Subjects
One hundred and fifty consecutive patients between sixteen and fifty years
of age (mean, twenty-nine years) who had an acute rupture of the anterior
cruciate ligament were treated in three university hospitals during the period
from May 1986 to April 1988. Three patients were lost to follow-up during the
first postoperative year and were excluded from the study. Patients who had
concomitant meniscal and/or medial collateral ligament injuries were included,
whereas those who had posterior cruciate or lateral collateral ligament
ruptures were excluded. After the anterior cruciate ligament injury was
confirmed by Lachman and pivot shift testing with the patient under general
anesthesia and by arthroscopy, treatment was randomized into one of three
surgical procedures: acute repair only (Group 1), acute repair with a ligament
augmentation device (Group 2), and acute repair with a bone-patellar
tendon-bone graft (Group 3). There were fifty patients in each group at the
end of the inclusion period.
The clinical results after sixteen years are described in the present
study. At the time of the sixteen-year follow-up evaluation, four patients had
died, eight patients had emigrated, and six patients had been lost to
follow-up for other reasons, leaving 129 patients (88%) who were available for
follow-up; fourteen of these patients were interviewed by telephone only.
Therefore, 115 patients (78%) had complete follow-up data including an
examination at sixteen years. Only the 103 patients who had not had anterior
cruciate ligament revision in the involved knee or anterior cruciate ligament
reconstruction in the contralateral knee were included in the comparisons of
the groups. This included thirty-three patients who had primary repair,
thirty-four who had repair with a ligament augmentation device, and thirty-six
who had repair with a bone-patellar tendon-bone graft.
Operative Procedures
Group 1: Primary Anterior Cruciate Ligament Repair
Primary repair is performed through a medial arthrotomy. Multiple loop
sutures are placed at different levels in the remnants of the anterior
cruciate ligament in an anteromedial to posterolateral direction. Two holes
are drilled at the attachment sites of the anterior cruciate ligament to the
femur and tibia with a 2-mm-diameter Steinmann pin. The sutures are tied over
the bone-bridge while tension is manually applied with the knee in 30° of
flexion.
Group 2: Anterior Cruciate Ligament Repair with a Ligament
Augmentation Device
After primary repair of the anterior cruciate ligament is completed, a
ligament augmentation device is inserted. A 2-mm-diameter Steinmann pin is
placed just anterior and medial to the attachment site of the anterior
cruciate ligament to the tibia. The hole is overdrilled with a 4-mm-diameter
cannulated reamer. The over-the-top position is identified and chamfered with
a curved rasp to produce a small groove. The ligament augmentation device is
passed through the tibial tunnel and is sutured to the periosteum of the
anteromedial aspect of the tibia with six, seven, or eight interrupted
nonabsorbable sutures. The anterior cruciate ligament remnants are sutured to
the ligament augmentation device with 2-0 Vicryl (polyglactin; Ethicon,
Somerville, New Jersey). The ligament augmentation device with the attached
anterior cruciate ligament is then routed over the top. With the knee in
30° of flexion and with manual tension applied, the construct is fixed to
the lateral femoral condyle with a belt-buckle staple technique. The ligament
augmentation device is fixed to the bone at only one end and is not used as a
prosthesis.
Group 3: Anterior Cruciate Ligament Repair with Bone-Patellar
Tendon-Bone Graft Augmentation
A bone-patellar tendon-bone graft is harvested from the central third of
the ligament. A 2-mm-diameter Steinmann pin is drilled from the medial aspect
of the proximal and anterior part of the tibia, emerging 5 mm anterior and
medial to the center of the anterior cruciate ligament tibial attachment. This
hole is overdrilled with a 10-mm-diameter cannulated reamer. With use of an
inside-out technique at maximum knee flexion, another Steinmann pin is placed
posterior and superior to the center of the femoral attachment of the anterior
cruciate ligament. This femoral hole is then overdrilled to 10 mm. A
notchplasty is performed. Tibial fixation is achieved with an interference
screw, and two sets of nonabsorbable sutures are passed through holes in the
bone block and tied around a cortical-bone screw and washer outside the
tunnel. The anterior cruciate ligament remnants are sutured to the graft. With
the knee in 30° of flexion and manual tension applied, the construct is
fixed in the femoral tunnel with an interference screw and reinforcing
sutures.
Rehabilitation
After the operation, all 150 patients were managed with a long leg cast
with the knee in 30° of flexion. The cast was worn for two weeks, after
which braces were worn for six weeks. No weight-bearing was permitted during
the first eight weeks after surgery.
Follow-up
The patients were evaluated at sixteen years by an independent observer
(J.O.D.). The results were compared with the five-year follow-up data. The
level of activity was evaluated with use of the Tegner activity
score16. The
functional state was graded according to the Lysholm
score17, which
emphasizes pain, instability, and locking. Physical examination included an
assessment of the range of motion and the evaluation of knee stability by
manual testing and with use of an arthrometer. The pivot shift was graded as
negative, trace positive (1+), moderate shift (2+), or subluxation (3+).
Anterior laxity with the knee in 20° of flexion was evaluated by the
Lachman test18 and
was graded as negative (0), slight (1+) when there was <5 mm of
displacement, moderate (2+) when there was 5 to 10 mm, or severe (3+) when
there was >10 mm compared with that of the normal knee. Instrumented
testing of anterior laxity was performed with a KT-1000 arthrometer
(MEDmetric, San Diego, California) at 20° of flexion and a load of 89 N
(20 lb) and was measured in millimeters. At sixteen years, the evaluation
tools were the same as those used at the initial and five-year evaluations so
that comparisons with the former results could be made.
Standing radiographs of both knees at 45° of flexion and with a 30°
angulation of the beam were obtained at the sixteen-year follow-up
examination, and the degree of osteoarthritis was graded with use of the
Ahlback
classification19 by
an independent experienced radiologist (A.T.V.).
Statistical Methods
The Statistical Package for the Social Sciences (version 11.0; SPSS,
Chicago, Illinois) was used for the statistical analyses. A nonparametric
analysis of variance (Kruskal-Wallis test) was used to determine the
significance of the overall difference in the outcome measures among the three
repair groups at a particular follow-up time-point. Paired comparisons among
the three groups at each time-point were made with use of the Mann-Whitney
test. The Wilcoxon test for paired data was used to test for the significance
of the changes in the outcome measures for a particular repair group between
any two time-points. Comparisons were considered significant when the p value
was <0.05.
Revisions
Eleven (24%) of the forty-five patients who had a primary repair (Group 1),
four (10%) of the forty-two patients who had repair with the ligament
augmentation device (Group 2), and one (2%) of the forty-two patients who had
repair with the bone-patellar tendon-bone graft (Group 3) underwent a revision
of the anterior cruciate ligament repair between the primary operation and the
sixteen-year follow-up evaluation. The rate of revision was ten times higher
in Group 1 compared with Group 3 (p = 0.003). With the numbers available, the
rate of revision was not significantly different between Group 1 and Group 2
(p = 0.066) or between Group 3 and Group 2 (p = 0.167). The survival of the
primary reconstructions over time in each group is shown in
Figure 1. Seven male and nine
female patients underwent revision surgery (p = 0.332). One patient in Group
1, five in Group 2, and five in Group 3 had an anterior cruciate ligament
reconstruction in the contralateral knee by sixteen years (p = 0.210). One
patient had both an anterior cruciate ligament revision in one knee and a
reconstruction in the other knee; thus, twenty-six patients were excluded from
the comparisons of the groups.
The patients who did not undergo a revision or a reconstruction on the
contralateral side underwent evaluation of activity level, functional level,
range of motion, stability, subjective knee function, and osteoarthritis.
Activity Level
The mean Tegner activity score for Group 1 was 6.7 points before the
injury, 5.0 points at five years, and 5.1 points at sixteen years (see
Appendix). The mean activity score for Group 2 was 6.2 points before the
injury, 5.0 points at five years, and 5.2 points at sixteen years. The mean
activity score for Group 3 was 6.5 points before the injury and 5.3 points at
five years, which was a significant difference (p = 0.008), and 5.6 points at
sixteen years. At sixteen years, no significant difference was detected among
the groups.
The percentages of patients in the three repair groups with an activity
score between 7 and 10 (representative of a high-performance athlete) ranged
from 49% to 62% before the injury, from 21% to 27% at the five-year
evaluation, and from 16% to 34% at sixteen years (see Appendix).
Functional Level
The mean Lysholm score for Group 1 (primary repair) was 88.3 points at five
years and 88.0 points at sixteen years. Ten (30%) of the thirty-three patients
in Group 1 had functional scores within the poor and fair categories (0 to 83
points) at sixteen years (see Appendix). The mean functional score for Group 2
(repair with a ligament augmentation device) was 91.4 points at five years and
85.0 points at sixteen years. Twelve (35%) of the thirty-four patients in
Group 2 had functional scores within the poor and fair categories at the
sixteen-year evaluation. The mean functional score for Group 3 (repair with
bone-patellar tendon-bone graft) was 93.3 points at five years and 90.0 points
at sixteen years. Five (14%) of the thirty-six patients in Group 3 had a
rating of poor or fair at sixteen years.
Significant differences in the mean functional scores among the three
repair groups were detected at five years but were not detected at sixteen
years. At the five-year follow-up evaluation, the patients in Group 3 had a
significantly higher mean functional score than the patients in Group 1 (p =
0.004). The main factors responsible for the relative reduction in the
functional scores for the patients in Group 1 were pain and instability. The
patients in Groups 2 and 3 did not experience that same level of instability
and pain.
Subjective Knee Function
The subjective knee function as graded by the patients was fair or poor for
five patients in Group 1, six patients in Group 2, and one patient in Group 3
at sixteen years (Fig. 2). The
difference between Groups 2 and 3 with respect to subjective knee function was
significant (p = 0.038), but no significant difference was detected between
Groups 1 and 2 (p = 0.783) or between Groups 1 and 3 (p = 0.068).
Range of Motion
In Group 1, two (5%) of forty-one patients had an extension deficit of
>5° at the five-year follow-up evaluation and the rate increased to
eight (30%) of twenty-seven patients at sixteen years (see Appendix). In Group
2, an extension deficit of >5° was found in six (14%) of forty-two
patients at five years and in eight (27%) of thirty patients at sixteen years.
In Group 3, eight (17%) of forty-eight patients had an extension deficit of
>5° at the five-year evaluation, and the rate increased to eight (23%)
of thirty-five patients at sixteen years. There was no significant difference
among the three groups at five years. Only one patient in Group 1 and one in
Group 3 had an extension deficit of >10° at five years. At sixteen
years, one patient in Group 1 and one in Group 2 had an extension deficit of
>10°.
In general, the patients in the three repair groups showed an improvement
in the ability to attain full flexion over the sixteen-year follow-up period.
However, one patient who had a primary repair and one managed with a
bone-patellar tendon-bone graft continued to demonstrate a flexion deficit
between 21° and 30° at sixteen years.
Stability
In Group 1, twenty-one (51%) of forty-one patients had a 2+ or 3+ result
(moderate to severe) on the pivot shift test at the five-year follow-up and
the rate decreased to five (18%) of twenty-eight patients at sixteen years
(see Appendix). By that time, eleven patients (27%) in the group had undergone
revisions and were excluded from further comparison. Fifteen (36%) of
forty-two patients in Group 2 had a 2+ or 3+ result on the pivot shift test at
five years and five (17%) of twenty-nine patients demonstrated the same degree
of instability at sixteen years. By that time, four patients in Group 2 had
undergone revisions and were excluded from further comparison. In Group 3,
four (8%) of forty-eight patients had a 2+ pivot shift at the five-year
follow-up examination, and two (6%) of thirty-five patients demonstrated the
same degree of instability at sixteen years. The knees of the patients in
Group 3 were significantly more stable than those in Groups 1 and 2 at five
years (p = 0.026); however, with the numbers studied, no difference was
detected at sixteen years (p = 0.088).
The Lachman test showed that eighteen (44%) of forty-one patients in Group
1 (primary repair) had 2+ or 3+ (moderate-to-severe) anterior instability at
the five-year follow-up examination and five (18%) of twenty-eight patients
demonstrated similar findings at sixteen years (see Appendix). Twelve (29%) of
forty-two patients in Group 2 (repair with a ligament augmentation device)
demonstrated moderate-to-severe anterior instability on the Lachman test at
five years, and ten (34%) of twenty-nine patients had similar findings at
sixteen years. In Group 3 (repair with a bone-patellar tendon-bone graft),
five (10%) of forty-eight patients had a 2+ or 3+ anterior instability on the
Lachman test at five years, and four (11%) of thirty-five patients had similar
findings at sixteen years. The knees in Group 3 were significantly more stable
than those in Group 1 (p = 0.03); however, there was no difference between the
knees in Groups 2 and 3 (p = 0.26) at the five-year follow-up evaluation. The
knees in Group 3 were more stable than those in Group 2 at the sixteen-year
follow-up examination (p = 0.026). The knees in Group 2 were significantly
more stable than those in Group 1 at five years (p = 0.05).
Anterior laxity as measured with the KT-1000 arthrometer showed that the
mean anterior laxity in Group 1 was 2.7 mm at five years and 1.9 mm at sixteen
years. The mean amount of anterior laxity for Group 2 did not change over the
follow-up period (2.3 mm at both five and sixteen years). The mean amount of
anterior laxity for Group 3 also did not change over the follow-up period (1.3
mm at five years and 1.4 mm at sixteen years). The knees in Group 3 were more
stable than those in Group 2 (p = 0.01) and Group 1 (p = 0.00) at five years;
however, there was no difference between the knees in Group 2 and those in
Group 1 (p = 0.27) (see Appendix). No difference was found in anterior laxity
between the knees in Group 1 and those in Group 2 (p = 0.55) or those in Group
3 (p = 0.50) at sixteen years. There was also no difference between the knees
in Groups 2 and 3 (p = 0.13) at sixteen years.
A difference in anterior laxity compared with the contralateral side (a
side-to-side difference) of =3 mm was seen at five years in twenty-seven
(66%) of forty-one patients in Group 1, twenty (48%) of forty-two patients in
Group 2, and ten (21%) of forty-eight patients in Group 3. At sixteen years,
thirteen (48%) of twenty-seven patients in Group 1, eleven (38%) of
twenty-nine patients in Group 2, and ten (30%) of thirty-three patients in
Group 3 had a side-to-side difference of =3 mm.
Osteoarthritis
Four patients in Group 1, one patient in Group 2, and four patients in
Group 3 had osteoarthritis of the knee sixteen years after surgery.
Figure 3 shows the distribution
of the osteoarthritis according to the Ahlback classification. Grade-3
osteoarthritis was seen in one patient in Group 1 but in no patient in Groups
2 or 3. Osteoarthritis developed in the involved knee in nine (11%) of the
eighty-five patients for whom data were available and in the contralateral
knee in three patients (4%) at sixteen years; the difference was significant
(p = 0.001).
At sixteen years, one (4%) of twenty-eight patients in Group 1 and six
(21%) of twenty-nine patients in Group 2 had patellofemoral crepitation during
concentric quadriceps work (p = 0.19). Seven (20%) of thirty-five patients in
Group 3 had patellofemoral crepitation, which was significantly more than that
in Group 1 (p = 0.038).
In the present study, we found that the patients who had a primary repair
of a ruptured anterior cruciate ligament had an unacceptably high revision
rate, and we believe that this procedure should not be performed. It is
difficult to interpret the outcome of repair procedures for the treatment of
anterior cruciate ligament ruptures because most published studies are
retrospective and nonrandomized and they compare different surgical
techniques, rehabilitation, and follow-up protocols. Sandberg and
Balkfors20
performed a prospective study comparing nonoperative cast treatment with
direct repair of the anterior cruciate ligament. They found that, although
there was no significant difference between the functional outcomes of the two
treatments, a higher percentage of patients who had nonoperative treatment had
a positive pivot shift. In long-term retrospective follow-up studies,
Fruensgaard et al.4
and Sommerlath et
al.3 reported that
conservative treatment and simple suture of acutely torn anterior cruciate
ligaments produced similarly poor outcomes.
The ligament augmentation device was developed by Kennedy et
al.21,22
to augment different types of anterior cruciate ligament reconstructions until
the biologic graft tissue remodeled to a sufficient strength. The ligament
augmentation device was first used by Schabus in the repair of acute anterior
cruciate ligament
ruptures6. The
patients who had repair with the ligament augmentation device in our study had
little change in activity and functional levels over the sixteen-year
follow-up period. At the five-year follow-up evaluation, the rate of
instability failures in this group had increased substantially, probably
because of ruptures of the ligament augmentation device itself. However, the
stability outcome for the group that had repair with the ligament augmentation
device was still significantly better than that for the group that had primary
repair. Four patients who had repair with the ligament augmentation device had
revision surgery because of gross instability at sixteen years. Five patients
underwent anterior cruciate ligament reconstruction in the contralateral knee.
These patients were excluded from the comparisons between the groups. Among
the patients who were included, those who had repair with a ligament
augmentation device had significantly less stable knees at the sixteen-year
follow-up evaluation than those who had repair with a bone-patellar
tendon-bone graft. In another
study8, no positive
long-term effects supporting the use of a ligament augmentation device in
anterior cruciate ligament reconstruction were found.
Augmentation with autogenous tissue has been performed with use of the
hamstring
tendons23,24,
the iliotibial band prepatellar
retinaculum25, and
the patellar
tendon26. Our
choice of bone-patellar tendon-bone graft for augmentation was based on animal
studies by Cabaud et
al.27 and clinical
studies by Clancy et
al.5. In other
studies28,29,
the results after arthroscopic reconstruction of the anterior cruciate
ligament with use of a bone-patellar tendon-bone graft appeared to be
satisfactory both at short-term and medium-term follow-up evaluations. In the
present study, the activity and functional levels at sixteen years were higher
for the patients who had repair with bone-patellar tendon-bone graft than for
those in the other two repair groups. Four patients (8%) who had repair with a
bone-patellar tendon-bone graft had a positive pivot shift sign at five years,
and two (6%) had a positive finding at sixteen years. A side-to-side
difference of =3 mm, as measured by the KT-1000 arthrometer, was found in
ten patients (21%) with a bone-patellar tendon-bone graft at the five-year
evaluation and in ten such patients (30%) at the sixteen-year evaluation.
These findings show that the bone-patellar tendon-bone graft did not stretch
during the sixteen years of this study. Only one patient in the bone-patellar
tendon-bone graft group underwent revision because of instability during the
sixteen years of follow-up. The Tegner activity score was higher in the group
that had repair with a bone-patellar tendon-bone graft at all time-intervals.
Five patients in this group underwent anterior cruciate ligament
reconstruction in the contralateral knee compared with only one patient in the
primary repair group. This indicates that the patients who had stable knees
probably performed more risky activities and thereby more often injured the
contralateral knee and/or stretched out the reconstructed anterior cruciate
ligament.
In the primary repair group, a positive pivot shift sign was demonstrated
in 51% of the patients at five years and in 18% at sixteen years. The reason
for the lower percentage at sixteen years is the high number of revisions done
in the group within the first five years. These patients were excluded from
further comparisons at sixteen years.
At the sixteen-year follow-up, there was no significant difference in range
of motion among the three groups. Arthrofibrosis following patellar tendon
repairs, when anterior cruciate ligament surgery is done acutely, has been
reported
previously30-32.
The prevalence of arthrofibrosis has been reported to be lower when surgery is
delayed33,
arthroscopic techniques are used, and accelerated rehabilitation is applied.
The early, limited range-of-motion problems in the group that had repair with
bone-patellar tendon-bone graft in the present study were probably related to
the acute, open surgical procedure and the conservative rehabilitation
protocol that was used.
Mok and Dowd reviewed the cases of thirty patients an average of 7.5 years
(range, five to eleven years) after an anterior cruciate ligament
reconstruction with a bone-patellar tendon-bone
autograft34. They
found that twenty-seven patients had maintained the good-to-excellent results,
according to the Lysholm scoring system, that they had achieved at the
two-year follow-up. In our study, we did not find any difference among the
groups with respect to subjective knee function at sixteen years. However, the
knees with the most laxity and functional problems in each group had been
revised and were excluded from further comparison.
In the present study, we found that the prevalence of osteoarthritis in the
injured knee was 11% after sixteen years. This percentage is surprisingly low,
especially when the long postoperative immobilization and slow rehabilitation
are taken into account. In a previous study by two of
us8, approximately
half of the 100 patients who had an anterior cruciate ligament reconstruction
were found to have osteoarthritis at eight years as verified by radiographic
examination with use of the Ahlback classification. The only differences
between the two studies were that the previous study of 100 patients involved
an aggressive rehabilitation protocol and arthroscopic surgery and the
patients were allowed an early return to sports. In both studies, the
radiographs were analyzed by independent experienced radiologists. Myklebust
et al.35, in a
study of handball players with a reconstructed anterior cruciate ligament,
found that 42% of the patients had gonarthrosis develop after eight years. The
high percentage of osteoarthritis in their study may be explained by a high
level of sports activity. In a similar
study36, with more
aggressive rehabilitation than that in our study, fifteen (24%) of sixty-two
patients had degenerative changes on radiographs after five years. In a
randomized study24,
radiographic assessment at five years revealed early osteoarthritic changes in
4% of the patients managed with hamstring tendon reconstruction and in 18% of
the patients who had patellar tendon reconstruction. Roos et
al.37 found that
patients who had anterior cruciate ligament injuries (isolated or combined
with an injury of the meniscus or collateral ligaments) showed the first
radiographic signs (joint space narrowing) of osteoarthritis at an average age
of approximately forty years, whereas patients who had an isolated meniscal
injury had the same stage of disease at an average age of approximately fifty
years. Both study groups displayed the first radiographic signs of
osteoarthritis at an average of approximately ten years after the injury and
showed increasingly serious arthroscopic and radiographic signs of joint
damage. We found a surprisingly low prevalence of osteoarthritis in our
patients after sixteen years, but the patients in our study had a lower
activity level compared with the handball players in the study by Myklebust et
al.
We obtained the best results when the anterior cruciate ligament repair was
augmented with a bone-patellar tendon-bone graft. In all outcome measures
other than for early, limited range of motion, the bone-patellar tendon-bone
graft augmentation technique was superior to primary repair and to repair with
the ligament augmentation device. On the basis of these data, we conclude that
open primary repair has an unacceptably high rate of revision and should not
be performed. It can be expected that approximately 10% of patients undergoing
anterior cruciate ligament reconstruction acutely will have osteoarthritis
develop in the reconstructed knee. We no longer perform any of these surgical
techniques as open procedures.
Tables showing the activity level, functional level, range of motion, and
stability for all knees that were evaluated are available with the electronic
versions of this article, on our web site at
(go to
the article citation and click on "Supplementary Material") and on
our quarterly CD-ROM (call our subscription department, at 781-449-9780, to
order the CD-ROM). ?