instrumented testing of anterior laxity was performed with a KT-1000 arthrometer (MEDmetric, San Diego, California) at 20 and an anterior load of 67 N and 89 N, and also using the Maximum Manual Displacement test, as described by Daniel et al (1985). We consider the injured-uninjured difference at 89 N pull to be the principal measurement value. The isokinetic strength of the quadriceps and hamstrings was also measured using the Biodex system (Biodex Corporation, Shirley, New York). The strength of the injured side was compared with that of the normal side throughout the follow-up period. The tests were carried out at 60 and 240 / sec and the peak torque and total work were measured. Statistical analysis. The Mann-Whitney, chi-squared and the unpaired t-tests were used. A p value of <0.05 was considered as significant. Results Of the two augmented and three non-augmented grafts which failed, four of the reruptures occurred with minor injuries, mainly hyperflexion, while the fifth was due to a twisting movement during a soccer game. At reoperation, we found that the femoral tunnel had been placed too far anterior in the first four patients; in the fifth both the tibial and femoral tunnels seemed to be anatomically correct. Before operation all the patients graded their subjective knee function as fair (57 patients) or poor (43 patients). At the two-year follow-up, 40 out of 45 patients in the augmented and 43 of 47 patients in the non-augmented group said that they had good or excellent function. The remaining patients in each group had fair function, and none was graded as poor. The mean Tegner activity score for the augmented group decreased from a preinjury level of 7.5 to 5.5 at six months, increasing to 6.2 at one year and remaining at the same level at two years. For the non-augmented group, the mean preinjury score was 7.3, decreasing to 5.2 at six months, and again increasing to 5.8 at one year and further to 6.0 at two years. There were no statistically significant differences in the Tegner activity level in the two groups at either the preinjury or follow-up examinations. Both groups improved their Lysholm functional score significantly. The mean score for the augmented group increased from 70 to 92 at six months, to 94 at one year and to 95 at two years and that for the non-augmented group from 68 to 93 at six months and at one year, and finally to Table II. Lysholm functional score in both groups preoperatively and at six months and one and two years postoperatively, by number and percentage of group Lysholm score* Augmented Non-augmented 0 to (31) 15 (29) 65 to (65) 35 (69) 84 to 94 2 (4) 1 (2) 95 to (0) 0 (0) 0 to 64 0 (0) 0 (0) 65 to 83 5 (11) 2 (5) 84 to (47) 25 (58) 95 to (42) 16 (37) One year (n = 44) (n = 46) 0 to 64 0 (0) 0 (0) 65 to 83 2 (5) 3 (7) 84 to (36) 21 (46) 95 to (59) 22 (48) 0 to 64 0 (0) 0 (0) 65 to 83 2 (4) 0 (0) 84 to (29) 19 (40) 95 to (67) 28 (60) * 0 to 64 = a poor outcome; 65 to 83 = fair; 84 to 94 = good; and 95 to 100 points = excellent 96 at two years. Preoperatively, 96% of the augmented and 98% of the non-augmented groups had functional scores in the poor and fair categories (0 to 83 points) while at the twoyear follow-up, 4% in the augmented and 0% in the nonaugmented groups were graded as fair or poor (Table II). There were no statistically significant differences between the two groups at any of the evaluation points. Range of movement. At the two-year follow-up, four patients in the augmented and six in the non-augmented group had a flexion deficit of more than 5 compared with the non-injured side. None of the patients had an extension deficit of 5 or more. Stability. The reconstruction was considered to have failed if the Lachman and pivot-shift tests were at least 2+, and there was more than 3 mm of laxity on the tested side than on the contralateral side as demonstrated by the KT-1000 device. Before operation all the patients had a 2+ or 3+ Lachman test. At six months, one patient in the non-augmented group had a 2+ instability. At one year, one patient in the augmented and two in the other group had a 2+ Lachman test while at two years all the examined patients were considered as stable (Table III). There were no statistically significant differences between the two groups at any time. All the patients had a positive pivot-shift sign before operation. All were stable after six months, while at one year two patients in the augmented and three in the nonaugmented group were graded as unstable. At two years, only one patient in the augmented group had a 2+ pivot-shift sign (Table IV). VOL. 78-B, No. 5, SEPTEMBER 1996

4 820 T. GRØNTVEDT, L. ENGEBRETSEN, T. BREDLAND Table III. Lachman test in both groups preoperatively and at six months and at one and two years postoperatively, by number and percentage of group Lachman test* Augmented Non-augmented 0 0 (0) 0 (0) 1 0 (0) 0 (0) 2 22 (45) 22 (43) 3 27 (55) 29 (57) 0 38 (84) 29 (68) 1 7 (16) 13 (30) 2 0 (0) 1 (2) One year (n = 43) (n = 45) 0 29 (68) 33 (73) 1 13 (30) 10 (22) 2 1 (2) 2 (5) 0 36 (80) 31 (66) 1 9 (20) 16 (34) 2 0 (0) 0 (0) * see text Table IV. Anterolateral rotatory instability (pivot-shift) test in both groups preoperatively and at six months and one and two years postoperatively, by number and percentage of group Rotatory instability test* Augmented Non-augmented 0 0 (0) 0 (0) 1 0 (0) 0 (0) 2 7 (14) 4 (8) 3 42 (86) 47 (92) 0 39 (87) 37 (86) 1 6 (13) 6 (14) 2 0 (0) 0 (0) One year (n = 44) (n = 45) 0 38 (86) (76) 1 4 (9) 8 (18) 2 2 (5) 1 (2) 3 0 (0) 2 (4) 0 31 (69) 32 (68) 1 13 (29) 15 (32) 2 1 (2) 0 (0) * see text Nine of the patients in each group had a side-to-side difference of 3 mm or less before operation when tested on the KT-1000 arthrometer at 89 N load. At operation, however, it was confirmed that all these patients had a complete rupture of the anterior cruciate ligament, and preoperative testing with maximum manual force showed that only one had anterior displacement of less than 4 mm. During the followup period the KT-1000 measurements showed an improvement in both groups, with no statistically significant difference between them (Table V), but the improvement from preoperative to postoperative values was significant (p < 0.001). At two years, one patient in each group had a sideto-side difference of 4 to 5 mm as measured by the KT-1000 with 89 N load, and accordingly they were considered as unstable. Before operation, 12 (25%) of the patients in the augmented and 15 (29%) in the non-augmented group had symptoms consistent with patellofemoral pain syndrome. At the two-year follow-up there were five (11%) in the augmented and four (9%) in the non-augmented group. Isokinetic testing. The Biodex tests showed that the mean deficit in quadriceps strength compared with the uninjured side at six months was approximately 25% in both groups, decreasing to about 15% at one year and 10% at two years, with similar values in peak torque and total work (Fig. 1). The hamstrings, however, showed normal and equal strength at six months and at one and two years. There were no statistically significant differences between the two groups concerning muscular strength at any interval. Table V. KT-1000 arthrometer results in both groups preoperatively and at six months and one and two years postoperatively, by number and percentage of group Anterior laxity (mm)* Augmented Non-augmented < 4 9 (18) 9 (18) 4 to 5 23 (47) 21 (41) > 5 17 (35) 21 (41) < 4 44 (98) 41 (95) 4 to 5 1 (2) 2 (5) > 5 0 (0) 0 (0) One year (n = 43) (n = 44) < 4 41 (95) 41 (93) 4 to 5 2 (5) 3 (7) > 5 0 (0) 0 (0) < 4 44 (98) 46 (98) 4 to 5 1 (2) 1 (2) > 5 0 (0) 0 (0) * see text Complications. In the augmented group, one patient had a superficial wound infection successfully treated with antibiotics, and another had a deep infection which required removal of the composite graft. In the non-augmented group, one patient needed arthroscopic debridement after eight months because of a decreased range of movement. At the one- and two-year follow-up, movement in this knee was normal. THE JOURNAL OF BONE AND JOINT SURGERY

5 ARTHROSCOPIC RECONSTRUCTION OF THE ACL WITH AND WITHOUT AUGMENTATION 821 Fig. 1 Difference in isokinetic quadriceps muscle strength (peak torque and total work, %) between the operated and non-operated side, measured in a Biodex machine. Discussion Since Kennedy and his colleagues introduced the ligament augmentation device in 1980, several reports have been published on the short-term effects of the use of the augmentation of various biological grafts both in animals and in man. McPherson et al 1985, in an animal study with a two-year follow-up, found the augmented group to be mechanically stronger, but this could not be verified statistically because of the small numbers involved. Most studies have used the augmentation device in combination with the Marshall-MacIntosh procedure. The quadriceps-tendon patellar-ligament graft used in this technique is one of the weakest for cruciate reconstruction (Noyes et al 1984). Poor results have been reported with this method even using an augmentation. Jonsson et al 1992 in a twoyear follow-up study found 100% recurrence of anterior laxity irrespective of the use of augmentation. Roth et al 1985 in a retrospective study of augmented and non-augmented knees, found that in subjective questioning, clinical testing and radiological examination the augmented group had better results. Objective testing with KT-1000 and Cybex isokinetic strength measurements, however, showed no difference between the two groups, and the patients with augmented knees had a follow-up two years shorter than those with no augmentation. Moyen et al 1992 published the results of a prospective, randomised two-year follow-up in which they compared the Marshall-MacIntosh technique alone and with augmentation with a Kennedy device. They found no significant functional or clinical difference between the two groups with respect to residual laxity or subjective symptoms. Muren et al 1995, in a similar study, came to the same conclusions. All these studies, however, used a conservative programme of rehabilitation. Noyes and Barber 1992 studied the effect of the ligament augmentation device on reconstructions using bone-patellar ligament-bone allografts. The patients were followed for a mean of 34 months, and the failure rate was found to be 29% in knees not augmented and 27% in the augmented group. They concluded that the addition of the ligament augmentation device did not improve the results of allograft reconstruction. Our results show a failure rate of 7% and a significant decrease in patellofemoral pain during the follow-up period. We had few complications. The Biodex tests showed that some patients did not regain full quadriceps function until one to two years after operation despite our aggressive rehabilitation programme, but the hamstrings recovered normal strength earlier. Our programme of aggressive rehabilitation was designed to maximise the possible protective effect of the augmentation device, but the similar results in the two groups suggest that the device made no difference to the result. Fixation of the device at both ends should reduce the load on the graft by approximately one third (Hanley et al 1989), but this protection made no difference. Our results are better than those in our previous reports and we obtained a normal range of movement (Engebretsen et al 1990). This is probably due to a combination of delayed surgery, a reduced surgical exposure, good graft placement and aggressive rehabilitation. Since the device was fixed at both ends, and thus may function as a prosthesis, stress protection may have occurred. This could lead to increased laxity when the ligament augmentation device eventually breaks. If it is taking most of the load of our aggressive rehabilitation, we consider that breakage would probably have happened before the two-year review. VOL. 78-B, No. 5, SEPTEMBER 1996

6 822 T. GRØNTVEDT, L. ENGEBRETSEN, T. BREDLAND 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. References 1. Butler DL, Grood ES, Noyes FR, et al. Mechanical properties of primate vascularised versus non-vascularised patellar tendon grafts: changes over time. J Orthop Res 1989;7: Clancy WG, Narechania RG, Rosenberg TD, et al. Anterior and posterior cruciate ligament reconstruction in Rhesus monkeys. J Bone Joint Surg [Am] 1981;63-A: Daniel DM, Stone ML, Sachs R, Malcom L. Instrumented measurement of anterior knee laxity in patients with acute anterior cruciate ligament disruption. Am J Sports Med 1985;13: Daniel DM, Stone ML, Riehl B. Ligament surgery: the evaluation of results. In: Daniel DM, Akeson WH, Connor JJ (eds). Knee ligaments: structure, function, injury and repair. New York: Raven Press, 1990: Engebretsen L, Benum P, Fasting O, Mølster A, Strand T. A prospective, randomized study of three surgical techniques for treatment of acute ruptures of the anterior cruciate ligament. Am J Sports Med 1990;18: Hanley P, Lew WD, Lewis JL, et al. Load sharing and graft forces in anterior cruciate ligament reconstructions with the ligament augmentation device. Am J Sports Med 1989;17: Jonsson H, Elmqvist L-G, Kärrholm J, Fugl-Meyer A. Lengthening of anterior cruciate ligament graft: roentgen stereophotogrammetry of 32 cases 2 years after repair. Acta Orthop Scand 1992;63: Kennedy JC, Roth JH, Mendenhall HV, Sanford JB. Presidential address: intraarticular replacement in the anterior cruciate ligament-deficient knee. Am J Sports Med 1980;8: Lysholm J, Gillquist J. Evaluation of knee ligament surgery results with special emphasis on use of a scoring scale. Am J Sports Med 1982;10: McPherson GK, Mendenhall HV, Gibbons DF, et al. Experimental, mechanical and histologic evaluation of the Kennedy ligament augmentation device. Clin Orthop 1985;196: Moyen BJL, Jenny J-Y, Mandrino AH, Lerat J-L. Comparison of reconstruction of the anterior cruciate ligament with and without a Kennedy ligament-augmentation device: a randomized prospective study. J Bone Joint Surg [Am] 1992;74-A: Muren O, Dahlstedt L, Dalen N. Reconstruction of old anterior cruciate ligament injuries: no difference between the Kennedy LAD-method and traditional patellar tendon graft in a prospective randomized study of 40 patients with 4-year follow-up. Acta Orthop Scand 1995;66: Noyes FR, Barber SD. The effect of ligament-augmentation device on allograft reconstructions for chronic ruptures of the anterior cruciate ligament. J Bone Joint Surg [Am] 1992;74-A: Noyes FR, Butler DL, Grood ES, Zernicke RF, Hefzy MS. Bio-mechanical analysis of human ligament grafts used in knee-ligament repairs and reconstructions. J Bone Joint Surg [Am] 1984;66-A: Roth JH, Kennedy JC, Lockstadt H, McCallum CL, Cunning LA. Polypropylene braid augmented and nonaugmented intraarticular anterior cruciate ligament reconstruction. Am J Sports Med 1985;13: Tegner Y, Lysholm J. Rating systems in the evaluation of knee injuries. Clin Orthop 1985;198: Van Kampen CL, Mendenhall HV, McPherson GK. Synthetic augmentation of biological anterior cruciate ligament substitutions. In: Jackson DW, Drez D, eds. The anterior cruciate deficient knee: new concepts in ligament repair. St Louis: CV Mosby, 1987: THE JOURNAL OF BONE AND JOINT SURGERY