Loss of knee extension has been reported by many authors to be a common
complication following anterior cruciate ligament
reconstruction1-4.
Many causes of this loss of extension have been identified, including
nonanatomic placement of the
graft5, impingement
of anterior-intercondylar notch scar tissue ("cyclops
lesion")6,
arthrofibrosis5,
immobilization in
flexion7, acute
anterior cruciate ligament
reconstruction4,
excessive graft
tension8, and
inadequate
rehabilitation9,10.
Clinically, we have seen patients with loss of extension following an anterior
cruciate ligament reconstruction that had been done with appropriate tunnel
placement within the femoral and tibial footprints of the native anterior
cruciate ligament. We suspected that these grafts may have been fixed under
high tension with the knee in flexion. Some of these patients have benefited
from notchplasty, but we have found that others have required graft excision
to obtain full extension. Even a very small loss of extension can be
functionally important during athletic activities as well as during activities
of daily living. Sachs et
al.11 reported that
a loss of =5° of extension causes an abnormal gait that can lead to
patellofemoral pain and quadriceps weakness.
Some surgeons recommend tensioning the anterior cruciate ligament graft in
some degree of knee flexion, with 20° to 30° being the most common
recommended
position12,13.
The optimum amount of force applied to the graft prior to fixation is a matter
of considerable debate, with most authors recommending between 20 and 90 N of
initial graft
tension12-19.
Experimental data suggest that fiber recruitment and tension in the anterior
cruciate ligament increase as the knee moves from flexion to
extension8,20,21.
If graft fixation is performed in 20° to 30° of knee flexion and if
the fixation is not rigid, the graft may migrate proximally in the tibial bone
tunnel or distally in the femoral tunnel, and full extension may still result.
If the anterior cruciate ligament graft is rigidly secured under high tension
in 30° of knee flexion, overconstraint may occur, with loss of knee
extension leading to a permanent flexion contracture. The unidirectional
collagen fibers of a reconstructed anterior cruciate ligament usually do not
resemble the multi-directional collagen fibers of the native anterior cruciate
ligament22,23.
If the unidirectional fibers of a reconstructed ligament are pulled taut and
the graft is rigidly fixed with the knee in flexion, loss of knee extension
may result
postoperatively3,8,20.
These concerns have led surgeons to recommend that the anterior cruciate
ligament graft be tensioned with the knee in full
extension3. In
theory, graft tensioning in full extension should not result in loss of knee
extension regardless of the amount of tension placed on the graft during
fixation.
Our hypothesis was that securing a bone-patellar tendon-bone graft in
30° of knee flexion during a reconstruction of the anterior cruciate
ligament will result in loss of knee extension when the graft is tensioned at
either 44 or 89 N. The aim of this in vitro study was to determine the effect
of the knee flexion angle and graft tension during fixation of a bone-patellar
tendon-bone graft on loss of knee extension following anterior cruciate
ligament reconstruction.
Pilot Study Analyzing Measurement of Knee Flexion Angle
Apilot study was conducted to compare three different methods of
measurement of the knee flexion angle: with use of a goniometer, on a lateral
knee radiograph, or on a lateral digital camera image. True lateral
radiographs and images were made by appropriately aligning the posterior
aspects of the femoral condyles (medial and lateral) with complete overlap. A
cadaver knee specimen was stripped of soft tissue, including fascia and
muscle, so that the tibial and femoral shafts were exposed. The posterior
aspect of the capsule, cruciate and collateral ligaments, and tendon
attachments about the knee were left intact. The femur and tibia were cut
transversely 20 cm from the knee joint line. The specimens were devoid of soft
tissue at the distal part of the femoral diaphysis and proximal part of the
tibial diaphysis, allowing accurate measurement of knee flexion angles. Two
unilateral external fixator pins were placed in a lateral-to-medial direction
in both the femur and the tibia, 10 and 15 cm from the tibiofemoral joint
line. An external fixator was used to position the cadaver specimen in 30°
of flexion with use of a goniometer, and rigid fixation was achieved by
connecting the pins in the femur to the pins in the tibia by means of carbon
fiber rods and crosslinks. Two of the authors (J.C.A. and C.P.) then measured
the knee flexion angle five times using each of the three methods noted above.
On the radiographs, the knee flexion angle was measured by drawing lines
parallel to the centers of the femur and tibia. On the digital camera images,
the knee flexion angle was measured with use of ImageJ software (National
Institutes of Health web site
[]).
All digital camera images were made from the medial side of the specimens so
that the fibula and the external fixator pins would not interfere with the
measurements. The centers of the femoral and tibial shafts were measured 10
and 15 cm proximal and distal to the knee joint line, respectively, with use
of the slight protrusions of the external fixator pins along the medial
femoral and medial tibial cortices as references. Lines were drawn through
these points, creating lines parallel to both the femoral and the tibial
shafts. The angle between these two intersecting lines was then measured with
the software. The results are listed in a table in the Appendix. The
interobserver difference was greatest with use of the goniometer method
(>4°), and it was <2° when either the lateral knee radiographs
or the digital camera images were used. The intraobserver standard deviations
were smaller for the measurements obtained with the camera images than they
were for the measurements made on the lateral knee radiographs. Consequently,
the camera image technique was chosen for measurement of the knee flexion
angles.
Knee Testing Before and After Reconstruction
In ten cadaver knees, the anterior cruciate ligament was reconstructed with
use of the central third of the patellar tendon with tibial and patellar bone
plugs and 7 × 20-mm interference screws for femoral fixation. Tibial
fixation was achieved with a custom-designed screw/washer device (Orthopaedic
Research Laboratories, Department of Orthopaedic Surgery, Ann Arbor, Michigan)
to allow for repeated tibial fixation. Specimens were stored at
—20°C and were thawed at room temperature for testing. The specimens
were prepared as they were in the pilot study, with removal of all
extracapsular soft tissue from the femoral and tibial shafts. The posterior
part of the capsule, ligaments, and tendons about the knee were preserved. The
femur, tibia, and fibula were transected 20 cm from the knee joint line, and
external fixator pins were placed in a lateral-to-medial direction in both the
tibia and the femur 10 and 15 cm from the knee joint line. First, holes were
drilled through both cortices in a medial-to-lateral direction in the femur
and tibia 10 and 15 cm from the joint line. The external fixator pins were
then placed in a lateral-to-medial direction so that their tips protruded from
the lateral cortex and were flush with the medial cortex of the femur and
tibia. This allowed reproducible measurement of knee flexion angles on the
digital images of the medial aspect of the specimens without interference from
the fibula and the external fixator pins
(Fig. 1-A). The tips of the
external fixator pins in the medial aspects of the femoral and tibial cortices
were used as reproducible reference points for measurement of knee flexion
angles on images made with the digital camera, as they were in the pilot
study.
A physical examination of each specimen was performed to evaluate the knee
range of motion, anteroposterior laxity (with the Lachman test and anterior
and posterior drawer tests), and stability to varus-valgus stress at 30°
of flexion. Medial and lateral parapatellar arthrotomies were performed to
facilitate direct inspection of the specimen as well as open anterior cruciate
ligament reconstruction. The articular cartilage of the patellofemoral,
medial, and lateral compartments of the knee was inspected, and the native
anterior cruciate ligament was visualized directly to check for evidence of
notch impingement from 90° of flexion to full extension. Specimens with
severe degenerative joint disease or ligament deficiencies were excluded.
Maximum knee extension with the anterior cruciate ligament intact was measured
in each specimen with use of gravity (inversion of the specimen so that the
patella faced the floor while the specimen was held by the femur and the tibia
was free). A true lateral image was made with use of a 5.0-megapixel digital
camera. Three 1200 × 1600-pixel images were made for each specimen. True
lateral images of the knee were made by placing a guide pin along the
posterior aspects of the medial and lateral femoral condyles and overlapping
the condyles so that the pin was seen on its axis. The true lateral images
were made of the medial side of the specimens, as they were in the pilot
study, to avoid interference from the fibula and the external fixator pins,
which protruded laterally (Fig.
1-A). This allowed reproducible lateral images to be made for each
specimen. The native anterior cruciate ligament was then excised from its
tibial and femoral attachment sites, and three lateral images were again made
for each specimen.
The anterior cruciate ligament was then reconstructed in each specimen with
use of a bone-patellar tendon-bone autograft, with the tunnels placed in the
tibial and femoral footprints of the native anterior cruciate ligament. The
graft was harvested with use of an oscillating saw to cut tibial and patellar
bone plugs measuring 30 mm in length and 10 mm in width. The central third of
the patellar tendon was 10 mm in width. The tibial guide pin was placed
approximately 2 mm medial and posterior to the center of the native anterior
cruciate ligament footprint and was overdrilled with an 11-mm cannulated
reamer. This large tibial tunnel was necessary to accommodate the method of
fixation of the tibial graft discussed below. The femoral guide pin was placed
in the native anterior cruciate ligament footprint at the 10 o'clock position
(right knee) or 2 o'clock position (left knee) against the posterior cortex of
the distal part of the femur with use of a 5-mm offset guide. A 10-mm reamer
was drilled 5 mm beyond the length of the bone plug over the guide pin.
In each specimen, the anterior cruciate ligament was reconstructed under
four different conditions (2 × 2 matrix): (1) the graft was fixed with
44 N (10 lb) of tension at 0° of knee flexion (extension), (2) the graft
was fixed with 44 N of tension at 30° of knee flexion, (3) the graft was
fixed with 89 N (20 lb) of tension at 0° of knee flexion (extension), and
(4) the graft was fixed with 89 N of tension at 30° of knee flexion. The
order of the testing conditions was randomized during a pilot study, and then
the testing was repeated with use of the sequence noted above. With use of
that sequence, the flexion angles measured under each of the four conditions
were within 2° of the values obtained during randomization. The specimens
were tested with use of the above sequence following the pilot study.
A hole was drilled in the tibial bone plug, and a number-5 Ticron suture
(Davis and Geck, Wayne, New Jersey) was passed through the hole. The suture
was passed through the hole in the end of the femoral guide pin, and the
tibial bone plug was advanced through the tibial tunnel into the femoral notch
and into the femoral tunnel. The tibial plug of the bone-patellar tendon-bone
autograft was secured in the femoral tunnel with a 7 × 20-mm
interference screw over a guidewire and reinforced with a number-5 Ticron
suture tied around a post proximal to the femoral tunnel on the lateral aspect
of the distal part of the femur. The tibial side of the graft was advanced
into the tibial tunnel and was secured with a special screw/washer device to
allow repeated tibial fixation (Fig.
1-B). We chose this device to ensure that graft fixation would not
be compromised by repeated fixation attempts, as the experimental protocol
involved testing the reconstructed knee under four different conditions. The
patellar bone block of the autograft was cemented into a threaded metal
cylinder, which was inserted through a smooth metal collar placed on the
outside surface of the tibial bone tunnel
(Fig. 2). The 11-mm tibial
tunnel accommodated a metal collar with an outside diameter of 11 mm and an
inside diameter of 10 mm. During our pilot study, we found that the collar was
necessary to prevent excessive friction between the threaded cylinder and the
cancellous bone surface of the tibial tunnel. With the collar in place, the
threaded inner cylinder pistoned freely within the metal collar. After
appropriate distal tensioning was achieved, the construct was locked in place
by adjusting a nut that threaded onto the inner cylinder, which rested on a
washer on the outer metal collar, which rested flush with the cortical surface
of the proximal part of the tibia (Fig.
1-B).
In all cases, the nut was secured by the same individual with maximal
two-finger torque applied by the thumb and index finger. A firm end point of
the nut on the threaded metal cylinder was achieved in all cases. Our goal was
to replicate anterior cruciate ligament reconstruction as it is performed in
patients. Most surgeons do not measure the exact amount of tension on the
anterior cruciate ligament graft during tibial fixation or measure the torque
on the interference screw. In order to verify the torque achieved in our
study, we measured two-finger torque applied by the same individual. A torque
wrench was used to measure the torque applied to the tibial construct ten
times. In every case, the torque was too low to be measured accurately
(<0.5 ft/lb [0.68 n-m]). Since a firm end point was achieved in every case,
we concluded that overconstraint is not likely during graft tensioning with
two-finger torque with use of our particular method of tibial fixation.
Each specimen was brought into 0° of extension and rigidly secured with
the unilateral external fixator. The anterior cruciate ligament graft was then
tensioned with 44 N of hanging weight attached to the graft through the
threaded tibial cylinder and was secured in place by tightening the nut
against the metal collar with two-finger torque as noted above. Following
graft fixation, the external fixator crosslink was removed, allowing free knee
motion. The knee was inverted and allowed to achieve maximal extension by
means of gravity. True lateral knee images were made by placing a guide pin
along the posterior aspects of the medial and lateral femoral condyles and
overlapping the condyles so that the pin was seen on its axis
(Fig. 3). This allowed
reproducible and accurate lateral images to be made for measurement. Three
lateral images were made of each specimen. The tibial fixation was then
removed by loosening the screw against the metal collar, and the knee was
flexed to 30° and rigidly secured by application and tightening of the
external fixator crosslinks to the carbon fiber bars connecting the pins in
the tibial and femoral cortex. The graft was tensioned with 44 N of hanging
weight, and the tibial fixation was again secured. Following graft fixation,
the external fixator crosslink was removed, allowing free knee motion. Three
true lateral images of the specimen were made as described above with maximum
extension obtained by means of gravity
(Fig. 3). The above procedure
was repeated with the graft tensioned with 89 N of hanging weight, first with
the knee in 0° and then with the knee in 30° of flexion. Three true
lateral images of each specimen were made under each of the different testing
conditions.
Image Analysis
Three images of each specimen were made under each of six conditions
(anterior cruciate ligament intact, anterior cruciate ligament excised, and
anterior cruciate ligament reconstruction under the four different
conditions). The clearest of the three lateral images made under each
condition that revealed appropriate alignment of the posterior aspects of the
femoral condyles (with use of the tibial guide pin on its axis to ensure
accuracy) was selected for analysis. The images were analyzed with use of
ImageJ software, as they were in the pilot study. The centers of the femoral
and tibial shafts were measured 10 and 15 cm proximal and distal from the knee
joint line, respectively, with use of the external fixator pins as a
reference. Lines were drawn through these points, creating lines parallel to
both the femoral and the tibial shaft. The angle between these two
intersecting lines was then measured with the software. Two of the authors
(J.C.A. and C.P.) independently measured the flexion angles on each image,
with a difference of <2° between the two authors' measurements on each
of the images. The two measurements on each image were averaged for each of
the six conditions, and the change in the knee flexion angle following the
anterior cruciate ligament reconstruction was calculated.
Gravity was used to obtain knee extension under each of the six conditions
(anterior cruciate ligament intact, anterior cruciate ligament excised, and
anterior cruciate ligament reconstruction under the four different conditions)
by inverting the specimen so that the patella faced the floor while the
specimen was held by the femur and the tibia was free. Thus, the extension
force exerted on each specimen by means of gravity was equal to the specimen's
weight and was proportional to its size. The average changes in the knee
flexion angles following each of the four different reconstructions are
presented in Table I. The
differences in knee flexion angles achieved following graft fixation were
compared between the different reconstruction conditions for each specimen.
First, the effects of the knee flexion angles were compared while the graft
tension was kept constant—i.e., the flexion angle obtained following
reconstruction with 44 N of tension at 0° of flexion was subtracted from
that obtained with 44 N of tension at 30° of flexion, and the flexion
angle obtained following reconstruction with 89 N of tension at 0° of
flexion was subtracted from that obtained with 89 N of tension at 30° of
flexion (Table II). Next, the
effects of tension were compared while the amount of knee flexion was kept
constant—i.e., the flexion angle following tensioning at 44 N in 0°
of flexion was subtracted from that following tensioning at 89 N in 0° of
flexion, and the flexion angle following tensioning at 44 N in 30° of
flexion was subtracted from that following tensioning at 89 N in 30° of
knee flexion (Table III).
Statistical Analysis
Sample Size Estimate
A loss of 5° of tibiofemoral extension was chosen as an estimate of a
relevant difference between the results of tensioning of the anterior cruciate
ligament graft at 0° of knee flexion and tensioning at 30° of flexion,
as a loss of =5° of extension causes an abnormal gait that can lead to
patellofemoral pain and quadriceps
weakness11. While
any loss of extension can be clinically relevant, our method of measurement of
knee flexion can accurately detect a 5° difference, as noted in our pilot
study. It was estimated that, with a p value of 0.05, a power of 0.8, and an
estimated difference of 5° between groups, eight specimens would be
required to detect a difference with use of the above criteria. On the basis
of our power analysis, ten specimens were tested in order to ensure
significance.
Statistical analysis of the four groups of anterior cruciate ligament
reconstructions was performed with use of the Student t test, with
significance defined as p < 0.05.
There was a significant difference between the groups fixed at 0° and
those fixed at 30° of knee flexion with both 44 N (p = 0.0003) and 89 N (p
= 0.0011) of tension. With use of 44 N of tension, the knees fixed at 30°
of knee flexion demonstrated more than a 12° increase in knee flexion
following anterior cruciate ligament reconstruction compared with the group
fixed at 0° of knee flexion. With use of 89 N of tension, there was a
similar increase in knee flexion (14°) in the group fixed at 30°
compared with that fixed at 0°. There was no significant difference
between the groups tensioned at 44 or 89 N at either 0° (p = 0.38) or
30° (p = 0.29) of knee flexion. Comparison of anterior cruciate ligament
reconstructions under the four different conditions allowed each specimen to
serve as its own internal control, as the only difference between groups was
the knee flexion angle and the graft tension during the final graft fixation.
There was no significant difference between the knee flexion angles of the
intact and disrupted anterior cruciate ligament groups (p = 0.80).
In this study, we specifically examined the effect of the knee flexion
angle and the tension of the graft during graft fixation because these
variables can be altered intraoperatively during reconstruction of the
anterior cruciate ligament. Bylski-Austrow et
al.8 examined the
effects of the knee flexion angle (0° or 30°), initial graft tension
(22 or 44 N), and femoral graft attachment site (anterior, posterior, or
distal edge of the femoral insertion of the native anterior cruciate ligament,
or the over-the-top position) on graft force and anteroposterior translation
of the tibia relative to the femur in six cadaveric knees that had been
reconstructed with a flexible cable. The anteroposterior tibiofemoral
translation of the reconstructed knees was closest to that of intact knees
when the graft had been secured at the distal edge of the femoral footprint of
the native anterior cruciate ligament with an initial tension of 44 N and with
the knee in full extension. Knees tensioned in 30° of flexion were
overconstrained regardless of the amount of tension applied during graft
fixation. Bylski-Austrow et al. focused on anteroposterior tibial translation
and the tension in the anterior cruciate ligament graft when a 100-N anterior
force was continuously applied to the proximal part of the tibia. The degree
of knee extension lost following anterior cruciate ligament reconstruction was
not reported. Furthermore, the anterior cruciate ligament graft was placed in
the anterior third of the tibial footprint of the anterior cruciate ligament
in all reconstructions whereas four femoral insertion sites (anterior,
posterior, distal, and over-the-top position) were used. We positioned the
tunnels within the central aspects of the tibial and femoral footprints of the
native anterior cruciate ligament.
Melby et al.20
evaluated the effects of the knee flexion angle (0° or 30°) and the
initial graft tension (18, 36, 54, 72, or 90 N) during graft fixation in eight
cadaveric knees in which the anterior cruciate ligament was reconstructed with
a polytetrafluoroethylene graft. The tibial tunnel was positioned in the
anterior half of the native anterior cruciate ligament footprint and the
femoral tunnel was placed in the over-the-top position. Anteroposterior tibial
motion, quadriceps force, and tension of the anterior cruciate ligament graft
were measured. The authors noted that increased graft tension increased the
quadriceps force required to achieve full knee extension following anterior
cruciate ligament reconstruction. Although they were able to achieve normal
anteroposterior femorotibial translation, the reconstructed knees demonstrated
posterior, lateral, and external tibial rotational subluxation at low loads.
Also, up to 26% greater quadriceps force was required to achieve full
extension of knees in which 90 N of initial graft tension had been applied at
30° of knee flexion. Melby et al. hypothesized that tensioning the graft
in flexion during graft fixation could be a contributing factor in the
clinical loss of knee extension following anterior cruciate ligament
reconstruction. However, the authors placed the femoral tunnel in the
over-the-top-position, which is not within the femoral footprint of the native
anterior cruciate ligament. Furthermore, they used a synthetic
(polytetrafluoroethylene) graft, which is stiffer than the native anterior
cruciate ligament.
Gertel et al.24
examined ten cadaveric knees that had been reconstructed with a bone-patellar
tendon-bone autograft. The graft was placed through the anteromedial aspect of
the tibial footprint of the native anterior cruciate ligament and the
posterosuperior aspect of the femoral footprint. Testing variables included
the magnitude of tension (22 or 67 N), the flexion angle during tensioning
(0° or 30°), and the direction of application of the tensioning force
(proximal, distal, or distal with a posterior force applied to the tibia). The
authors measured graft tension and joint motion with and without a 100-N
anterior force applied to the proximal part of the tibia at 0°, 30°,
60°, and 90° of flexion for each of the twelve permutations noted
above. Graft forces under all of the twelve different conditions tested were
significantly greater than the forces in the intact anterior cruciate
ligament. Graft forces were the highest when the graft had been tensioned at
30° of flexion (rather than full extension). Joint motion and residual
graft force following fixation were not significantly affected by the
magnitude of the graft tension applied during the initial graft fixation in
that study.
Nabors et al.3
performed a cadaveric and clinical study specifically to examine tensioning of
anterior cruciate ligament grafts in full extension. The cadaveric arm of the
study involved reconstruction of the anterior cruciate ligament with a
bone-patellar tendon-bone graft in seven cadaveric knees. The tunnels for
reconstruction were placed within the central aspects of the footprints of the
anterior cruciate ligament on the femur and tibia. Graft fixation was
performed in full extension with a maximal one-handed pull by the senior
author. Anterior laxity, graft set force, and graft tension were subsequently
measured. The average graft set force (68 N) measured prior to fixation was
significantly higher than the average graft tension (18 N) measured
immediately following interference screw fixation of the graft. The average
anterior laxity was 1 mm greater than that in the intact knee, with an 89-N
anterior force at 30° of knee flexion. While the cadaveric portion of this
study demonstrated that an anterior cruciate ligament graft is not excessively
lax when secured in full extension, the study did not address the issue of
postoperative loss of extension if the graft is secured in flexion. The
clinical arm consisted of a prospective study of fifty-seven patients who
underwent reconstruction of the anterior cruciate ligament with a
bone-patellar tendon-bone autograft tensioned in full extension. At the time
of follow-up, at a minimum of two years postoperatively, the result of
instrumented knee laxity testing (with a KT-1000 device) at 89 N had improved
to 0.8 mm, compared with 7.5 mm preoperatively, and 83% of the patients rated
knee function as nearly normal or normal. One patient had a significant
(defined as >3°) postoperative flexion contracture of 5°. A flexion
contracture occurred in 2% (one) of the fifty-seven patients in this study,
whereas the prevalence of flexion contracture with tensioning performed at
30° of knee flexion in other studies has ranged from 11% to
24%11,25.
The authors concluded that graft tensioning in full extension is associated
with a low prevalence of flexion contractures and excellent functional
results. The authors of the above studies examined the effects of the knee
flexion angle and other variables during graft fixation on knee biomechanics,
specifically anteroposterior translation of the tibia relative to the femur.
No prior study of which we are aware has specifically examined the
relationship between the knee flexion angle during graft fixation and loss of
knee extension following reconstruction of the anterior cruciate ligament.
There are many possible causes of loss of knee extension following
reconstruction of the anterior cruciate
ligament26,27.
The knee flexion angle during graft fixation, tunnel placement, and graft
tension are under the surgeon's control intraoperatively. In our study, we
analyzed the knee flexion angle and graft tension because there is still much
debate about the optimum parameters for these variables and some authors have
recommended graft fixation in some amount of knee
flexion12,13.
Studies have suggested that an anterior cruciate ligament graft tensioned at
30° of flexion is
overconstrained8,20.
Since the graft is under maximum tension in full extension, it is logical that
the amount of tension placed on the graft in this study (44 or 89 N) was not
critical. While anterior cruciate ligament grafts that are overconstrained by
fixation in knee flexion may stretch out following reconstruction, permanent
knee flexion contractures may still develop as postoperative scarring and
posterior capsular contracture may occur before the graft loosens or if it
does not. One in vivo study of a rabbit model demonstrated that higher graft
tensions did not have a detrimental effect on the pullout strength of anterior
cruciate ligament grafts or on histomorphometric graft
remodeling28,
although in another study, of a dog model, anterior cruciate ligament grafts
secured at higher tension demonstrated poorer vascularity and focal myxoid
degeneration29.
These results followed the use of bone-patellar tendon-bone grafts with rigid
fixation. Other anterior cruciate ligament reconstructions may behave
differently.
Our cadaveric study demonstrated a relationship between the knee flexion
angle during graft fixation and loss of extension following anterior cruciate
ligament reconstruction with placement of the tibial and femoral tunnels
within the femoral and tibial footprints of the native anterior cruciate
ligament. Increasing the graft tension from 44 to 89 N at 0° of knee
flexion did not result in a significant loss of extension. At 30° of
flexion, the increased fixation tension did cause a small increase in the loss
of extension, of approximately 4°, although this difference was not
significant (p = 0.29).
Several limitations must be kept in mind when these results are
interpreted. Some of the specimens had contractures of the remaining soft
tissues, including the posterior aspect of the capsule, which may have
contributed to knee flexion during gravity extension in some specimens. When
the specimens were reconstructed, they were placed in 0° of extension and
held firmly in place by the application of a bar clamp to the unilateral
external fixator. Following graft fixation, the bar clamp was removed,
allowing free knee motion. The knee flexion angle was then measured during
gravity extension. The posterior aspect of the capsule and the other
soft-tissue contractures may have stretched out somewhat during knee extension
to 0° during tibial fixation. This could contribute to a difference in the
amount of knee flexion observed between specimens during gravity extension
prior to anterior cruciate ligament reconstruction. Furthermore, each specimen
has unique properties, including size, osseous anatomy, and native cruciate
and collateral ligament morphology. This is why the differences in the knee
flexion angle following anterior cruciate ligament reconstruction were
compared for each specimen under the four different testing conditions. This
allowed each specimen to serve as its own internal control so that the desired
variables (the knee flexion angle and graft tension) could be most effectively
isolated.
We found a clinically and statistically significant loss of knee extension
following the anterior cruciate ligament reconstructions in which the grafts
were tensioned at 30° compared with that in the groups tensioned at
0°. There was no significant difference in the observed loss of extension
following anterior cruciate ligament reconstruction between the groups
tensioned at 89 N and the groups tensioned at 44 N.
In retrospect, we could have randomized the tension and flexion angle
during testing. The reason why we did not do so was that randomization had no
effect on the measured flexion angles in our pilot study. Loss of extension
was observed when the anterior cruciate ligament graft had been tensioned at
30° of flexion regardless of the order of testing. If the order of testing
had an effect on the measured knee flexion angle, we would expect to see a
pattern of either increased or decreased knee flexion as testing proceeded.
This was not observed.
Cadaver studies of surgical procedures must be interpreted with caution
because soft-tissue healing and graft remodeling do not occur. It is unknown
how these factors would affect knee extension over time in vivo. Our results
are valid for these particular reconstruction and tensioning parameters only.
We also did not examine the relationship between the anteroposterior
tibiofemoral position and the anterior cruciate ligament graft tension during
fixation, as these parameters had been adequately analyzed in previous
studies8,20,24.
The goal of our study was to evaluate the loss of knee extension following
anterior cruciate ligament reconstruction with femoral and tibial tunnel
placement within the native femoral and tibial footprints and with use of
different knee flexion angles during graft fixation. The results indicate that
appropriate placement of the anterior cruciate ligament graft within the
native femoral and tibial footprints may not ensure full extension if the
graft is fixed in 30° of knee flexion. In conclusion, fixation of the
graft in full knee extension may help to decrease the loss of knee extension
following anterior cruciate ligament reconstruction.
A table showing the pilot study data is 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). ?