Arthroscopy: The Journal of Arthroscopic and Related Surgery 8(4):453-458

Published by Raven Press, Ltd. 0 1992 Arthroscopy Association of North America

The Biomechanical Effects of Geometric Configuration of

Bone-Tendon-Bone Autografts in Anterior Cruciate
Ligament Reconstruction

Jeffrey D. Shapiro, M.D., Bruce T. Cohn, M.D., Douglas W. Jackson, M.D.,

Paul D. Postak, B.S., Richard D. Parker, M.D., and
A. Seth Greenwald, D. Phil. (OXON)

Summary: This study provides biomechanical support for a new technique of

autograft anterior cruciate ligament reconstruction featuring circular bone
plugs and endosteal interference tit fixation. Six matched pairs of fresh frozen
human knees were utilized. Femoral interference tit pull-out strength was de-
termined from material-testing-machine-generated oscillograph recordings at a
strain rate of 100%/s. Circular bone plugs, obtained with a circular oscillating
saw, provided 19.9% greater interference fit pull-out strength compared with
identically fixed trapezoidal bone plugs. Different geometric defects were com-
pared in three- and four-point bending on an Instron machine with frozen
patellae and an artificial bone composite. Circular defects have 107% greater
strength than matched trapezoidal patellar defects in three-point bending. In a
bone composite, circular defects are 53% stronger than triangular and 25%
stronger in four-point bending than trapezoidal defects. A new technique of
harvesting bone plugs with endosteal interference fit fixation is described and
biomechanically supported. To date, this technique has been performed on
over 500 cases clinically without evidence of patellar fracture or fixation fail-
ure. This study demonstrates the efficacy of this simple and reproducible tech-
nique compared with previously reported procedures. Key Words: Autograft-
Anterior cruciate ligament, reconstruction.

Autograft intraarticular reconstruction of the an-
terior c&iate ligament has become the gold stan-
dard with which other procedures are often com-
pared (1,2,3,4). Bone-patellar tendon-bone au-
tografts have been shown to have greater strength
than other commonly used biologic tissues (5). In
addition, bone plugs enhance eventual bone healing
within osseous tunnels (6,7). Concern, therefore,
has focused on the initial means of fixation, with
From the Sports Medicine Center of Metro Detroit (J.D.S.),
Troy, Michigan; the Southern California Center for Sports Med-
icine (D.W.J.), Long Beach, California; and Mt. Sinai Sports
Medicine Institute (B.T.C., R.D.P.) and Orthopaedic Research
(P.D.P., A.S.D.), Mt. Sinai Medical Center, Cleveland. Ohio.
U.S.A.
Address correspondence and reprint requests to Dr. Jeffrey D.
Shapiro, Sports Medicine Center of Metro Detroit, 755 W. Big
Beaver, Suite 1300, Troy, MI 48084, U.S.A.
cancellous screws providing superior interference
fit of bony anchors (8).
Jackson et al. have described a new technique of
harvesting circular bone plugs (9,lO). We hypothe-
sized that (a) circular bone plugs provide more se-
cure fixation within circular osseous tunnels than
other geometric configurations; (b) circular plugs
have decreased stress concentration effects on the
patella; and (c) circular grafts are easily obtained,
are reproducible, and require no graft manipulation
prior to insertion within bone tunnels.
In addition to providing a simpler, more repro-
ducible technique, circular bone plugs provide im-
mediate initial stability. Clinically, earlier and more
aggressive postoperative rehabilitation may be in-
stituted with greater confidence than with previ-
ously described autograft fixation techniques.

453
454 J. D. SHAPIRO ET AL.

MATERIALS AND METHODS

Six matched pairs of fresh frozen human knees

with a mean age of 73.5 years were utilized in the
study (Table 1). The cadavers were refrigerated at
4°C prior to autopsy.
All 12 specimens were harvested within 24 h of
death and frozen at - 20°C. They were thawed 24 h
before testing (11).
In six knees, a central one-third patellar bone-
patellar tendon-tibia1 tubercle autograft was ob-
tained with an 11 mm circular oscillating saw. The
central 11 mm of patellar tendon was sharply dis-
sected in all knees. The bone plugs were obtained in
FIG. 1. Once cortical penetration has been achieved, the circu-
continuity with the central portion of the patellar lar oscillating saw progresses perpendicular to the longitudinal
tendon. The saw was initially angled approximately axis of the tibia.
45” to penetrate the tibial and patellar cortical bone.
Since the excursion of the teeth of the circular saw bone plug, and a cannulated 9-mm-wide, 25mm-
was one fifth of the radius of the circle, it was nec- long, Kurosaka screw provided interference fit via
essary to rotate the oscillator back and forth to ef- the intercondylar notch. The screw was inserted
fectively cut bone. After cortical penetration was just past the bone-tendon interface. Care was taken
achieved, the saw was slowly redirected toward the to avoid injury to the tendinous portion of the graft
longitudinal axis of tibia or patella (Figs. 1 and 2). and prevent advancement of the bone plug (12).
Bone plugs were approximately 25 mm in length. In the opposite six knees, bone plugs were ob-
The remaining patella was then excised and the tained with a trapezoidal 11-mm graft template and
surrounding soft tissues were sharply removed. A a straight oscillating saw (Concept, Inc., Largo,
tibia1 aimer was utilized to direct accurate place- FL, U.S.A.). The bone plugs were then trimmed to
ment of a 12-mm tibia1 tunnel with a cannulated enable passage through a 12-mm-diameter drill
drill. A guide wire was reintroduced through the sleeve. The autografts were then inserted and fixed
tibia1 tunnel to the anatomic origin of the anterior identically to their matched pairs.
cruciate ligament on the distal femur. The guide The tibia1 bone plugs were fixed to large rods with
wire was advanced and followed by a 12-mm can- a template of methylmethacrylate. An intramedul-
nulated drill. The femur was then dissected from lary rod was then cemented into the femoral canal.
its surrounding soft-tissue envelope. A 0.062
Kirschner wire was utilized to create a hole in the
bone plug and a #5 Ethibond suture was passed for
graft manipulation. These 1l-mm bone plugs re-
quired no trimming to pass through a 12-mm-
diameter drill sleeve. The bone plug was then pulled
into the femoral tunnel until the bone-tendon inter-
face reached the entrance of the femoral tunnel. A
small guide wire was introduced adjacent to the

TABLE 1. Knee data

Specimen
no. Age Sex Cause of death
1 73 M Cardiac arrest
2 50 F Pulmonary embolism
3 97 F Cardiac arrest
4 88 F Cerebrovascular accident
5 54 F Renal-cell carcinoma
6 69 F Cardiac arrest
FIG. 2. Circular patellar bone plug with central patellar tendon.

Arthroscopy, Vol. 8, No. 4, 1992

 GEOMETRY OF ACL AUTOGRAFTS 455

FIG. 3. Test fixture secured at

45” of flexion.

FIG. 5. Four-point bending of triangular defect with calibration

measurements.

on an Instron machine. The patellae were mounted

on a vise with metal bars placed anteriorly, 12 mm
medial and lateral from the center of each bone
trough. Uniaxial impulse load to failure was then
The test fixture was secured on a material testing
applied at a rate of 8.5 mm/s to the opposite artic-
machine (M.T.S. Model 810.11) at 45” of knee flex-
ular surface (Fig. 6).
ion (Fig. 3). Moistened dressing sponges were uti-
Nondimensionalized paired data were than ana-
lized to prevent graft dehydration (13). Uniaxial
lyzed to determine statistical significance using a
load to failure was performed at a strain rate of
two-tailed t analysis.
100%/s. The ultimate load to failure was obtained
from force-elongation curves on oscillograph re-
cordings ( 14). RESULTS
An artificial bone composite was tested in four-
point bending to determine stress concentrations of Graft pull-out strength
different geometric defects. Circular, trapezoidal, Individual results for ultimate load to failure are
and triangular defects were created, measured, and listed in Table 2. Graft pull-out and graft-plus-screw
tested. Four samples of each geometric configura- pull-out were considered failures of fixation. These
tion were utilized. The samples were carefully mea-
sured for depth of cut and amount of remaining
material to standardize for calculations. Ultimate
load to failure was recorded and compared (Figs.
4 and 5).
The six matched pairs of enucleated patellae were
individually tested for three-point bending strength

FIG. 4. Trapezoidal, circular, and triangular defects in bone

composite with calibration markings. FIG. 6. Three-point patellar bending of circular defect.

Arthroscopy, Vol. 8. No. 4. 1992

456 J. D. SHAPIRO ET AL.

TABLE 2. Femoral pull-out strength: ultimate load found to have 24% greater strength than a trapezoi-
to failure dal defect (p < 0.03) (Fig. 7).
Specimen Mode of
no. Age Sex Trapezoidal Circular failure
DISCUSSION
: 73
50 M
F 703.6
845.8 2 8
-c- 972.3
885.4 +
k 8 Ligament
Pull-out Surgical reconstruction of the anterior cruciate
3 97 F 695.6 f 8 837.9 * 8 Pull-out
4 88 F 695.6 2 8 837.9 f 8 Ligament ligament is undergoing constant evolution. Noyes et
5 54 F 822.1 2 8 1,043.5 + 8 Li&rnent al. have shown that autograft reconstructions with
6 69 F 387.4 k 8 498.0 + 8 Pull-out bone-patellar tendon-bone specimens have signifi-
All values are given in Newtons. cantly increased maximum load to failure compared
with semitendinosus, gracilis, fascia lata, iliotibial
tract, and patellar retinacular tissues (5). Further-
occurred in specimens 2, 3, and 6. In specimens 1, more, bone-bone fixation encourages eventual
4, and 5, failure occurred at the midsubstance of the union of the graft attachment site to both the femur
patellar tendon graft. Both specimens in each pair and the tibia. Ultimately, patellar tendon grafts re-
failed at the same location, so ultimate load to fail- gain approximately 80% of their original strength by
ure of fixation was calculated from specimens 2, 3, a process of revascularization (1,6,7). Prior to in-
and 6. Circular bone plugs were found to provide corporation of bone plugs within their tunnels, fix-
19.9% more strength than their trapezoidal matches ation is of the utmost importance. Kurosaka et al.
(p < 0.006). compared different methods of fixation of various
autogenous tissues in anterior cruciate ligament re-
Three-point bending of patellae construction. Patellar bone-patellar tendon-tibia1
In three-point bending, all 12 patellae fractured specimens fixed with either a 6.5-mm A-O cancel-
within the circular or trapezoidal bone troughs. lous screw or a custom 9.0-mm screw had greater
Bending strengths were averaged and statistical sig- strengths with greater linear load to failure than sta-
nificance was calculated. Data for each specimen ples or sutures over buttons for fixation (8). This
are reported in Table 3. Circular defects were found fixation technique, as originally described by Lam-
to have 107% greater strength than trapezoidal de- bert (15), involves interference fit of a bone plug
fects (p < 0.05). with a 30-mm-long, 6.5-mm A-O cancellous screw
The difference in bending strength between bone inserted extraarticularly through the exit of the lo-
and artificial bone composite may be due to patellar mm femoral tunnel (8).
anatomy or the imprecision of cuts when creating Interference fit is dependent on the geometry of
trapezoidal defects. Both tests, however, demon- the bone plug in addition to methods of fixation.
strated statistical significance. Traditionally, bone plugs have been obtained with
an oscillating saw using straight fine blades yielding
Four-point bending of artificial bone composite triangular bone plugs (16,17). Clancy (18) has de-
Values recorded for uniaxial impulse load to fail-
ure were averaged, pooled, and compared for sta- Ultimate Stress To Failure
Newtons
tistical significance. A circular defect was found to / sq. mm In Bone Composite (4 pt. Bending)
have 53% greater strength than a triangular defect 4,000 1
(p < 0.002). Furthermore, a circular defect was

TABLE 3. Ultimate load to patella fracture

Specimen
no. Age Sex Trapezoidal Circular
1 73 M 3,759 + 22 3,759 + 44
2 50 F 578 2 22 2,135 + 22
3 97 F 1,201 * 11 1,646 + 22
4 88 ; 623 2 11 1,068 + 11
5 54 1,423 + 11 3,736 + 22 Triangular Trapazoidal Circular
6 69 F 712 + 44 1,423 k 44
FIG. 7. Ultimate stress to failure in bone composite (four-point
All values are given in Newtons/millimeter (2). bending).

Arthroscopy, Vol. 8, No. 4. 1992

 GEOMETRY OF ACL. AUTOGRAFTS
scribed a technique that produces a thin rectangular
bone plug measuring 10 mm wide and 4 mm deep
(8,12,19). Recently, templates have been developed
to produce trapezoidal grafts to enhance interfer-
ence fit and decrease the stress concentration in
remaining patella. However, these grafts all require
manipulation of the bone plugs to allow passage
through circular tunnels in the femur and tibia, and
this reshaping necessarily alters the biomechanical
strength and produces stress risers within the bone
plugs.
We have described a new technique of harvesting
bone plugs utilizing a circular oscillating saw that
produces circular bony anchors for fixation within
circular tunnels (9,lO). The technique is quick and
simple, reproducibly providing consistently shaped
and sized bone plugs. Compared with trapezoidal
bone plugs, circular bone plugs secured with inter-
ference fit fixation showed significantly higher max-
imum tensile strength to failure. We found that 1l-
mm circular bone plugs fit snugly into 12-mm bone
tunnels without the need for graft manipulation.
The interference fit obtained with circular grafts
and 9-mm screws is significantly stronger than the
values reported by Kurosaka et al. utilizing an iden-
tical strain rate (8).
Osteopenia associated with bone aging tends to
weaken the bony beds available for graft fixation
and subsequent pull-out studies. Kurosaka et al. (8)
reported a maximum tensile strength of 475.8 N for
fixation of bone plugs with a 9-mm Kurosaka screw.
The values were obtained from cadavers with a
mean age of 20.3 years. In our study of cadavers
with a mean age of 73.5 years, the average value of
femoral fixation failure was 846 N. No previous
studies have reported this strength of bony fixation.
Clinical judgment is necessary when selecting
femoral tunnel and bone plug diameters. Concerns
of femoral stress concentration effects and in-
creased postoperative pain secondary to large fem-
oral tunnels have led many surgeons toward smaller
diameter tunnels. In addition, the creation of larger
endosteal femoral tunnels increases the possibility
of posterior cortical perforation with anatomic tun-
nel placement. Although we have tested 11-mm
plugs within 12-mm tunnels, the comparison of dif-
ferent geometric configurations remains valid at any
given diameter. Sizes, therefore, remain at the dis-
cretion of the individual surgeon.
Improved technology of autogenous anterior cru-
ciate ligament reconstruction has progressively de-
creased postoperative morbidity and length of hos-
pitalization of patients. In an effort to eliminate a
lateral femoral incision and limit perforation of the
lateral distal femoral cortex, Rosenberg et al. (per-
sonal communication) described a technique that
allows “endosteal” fixation of bone plugs within
their femoral tunnels. In this procedure, interfer-
ence fit fixation is provided by introducing a screw
into femoral tunnels intraarticularly via the inter-
condylar notch. We found that this technique pro-
vided significantly greater pull-out strength com-
pared with previously reported results utilizing the
technique of Lambert (15). Furthermore, the initial
strength of fixation was found to exceed that of pa-
tellar tendon, often resulting in tears of the midsub-
stance of the reconstructed grafts. This finding con-
trasts with findings in previous studies, in which
graft failure occurred consistently at fixation sites
(8,19,20). Clinically, rigid graft fixation allows ag-
gressive early postoperative muscle strengthening
and range-of-motion exercises after anterior cruci-
ate ligament reconstruction. Continuous passive
motion machinery can thus be utilized with impu-
nity. Muscle atrophy and decreased range of motion
can be thus addressed and counteracted without
screw failure as reported by Burks et al. (19,21).
Intraoperative and postoperative patella fracture
is rarely reported in the literature (22). There is still
concern, however, about damage produced in the
patellofemoral joint by patellar bone-tendon-tibia
autograft reconstruction. Alterations in patellofem-
oral tracking, rupture of patellar tendon, patello-
femoral pain, and patellar fracture are all potential
problems. Although larger autografts have greater
strength, a larger patellar bone plug will inevitably
increase the rate of complications. In an artificial
bone composite, circular defects were found to re-
quire greater forces to fracture in four-point bend-
ing than in triangular and trapezoidal contigura-
tions. These findings were reproduced in fresh fro-
zen patella comparing circular and trapezoidal
defects in three-point bending. In obtaining triangu-
lar or trapezoidal bone plugs, saw penetration to
variable depths beyond the desired geometry is in-
evitable. In tests on an artificial bone composite,
these iatrogenic errors were minimized. In a clinical
setting, however, the patellar stress concentration
differences may be even greater than reported in
this study. With a circular oscillating saw, penetra-
tion beyond a desired depth is nearly impossible. In
addition, the simplicity and reproducibility of har-
vesting circular bone plugs enhances the biome-
chanical advantages previously described.
Arthroscopy. Vol. 8. No. 4, 1992
 J. D. SHAPIRO
In summary, we found that circular bone plugs
provided 19.9% stronger femoral interference fit
strength than trapezoidal bone plugs. We have also
shown that endosteal fixation of circular bone plugs
is extremely rigid compared to previously reported
values. In fact, the strength exceeded that of the
patellar tendon in 50% of specimens. This also con-
trasts with previous reports stating that all failures
occurred at fixation sites.
Circular bone defects decrease stress concentra-
tion effects on the patella by 107% compared with
trapezoidal configurations. This significantly less-
ens the possibility of intraoperative and postopera-
tive patella fracture after autogenous anterior cru-
ciate ligament reconstruction.
The technique of harvesting circular bone plugs
in autograft anterior cruciate ligament reconstruc-
tion has been utilized in over 500 clinical cases to
date. There have been no cases of patellar fracture
or clinical fixation failures. This technique is sim-
ple, reliable, and biomechanically proven to en-
hance clinical results of anterior cruciate ligament
reconstruction.
REFERENCES
1. Butler DL. Anterior cruciate ligament: its normal response
and replacement. J Ormop Res 1989;7:910-21.
2. Feagin J. The crucial ligaments. New York: Churchill Liv-
ingstone, 1988.
3. Jackson DW, Jennings LD. Arthroscopically assisted recon-
struction of the anterior cruciate ligament using a patella
tendon bone autograft. Clin Sports Med 1988;7:785-800.
4. Kochan A, Markolf KL, More RC. Anterior-posterior stiff-
ness and laxity of the knee after major ligament reconstruc-
tion. J Bone Joint Surg 1984;66A:1460-5.
5. Noyes FR, Butler DL, Grood ES, et al. Biomechanical anal-
ysis of human ligament grafts used in knee ligament repairs
and reconstructions. J Bone Joint Surg 1984;66A:344-52.
6. Amoczky SP, Tarvin GB, Marshall JL. Anterior cruciate
ligament replacement using patellar tendon. J Bone Joint
Surg 1982;64A:217-24.
Arthroscopy, Vol. 8, No. 4, 1992
ET AL.
7. Clancy WG, Narchania RG, Rosenberg TD, et al. Anterior
and posterior cruciate ligament reconstruction in Rhesus
monkeys. J Bone Joint Surg 1981;63A:1270-84.
8. Kurosaka M, Yoshiya S, Andrish JT. A biochemical com-
parison of different surgical fixation techniques of graft fix-
ation in anterior cruciate ligament reconstruction. Am J
Sports Med 1987;15:225-9.
9. Jackson DW, Cohn BT, Morrison DS. A new technique for
harvesting the patella tendon in patients undergoing anterior
cruciate ligament reconstruction. Orthopedics 1990;13:
165-7.
10. Jackson DW. Schaefer RK. A new technique for harvesting
patella tendon autografts for anterior cruciate ligament re-
construction. American Academy of Orthopaedic Surgeons
Videotape. Presented at the 57th Annual Meeting, New Or-
leans, LA, U.S.A., 1990.
11. Noyes FR, Grood ES. The strength of the anterior cruciate
ligament in humans and Rhesus monkeys. J Bone Join? Surg
1976;58A:1074-82.
12. Matthews LS, Soffer SR. Pitfalls in the use of interference
screws for anterior cruciate ligament reconstruction: brief
report. Arthroscopy 1989;5:225-6.
13. Noyes FR, DeLucas JL, Torvik PJ. Biomechanics of ante-
rior cruciate ligament failure: an analysis of strain-rate sen-
sitivity and mechanisms of failure in primates. J Bone Joint
Surg 1974;56A:236-53.
14. Kennedy JC, Hawkins RJ, Willis RB, et al. Tension studies
of human knee ligaments. J Bone Joint Surg 1976;58A:
350-5.
15. Lambert KL. Vascularized patellar tendon graft with rigid
internal fixation for anterior cruciate ligament insufficiency.
C/in Orthop 1983;172:85-9.
16. Jones K. Reconstruction of the anterior cruciate ligament. J
Bone Joint Surg 1963;45A:925-32.
17. Wang JB, Hewson GF. Anterior cruciate ligament recon-
struction using the lateral one third of the patellar tendon.
Tech Orthop 1988;2(4):23-7.
18. Clancy WG. Arthroscopic anterior cruciate ligament recon-
struction with patellar tendon. Tech Orthop 1988;2(4): 13-22.
19. Burks R, Daniel D, Losse G. The effect of continuous pas-
sive motion on anterior cruciate ligament reconstruction sta-
bilitv. Am J Soorts Med 1984:12:323-6.
20. Butler DL. Evaluation of fixation methods in cruciate liga-
ment replacement. Am Acad Orthop Surg Instruct Course
Lect 1987:36:173-8.
21. Bessette GC, Hunter RE. The anterior cruciate ligament.
Orthopedics 1990;13:55142.
22 McCarroll JR. Fracture of the patella during a golf swing
following reconstruction of the anterior cruciate ligament: A
case report. Am J Sports Med 1983;11(1):26-27.