The Effect of Soft-Tissue Graft Fixation in Anterior Cruciate

Ligament Reconstruction on Graft-Tunnel Motion Under

Anterior Tibial Loading

Eiichi Tsuda, M.D., Yukihisa Fukuda, M.D., John C. Loh, M.D.,

Richard E. Debski, Ph.D., Freddie H. Fu, M.D., and Savio L-Y. Woo, Ph.D., D.Sc.

Purpose: To compare the motion of an anterior cruciate ligament (ACL) replacement graft within the
femoral bone tunnel (graft-tunnel motion) when a soft-tissue graft is secured either by a titanium button
and polyester tape (EndoButton fixation; Acufex, Smith & Nephew, Mansfield, MA) or by a biodegrad-
able interference screw (Biointerference fixation; Endo-fix; Acufex, Smith & Nephew) An additional
purpose was to evaluate the effect of the graft-tunnel motion on the kinematics of ACL-reconstructed
knees and in situ force of the ACL replacement graft. Type of Study: Biomechanical experiment using
an in vitro animal model. Methods: ACL reconstruction with a flexor tendon autograft was performed in
8 cadaveric knees of skeletally mature goats. The knee kinematics and the in situ force in the ACL
replacement graft in response to anterior tibial loads were evaluated using the robotic/universal force-
moment sensor testing system. The longitudinal and transverse graft-tunnel motion during anterior tibial
loading was determined based on radiographic measurements parallel and perpendicular to the femoral
bone tunnel, respectively. Results: In response to an anterior tibial load of 100 N, the longitudinal
graft-tunnel motion for EndoButton fixation and Biointerference fixation was 0.8 ⫾ 0.4 mm and 0.2 ⫾ 0.1
mm, respectively (P ⬍ .05), whereas the transverse graft-tunnel motion was 0.5 ⫾ 0.2 mm and 0.1 ⫾ 0.1
mm, respectively (P ⬍ .05). Furthermore, the anterior tibial translation for EndoButton fixation (5.3 ⫾ 1.2
mm) was also significantly larger than that for Biointerference fixation (4.2 ⫾ 0.9 mm) (P ⬍ .05). With
both fixations, however, no significant difference between the in situ forces in the ACL replacement graft
and that in the intact ACL could be detected. Conclusions: EndoButton fixation of a soft-tissue graft via an
elastic material resulted in significantly larger graft-tunnel motion, and consequently, greater anterior knee laxity
compared with more rigid fixation using an interference screw closer to the intra-articular entrance of the bone
tunnel. In terms of force distribution, the ACL replacement graft in both fixations still functioned as a primary
restraint to an anterior tibial load close to the intact ACL. Key Words: Anterior cruciate ligament—
Reconstruction—Graft fixation—Graft-tunnel motion—Knee kinematics—In situ force.

A current subject of debate related to anterior cru-

ciate ligament (ACL) reconstruction is the selec-
tion of conservative versus aggressive postoperative
From the Musculoskeletal Research Center, Department of Or-
thopaedic Surgery, University of Pittsburgh Medical Center, Pitts-
burgh, Pennsylvania, U.S.A.
Supported by National Institutes of Health Grant No. AR39683
and the Musculoskeletal Research Center, University of Pittsburgh
Medical Center, Pittsburgh, Pennsylvania.
Address correspondence and reprint requests to Savio L-Y. Woo,
Ph.D., D.Sc., Musculoskeletal Research Center, 210 Lothrop St,
E161-BST, P.O. Box 71199, Pittsburgh, PA 15213, U.S.A. E-mail:
decenzod@msx.upmc.edu
© 2002 by the Arthroscopy Association of North America
0749-8063/02/1809-3092$35.00/0
doi:10.1053/jars.2002.36112
rehabilitation. Since Shelbourne and Nitz1 advocated
accelerated rehabilitation protocols for the knee recon-
structed using a patellar tendon autograft, similar pro-
tocols have been extended to ACL-reconstructed
knees with a hamstring graft.2-4 Concerns about
whether early aggressive rehabilitation would subject
the ACL replacement graft to excessive graft-tunnel
motion before its biological incorporation within the
bone tunnel have arisen because the literature suggests
that graft incorporation normally takes 8 to 12 weeks
after reconstruction surgery.5,6 Thus, the fixation site
for a hamstring graft could be compromised secondary
to repetitive loading.
Recent biomechanical studies have shown that teth-
ering a hamstring graft via polyester tape connected to

960 Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 18, No 9 (November-December), 2002: pp 960 –967
 ACL GRAFT-TUNNEL MOTION
a titanium button (EndoButton fixation) (Acufex,
Smith & Nephew, Mansfield, MA) allows a sizable
amount of motion telescoping parallel to the bone
tunnel (longitudinal graft-tunnel motion) during cyclic
tensile loading.7 In addition, fixation of the graft far
away from the original ACL insertion would allow the
graft to move back and forth within the bone tunnel in
the sagittal plane (transverse graft-tunnel motion).8-10
These excessive graft-tunnel motions may not only
delay graft incorporation within the bone tunnel, but
also alter the biomechanical function of the ACL-
reconstructed knee.11,12 In contrast, graft fixation with
a biodegradable interference screw (Biointerference
fixation; Endo-fix; Acufex, Smith & Nephew) enables
the graft to be secured closer to the anatomic location
of the original ACL attachment, thus reducing both
longitudinal and transverse graft-tunnel motion.13-15
The first objective of this study was to compare the
longitudinal and transverse graft-tunnel motion in re-
sponse to anterior tibial loads between EndoButton
fixation and Biointerference fixation. The hypothesis
was that EndoButton fixation would result in larger
longitudinal graft-tunnel motion than Biointerference
fixation because the majority of graft-tunnel motion
would come from the polyester tape rather than the
graft. Furthermore, differences in the graft fixation
location at the femur led to another hypothesis: that
EndoButton fixation would result in larger transverse
graft-tunnel motion than Biointerference fixation. The
second objective was to measure the corresponding
knee kinematics and the in situ force in the ACL
replacement graft. It was further hypothesized that the
larger graft-tunnel motion following EndoButton fix-
FIGURE 1. Photograph and
schematic illustration of the ro-
botic/UFS testing system with a
goat knee mounted for testing.
961
ation would result in greater anterior tibial translation
and lower in situ force than Biointerference fixation in
the ACL replacement graft.
METHODS
A goat model was chosen for this study because it
allowed for more accurate and consistent surgical
reconstruction than smaller animal models such as
rabbits or dogs. Previous in vivo and in vitro experi-
ments have shown that ACL reconstruction in the goat
model can effectively restore knee stability, represent-
ing a significant decrease in anterior tibial translation
compared with the ACL-deficient knee.16-18 Eight
hind limbs were obtained from skeletally mature
Spanish-bred female goats (40-50 kg). The hind limbs
were immediately frozen after removal and stored at
⫺20° C until 1 day before testing.19 After the legs
were thawed at room temperature for 24 hours, the
deep flexor digitorum tendon was harvested as a re-
placement graft for the ACL. The harvested tendon
was cut to a length of approximately 20 cm and folded
at the midpoint. The looped graft was trimmed so it
could pass through a 5-mm graft sizer. Pretension of
30 N was applied for 15 minutes.
The surrounding skin and muscles were dissected
10 cm proximally and distally away from the knee
joint line, and the exposed proximal femur and distal
tibia were potted in cylinders of epoxy compound.
The femoral cylinder was fixed to the base of the
robotic manipulator (Puma Model 762; Unimate,
Pittsburgh, PA) through an adjustable clamp (Fig 1).
The tibial clamp was attached to the end-effector of
962 E. TSUDA ET AL.
TABLE 1. Outline of Testing Protocol and Data Acquired
Protocol
Intact ACL
Path of passive flexion-extension
Apply load (33 N, 67 N, and 100 N AP)
ACL transection
Repeat kinematics (K1)
Apply load (33 N, 67 N, and 100 N AP)
ACL reconstruction (first fixation)
Apply load (33 N, 67 N, and 100 N AP)
Radiographs under anterior tibial loads of 0 N, 33 N, 67 N, 100 N, and 0 N
Graft release
Repeat kinematics (K3)
ACL reconstruction (second fixation)
Apply load (33 N, 67 N, and 100 N AP)
Radiographs under anterior tibial loads of 0 N, 33 N, 67 N, 100 N, and 0 N
Graft release
Repeat kinematics (K4)
the robotic manipulator through the universal force-
moment sensor (UFS, Model 4015; JR3, Woodland,
CA). This robotic/UFS testing system is capable of
operating in position-control mode, whereby the tibia
is moved to a desired position and orientation while
the resulting forces are measured.20 The system can
also be operated in force-control mode. With force-
position feedback between the UFS and the robotic
manipulator, a desired force target could be achieved
while the resulting changes in knee kinematics were
recorded.21-23
A summary of the testing protocol is shown in
Table 1. First, the path of passive flexion-extension of
the intact knee was determined in 1° increments by
minimizing all external forces and moments. This
position was used as the starting point during the
ensuing experiments. Anterior tibial loads of 33 N, 67
N, and 100 N were then applied to the intact knee at
60° of knee flexion, and the knee kinematics (K1)
were recorded.
Subsequently, the ACL was transected under direct
visualization through a lateral parapatellar arthrotomy.
The robotic manipulator reproduced the previously
recorded kinematics of the intact knee in response to
the anterior tibial loads (K1), while the UFS measured
a new set of forces and moments. Using the principle
of superposition,22,24 the change in the force vector
measured before and after ACL transection repre-
sented the in situ force in the ACL. The anterior tibial
loads were then applied to determine the kinematics of
the ACL-deficient knee (K2).
Finally, ACL reconstructions were performed on
the same knee using the prepared graft. EndoButton
Data Acquired
Kinematics of intact knee (K1)
In situ force of intact ACL
Kinematics of ACL-deficient knee (K2)
Kinematics of ACL-reconstructed knee (K3)
Radiographs 1, 2, 3, 4, and 5
In situ force of ACL graft in first fixation
Kinematics of ACL-reconstructed knee (K4)
Radiographs 1, 2, 3, 4, and 5
In situ force of ACL graft in second fixation
fixation and Biointerference fixation procedures were
used in a varied order (Fig 2). Before graft placement,
4 radiographic markers were placed into the looped
graft. A 1-mm diameter metal chip was completely
buried and sutured into 2-mm long slits that were 5
mm and 17 mm from the looped end of the graft for
proximal and distal markers, respectively. Marker mi-
gration within the tendon was prevented by using 2
circumscribing sutures placed at locations correspond-
ing to just proximal and distal of each marker.
Bone tunnels (5 mm in diameter) were drilled at the
femoral and tibial insertions of the transected ACL.
For drilling the tibial bone tunnel, the guide wire was
placed between the insertions of the anteromedial and
posterolateral bundles. During the procedures for the
femoral bone tunnel, a femoral marker consisting of a
metal pin of known length was inserted parallel to the
femoral bone tunnel (Fig 2). This stationary femoral
marker served as a reference to determine the relative
position of the graft markers on radiographs (Fig 3). A
portable x-ray machine was set up to take lateral
radiographs of the ACL-reconstructed knee. The x-ray
cassette was aligned parallel to the femoral bone tun-
nel, and the gantry of the machine was adjusted so that
the x-ray beam was perpendicular to the cassette. The
cassette holder and the gantry of the machine were
fixed throughout testing.
After the length of the femoral and tibial bone
tunnels were measured, a 20-mm portion of the looped
graft was placed into the femoral bone tunnel. For
EndoButton fixation, the graft was connected to a
titanium button (EndoButton) via a single looped 5-
mm wide polyester tape (EndoButton tape; Acufex,
 ACL GRAFT-TUNNEL MOTION
FIGURE 2. Placement of ra-
diographic markers during
ACL reconstruction using (A)
EndoButton fixation and (B)
Biointerference (Endo-fix) fix-
ation.
Smith & Nephew) tied with 2 square knots (Fig 2A).
For Biointerference fixation, a biodegradable interfer-
ence screw 5.5 mm in diameter and 15 mm long
(Endo-fix) was driven in an inside-out direction be-
yond the distal graft markers (Fig 2B). To minimize
the possibility that graft rotation during the fixation
963
causes an overlapping of the graft markers on the
lateral radiograph, the interference screw was inserted
between 2 strands of the looped graft so that each
strand filled the anterior and posterior half of the
femoral tunnel. Although insertion of the oversized
bioabsorbable interference screw might damage the
FIGURE 3. (A) The longitudi-
nal and transverse graft-tunnel
motion. (B) Radiographic mea-
surement: D1 and D2, longitu-
dinal graft-tunnel motion; D3
and D4, transverse graft-tunnel
motion; E1 and E2, graft elon-
gation.
964 E. TSUDA ET AL.
soft tissue graft and enlarge the bone tunnel, no sig-
nificant change in the graft-tunnel motion for the
EndoButton fixation before and after Biointerference
fixation was detected in a preliminary study.
The other end of the graft was secured on the
anteromedial cortex of the tibia with a soft tissue
spiked washer (Linvatec, Largo, FL) and a bicortical
screw 4.0 mm in diameter, applying an initial graft
tension of 17 N to each strand of the looped graft at
60° of knee flexion. The anterior tibial loads of 33 N,
67 N, and 100 N were applied to the ACL-recon-
structed knee to determine the kinematics for each
fixation (K3 or K4). Sequentially, 5 lateral radio-
graphs at 60° of knee flexion were taken for each
fixation: 0 N (radiograph 1), 33 N (radiograph 2), 67
N (radiograph 3), 100 N (radiograph 4), and 0 N again
(radiograph 5) (Table 1).
The graft was then released at the tibial fixation site,
and the recorded kinematics of ACL-reconstructed
knee (K3 or K4) were repeated while a new set of
forces and moments were measured for each fixation.
The change in the force vector measured before and
after graft release represented the in situ force in the
ACL replacement graft.
The radiographs were scanned and the data was
analyzed using a personal computer. The radiological
magnification coefficient was calculated from the di-
rect measurement with a dial caliper and radiographic
measurement of the femoral marker. The longitudinal
graft-tunnel motion was defined as the change in dis-
tance from the distal end of the femoral marker to the
center of the distal graft marker, parallel to the fem-
oral marker (D1 and D2, Fig 3). The transverse graft-
tunnel motion was defined as the change in perpen-
dicular distance from the femoral marker to the center
of the distal graft marker (D3 and D4, Fig 3).
Using the position of the graft marker when 0 N
load (radiograph 1) was applied as a reference posi-
tion, the longitudinal and transverse graft-tunnel mo-
tion in response to anterior tibial loads of 33 N, 67 N,
and 100 N were determined based on the displacement
of the graft marker in radiographs 2, 3, and 4. Simi-
larly, the residual longitudinal and transverse graft
displacement after anterior tibial loading were calcu-
lated from radiographs 1 and 5. Furthermore, the graft
elongation within the femoral bone tunnel and its
contribution to the longitudinal graft-tunnel motion
were determined. Graft elongation was defined as the
change in distance between proximal and distal graft
markers parallel to the femoral marker (E1 and E2,
Fig 3). All radiographic measurements were corrected
with the radiologic magnification coefficient and av-
eraged in both strands. In a preliminary study, the
difference between the radiographic and dial caliper
measurements was found to be less than 0.05 mm;
thus the accuracy of this method was determined to be
0.05 mm.
Because the entire test was performed in the same
specimen, statistical analysis was performed using a
2-factor repeated-measures analysis of variance
(ANOVA). The 2 factors investigated were the con-
ditions of the knee and the magnitude of the anterior
tibial load. The dependent variables evaluated were
longitudinal and transverse graft-tunnel motion, ante-
rior tibial translation, and the in situ force in the ACL
or ACL replacement grafts. Significance was set at
P ⬍ .05.
RESULTS
In response to anterior tibial loads of 33 N, 67 N,
and 100 N, distal graft markers were found to displace
distally and anteriorly within the femoral bone tunnel.
The graft-tunnel motion for both EndoButton and
Biointerference fixations increased in proportion to
the increase in the magnitude of anterior tibial load. In
response to anterior tibial loads of 67 and 100 N, the
longitudinal graft-tunnel motions for EndoButton fix-
ation were 0.5 ⫾ 0.3 mm and 0.8 ⫾ 0.4 mm, respec-
tively. The longitudinal graft-tunnel motions for
Biointerference fixation were significantly reduced, to
0.1 ⫾ 0.1 mm and 0.2 ⫾ 0.1 mm, respectively (Fig 4).
Elongation of the graft was smaller in comparison
to longitudinal graft-tunnel motion. For EndoButton
fixation, the elongation was 0.2 ⫾ 0.2 mm in response
FIGURE 4. Longitudinal graft-tunnel motion for EndoButton fix-
ation and Biointerference fixation under anterior tibial loads. As-
terisk indicates P ⬍ .05 when compared with Biointerference
fixation.
 ACL GRAFT-TUNNEL MOTION
FIGURE 5. Transverse graft-tunnel motion for EndoButton fixa-
tion and Biointerference fixation under anterior tibial loads. Aster-
isk indicates P ⬍ .05 when compared with Biointerference fixation.
to an anterior tibial load of 100 N, which accounted
for less than 25% of the longitudinal graft-tunnel
motion. In other words, more than 75% of the longi-
tudinal graft-tunnel motion could be attributed to the
polyester tape and titanium button construct. Graft
elongation in Biointerference fixation was even
smaller (⬍0.05 mm), less than the accuracy of the
radiographic measurement system.
For EndoButton fixation, transverse graft-tunnel
motions in response to anterior tibial loads of 67 and
100 N were 0.3 ⫾ 0.2 mm and 0.5 ⫾ 0.2 mm,
respectively. For Biointerference fixation, these mo-
tions were 0.1 ⫾ 0.1 mm and 0.1 ⫾ 0.1 mm, respec-
tively, representing a significant difference between
the 2 fixations (Fig 5). Residual longitudinal and
transverse graft displacements after anterior tibial
loading were less than 0.1 mm in both fixations,
representing no significant difference between the 2
fixation methods.
In terms of the knee kinematics, a significant in-
crease in anterior tibial translation, ranging from 11
mm to 15 mm, occurred after transection of the ACL
(Table 2). Following ACL reconstruction using
TABLE 2. Anterior Tibial Translation of Intact ACL-Deficient and ACL-Reconstructed Knee
in Response to Anterior Tibial Loads
Anterior Tibial Load Intact Knee ACL-Deficient Knee
33 N 1.2 ⫾ 0.4 12.7 ⫾ 3.2*
67 N 2.4 ⫾ 0.4 17.2 ⫾ 2.1*
100 N 3.6 ⫾ 0.5 18.5 ⫾ 2.3*
NOTE. Values are given as mean ⫾ standard deviation (mm).
*Significantly larger compared with 3 other knee conditions (P ⬍ .05).
†Significantly larger compared with the intact knee (P ⬍ .05).
‡Significantly larger compared with ACL-reconstructed knee using biointerference fixation (P ⬍ .05).
965
Biointerference fixation, the anterior tibial translation
was restored to within 0.6 mm of the intact knee,
although that increase was still statistically significant
with an anterior tibial load of 100 N. Similarly, in
response to anterior tibial loads of 67 and 100 N, the
anterior tibial translations for EndoButton fixation
were 1.0 mm and 1.7 mm greater than that of the intact
knee, respectively, and were also significantly larger
than the translations in Biointerference fixation.
In contrast, the in situ force in the ACL replacement
graft in either EndoButton or Biointerference fixation
was regained to more than 90% of the in situ force in
the intact ACL. No significant difference in in situ
force was detected between the ACL replacement
grafts in EndoButton and in Biointerference fixation
and the intact ACL (Table 3).
DISCUSSION
Using a goat knee model, the longitudinal and trans-
verse graft-tunnel motion in response to anterior tibial
loading has been quantified and compared between
EndoButton fixation and Biointerference fixation. The
robotic/UFS testing system was used to apply the
anterior tibial load and evaluate the mechanical func-
tion of intact ACL and ACL replacement grafts; in
other words, to evaluate multiple degree of freedom
knee kinematics and the in situ forces in the tis-
sue.12,17,24-26 The testing system enabled different
knee states (intact, ACL-deficient, or ACL-recon-
structed) to be tested using the same knee. Therefore,
the interspecimen variability can be minimized, and
the statistical power increased.27
The data obtained confirmed the hypothesis for the
first objective: that EndoButton fixation would lead to
larger longitudinal graft-tunnel motion than Biointer-
ference fixation. The longitudinal graft-tunnel motion
for EndoButton fixation was approximately 4 times
ACL-Reconstructed Knee
Endobutton Fixation Biointerference Fixation
1.2 ⫾ 1.4 0.8 ⫾ 0.9
3.4 ⫾ 1.0‡‡ 2.8 ⫾ 1.1
5.3 ⫾ 1.2†‡ 4.2 ⫾ 0.9†
966 E. TSUDA ET AL.
TABLE 3. In Situ Force in Intact ACL and ACL
Replacement Grafts in Response to Anterior Tibial Loads
ACL Replacement Graft
Anterior Tibial Intact Endobutton Biointerference
Load ACL Fixation Fixation
33 N 32 ⫾ 6 31 ⫾ 9 30 ⫾ 7
67 N 66 ⫾ 8 65 ⫾ 6 67 ⫾ 7
100 N 104 ⫾ 11 104 ⫾ 12 104 ⫾ 8
NOTE. Values are given as mean ⫾ standard deviation (N).
larger than that for Biointerference fixation when the
reconstructed knee was subjected to an anterior tibial
load of 100 N. Furthermore, the majority of the lon-
gitudinal graft-tunnel motion resulted from the poly-
ester tape and its interfaces with the graft and the
titanium button. These results correspond to the find-
ings in previous studies that showed the mechanical
behavior of the hamstring graft under cyclic load-
ing.7,11
In addition, the study measured the transverse graft-
tunnel motion in response to an anterior tibial load. It
was found that transverse graft-tunnel motion for En-
doButton fixation was approximately 5 times greater
than that for Biointerference fixation, confirming the
hypothesis for the first objective. Transverse graft-
tunnel motion has typically been observed when using
a bone–patellar tendon– bone graft because substantial
space near the intra-articular entrance of the bone
tunnel is caused by a diameter mismatch between the
bone tunnel and the tendinous portion of the
graft.10,12,28 In this study, the looped graft was
trimmed to fit snugly into the femoral bone tunnel;
therefore, the amount of transverse graft-tunnel mo-
tion did not exceed 10% of the tunnel diameter. How-
ever, there was a significant increase in the transverse
graft-tunnel motion when the graft fixation was far
from the intra-articular entrance of the bone tunnel.
In terms of biomechanical function of ACL replace-
ment graft, the results partially support the hypothesis
for the second objective. The anterior tibial translation
for EndoButton fixation was significantly larger than
that for Biointerference fixation. Additionally, be-
cause the same graft and the same knee were used for
both fixations, the larger graft-tunnel motion in En-
doButton fixation compared with Biointerference fix-
ation was directly reflected in the greater anterior
tibial translation. In contrast, no significant difference
was seen between the in situ force in the ACL replace-
ment graft fixed with an EndoButton fixation or
Biointerference fixation and the in situ force in intact
ACL. These results suggest that the ACL replacement
graft in both fixations still functioned during the ap-
plication of an anterior tibial load. However, larger
graft-tunnel motion with EndoButton fixation compro-
mised the knee kinematics, as evidenced by larger
anterior tibial translation.
Because of the tibial motion in 5 degrees of freedom
during application of the anterior tibial load, graft-
tunnel motion within the tibial bone tunnel could not
be measured using the radiographic system. There-
fore, the ACL replacement graft was secured to the
tibia using the same fixation technique (the screw and
washer) in EndoButtton and Biointerference fixation.
Use of the bioabsorbable interference screw for both
femoral and tibial fixation would more closely mimic
the clinical setting and may cause changes in the
biomechanical function of ACL replacement graft.
This cadaveric study using the goat model demon-
strated that a soft-tissue graft tethered via an elastic
material would result in significantly larger graft-tun-
nel motion. The results may be may be altered using a
human cadaveric model. However, it is reasonable
that the graft-tunnel motion during aggressive rehabil-
itation could be larger before soft-tissue graft incor-
poration within the bone tunnel is complete. Whether
this larger graft-tunnel motion increases the risk for
graft failure at the fixation site needs to be investi-
gated. Furthermore, whether the larger transverse
graft-tunnel motion can cause greater bone tunnel
enlargement following fixation28-30 could be examined
applying this model. The methodology developed and
quantitative data obtained from this study can serve as
a foundation for future in vivo evaluations of the
effects of graft-tunnel motion on soft tissue graft in-
corporation and the biomechanical properties of the
newly formed ligament-bone interface. The scientific
data obtained should help in the consideration and
determination of rehabilitation protocols when using
soft-tissue grafts.
Acknowledgment: The authors thank Theodore W.
Rudy, M.A., Jennifer Zeminski, B.S., and Yasuhiko Wa-
tanabe, M.D. for their technical assistance.
REFERENCES
1. Shelbourne KD, Nitz P. Accelerated rehabilitation after ante-
rior cruciate ligament reconstruction. Am J Sports Med 1990;
18:292-299.
2. Brown CH, Sklar JH. Endoscopic anterior cruciate ligament
reconstruction using doubled gracilis and semitendinosus ten-
dons and endobutton femoral fixation. Tech Orthop 1998;13:
281-298.
3. Howell SM, Hull ML. Aggressive rehabilitation using ham-
 ACL GRAFT-TUNNEL MOTION
string tendons: Graft construct, tibial tunnel placement, fixa-
tion properties, and clinical outcome. Am J Knee Surg 1998;
11:120-127.
4. MacDonald PB, Hedden D, Pacin O, Huebert D. Effects of an
accelerated rehabilitation program after anterior cruciate liga-
ment reconstruction with combined semitendinosus-gracilis
autograft and a ligament augmentation device. Am J Sports
Med 1995;23:588-592.
5. Grana WA, Egle DM, Mahnken R, Goodhart CW. An analysis
of autograft fixation after anterior cruciate ligament recon-
struction in a rabbit model. Am J Sports Med 1994;22:344-351.
6. Rodeo SA, Arnoczky SP, Torzilli PA, et al. Tendon-healing in
a bone tunnel: A biomechanical and histological study in the
dog. J Bone Joint Surg Am 1993;75:1795-1803.
7. Höher J, Livesay GA, Ma CB, et al. Hamstring graft motion in
the femoral bone tunnel when using titanium button/polyester
tape fixation. Knee Surg Sports Traumatol Arthrosc 1999;7:
215-219.
8. Fu FH, Bennett CH, Lattermann C, Ma CB. Current trends in
anterior cruciate ligament reconstruction. Part 1: Biology and
biomechanics of reconstruction. Am J Sports Med 1999;27:
821-830.
9. Jørgensen U, Thomsen HS. Behavior of the graft within the
bone tunnels following anterior cruciate ligament reconstruc-
tion, studied by cinematic magnetic resonance imaging. Knee
Surg Sports Traumatol Arthrosc 2000;8:32-35.
10. Morgan CD, Kalmam VR, Grawl DM. Isometry testing for
anterior cruciate ligament reconstruction revisited. Arthros-
copy 1995;11:647-659.
11. Hḧer J, Scheffler SU, Withrow JD, et al. Mechanical behavior
of two hamstring graft constructs for reconstruction of the
anterior cruciate ligament. J Orthop Res 2000;18:456-461.
12. Ishibashi Y, Rudy TW, Livesay GA, et al. The effect of
anterior cruciate ligament graft fixation site at the tibia on knee
stability: Evaluation using a robotic testing system. Arthros-
copy 1997;13:177-182.
13. Brand J, Weiler A, Caborn DM, et al. Graft fixation in cruciate
ligament reconstruction. Am J Sports Med 2000;28:761-774.
14. Jansson KA, Harilainen A, Sandelin J, et al. Bone tunnel
enlargement after anterior cruciate ligament reconstruction
with the hamstring autograft and endobutton fixation tech-
nique: A clinical, radiographic and magnetic resonance imag-
ing study with 2 years follow-up. Knee Surg Sports Traumatol
Arthrosc 1999;7:290-295.
15. Pinczewski LA, Clingeleffer AJ, Otto DD, et al. Integration of
hamstring tendon graft with bone in reconstruction of the
anterior cruciate ligament. Arthroscopy 1997;13:641-643.
16. Jackson DW, Grood ES, Goldstein JD, et al. A comparison of
patellar tendon autograft and allograft used for anterior cruci-
ate ligament reconstruction in the goat model. Am J Sports
Med 1993;21:176-185.
967
17. Ma CB, Papageorgiou CD, Debski RE, Woo SL-Y. Interaction
between the ACL graft and MCL in a combined ACL and
MCL knee injury using a goat model. Acta Orthop Scand
2000;71:387-393.
18. Ng GY, Oakes BW, Deacon OW, et al. Long-term study of the
biochemistry and biomechanics of anterior cruciate ligament-
patellar tendon autografts in goats. J Orthop Res 1996;14:851-
856.
19. Woo SL-Y, Orlando CA, Camp JF , Akeson WH. Effects of
postmortem storage by freezing on ligament tensile behavior.
J Biomech 1986;19:399-404.
20. Fujie H, Mabuchi K, Woo SL-Y, et al. The use of robotics
technology to study human joint kinematics: A new method-
ology. J Biomech Eng 1993;115:211-217.
21. Fujie H, Livesay GA, Woo SL-Y, et al. The use of a universal
force-moment sensor to determine in-situ forces in ligaments.
A new methodology. J Biomech Eng 1995;117:1-7.
22. Livesay GA, Fujie H, Kashiwaguchi S, et al. Determination of
the in situ forces and force distribution within the human
anterior cruciate ligament. Ann Biomed Eng 1995;23:467-474.
23. Rudy TW, Livesay GA, Woo SL-Y, Fu FH. A combined
robotic/universal force sensor approach to determine in situ
forces of knee ligaments. J Biomech 1996;29:1357-1360.
24. Sakane M, Fox RJ, Woo SL, et al. In situ forces in the anterior
cruciate ligament and its bundles in response to anterior tibial
loads. J Orthop Res 1997;15:285-293.
25. Kanamori A, Woo SL-Y, Ma CB, et al. The forces in the
anterior cruciate ligament and knee kinematics during a sim-
ulated pivot shift test: A human cadaveric study using robotic
technology. Arthroscopy 2000;16:633-639.
26. Papageorgiou CD, Gil JE, Kanamori A, et al. The biomechani-
cal interdependence between the anterior cruciate ligament
replacement graft and the medial meniscus. Am J Sports Med
2001;29:226-231.
27. Woo SL-Y, Debski RE, Withrow JD, Janaushek MA. Biome-
chanics of knee ligaments: Current concepts. Am J Sports Med
1999;27:533-543.
28. L’Insalata JC, Klatt B, Fu FH, Harner CD. Tunnel expansion
following anterior cruciate ligament reconstruction: A com-
parison of hamstring and patellar tendon autografts. Knee Surg
Sports Traumatol Arthrosc 1997;5:234-238.
29. Clatworthy MG, Annear P, Bulow JU, Bartlett RJ. Tunnel
widening in anterior cruciate ligament reconstruction: A pro-
spective evaluation of hamstring and patella tendon grafts.
Knee Surg Sports Traumatol Arthrosc 1999;7:138-145.
30. Webster KE, Feller JA, Hameister KA. Bone tunnel enlarge-
ment following anterior cruciate ligament reconstruction: A
randomized comparison of hamstring and patellar tendon
grafts with 2 years follow-up. Knee Surg Sports Traumatol
Arthrosc 2001;9:86-91.