In 1990, we reported the results for the first forty-seven consecutive patients who had had a reconstruction with an allograft for an acute rupture of the anterior cruciate ligament at our institution31. To our knowledge, there have been no other reports in the English-language literature on the use of allografts in skeletally mature patients solely for the reconstruction of an acute rupture of this ligament; other investigators have reported the results of reconstructions with an allograft for both acute and chronic ruptures21,23,38,41,42 or for only chronic ruptures4,20,22,27,28,40. Also, although one study documented the results of reconstruction of the anterior cruciate ligament with an allograft at a mean of nearly four years postoperatively35, in other such investigations the mean duration of follow-up has been less than four years14,20-23. We recently reported that abnormal anterior-posterior displacements occurred as late as four years postoperatively in four (5 per cent) of eighty-four patients who had had reconstruction of the anterior cruciate ligament4; therefore, we believed that a longer-term follow-up study was required. Importantly, we could not find a report on this subject in which the early postoperative results were compared, for the same population of patients, with those that were assessed several years later. An analysis of any change in results over the long term is critical to an understanding of the ability of allografts to provide long-term stability and function of the knee.
The purpose of the current study was to determine the results of operative treatment with use of one of two types of allograft for acute ruptures of the anterior cruciate ligament, and to assess whether there had been any change in these results between the early (two to four-year) and the later (five to nine-year) follow-up evaluation.
*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. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was the Cincinnati Sportsmedicine Research and Education Foundation.
†Sportsmedicine Research Department, Deaconess Hospital, 311 Straight Street, Cincinnati, Ohio 45219.
Patients
Ninety-six patients who had had an acute, complete rupture of the anterior cruciate ligament were operated on by the senior one of us (F. R. N.) between November 1982 and April 1987. Twenty-one of these patients did not meet the criteria for inclusion in this study: thirteen had had a bilateral rupture of the anterior cruciate ligament, five had had a traumatic rupture of the allograft (three fascia lata and two bone-patellar ligament-bone allografts) within the first postoperative year, one had had a rupture of the posterior cruciate ligament, and two had died of unrelated causes before the minimum five-year follow-up interval. Seven additional patients were not available for the later evaluation.
Thus, sixty-eight (91 per cent) of the seventy-five patients who met the criteria for inclusion were available. They had been followed for a mean of seven years (range, five to nine years) postoperatively. These patients included forty-four of the forty-seven from our previous report31 and twenty-four subsequent patients who met the criteria for inclusion. There were thirty-six male and thirty-two female patients, and their mean age at the time of the operation was twenty-three years (range, fourteen to fifty-one years). Sixty-five patients (96 per cent) had sustained the injury of the knee during sports activities. No patient had a history of trauma to the knee before the injury of the anterior cruciate ligament. The mean interval between the original injury and the index operation was three weeks (range, one to twelve weeks).
Forty-one patients had had diagnostic arthroscopy before the index operation; partial meniscectomy had been performed during ten of the arthroscopic procedures and meniscal repair, during eight. All sixty-eight patients had had a fully positive pivot-shift test preoperatively, and all were found to have a complete rupture of the anterior cruciate ligament at the time of the index operation.
Evaluation
The objective evaluation consisted of examination with a KT-1000 arthrometer (MedMetric, San Diego, California) at eighty-nine newtons at the early evaluation and at eighty-nine and 134 newtons at the later evaluation. The arthrometric examination at the early evaluation was done by two experienced examiners; one, whose reliability has been reported previously43, performed all of the tests at the later evaluation. The differences in the measurements between the affected knee and the contralateral, normal knee were used for analysis.
A comprehensive examination of the knee, which included a rating of general, tibiofemoral, patellofemoral, and alignment-related factors, was performed. Subluxation was recorded in millimeters for all motions of both knees29. The result of the pivot-shift test was graded on a scale of 0 to 3, with 0 indicating no pivot shift; grade 1, a slip; grade 2, a jerk with gross subluxation; and grade 3, gross subluxation with impingement of the posterior aspect of the lateral side of the tibial plateau against the femoral condyle.
A classification system for determining the function of the anterior cruciate ligament on the basis of arthrometric data has been described previously28. This system was devised for patients who had been operated on for a chronic rupture of the ligament and for whom the preoperative and postoperative displacement values could be compared. In the current study, preoperative testing could not be performed in many patients because of swelling and pain in the knee after the acute injury. Therefore, a new classification system was developed for this group. A knee that had less than a three-millimeter increase in displacement and a negative pivot-shift test was classified as functional; a knee that had a three to 5.5-millimeter increase and a negative pivot-shift test, as partially functional; and a knee that had either more than a 5.5-millimeter increase in displacement or a grade of 2 or 3 on the pivot-shift test, as having had failure of the graft. When using this classification system, we recognized that, along with the reconstructed anterior cruciate ligament, secondary restraints may provide a minimum restraining force.
The appearance of the articular cartilage at the time of the index operation was classified according to a system that was described previously30. A lesion was considered noteworthy if it was at least fifteen millimeters in diameter and there was fissuring and fragmentation of more than one-half of the involved articular surface, or if any subchondral bone was exposed.
An independent examiner assessed the radiographs (anteroposterior; lateral, with the knee flexed 30 degrees; and patellofemoral axial), made at the later examination for all sixty-eight patients, for any narrowing of the patellofemoral or tibiofemoral joint space. Narrowing was classified as either none or mild, moderate (loss of less than one-half of the total joint space), or severe (loss of one-half or more of the total joint space). Additionally, the femoral and tibial bone tunnels were measured on the anteroposterior and lateral radiographs of fifty-six patients (82 per cent). The method used was similar to that reported by Fahey and Indelicato14.
The subjective evaluation and over-all rating scheme have been described in detail previously31. All patients completed a written questionnaire and were then interviewed. An over-all rating of excellent, good, fair, or poor was determined on the basis of twenty factors in the manner described previously31. For a result to be rated excellent, all but one score had to be in that category, with the remaining score being in the good category. A good result was one in which all scores were in the excellent or good category with none in the fair or poor category. A fair result was given when any score was fair and a poor result, when any score was poor. This is a rigorous system, designed to determine the impairment of the knee related to the initial injury, the operation, and subsequent trauma. Knees that have moderate narrowing of the joint space receive at least a fair rating, and those that have severe narrowing receive a poor rating.
Operative Procedure
The operative procedures that were used for the intra-articular placement of the allograft have been described previously31. Fifty-eight operations (85 per cent) were performed through a limited medial arthrotomy, just adjacent to the patellar ligament, to allow placement of the allograft without dislocation of the patella. In the remaining ten knees, the allograft was placed with the assistance of an arthroscope with use of a procedure31 initiated in 1986.
Forty patients, who were operated on between November 1982 and March 1986, received a fascia lata allograft. Thirty-two of these grafts were freeze-dried and secondarily sterilized with ethylene oxide, and the remaining eight were freeze-dried only. The remaining twenty-eight patients, who were operated on between January 1986 and April 1987, received a fresh-frozen bone-patellar ligament-bone allograft. All grafts were obtained from tissue banks certified by the American Association of Tissue Banks. None of the bone-patellar ligament-bone allografts had gamma irradiation before implantation, as we did not begin to use that method of sterilization until mid-1987. We continue to recommend its use27,28.
Sixty-four additional procedures were performed concurrently with the index operations. These procedures included nineteen repairs of the medial collateral ligament, eighteen iliotibial-band extra-articular procedures, and twenty-seven meniscal repairs. With use of statistical methods similar to those reported previously27,28,31, we analyzed the effect of these additional procedures on anterior-posterior displacement and on the over-all rating at the latest follow-up evaluation. With the numbers available, we could detect no significant effect of the extra-articular procedure on anterior-posterior displacement (p = 0.45) or the over-all rating (p = 0.94), of the repair of the medial collateral ligament on anterior-posterior displacement (p = 0.21) or the over-all rating (p = 0.10), or of the meniscal repair on anterior-posterior displacement (p = 0.53) or the over-all rating (p = 0.76). These findings led us to believe that these patients could be included in the current study without biasing the results.
A postoperative program of immediate motion of the knee and rehabilitation exercises, as has been described in detail previously31,32, was used for all patients. The program was designed to include several months of protection of the allograft to allow for healing, which we believe is delayed compared with healing of autogenous patellar-ligament grafts10. Continuous passive motion of the knee was begun on the day of the operation. A Bledsoe brace (Medical Technology, Grand Prairie, Texas) was worn by all patients until full weight-bearing was begun. Partial weight-bearing was begun by the seventh postoperative day and was gradually increased to full weight-bearing at approximately the eighth postoperative week for all patients except those who had had a concomitant repair of the medial collateral ligament. For those patients, toe-touch weight-bearing only was permitted for the first four postoperative weeks, after which there was a slow progression to full weight-bearing by the twelfth week.
All patients except those who had had a concomitant repair of the medial collateral ligament were allowed a range of motion of 0 to 90 degrees immediately. The range of flexion was increased to 120 degrees by the second postoperative week and to 135 degrees by the third postoperative week. Patients who had also had a repair of the medial collateral ligament were allowed 10 to 90 degrees of flexion for the first four postoperative weeks to protect the soft-tissue suture lines. From the fifth to the eighth week, the motion was gently increased to 0 to 110 degrees and thereafter, to 0 to 130 degrees by the twelfth week.
The exercises in the rehabilitation program were similar for all patients. Muscle-strengthening exercises were begun the day after the operation. Other exercises included mobilization of the patella, straight-leg-raising, isometric exercises, and electrical muscle stimulation. In the third postoperative month, the patients were encouraged to begin a swimming program and a progressive-resistance exercise program in a protected range of 30 to 90 degrees unless symptoms occurred with these activities. By the tenth to the twelfth postoperative month, a progressive program of running was initiated if the knee was asymptomatic with this activity, the strength of the quadriceps was at least 70 per cent that of the contralateral limb, and there was no damage to the articular cartilage that would contraindicate this activity.
Statistical Analysis
The results for the patients who had received a fascia lata allograft were analyzed separately from those for the patients who had received a bone-patellar ligament-bone allograft. All analyses were conducted with StatView II software (Brain Power, Calabasas, California). Paired two-tailed Student t tests and chi-square tests with Yates continuity correction were used to determine significant differences between these two groups for the parameters that were rated preoperatively, at the short-term follow-up evaluation, and at the later follow-up evaluation. The level of significance was set at p = 0.05. Since there were 125 comparisons in the study, there is a 99.8 per cent chance (1 - [1 - 0.05n]) that at least one comparison may appear to show a spuriously significant difference18.
An analysis of any change in the results between the early and the later follow-up evaluation was performed. To be included, a patient had to have been evaluated in both the early (two to four-year) and the later (five to nine-year) postoperative period. The only exceptions were three patients who had had an excellent result one year postoperatively; these patients were not able to return for the two to four-year evaluation but continued to have an excellent result at the seven to nine-year evaluation. These patients were included in the analysis of deterioration. Of the sixty-eight patients, only one—who had had a fair long-term over-all rating but had not been available for the short-term evaluation—could not be included. The mean interval between the early and the later evaluation for the sixty-seven patients was fifty-six months (range, twenty to ninety-six months): for six patients (9 per cent), it was twenty months to three years; for twelve (18 per cent), more than three to four years; for thirty-seven (55 per cent), more than four to six years; and for twelve (18 per cent), more than six to eight years.
For the analysis of changes in the results over time, we selected a subset of five factors, including four of the twenty that were used to determine the over-all rating. A stepwise linear-regression model performed on a subset of forty-seven patients from our previous study31 revealed the combination of pain, jumping, KT-1000 arthrometric data, and patellofemoral crepitus to be significant (r2 = 0.98) in the determination of the over-all rating26. Therefore, these four variables and the over-all rating were used for this analysis.
Deterioration of the result of the KT-1000 arthrometric examination was defined as an increase in anterior-posterior displacement of more than two millimeters as compared with the short-term follow-up value and an increase of at least three millimeters as compared with the value for the contralateral limb. Deterioration of the result with regard to patellofemoral crepitus consisted of an increase from none or mild at the early follow-up evaluation to moderate or severe at the later evaluation. The jumping score was considered to have deteriorated if it had decreased at least one level (on a four-level gradient scale) between the two follow-up evaluations. The pain score was considered to have deteriorated if it had decreased at least one level (on a six-level gradient scale), unless it had decreased from level 10 (a normal knee) to level 8 (no pain with running, twisting, or turning activities) for patients who had not returned to highly strenuous activities. The over-all rating was considered to have deteriorated when a knee that had had an excellent or good rating at the early follow-up evaluation had a fair or poor over-all rating at the later evaluation.
Objective Evaluation
Testing with the KT-1000 arthrometer at eighty-nine newtons at the early evaluation revealed that fifty-two (78 per cent) of sixty-seven patients had less than a three-millimeter increase in displacement, fourteen (21 per cent) had between three and 5.5 millimeters, and one (1 per cent) had more than 5.5 millimeters. At the later evaluation, with testing at eighty-nine newtons, fifty-four (79 per cent) of the sixty-eight patients had less than a three-millimeter increase in displacement, thirteen (19 per cent) had between three and 5.5 millimeters, and one (1 per cent) had more than 5.5 millimeters. With testing at 134 newtons, forty-nine (77 per cent) of the sixty-four patients tested had less than a three-millimeter increase in displacement, thirteen (20 per cent) had between three and 5.5 millimeters, and two (3 per cent) had more than 5.5 millimeters. With the numbers available, there was no significant difference in displacement between the knees that had a fascia lata allograft and those that had a bone-patellar ligament-bone allograft at either the early evaluation (1.8 ± 2.1 and 0.9 ± 1.9 millimeters, respectively [mean and standard deviation]; p = 0.09) or the later evaluation (1.0 ± 2.9 and 0.7 ± 3.3 millimeters, respectively; p = 0.75) (Fig. 1).
With the numbers available, no significant deterioration was found with regard to the displacement values between the early and the later evaluation (p = 0.62; Table I). In eight patients (12 per cent), displacement increased between the two evaluations; however, in only three patients did this increase result in failure of the graft (a difference of more than 5.5 millimeters between the involved and the contralateral limb).
A repair of the medial collateral ligament, performed concomitantly with the index operation, had a significant effect on the KT-1000 arthrometric values (p = 0.02; Table II). When tested at the 134-newton force level, ten of the seventeen patients who had had such a repair had less than a three-millimeter increase in displacement as compared with thirty-nine (83 per cent) of the forty-seven patients who had not had such a repair. With the numbers available, the concomitant performance of a meniscal repair or an extra-articular procedure with the index operation had no significant effect on deterioration of the KT-1000 arthrometric values (p = 0.35 and 0.33, respectively).
Examination of the Knee and Functional Rating
Before the index operation, all of the patients had at least a grade-2 pivot shift. At the most recent follow-up examination, sixty-six (97 per cent) had a grade of 0 or 1 and two (3 per cent) had a grade of 2.
With use of both the KT-1000 arthrometric and the pivot-shift-test data at the early evaluation, forty-eight (75 per cent) of sixty-four allografts were classified as functional; fourteen (22 per cent), as partially functional; and two (3 per cent), as having failed. (Four grafts could not be classified because of incomplete data.) At the later evaluation, fifty (74 per cent) of the sixty-eight allografts were classified as functional; thirteen (19 per cent), as partially functional; and five (7 per cent), as having failed. Three of the grafts that had been rated as functional at the early evaluation gradually stretched and failed.
Preoperative valgus stress-testing at 25 degrees of flexion of the knee had revealed that seven knees had five more millimeters of opening of the joint space medially than the contralateral knee and twelve had ten more millimeters of opening. These nineteen knees all had a repair of the medial collateral ligament. At the later evaluation, none of these knees had more than a two-millimeter increase in opening of the joint space medially as compared with the contralateral knee.
None of the sixty-eight knees had an increase in posterior translation or external tibial rotation as compared with the contralateral knee. In four knees, preoperative varus stress-testing had revealed a three to five-millimeter increase in the opening of the joint space laterally as compared with the contralateral knee. At the later evaluation, two of these knees still had a three to five-millimeter increase and two did not.
Preoperatively, sixty-seven (99 per cent) of the patients had had no or only mild palpable patellofemoral crepitus, and one had had moderate crepitus. At the early evaluation, sixty-two (93 per cent) of sixty-seven patients had no patellofemoral crepitus and five (7 per cent) had moderate crepitus. At the later evaluation, fifty-one (75 per cent) of the sixty-eight patients had no crepitus, sixteen (24 per cent) had moderate crepitus, and one (1 per cent) had severe crepitus. Twelve (19 per cent) of the sixty-two patients who had normal crepitus at the early evaluation had had an increase in palpable patellofemoral crepitus between the early and the later evaluation; in seven of these twelve, the increase was accompanied by an increase in symptoms related to the patellofemoral joint, which resulted in a fair or poor over-all rating. There was no significant difference, with the numbers available, between the patients who had received a fascia lata allograft and those who had received a bone-patellar ligament-bone allograft, with regard to the deterioration of the values for patellofemoral crepitus (18 per cent [seven and five patients, respectively] had deterioration in each group; p = 0.88) (Table I). There was also no significant difference, between the two allograft groups, with regard to the percentage of patients who had each classification of patellofemoral crepitus (normal, moderate, and severe) at either the early (p = 0.38) or the later evaluation (p = 0.63) (Fig. 2). The rate of conversion from no or mild crepitus preoperatively to moderate or severe crepitus at the most recent evaluation was 25 per cent (seventeen of sixty-seven knees).
At the later evaluation, no patient had moderate or severe tibiofemoral crepitus. Four patients had a slight joint effusion that did not need to be aspirated.
At the later evaluation, the first analysis of the radiographs showed that all knees had no or only mild narrowing of the patellofemoral and tibiofemoral joint spaces. The second analysis, the measurement of the tibial and femoral tunnels (performed for fifty-six knees), showed a mean width of the tibial tunnel of 10.4 ± 2.6 millimeters (range, six to twenty-two millimeters). Forty-four knees (79 per cent) had no enlargement of the tibial tunnel compared with the size (ten to eleven millimeters) that had been produced during the operation, nine (16 per cent) had an increase of one to two millimeters, two (4 per cent) had an increase of three to four millimeters, and one knee (2 per cent) had a noteworthy increase of eleven millimeters. The mean width of the femoral tunnel at the later evaluation was 10.3 ± 0.9 millimeters (range, seven to twelve millimeters). Fifty-two knees (93 per cent) had no increase in the size of the femoral tunnel compared with the size that had been produced during the reconstruction, and four (7 per cent) had an increase of one to two millimeters.
Articular Cartilage at the Index Operation
At the time of the index operation, eleven (16 per cent) of the sixty-eight patients had an abnormal surface of articular cartilage: four, on the undersurface of the patella; two, on the medial femoral condyle; and six, on the lateral femoral condyle. (One patient had two abnormal joint surfaces—on the undersurface of the patella and on the lateral femoral condyle.) Seven (25 per cent) of the twenty-eight patients who had a bone-patellar ligament-bone allograft had abnormal surfaces, compared with four (10 per cent) of the forty who had a fascia lata allograft.
Subjective Evaluation and Return to Activities
At the later evaluation, a significant difference was found, between the patients who had a fascia lata allograft and those who had a bone-patellar ligament-bone allograft, with regard to the mean scores for pain (p = 0.009), swelling (p = 0.006), and full giving-way (p = 0.04) and also with regard to those for functional limitations with stair-climbing (p = 0.006), jumping (p = 0.04), and twisting or turning (p = 0.006) (Table III). Of the twenty-eight patients who had a bone-patellar ligament-bone allograft, only two (7 per cent) had pain with activities that required running and twisting or turning (Fig. 3). The rest of the group was able to perform either these activities or jumping and hard pivoting without pain. Of the thirty-nine patients who had a fascia lata allograft who were rated at the later evaluation, three (8 per cent) had pain with any sports activity and eight (21 per cent) had pain with running and twisting or turning. All but one patient who had a bone-patellar ligament-bone allograft had no giving-way with activities that involved jumping and hard pivoting. However, of the patients who had a fascia lata allograft, three (8 per cent) had giving-way with running and twisting or turning and five (13 per cent), with jumping and hard pivoting. Swelling with sports activity was noted by two patients (7 per cent) who had a bone-patellar ligament-bone allograft and by eleven (28 per cent) who had a fascia lata allograft.
Only two patients (7 per cent) who had a bone-patellar ligament-bone allograft had moderate-to-severe problems with running at the later evaluation, compared with five patients (13 per cent) who had a fascia lata allograft. Two patients (7 per cent) who had a bone-patellar ligament-bone allograft had moderate-to-severe problems with activities involving twisting or turning, compared with ten patients (25 per cent) who had a fascia lata allograft. Only two (3 per cent) of the entire series of sixty-eight patients had moderate difficulties with walking and stair-climbing; all others had no problems with activities of daily living.
For eight patients (12 per cent), the subjective pain score had decreased at least one level from the early to the later follow-up evaluation. There was no significant difference, with the numbers available, between the patients who had a bone-patellar ligament-bone allograft and those who had a fascia lata allograft, with regard to the prevalence of deterioration of the pain score (p = 0.30; Table I and Fig. 3). In eight other patients, the pain score improved at least one level from the early to the later evaluation.
For fourteen patients (21 per cent), the subjective jumping score decreased at least one level from the early to the later evaluation. There was no significant difference, with the numbers available, between the patients who had a bone-patellar ligament-bone allograft and those who had a fascia lata allograft, with regard to the prevalence of deterioration of the jumping score (p = 0.25; Table I and Fig. 4). Fifteen other patients had improvement of at least one level in the jumping score between the early and the later evaluation.
A subset analysis was performed to determine the effect of a repair of the medial collateral ligament or an iliotibial-band extra-articular procedure on symptoms and limitations. No deleterious effect was found with regard to pain (p = 0.12), swelling (p = 0.71), or giving-way (p = 0.33). However, patients who had had either of these additional procedures had significantly lower scores than those who had had only the allograft reconstruction with regard to kneeling (p = 0.04), running (p = 0.04), and jumping (p = 0.05).
There was no significant difference, with the numbers available, between the two allograft groups with regard to the types of sports activities in which the patients participated (Table IV), either preoperatively (p = 0.40) or at the time of the later follow-up evaluation (p = 0.74). Before the injury, twenty-two patients (79 per cent) who had a bone-patellar ligament-bone allograft and thirty-five patients (88 per cent) who had a fascia lata allograft had participated in strenuous sports that involved jumping, pivoting, and cutting. At the later evaluation, ten patients (36 per cent) who had a bone-patellar ligament-bone allograft and twelve patients (30 per cent) who had a fascia lata allograft had returned to these strenuous activities. Approximately one-half of the patients had decreased their level of sports activity but were asymptomatic at their current level. Only five patients (7 per cent) in the entire population were participating despite symptoms and limitations, against medical advice.
Over-All Rating
At the early evaluation, the result was rated excellent for thirteen (21 per cent) of sixty-three patients, good for twenty-five (40 per cent), fair for twenty-one (33 per cent), and poor for four (6 per cent). (The results for four patients could not be rated because of incomplete data.) At the later evaluation, the result was rated excellent for fifteen (22 per cent) of the sixty-eight patients, good for thirty (44 per cent), fair for eighteen (26 per cent), and poor for five (7 per cent). No significant difference was found, with the numbers available, between the two allograft groups with regard to the over-all rating at either the early (p = 0.06) or the later evaluation (p = 0.31) (Fig. 5).
Nine patients (13 per cent) had a deterioration of the over-all rating from excellent or good at the early evaluation to fair or poor at the later evaluation. There was no significant difference, between the two allograft groups, with regard to the prevalence of deterioration of the over-all rating (p = 0.58; Table I). Six other patients (9 per cent) had an increase in the over-all rating from fair at the early evaluation to excellent or good at the later evaluation.
Of the twenty-three patients who had an over-all rating of fair or poor at the later evaluation, thirteen (57 per cent) had problems related to the patellofemoral joint, which had a deleterious effect on the rating. A significant relationship was found between deterioration of the over-all rating and deterioration of the result with regard to patellofemoral crepitus (p = 0.04). Two other patients had deterioration of the articular cartilage, noted both during the index operation and on follow-up arthroscopy; both had swelling with any activity, resulting in an over-all rating of fair. One patient had a grade-2 pivot shift and a six-millimeter increase in displacement on testing with the KT-1000 arthrometer, which resulted in an over-all rating of poor. The remaining seven patients continued to have moderate-to-severe symptoms and functional limitations with sports activities, which resulted in a fair or poor rating.
Significant relationships were found between deterioration of the over-all rating and deterioration of the pain score (p = 0.0001) and the jumping score (p = 0.0001). No significant relationship was found, with the numbers available, between deterioration of the over-all rating and deterioration of the score on testing with the KT-1000 arthrometer (p = 0.62).
We also calculated an over-all rate of failure, on the basis of the group of sixty-eight patients as well as the five patients (two who had received a bone-patellar ligament-bone allograft and three, a fascia lata allograft) in whom the graft had ruptured in the first postoperative year. The over-all rate of failure for these seventy-three patients was 10 per cent (seven patients) for the early period and 14 per cent (ten patients) for the later period.
Postoperative Complications
There was no clinical evidence of rejection of an allograft or transmission of disease at the time of the later evaluation.
Four patients who had a bone-patellar ligament-bone allograft and nine who had a fascia lata allograft had an early postoperative limitation of motion of the knee, which was treated with gentle manipulation with the patient under anesthesia. In two patients who had a bone-patellar ligament-bone allograft and in one who had a fascia lata allograft, an arthroscopic release of contracted scar tissue was performed. Therefore, following our previously described guidelines for complications related to motion32, the prevalence of reoperation for postoperative arthrofibrosis was 4 per cent (three of sixty-eight knees). At the later evaluation, thirty-nine patients who had a fascia lata allograft and twenty-three who had a bone-patellar ligament-bone allograft had a full range of motion of the knee. Six patients (9 per cent) lacked between 1 and 5 degrees of full extension. With the numbers available, the additional procedures performed with the index operation had no significant effect on complications related to motion (p = 0.41).
Thirteen patients who had a fascia lata allograft and sixteen who had a bone-patellar ligament-bone allograft had arthroscopy and removal of prominent, painful staples and cancellous-bone screws, at a mean of fifteen months (range, five to thirty-eight months) postoperatively. This complication has been discussed previously31. In late 1987, we began to use a larger, nine-millimeter-diameter interference cancellous-bone screw, which has allowed us to discontinue the use of staples.
Additional operative procedures were performed in seven patients who had a fascia lata allograft. Arthroscopy and partial meniscectomy was done for new meniscal symptoms in two patients, at thirty-one and sixty-one months postoperatively. Arthroscopy for subsequent patellofemoral symptoms was done in five patients, thirteen to eighty-three months postoperatively.
The current study provides data that, for the first time, to our knowledge, answer questions concerning the ability of allografts to provide function and stability of the knee at a mean of seven years postoperatively. When the results are assessed, two factors must be taken into account. First, our study included patients in whom an additional operative procedure (a repair of the medial collateral ligament, an iliotibial-band extra-articular procedure, or a meniscal repair) had been performed concomitantly with the index operation. If these patients had been excluded, the size of the series would have been reduced to such an extent that the study could not have been performed. Importantly, we found no significant effect of these additional procedures on anterior-posterior displacement, the over-all knee rating, or the prevalence of postoperative complications related to motion of the knee. Therefore, the inclusion of these patients did not bias our results. We caution clinicians, however, that it might be difficult to restore postoperative motion of the knee when these additional procedures have been performed, and we suggest that the therapist closely monitor these patients for early postoperative limitation of extension or flexion. According to guidelines that were described previously, intensive treatment should be instituted early in the postoperative period to avoid a permanent limitation of motion or patella infera32,33.
Second, the operative procedure and techniques for sterilization of allografts have evolved through a natural learning process since the conclusion of this study (in mid-1987). Since 1988, we have performed all reconstructions of the anterior cruciate ligament with the assistance of an arthroscope and have used allografts that have been sterilized with gamma irradiation. Because of the rapid proliferation of information with regard to reconstruction of the anterior cruciate ligament, it is not feasible to perform a long-term study on operative techniques that are considered to be state of the art five to nine years after the study's inception. This is, of course, a problem with many operative procedures and not just those related to the anterior cruciate ligament. However, there are important findings in the present study that are of clinical relevance to the procedures performed currently.
One of the most noteworthy findings of our study was the apparent ability of the allograft to provide stability for at least five to nine years postoperatively. Only three patients (4 per cent) had an increase in displacement between the early and the later evaluation that resulted in failure of the graft. Five additional patients had a slight increase in displacement but still had a partially functioning graft according to the KT-1000 arthrometric and pivot-shift-test criteria. Both our previous study4 and the current one therefore showed that a small number (4 to 5 per cent) of allografts may stretch slightly over time. However, this stretching amounted to only two to three millimeters in most of the knees and did not inevitably lead to a failure or a poor result.
We previously reported that the incorporation of the iliotibial-band extra-articular procedure significantly reduced anterior-posterior displacements after reconstruction of the anterior cruciate ligament with an allograft (p < 0.01)27. That study included patients who had a chronic rupture, and we hypothesized that the extra-articular procedure could provide support to the healing allograft by restoring the secondary restraints provided by the lateral iliotibial band. In the current study, however, the extra-articular procedure did not yield any appreciable benefit with regard to reducing anterior-posterior displacements. We have therefore concluded that this additional procedure does not provide any additional benefit for a patient who has an acute rupture of the anterior cruciate ligament, and we no longer perform it.
The over-all rate of failure of the graft in the current study (10 per cent [seven of seventy-three patients] at the early evaluation and 14 per cent [ten of seventy-three patients] at the later evaluation) was somewhat higher than the 4 per cent rate that we reported earlier for patients who had an acute rupture of the anterior cruciate ligament and were managed with an allograft31. In that report, bone-patellar ligament-bone allografts were associated with significantly lower values for anterior-posterior displacement than were fascia lata allografts (p < 0.05). The force level of eighty-nine newtons used in that study, however, was lower than the 134 newtons used in the present investigation. In the current study, at the early follow-up evaluation, at which the eighty-nine-newton force was used, there was a statistical trend (p = 0.09) between the two groups with regard to displacement values, which supported our earlier findings. This trend was not found, however, at the later follow-up examination. We believe that the higher level of force used at the later evaluation in the current investigation accounted for the difference in the results for displacement between the previous and the current study. We could not find any other reports that documented rates of failure of allografts used for acute ruptures of the anterior cruciate ligament with which to compare our results. Rates of failure that have been reported after reconstruction of acute ruptures with autogenous patellar-ligament grafts have not been established at high levels of force (134 newtons) or at the time of long-term follow-up. Because of these discrepancies, we urge caution in interpreting the reported rates (0 to 2 per cent8,13,39).
An increase in the size of the femoral and tibial bone tunnels has been reported previously after reconstruction of the anterior cruciate ligament with an allograft. Fahey and Indelicato found a mean increase of 1.2 millimeters in the diameter of the tibial tunnel one year after reconstruction with a fresh-frozen patellar-ligament allograft14. Forty (46 per cent) of eighty-seven knees had an enlargement of the tibial tunnel of more than 2.8 millimeters, which was greater than one standard deviation of increase compared with the size recorded during the reconstruction. Linn et al. reported that fourteen (54 per cent) of twenty-six knees that had received a fresh-frozen Achilles-tendon allograft had an enlargement of the tibial tunnel of four millimeters or more compared with the original size of the tunnel at the time of the operation, and that ten of twelve knees had a similar increase in the size of the femoral tunnel, at a mean of two and one-half years postoperatively23. No association was found, in either of those investigations, between an enlargement of the tunnel postoperatively and failure of the allograft. In the current study, only one patient (1 per cent) had a noteworthy enlargement of the tibial tunnel (to twenty-two millimeters, compared with eleven millimeters at the reconstruction). Two other patients (4 per cent) had a slight (three to four-millimeter) increase, but neither increase was associated with an increase in anterior-posterior displacement postoperatively. Therefore, although we found a lower prevalence of enlargement of the tibial and femoral tunnels than has been reported by others14,23, our data support the earlier conclusions that the increase does not appear to be deleterious to healing of the allograft or to the eventual clinical result.
Five factors—pain, KT-1000 arthrometric data, jumping, patellofemoral crepitus, and the over-all rating—were chosen for the analysis of deterioration on the basis of the results of a stepwise linear-regression model performed on a subset of forty-seven patients from our previous study31. In that model, the combination of the first four factors resulted in the ability to predict the over-all rating for 98 per cent of the population26. In the current study, a significant relationship was found between deterioration of the individual factors of pain, jumping, and patellofemoral crepitus and deterioration in the over-all rating. Our rating system has been found to be one of the strictest reported to date2,5; the findings of the current study continue to support these conclusions.
One unresolved problem after reconstruction of the anterior cruciate ligament is the increase in patellofemoral crepitus, indicative of early patellofemoral osteoarthrosis. Roberts et al. reported an increase in patellofemoral crepitus from none to mild in eighteen (58 per cent) of thirty-one knees and from none to marked in four (13 per cent), at a mean of two years after arthroscopic reconstruction with a bone-patellar ligament-bone allograft for a chronic rupture of the anterior cruciate ligament38. They found no association between crepitus and symptoms related to the patellofemoral joint. Engebretsen et al. reported that twenty-seven (54 per cent) of fifty knees had patellofemoral crepitus one year after reconstruction with a patellar-ligament autogenous graft, but they did not state how many of the knees had had crepitus preoperatively or if the crepitus was associated with the patellofemoral symptoms13. In our previous report on reconstruction of the anterior cruciate ligament for an acute rupture, ten (21 per cent) of the forty-seven patients who had had no or only mild patellofemoral crepitus preoperatively had moderate crepitus at a mean of forty months postoperatively31. In the current study, seventeen (25 per cent) of sixty-seven patients who had had no or mild crepitus preoperatively had moderate or severe crepitus at the time of the later follow-up evaluation. Only two of these patients had had major damage to the articular cartilage at the time of the index operation. In nine of these patients, the patellofemoral crepitus was accompanied by symptoms, which resulted in an over-all rating of fair or poor. Several possible causes may be proposed for patellofemoral crepitus that increases with time. The effects of the initial injury, the magnitude of the operative procedure, the effects of disuse6, or the fact that the patient has remained athletically active for many years after the operation may alter the normal homeostasis of the articular cartilage and contribute to the gradual deterioration by as yet unexplained mechanisms. Although moderate crepitus is not necessarily associated with symptoms, we believe that it may eventually lead to pain and swelling of the joint and a reduction in the level of activity. We advise therapists to monitor the patellofemoral joint carefully during the rehabilitation program and to modify the program for any patient who has demonstrable crepitus and cracking. The use of closed-kinetic-chain exercises24 and the avoidance of both stationary cycling and progressive resistance exercises with heavy weights are recommended to decrease symptoms of the patellofemoral joint.
The rehabilitation program that we use after reconstruction with an allograft incorporates specific delays in achieving full weight-bearing without crutches and in the implementation of various exercises. These delays were instituted to account for the delay in maturation that we believe the allograft undergoes after implantation. Our rehabilitation program after reconstruction of the anterior cruciate ligament with an autogenous graft provides a more progressive approach3,10,11,25. Full weight-bearing without crutches is allowed by the fifth or sixth postoperative week, provided that arthrometric testing does not show an increase in anterior-posterior displacement of more than two millimeters as compared with that of the contralateral limb. Closed-kinetic-chain resistance exercises are initiated as early as the second or third postoperative week. A running program is begun by the twentieth postoperative week for patients who do not have pain, swelling, giving-way, or an increase in patellofemoral crepitus. Importantly, the program is continually monitored every two to four weeks with an arthrometric examination; modifications are implemented for any patient who has an increase in anterior-posterior displacement of more than two millimeters compared with that of the contralateral limb. These modifications, and the program in general, have been discussed in detail elsewhere3,10,11,25.
The immediate program for motion of the knee was effective, as only three (4 per cent) of the sixty-eight patients needed a reoperation for early postoperative arthrofibrosis. Shelbourne et al. reported that twenty-one (14 per cent) of 155 knees needed an arthroscopic lysis of adhesions39, and Engebretsen et al. reported that four (8 per cent) of fifty knees had a reoperation for the same condition13, after reconstruction of the anterior cruciate ligament for an acute rupture. On the basis of these findings, both sets of authors modified their rehabilitation program to encourage earlier full extension of the knee, and they reported a subsequent reduction in the rate of reoperation for this complication. Since 1982, we have advocated immediate motion of the knee after reconstruction of the anterior cruciate ligament with either an autogenous graft or an allograft. This program, which includes immediate full extension of the knee31,34, proved effective in an earlier study, in which only three (4 per cent) of sixty-nine knees needed an arthroscopic release of contracted tissues after reconstruction with an allograft for an acute rupture of the anterior cruciate ligament32.
We remain concerned about the possibility of transmission of disease and sterilization of the graft. Allografts should be obtained only from tissue banks certified by the American Association of Tissue Banks, which has strict standards for procurement of tissue, selection of donors, and serological testing1. In addition to adhering to these standards, we obtain other tests to detect the human immunodeficiency virus, including the P24 core protein antigen test, which is now used at many tissue banks although it has not yet been advocated by the American Association of Tissue Banks, and the polymerase chain reaction test12. These tests reduce the time-period of seronegativity to an estimated one to two weeks or less and, used in combination with gamma irradiation, they should reduce what is already a very small risk of transmission of disease to almost no risk.
We ceased using ethylene oxide for secondary sterilization of allografts in 1986 because of early reports of deleterious effects on allogeneic tissue22,36,37. In mid-1987, gamma irradiation became our standard method of sterilization. The American Association of Tissue Banks currently recommends a dose of 2.5 megarads (25,000 gray) for sterilization of soft-tissue allografts1. However, recent reports have shown that this level does not always completely inactivate the human immunodeficiency virus in cell cultures9 or in bone-patellar ligament-bone grafts16. Fideler et al. concluded that a minimum of 3.0 megarads (30,000 gray) may be required to inactivate the human immunodeficiency virus totally in bone-patellar ligament-bone allografts16. Their work supported earlier findings by Conway et al., who reported that as much as 3.6 megarads (36,000 gray) was required to inactivate the virus in a cultured osteocyte model9. Since 1990, we have used 4.0 megarads (40,000 gray) for irradiation of bone-patellar ligament-bone allografts. Studies were conducted in our laboratory to assess radiation-dose-dependent alterations in mechanical and material properties in human and goat bone-patellar ligament-bone units7,15,17,19,37. Gibbons et al. found no significant reduction in the initial material properties of goat patellar ligament-bone units after irradiation with 2.0 megarads (20,000 gray) compared with those found in fresh-frozen tissue17. Butler et al. found no significant decrease (p > 0.05) in the mechanical properties of bone-patellar ligament-bone units six months after reconstruction of the anterior cruciate ligament in goats7. Rasmussen et al. found no significant differences in static or cyclic creep between controls and bone-patellar ligament-bone units from human cadavera that had been treated with 4.0 megarads (40,000 gray); however, significant reductions in stiffness (26 per cent; p < 0.03) and maximum force (40 per cent; p < 0.001) were reported37. The long-term biological remodeling and mechanical properties of bone-patellar ligament-bone allografts, and whether deleterious alterations are incurred with higher levels of gamma irradiation, remain to be investigated.
The results of the current study showed that eight (12 per cent) and fourteen (21 per cent) of the patients had a decrease in the subjective scores for pain and jumping (or landing on the involved limb), respectively. For many patients, the decreases represented a decline in participation in sports involving jumping, pivoting, and cutting as well as a loss of the patients' confidence in their ability to perform these activities. As patients graduated from school or grew older, many decreased either their level of sports activity or their frequency of participation. Either of these factors can result in lower confidence and, consequently, in a lower score on the subjective evaluation. Conversely, eight (12 per cent) and fifteen (22 per cent) of the patients had an improvement in the scores for pain and for jumping and landing, respectively. For these patients, the residual limitations that had been noted at the time of the two to four-year follow-up evaluation had resolved or decreased over time. These data offer hope, to patients who still have symptoms and limitations at the time of the early evaluation, that improvement may occur even many years after the reconstruction. In our interviews with the patients, many of them mentioned a gradual loss of fear of reinjury or an increased confidence in the ability to perform more difficult activities involving the knee that had continued to improve with time.
We currently recommend reconstruction with an autogenous bone-patellar ligament-bone graft for acute ruptures of the anterior cruciate ligament because of concerns about the cost of allografts and the transmission of disease. Because of these concerns, we do not recommend the widespread use of allografts. However, the current study shows that a favorable result may be achieved with an allograft, if the surgeon and the patient choose this approach.