In this prospective study we investigated 70 patients with chronic anterior instability who underwent anterior cruciate ligament reconstruction with a Dacron prosthesis ( Stryker Corp., Kalamazoo, MI, USA). Follow-up intervals were 2, 5 and 7-11 (mean 9) years. Fiftyfive (79%) patients could be followed to 9 years. Within the follow-up time 47% of the prostheses were diagnosed as ruptured, the annual rupture frequency was 5%. A second reconstruction was performed in 20 patients (29%) and five patients had a third reconstruction. Only 9 of 29 patients with a ruptured prosthesis did not want further surgery. Fifty-two patients (74% of the original group) underwent a radiographic examination. Eighty-three percent had radiological changes in the operated knee, 40% with joint space reduction up to 50%. Patients with medial instability as well as patients in the age group 30 years or more at the index operation showed more severe osteoarthritic changes.

Summarizing the functional results of the remaining patients with a ligament in place after 9 years, only 14% of the original group had acceptable stability and knee function.

The results of the present study confirm that artificial ligaments are not acceptable for the treatment of the ACL-deficient knee. The question of whether ACL reconstruction can prevent osteoarthrosis has to be answered negatively, because the rate of radiological osteoarthitis was increased 10 times compared to the normal knee. It seems from previously published long term studies that other reconstructions are not better in this regard.

Introduction:

In 1993 we reported the 5 year results of ACL reconstruction using the Stryker-Meadox dacron ligament prosthesis in a prospective study of 70 patients (10). We found that the immediate postoperative stability was gradually lost and the rupture rate was 23%. The results with the dacron prosthesis were not optimal and the failure rate was unacceptable. Similar experiences with other prostheses were also published (3,6,8,9,11,20 ) and ligament prostheses were abandoned.

Nevertheless it is extremely important to evaluate failed techniques in a long term perspective in order to analyze the potential problems caused by the procedure. Some authors have claimed that prosthetic ACL substitution was an " iatrogenic model of degenerative arthritis in the human knee" (13). The present study is a continued follow up of the previously reported prospective patient material analyzing the frequency of osteoarthrosis 7-11 years after the operation.

The original study group ( N=70, 24 women and 46 men, mean age at surgery 25.4 ± 6 years) was selected from a larger population of patients with chronic anterior instability after sport accidents. In spite of a rehabilitation program they had instability preventing participation in work and/or sports. Patients with osteoarthrosis on standing radiographs or posterior instability were excluded.

Combined medial instability was present in 49%, previous meniscectomy in 37% and previous ACL surgery in 32%. Eighty-six percent had been active in competitive sports before injury, but before surgery 61% had problems with activities of daily life and could not participate even in physical fitness activities. The mean time between injury and the index operation was 44 ± 39 months. The patients were fully informed about the potential risks and possible benefits with the prosthesis and made their choice between the prosthesis and autologous soft tissue procedures. All patients could be followed to the 2-year examination, 69 (99%) were examined after 5-years and 55 patients (79%) after 7-11 years ( mean 9 years). At that time three patients had emigrated, one was excluded because of pregnancy, three because of knee surgery for osteoarthrosis and eight patients could not be traced or refused to be examined.

Statistics.

Student’s T-test was used for continuous variables and the chi-square test for categorical data. Relations between variables were analyzed by standard regression methods.

The operation was performed by a miniarthrotomy as previously described (10). Rehabilitation was usually complete between 6 and 12 months and the patients were allowed to go back to their desired activities.

A standard manual knee examination was performed preoperatively and at each follow-up interval.

Knee stability. A laxity tester (Orthopedic Systems Inc.; Hayward, CA, USA) was used to document antero-posterior (AP) displacement of the tibia at a 90N load before surgery, after fixation of the ligament, and at each follow-up interval on both the injured and the uninjured knee. The total displacement ( anterior plus posterior) is reported as well as the injured-to-uninjured knee differences for total AP displacement (T difference).

Knee stability was also documented using standard manual tests graded 0 to 3 ( 0= normal).

The ligament rupture rate was based on arthroscopic diagnosis of ligament rupture and revision surgery.

For evaluation of subjective knee function, the Lysholm knee score was used (16). A level of <84 points corresponds to knee problems in daily life, 84-94 points to occasional or light problems in sports but no problems in daily life. Athletic participation on a low level is possible. More than 94 points indicates subjectively normal knee function.

The work and sports activities were graded with the Tegner activity scale (22). Levels 7 to 10 consist of competitive sports, levels 4 to 6 represent physical fitness activities and moderate to strenuous work activities, levels 1 to 3 reflect light work and activities of daily living. Level 0 means permanent disability because of knee problems. The final rating was composed of all aspects of evaluation. The different levels are listed in Table I.

At the 9-year follow-up, 52 patients (74% of the original group) underwent a radiographic examination of both knee joints including weight-bearing, antero- posterior and lateral roentgenograms at 30∞ of knee flexion. A combined scale, based on Fairbanks signs (7) and Ahlbäcks radiologic classification of arthritis (2), was used for evaluation: Grade 0: No changes; Grade I: Sharpening of the edges, beginning formation of osteophytes, sclerosis and flattening of the condyles, but no joint space reduction (Fairbanks signs); Grade II: Grade I plus joint space reduction up to 50 percent (Ahlbäck, Grade I); Grade III: joint space reduction more than 50 percent up to joint space obliteration (Ahlbäck, Grade II).

The postoperative course was uneventful in 66 patients. No patient had difficulties regaining full range of motion. In no case postoperative infection occurred and there were no cases of clinical synovitis. The complications are listed in table II and have been partly described previously (10).

Forty-six patients had arthroscopy during the 9 year follow-up period for diagnostic purposes; 32 of these had staple removal. Twenty-eight patients had meniscus problems that were treated arthroscopically. There were 16 medial and 8 lateral meniscectomies and 4 medial meniscal repairs. At the 9 year follow-up, 48% of the menisci were still intact. Three patients had repeat surgery for gonarthrosis.

In one patient the prosthesis was removed after 8 months because of a synovial fistula. She was seen at 5 and 9 years with an unstable knee, but otherwise no complications. She did not want further surgery. Two patients had the prosthesis removed in connection with secondary knee surgery to alleviate persistent, severe pain. One was treated with knee fusion and the other with a total knee prothesis.

In the analysis of rupture rate these patients were excluded as well as the patient lost to follow up after 2 years. Within the 9 years follow-up time 29 of the remaining 66 prostheses were diagnosed as ruptured (44%). The ligament rupture rate was calculated as the number of ruptured ligaments per time period expressed as a fraction of the available ligaments at the beginning of that time period (Figure 1). The annual rupture frequency was 5%. Nine patients were satisfied with their knee function in spite of the ruptured prosthesis. A new reconstruction was performed in 20 cases (30%), in 9 using a polyethylene-implant, in 6 cases another dacron-implant and in 5 cases bone-patellar-tendon-bone strip. Five patients were reoperated a third time because the second-implant (3 polyethylene, 2 dacron) had ruptured.

A number of pre- and intra-operative factors, such as poor tunnel placement, untreated medial instability and previous meniscal or ligament surgery, led to a significant increase of the 5-year rupture rate (10), but the effect had disappeared at the 9-year follow up. Sex, preinjury- and postoperative activity levels had no effect on the rupture rate.

Radiologic evaluation:

Out of the 52 evaluated patients, 83% had radiologic changes in the operated knee, 40% with joint space reduction up to 50% ( Grade II). In the non-operated knee, only 4% had radiological signs Grade II (p< 0.001). The ligament status ( removed or remaining) had no effect on the rate of radiological osteoarthitis. Out of the 21 patients with grade II radiological osteoarthritis one had high tibial osteotomy and another a total knee prosthesis 6 and 10 years after the reconstruction respectively.

Patients with medial instability at the index operation had a higher frequency of osteoarthritis grade II (50% vs. 20%, p<0.05). Osteoarthritis grade II was also two times more frequent in the age group 30 years or more at the index operation ( 67% vs. 33%, p<0.05). There was no significant correlation between radiological osteoarthrosis and various factors like sex, preinjury activity, preoperative time, number of operations, and postoperative activity levels.

Knee stability could be evaluated in 54 of the 55 patients seen at 9 years. One patient had a total knee prothesis and was not tested.

Instrumented tests: For the injured knee, the AP displacement diminished significantly after operation but then increased again mainly during the first two years (p<0.01). The AP displacement of the uninjured knee also increased. (Figure 2). At 5 years 63% of the remaining patients had £ 2 mm T difference. At the 9 years follow-up, 69% of the remaining patients had £ 2 mm difference, 13% had ≥ 5 mm T difference. The distribution of laxity values in the remaining group at nine years is given in table III. There was a significant correlation between the total AP displacement of the injured knee and that of the uninjured knee at the 5-year follow-up, ( r= 0.64 for women and 0.54 for men, p< 0.001), and also at the 9-year follow-up (r=0.37, p<0.05).

Manual tests. Preoperatively, the Lachman test was increased by 1 grade in 3 patients, by 2 grades in 66 patients, and by 3 grades in 1 patient. The pivot shift was increased by 1 grade in 9 patients, by 2 grades in 60 patients, and by 3 grades in 1 patient. These tests were normalized by the operation. During the follow-up, an increase was observed and the 9-year distributions are shown in Figure 3.

There was a significant correlation between manual stability tests and displacement measurements (total AP displacement/Lachman, r = 0.61; total AP displacement/pivot shift, r = 0.68, p<0.001). There was no correlation between total AP displacement or T-difference and subjective symptoms or activity level.

The functional results were evaluated in all 55 patients at 9 years. In 28 of them the ligament had been removed because of rupture and in one case because of a synovial fistula. Therefore 27 of the examined patients still had the ligament in place.

Subjective Symptoms: The Lysholm- scores improved from a mean of 75± 13 before operation to 92± 11 at 12 months (p< 0.001). The mean scores remained the same up to 9 years (92± 12). Of the 27 patients with a ligament in place, 18 had ≥ 95 points, 6 between 84 and 94 points, and 3 < 84 points. In the group with removed ligaments only 5 had ≥ 95, 8 between 84 and 94 and 15 <84 ( p<0.003).

Activity: The mean activity level increased from the preoperative level (mean, 2± 2) to level 5 at 2 and 5 years, at 9 years it was still significantly increased compared to the preoperative level ( mean 4± 2; p< 0.004). At 9 years, 90% of the patients had lowered their activity level, 7% had an unchanged level, and 3% had a higher level than before injury. About 50% of the patients were active on a recreational level. There was no difference in activity wether the ligament had been removed or not.

There was no correlation between the activity levels at 2, 5, and 9 years and various factors like sex, age, preinjury activity, or preoperative time. There was a relation between the Lysholm and Tegner scores; patients with low Lysholm scores had also low activity levels.

Final Rating: Of the remaining 27 patients with ligaments in place ( 39%), in the 9 year follow-up, 10 were in the "excellent" group (14% of the initial 70), 6 patients were classified as "good" (9% of 70), 11 were in the fair/poor group. Summarizing the results at 9 years, only 14% patients had acceptable stability and knee function. Twenty-three patients (33%) had secondary surgery because of ligament rupture (n=20) or osteoarthritis (n=3).

The present follow up confirms the conclusion of the previous report (10) that artificial ligaments are not acceptable for the treatment of the ACL-deficient knee. It also expands the information gained by the 5 year follow-up. It is important to note that the intraoperative technical details seem to play a role only for the short term survival of the substitute. The ruptures were diagnosed at a steady rate with fatigue failures replacing the early failures from technical errors. Our ligament survival curve shows that the 50% rupture/removal limit is reached around 9 years, but estimation on the 5 year data indicated almost 12 years half life (10). However in vitro tests of the Stryker Dacron prosthesis (8) showed an even shorter survival time than the number of cycles necessary for 9 years of use. During the long term follow up the fatigue limit of the prosthesis was reached and therefore probably all prostheses were ruptured irrespective of the subjective and objective outcome. In fact a pathologic T-difference could be demonstrated at 9 years in 30% of the patients with remaining ligaments.

A long term follow up will show a false improvement in subjective results because only the patients who can accept their knee status will remain, the others will have tried other treatment. The small group with remaining ligaments at the 9 year follow up represents subjects who are able to cope with their instability problems and who have lower demands on the knee joints. At 5 years 23% were in the lowest activity levels (0-3) but at 9 years 40% of the remaining group had reduced their activities to that level but nevertheless one third of them had low subjective scores indicating unsatisfactory compensation for their problem. It is interesting to note that only 14% of the patients finally had a subjectively acceptable result. They could maintain the desired activity level without symptoms, had stability measurements within the normal range and only minor changes on the x-ray films. These patients would probably have been successful with any treatment. A similar frequency of acceptable results has been presented with conservative treatment alone ( 4,23). During the follow up time several patients left the study group because of various complications leaving only 47% of the original group ( excluding lost to follow up). The complication and reoperation rates are quite high (53%). About 30% of the original group had repeat ligament reconstruction which is higher than with other types of reconstructions ( 1,5,14,17,18,24). In fact only one third of the patients with ruptured ligaments decided not to have another reconstruction. Even if the majority of the reoperations occurred during the first 5 years one cannot expect that instability problems will not recur if the ligament ruptures later as shown by the lower subjective knee function scores in the group with ruptured ligaments at nine years. In comparison Aglietti et al (1) reported 11% ( no repeat ligament surgery ) and Johnson et al (12) 23% complications including 7% repeat ligament surgery after reconstruction with a patellar tendon strip. Pritchard et al (18) using allografts had 27% complications, but no repeat ligament surgery. In all studies the complication and reoperation rates are lower than in our material.

It is interesting to note that irrespective of the ligament status ( prosthesis removed or remaining) the distribution of laxity differences at 9 years was very similar to what was found by Aglietti et al (1) after patellar tendon reconstruction in a 4 to 10 year follow up. He demonstrated a laxity difference of 0-3 mm in 64% of his patients with 90 N load which is similar to our finding of 74% in the same range. Pritchard et al (18) using allograft reported 76% 0-3 mm difference after a similar follow up time.

The rate of osteoarthrosis in the operated knee was high but not higher than what has been found previously after primary repair (12,21) or other reconstructions (1,18). It seems that the prosthesis did not accelerate the arthritic process. Recent studies suggest that the initial trauma may cause the major increase in osteoarthrosis risk and thereafter the increase is only moderate ( 15, 19). In comparison to the normal knee the rate of grade II changes was increased about 10 times which is higher than after meniscectomy at a comparable time interval(19). It is also interesting to note that a knee with more severe instability at the time of operation had a higher frequency of osteoarthritic changes which stresses the importance of the initial trauma for the development of osteoarthrosis. The degree of osteoarthrosis was however low, most patients having only Grade II ( up to 50% joint space reduction) but in three cases the osteoarthritis lead to major surgery.

The long term results with this prosthetic ligament are clearly inferior in comparison to other reconstruction methods. The question of whether ACL reconstruction can prevent osteoarthrosis has to be answered negatively, and it seems from previously published long term studies that other reconstructions are not better in this regard.

Acknowledgments

This study was supported by grants from the Swedish National Center for Research in Sports, the Östergötland County Council Research fund and by a scholarship from the Swedish Institute.

1. Aglietti P, Buzzi R, D´Andria S, Zaccherotti G: Long-term study of anterior cruciate ligament reconstruction for chronic instability using the central one-thrid patellar tendon and a lateral extraarticular tenodesis. Am J Sports Med 20: 38-45, 1992.
2. Ahlbäck S: Osteoarthrosis of the knee. A radiographic investigation. Acta Radiol (Suppl) 277: 7-72, 1968.
3. Ahlfeldt SK, Larson RL, Collins HR: Anterior cruciate reconstruciton in the chronically unstable knee using an expanded polytetrafluoroethylene (PTFE) prosthetic ligament. Am J Sports Med 15: 326-330, 1987.
4. Andersson C, Odensten M, Good L, 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 Joint Surg 71-A: 965-974, l989.
5. Clancy WG, Nelson DA, Reider B, Narechiana RG: Anterior cruciate ligament reconstruction using one-third of the patellar ligament, augmented by extraarticular tendon transfers. J Bone Joint Surg 64A: 352-359, 1982.
6. Dandy DJ, Gray AJR. Anterior cruciate ligament reconstruction with the Leeds-Keio prosthesis plus extra-articular tenodesis. Results after six years. J Bone Joint Surg 76-B: 193-197, 1994.
7. Fairbank TJ: Knee joint changes after meniscectomy. J Bone Joint Surg 30B: 664-70, 1948.
8. Friedman MJ: Prosthetic anterior cruciate ligament.Clin Sports Med 10: 499-513, 1991.
9. Fox JM, Karzel RP, Diefendorf DR: Four years experience with the Goretex prosthetic ligament in anterior cruciate deficient knees. Proc 57th annual meeting of the American Academy of Orthopedic Surgeons, pp 221, New Orleans, 1990.
10. Gillquist J, Odensten M: Reconstruction of old anterior cruciate ligament tears with a Dacron prosthesis. Am J Sports Med 21: 358-366, 1993.
11. Indelicato PA, Pascale MS, Huegel MO: Early experience with the Gore-Tex polytetrafluoroethylene anterior cruciate ligament prosthesis. Am J Sports Med 17: 55-62, 1989.
12. Johnson RJ, Eriksson E, Häggmark T, et al: Five to ten year follow-up evaluation after reconstruction of the anterior cruciate ligament. Clin Orthop183: 122-140. 1984.
13. Klein W, Jensen KU: Synovitis and artificial ligaments. Arthroscopy 8: 116-124, 1992.
14. Kornblatt I, Warren RF, Wickiewitz TL: Long term follow up of anterior cruciate ligament reconstruction using the quadriceps tendon substitution for chronic anterior cruciate ligament insufficiency. Am J Sports Med 16: 444-448, 1988.
15. Lohmander LS, Dahlberg L, Ryd L, Heinegård D. Increased level of proteoglycan fragments in knee joint fluid after injury. Arthritis Rheum 32: 1434-1442, 1989.
16. Lysholm J, 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.
17. Noyes FR, Barber SD, Mangine RE: Bone-patellar ligament-bone and fascia lata allografts for reconstruction of the anterior cruciate ligament. J Bone Joint Surg 72-A:1125-1136, 1990.
18. Pritchard JC, Drez D, Moss M, Heck S. Long- followup of anterior cruciate ligament reconstructions using freeeze-dried fascia lata allografts. Am J Sports Med 23: 593-596, 1995.
19. Rockborn P, Gillquist J: Outcome of arthroscopic meniscectomy. The role of preexisting cartilage fibrillation in a 13 year follow-up of 60 patients. Int J Sports Med in press, 1996
20. Sledge SL, Steadman JR, Silliman JF, et al: Five-year results with the Gore-Tex anterior cruciate ligament prosthesis. Am J Knee Surg 5: 65-70, 1992.
21. Sommerlath K, Lysholm J,Gillquist J: The long term course after treatment of acute anterior cruciate ligament ruptures. Am J Sports Med 19: 156-161, 1991
22. Tegner Y, Lysholm J: Rating systems in the evaluation of knee ligament injuries. Clin Orthop 198: 43-49, 1985.
23. Tegner Y, Lysholm J, Lysholm M, Gillquist J:Strengthening exercises for old cruciate ligament tears. Acta Orth Scand 57: 130-134, 1986.
24. Zarins B, Rowe CR: Combined anterior cruciate-ligament reconstruction using semitendinosus tendon and iliotibial tract. J Bone Joint Surg 68-A:160-177, 1986.

Correspondence to:

Wolfgang Maletius, MD

Department of Sports Medicine

Faculty of Health Sciences, Linköping University

58185 Linköping, Sweden.

Figure 1.

Ligament survival with time. The slope of the line corresponds to a rupture rate of 5% per year. Almost half of the ligaments had been removed at 9 -11 years.

Figure 2.

Development of the total AP displacement for the injured (∞ ) and the uninjured (∑ ) knee over the 9-11 year follow-up.

Figure 3.

Distribution of pivot shift grades before, at five years and 9-11 years follow up.

Table I:

Failure in one item moved the patient to a lower category.

Table II. Complications.

Table III. Distribution of laxity difference in millimeters I-N at 90N load at 9 years postop.

Figure 1.

Figure 2.

Figure 3.