Biomechanical Comparison of Anatomic

and Extra-Anatomic Reconstruction

Techniques Using Local Grafts for Chronic
Instability of the Acromioclavicular Joint
Malo Le Hanneur,*z§ MD, Andrew Thoreson,z MS, Damien Delgrande,k MD, Thibault Lafosse,§ MD,
Jean-David Werthel,k MD, Philippe Hardy,yk MD, PhD, and Bassem Elhassan,z MD
Investigation performed at the Biomechanics Laboratory, Department of Orthopedics,
Mayo Clinic, Rochester, Minnesota, USA

Background: Anatomic reconstruction techniques are increasingly used to address cases of acromioclavicular (AC) joint chronic
instability. These usually involve an additional surgical site for autograft harvesting or an allograft.
Purpose: To describe a triple-bundle (TB) anatomic reconstruction using on-site autografts, the semiconjoint tendon (SCT) and
the coracoacromial ligament (CAL), and compare its primary stability to the native AC joint ligamentous complex and to a modified
Weaver-Dunn (WD) reconstruction.
Study Design: Controlled laboratory study.
Methods: Intact AC joints of 12 paired cadaveric shoulders were tested for anterior, posterior, and superior translations under
cyclic loading with a servo-hydraulic testing system. One shoulder from each pair was randomly assigned to the TB group, where
2 SCT strips were used to reconstruct the coracoclavicular ligaments while the distal end of the CAL was transferred to the distal
extremity of the clavicle to reconstruct the AC ligaments; the other shoulder received a modified WD reconstruction. After recon-
struction, the same translational testing was performed, with an additional load-to-failure test in the superior direction.
Results: In both the TB and the WD groups, no significant differences were found before and after reconstruction in terms of joint
displacements after cyclic loading, in all 3 directions. Compared with the WD reconstruction, the TB repair resulted in significantly
lower displacements in both the anterior (ie, 2.59 6 1.08 mm, P = .011) and posterior (ie, 10.17 6 6.24 mm, P = .014) directions, but
not in the superior direction. No significant differences were observed between the 2 reconstructions during the load-to-failure test-
ing, except for the displacement to failure, which was significantly smaller (ie, 5.34 6 2.97 mm) in the WD group (P = .037).
Conclusion: Anterior, posterior, and superior displacements after an anatomic reconstruction of the AC joint complex using the
SCT and CAL as graft material were similar to those of native AC joints and significantly smaller in the axial plane than those of AC
joints after a WD repair.
Clinical Relevance: An anatomic reconstruction is achievable using the CAL and the SCT as on-site graft materials, providing
satisfactory initial stability and thereby allowing earlier mobilization.
Keywords: acromioclavicular joint; anatomic reconstruction; biceps; conjoint tendon; coracoacromial ligament

In addition to the dynamic stability ensured by the deltoid attributed to the conoid ligament medially, and in axial
and trapezius muscles, the static stability of the acromiocla- compression of the clavicle toward the acromion process,
vicular (AC) joint relies primarily on 3 ligamentous struc- attributed to the trapezoid ligament laterally.13
tures.11 Considered as one functional unit, the AC Despite these anatomic constraints, the AC joint is com-
ligaments represent the major restraint to large displace- monly subject to traumatic dislocations, due to its particu-
ments in the posterior direction. The coracoclavicular (CC) larly exposed location. Indeed, such separations represent
ligaments play a primary role in restraining large displace- 3.2% of all injuries involving the shoulder girdle and are
ments in the anterior and superior directions, which can be primarily caused by a direct impact to the shoulder, espe-
cially during contact sports.3 In high-grade injuries, both
AC and CC ligaments fail and surgical management is usu-
The American Journal of Sports Medicine ally recommended (ie, reduction and stabilization).32 Addi-
1–9 tionally, in cases of delayed diagnosis or failure of primary
DOI: 10.1177/0363546518770603
Ó 2018 The Author(s) joint stabilization, inflammatory changes in ligamentous

1
2 Le Hanneur et al The American Journal of Sports Medicine

tissues decrease their healing potential over time.7,8,23,25 39-92 years). Specimens were thawed overnight at room
Subsequently, after a preoperative delay ranging between temperature. The glenohumeral joint was disarticulated,
10 and 21 days, an additional biological graft is and all cutaneous, fatty, and muscular tissues surrounding
recommended.1 the AC joint and the CC interval were removed except for
Early AC ligamentoplasty techniques were performed the AC and CC ligaments, which were left intact, along
using local grafts, pedicled anteriorly onto the coracoid pro- with the conjoint tendon and the CAL. Before the native
cess and secured posteriorly onto the distal clavicle. In joint was tested, the specimens were visually ascertained
1942, Vargas40 reported on the use of the lateral part of the to be free of any previous alterations of the AC and CC lig-
semiconjoint tendon (SCT) corresponding to the short head aments, conjoint tendon and CAL, scapula, and clavicle.
of the biceps tendon; 30 years later, Weaver and Dunn42 pro- Before reconstruction, the AC joints were dissected and
posed an intramedullary fixation of the coracoacromial liga- inspected for degenerative disease. Specimens presenting
ment (CAL) into the shaft of the distal clavicle. However, any signs of such alterations were excluded from the study.
such reconstructions do not restore the complex anatomic Throughout all phases of preparation and testing, the
and biomechanical properties of the AC static stabilizers; specimens were kept moist by use of a 0.9% saline solution.
subsequently, inconsistent radioclinical outcomes have been The proximal quarter of the clavicle shaft and the inferior
reported.10,39,43 New anatomic techniques have been devel- third of the body of the scapula were trimmed with an
oped more recently, reconstructing the native ligamentous oscillating saw to facilitate further potting and mounting.
lines of action close to the original anatomic structure, and The scapula was potted in a 15-cm (length) 3 5-cm (width)
have been shown to provide better biomechanical and clinical 3 8-cm (height) custom block mold with urethane resin
results.26,28,38,41 However, tendon grafts currently used to (Smooth-Cast 65D; Smooth-On Inc), such that the glenoid
perform such reconstructions (ie, hamstrings, long toe exten- articular surface was perpendicular to the floor along
sor, wrist flexors) usually require additional surgical sites with its inferior-superior axis. The clavicle was potted in
that increase the graft’s morbidity.14,24 Use of an allograft a 5-cm (diameter) 3 6-cm (length) acrylic tube such that
may prevent such limitations. However, this material is its long axis was centered into the pipe. The potted scapula
expensive, may not be routinely available in some countries, was then mounted onto a vise that was secured to the base
and adds the potential risk of disease transmission.2,6 of the servo-hydraulic test frame (858 Mini Bionix II; MTS
The purpose of this study was to assess the biomechanical Systems), while the potted clavicle was secured to the sys-
properties (ie, translational displacements, ultimate load to tem’s linear actuator with a custom aluminum clamp. Both
failure, maximal displacement to failure, and stiffness) of an scapular and clavicular clamps allowed rotations of each
anatomic triple-bundle (TB) reconstruction performed with part of the specimen around 3 axes (ie, posterior-anterior,
the SCT and the CAL and to compare them with the native medial-lateral, and inferior-superior); the specimen was
AC ligamentous complex and with an extra-anatomic recon- successively placed in such a way to generate 3 clavicular
struction also using an on-site graft, that is, the Weaver- translation directions, assessed relative to the scapula,
Dunn (WD) procedure. Our hypothesis was that there would which included anterior, posterior, and superior orienta-
be no significant difference in the biomechanical parameters tions (Figure 1). To test the anterior laxity, the distal clav-
between the TB ligamentoplasty and the native joint and icle was positioned so that its superior surface was
that both would be superior to the WD reconstruction. perpendicular to the test machine base and its anterior
edge was parallel to the base; the inferior-superior axis of
the glenoid was positioned parallel to the base, with the
METHODS scapular plane forming an angle with the base ranging
from 30° to 45°, depending on the specimen’s anatomic fea-
Specimen Preparation tures; the linear actuator was then displaced upward to
load the joint. For posterior laxity testing, the specimen
Twelve paired, fresh-frozen cadaveric shoulders were position was the same as in the anterior testing, but the
obtained by our institutional anatomic bequest program actuator was displaced downward. To test the superior lax-
with Biospecimens Committee approval. The mean age of ity, the inferior-superior axis of the glenoid was positioned
the donors (4 men, 2 women) was 76.4 6 18.1 years (range, perpendicular to the test machine base, and the superior

*Address correspondence to Malo Le Hanneur, MD, Department of Orthopedics and Traumatology – Service of Hand, Upper Limb and Peripheral Nerve
Surgery, Georges-Pompidou European Hospital, 20 rue Leblanc, 75015 Paris, France (email: [email protected]).
y
Author deceased.
z
Biomechanics Laboratory, Department of Orthopedics, Mayo Clinic, Rochester, Minnesota, USA.
§
Department of Orthopedics and Traumatology – Service of Hand, Upper Limb and Peripheral Nerve Surgery, Georges-Pompidou European Hospital
(HEGP) – Assistance Publique – Hôpitaux de Paris (APHP), Paris, France.
k
Department of Orthopedics and Traumatology, Ambroise-Paré Hospital – Assistance Publique – Hôpitaux de Paris (APHP), Boulognes, France.
One or more of the authors has declared the following potential conflict of interest or source of funding: M.L.H has received an institutional grant of
e22,000 from the French orthopedic society (SOFCOT) as a financial support during his research fellowship at Mayo Clinic. The Arthrex products (ie, Fiber-
Wire No. 2 sutures, Bio-Tenodesis screws, TightRope devices) were donated by Arthrex Inc (Naples, Florida). The cost of the cadaveric specimens was
supported by the Mayo Clinic Orthopedics Department. Testing for this study was supported by the Mayo Clinic Materials and Structural Testing Core Lab-
oratory (Timothy Hewett, PhD). Data analysis for this study was supported by the Mayo Clinic Assistive and Restorative Technology Laboratory (Kristin
Zhao, PhD).
AJSM Vol. XX, No. X, XXXX AC Joint Triple-Bundle Repair Using Local Grafts 3

Figure 1. Intact potted left specimen set in the MTS servo-hydraulic testing system. (A) The specimen was set with the anterior
edge of the distal clavicle parallel to the floor so that the scapula formed a 30° angle with the MTS base; anterior displacement
was tested first by moving the actuator upward (a), and then posterior displacement was tested by moving the actuator down-
ward (b). (B) Superior displacement was evaluated after setting the specimen with the glenoid surface perpendicular to the base,
with the actuator moving upward.

surface of the distal clavicle was parallel to the base. The Surgical Reconstructions
relative positions of the acromion and the clavicle were
adjusted to reproduce the anatomic features through The TB group received an anatomic triple-bundle recon-
visual assessment and palpation; once anatomic position- struction, using the SCT and the CAL as grafts to replace
ing was achieved, the MTS machine was tared and both the CC and AC ligaments, respectively (Figure 2).
clamps were securely tightened. The lateral part of the conjoint tendon was used for the
An optical marker tracking system (Metria Innovation CC ligament reconstruction. The anterior and superficial
Inc) was used to assess 3D displacements during testing, layer of the SCT (ie, tendinous layer) was isolated from
to more directly assess changes in the AC joint space dur- the muscle fibers that were sharply dissected away, to
ing loading (ie, accuracy of 1/2500 of the measurement dis- obtain a 12-mm (width) 3 80-mm (length) tendinous
tance27; in this experiment, the camera was positioned band pedicled proximally onto the lateral part of the cora-
0.5 m away from the markers, yielding a 0.2-mm accuracy). coid tip.35 In a preliminary anatomic study, these dimen-
Each specimen received 2 markers, 1 marker on the clavi- sions were determined to be sufficient for the
cle side and 1 marker on the scapula side. The markers reconstruction.22 This lateral SCT band was then divided
were fixed to the resin blocks, and not directly to the bones, longitudinally into two 6 mm–wide strips, and each of
to avoid the inevitable weakening of the specimen that them was prepared with a No. 2 suture (FiberWire;
would be caused by the insertion of posts for mounting Arthrex) in a Krackow fashion. Two 4-mm tunnels were
the markers. This was deemed acceptable considering drilled in the distal clavicle at the insertion sites of the
that the bending of the bony structures was previously CC ligaments. Based on the anatomic study by Rios and
demonstrated to be negligible with loads less than 70 N.4 colleagues,31 the medial tunnel was drilled 35 mm medial
One digitizing camera along with built-in software mea- to the lateral edge of the clavicle, in the posterior half of
sured the 3D position and orientation of each marker, the clavicle; the lateral tunnel was drilled 25 mm medial
with automatic calculation of 3D displacements of one to the clavicle lateral edge, centered in the anteroposterior
marker over the other. thickness of the clavicle. An additional 5-mm tunnel was
Intact specimens were first tested to assess their phys- created, centered through the coracoid base. The 2 SCT
iological laxity in all 3 directions. The AC and CC liga- strips were then flipped below the coracoid process, shut-
ments were then excised, and one shoulder of each pair tled throughout its base, blocked with a 4-mm (diameter)
was randomly assigned to a reconstruction group (ie, TB 3 10-mm (length) tenodesis screw (Bio-Tenodesis Screw;
or WD), while the contralateral shoulder received the other Arthrex), and then separated with one strip inserted into
reconstruction. Specimens were tested again after recon- each clavicular tunnel. Regarding the AC bundle recon-
struction, with the same translational testing and an addi- struction, the distal 10 mm of the clavicle was resected.
tional load-to-failure testing in the superior direction. Two 1.6-mm holes were drilled into the superior cortex of
4 Le Hanneur et al The American Journal of Sports Medicine

Figure 3. Dissection of the left shoulder showing the triple-

bundle reconstruction of the acromioclavicular joint, with
two 6-mm strips of semiconjoint tendon (black arrows)
flipped inferiorly and posteriorly under the coracoid process
(a), passed through the coracoid base and exiting centered
on its superior surface (b), and finally passed through 2 cla-
vicular tunnels (c) corresponding to the insertion site area
of the conoid (medial) and trapezoid (lateral) coracoclavicular
ligaments. In addition, the coracoacromial ligament (yellow
arrow) was detached from the coracoid and transferred intra-
Figure 2. Drawings representing (A) anterior and (B) antero- medullarily into the distal clavicle shaft.
superior views of the anatomic triple-bundle reconstruction
of the right acromioclavicular joint ligamentous complex,
with 2 semiconjoint tendon strips, one medial (green) and sutures. Once the clavicle was placed in a reduced position
one lateral (blue), used to reconstruct the conoid and trape- (ie, superior surfaces of the distal clavicle and anterior
zoid coracoclavicular ligaments, respectively, while the cora- acromion brought together flush), a 2.4-mm pin was used
coacromial ligament (white) was used to reconstruct the to drill through the distal clavicle and the coracoid base
acromioclavicular ligaments. starting 20 mm from the edge of the distal clavicle excision
(ie, 30 mm from the native lateral edge of the clavicle). A
4-mm cannulated drill was then used to overdrill the guide
the clavicle through the medullary canal, 10 mm from the pin. A suture passer was used to pass the oblong button
distal end and with a 10 mm–wide bone bridge preserved through the clavicle and the coracoid, to be reoriented
between the 2 holes. The CAL was detached from the lat- and stabilized under the coracoid base while the round but-
eral aspect of the coracoid process and retained on the acro- ton was applied on the superior aspect of the distal clavicle.
mion. The CAL free end was then prepared with a No. 2 Both the clavicular and coracoid buttons’ orientations were
suture in a Krackow fashion and transferred to the verified before tensioning the device. Tightening of the
remaining distal clavicle, with the two free ends of the reconstruction started with the adjustable loop of the dou-
suture passed into the drill holes. The clavicle was then ble-button device and was completed with the sutures of
reduced to be flush with the acromion, with adequate ten- the WD, using the same landmarks as in the TB recon-
sion on the construct. Both SCT strips were secured into struction to obtain satisfactory tensioning (Figure 5).
their respective clavicular tunnels with 4 3 10–mm tenod-
esis screws, and the CAL suture ends were tied together
over the distal clavicle (Figure 3). Biomechanical Protocol
The WD group received a WD reconstruction with
Zooker’s modification (Figure 4).45 To evaluate the laxity of the native joints and of the recon-
The distal clavicle preparation was the same as previ- struction techniques, translational testing of all specimens
ously described for the TB group. Retained onto the cora- before and after reconstruction was performed in the ante-
coid tip, the CAL was detached from the acromion to be rior, posterior, and superior directions. Preconditioning was
transferred into the medulla of the remaining distal clavi- performed by cycling the AC joint between 0 and 25 N over
cle; a Krackow stitch was used to weave a No. 2 suture 10 cycles; the specimens were then loaded to an amplitude
through the CAL, the ends of which were passed through of 70 N over 1000 cycles at a frequency of 1 Hz. Displacement
the drill holes in the distal clavicle. A double-button device at peak force was documented at 1 and 1000 cycles. At the
(TightRope; Arthrex) was then used to augment the WD conclusion of cyclic testing, reconstructed joints were loaded
AJSM Vol. XX, No. X, XXXX AC Joint Triple-Bundle Repair Using Local Grafts 5

Figure 5. Dissection of the right shoulder showing a modified

Weaver-Dunn reconstruction augmented with the TightRope
device (black arrow), with the coracoacromial ligament
(yellow arrow) detached distally, reinforced with a No. 2
FiberWire suture, and transferred intramedullarily into the dis-
tal clavicle shaft.

RESULTS

Translational Testing
Figure 4. Drawings showing (A) anterior and (B) anterosupe-
rior views of the Weaver-Dunn reconstruction of the right After 1 loading cycle, significant differences were noted in
acromioclavicular joint ligamentous complex, with the mean anterior, posterior, and superior displacement (P =
TightRope augmentation. .0029, P = .0276, and P \ .001, respectively). After 1000
loading cycles, significant differences were noted in mean
to failure in the superior direction at a constant distraction anterior and posterior displacement but not superior dis-
rate of 1 mm/s to assess the maximal tensile loading capacity placement (P = .011, P = .024, and P = .065, respectively).
and the displacement to failure of each technique; the corre- Post hoc testing showed that mean joint displacement of
sponding stiffness was calculated from the slope of the linear both the TB and WD groups was not significantly different
region of the force-displacement curve. Failure was defined before and after the reconstructions in either the anterior,
as construct breakage with interruption of the linear progres- posterior, or superior directions after cyclic loading (Table 1).
sion of the slope of the force-displacement curve. Data The comparison between the 2 repair groups demon-
analysis was conducted with a custom MatLab algorithm strated significantly smaller displacements after cyclic
(MathWorks). loading in the anterior and posterior directions after the
TB reconstruction (P = .011 and P = .014, respectively);
no significant difference was observed in the superior
direction between the 2 constructs.
Statistical Analysis
An a priori power analysis showed that a sample size of 6 Load-to-Failure Testing
specimens in each group could detect differences in means
that were 1.8 SDs or larger, with 80% power and 95% con- TB reconstructions had significantly greater displacement
fidence. An analysis of variance was performed to compare to failure than WD reconstructions (P = .037, Table 2).
mean values of the intact and reconstructed groups; Tukey However, no significant difference was found in the ulti-
honestly significant difference test was applied if signifi- mate tensile load or stiffness between the 2 constructs.
cant differences were identified. Anterior, posterior, and The modes of failure of the TB reconstruction entailed rup-
superior displacement means were compared at 1 cycle of ture of the SCT strips in the CC interval in 4 specimens,
loading and at 1000 cycles of loading. Mean ultimate ten- pullout of the coracoid screw in 1 specimen, and coracoid
sile load, displacement to failure, and stiffness were com- fracture in 1 specimen. In the WD group, 5 specimens
pared after 1000 cycles of loading. Results were failed because the oblong button of the adjustable loop
presented as mean 6 SD. The level of significance was pulled through the coracoid, and 1 specimen failed due to
defined as P \ .05 for all tests. suture breakage of the loop; failures of the CAL transfer
6 Le Hanneur et al The American Journal of Sports Medicine

TABLE 1
Translational Testinga

TB Group WD Group TB vs WD

Directions of Displacement Intact TB P Value Intact WD P Value TB WD P Value

Anterior displacement, mm
Preloading 3.11 6 1.16 1.09 6 0.18 .004 2.08 6 0.68 2.87 6 1.09 .419 1.09 6 0.18 2.87 6 1.09 .010
Postloading 4.71 6 2.08 2.80 6 0.87 .076 3.65 6 0.77 5.39 6 0.94 .119 2.80 6 0.87 5.39 6 0.94 .011
Posterior displacement, mm
Preloading 8.30 6 4.80 3.14 6 2.44 .243 6.32 6 3.35 11.83 6 6.69 .196 3.14 6 2.44 11.83 6 6.69 .019
Postloading 9.14 6 4.65 3.84 6 2.50 .316 7.99 6 4.85 14.01 6 7.49 .218 3.84 6 2.50 14.01 6 7.49 .014
Superior displacement, mm
Preloading 2.60 6 0.54 0.80 6 0.30 .005 2.64 6 1.37 1.04 6 0.67 .014 0.80 6 0.30 1.04 6 0.67 .967
Postloading 3.84 6 1.42 2.65 6 2.34 .568 3.25 6 1.40 1.33 6 0.60 .178 2.65 6 2.34 1.33 6 0.60 .477

a
Mean postloading posterior displacements after triple-bundle (TB) anatomic repair were less than in the intact state or after modified
Weaver-Dunn (WD) repair; mean postloading superior displacements after WD reconstruction were less than in the intact state. All data
are reported as mean 6 SD. Displacement values were recorded after 1 cycle (ie, preloading) and after 1000 cycles (ie, postloading); compar-
isons were made before and after reconstruction for each specimen and then between reconstructions for each pair of shoulders.

TABLE 2
Load-to-Failure Testinga

TB WD P Value

Ultimate tensile load, N 472 6 123 516 6 371 .787

Displacement to failure, mm 10.47 6 4.6 5.13 6 2.32 .037
Stiffness, N/mm 85.05 6 23.56 144.71 6 86.02 .155

a
Mean displacement to failure was greater after triple-bundle (TB) reconstruction (which entailed triple-bundle anatomic repair, using the
semiconjoint tendon and the coracoacromial ligament) than after Weaver-Dunn (WD) reconstruction (which entailed modified WD repair
with TightRope augmentation). All data are presented as mean 6 SD.

occurred after the double-button device failure in all cases suture anchors have also been described.10,15,16,29 More
and were not considered since the load value was inferior recently, Zooker and colleagues45 reported on the biome-
to the coracoid failures or suture breakage. chanical properties of the double-button augmentation
method, which demonstrated greater stability over tape
cerclage in both the superoinferior and anteroposterior
DISCUSSION directions. Additionally, the investigators compared trans-
lations between intact and reconstructed specimens,
In this study, we described a reconstructive technique to demonstrating a better superior stability after WD recon-
address chronic cases of AC joint separations by using on- struction with double-button augmentation than in the
site grafts (ie, SCT and CAL) to reconstruct the CC and intact group, with a mean translation similar to our find-
the AC ligaments, respectively, and compared its biome- ings (ie, 2.1 6 0.1 mm of translation after 2000 cycles).
chanical properties with a modified WD technique aug- Zooker and colleagues45 also reported greater horizontal
mented with a double-button device. Translational testing stability in the intact group than after reconstruction,
demonstrated significantly less mean displacement in the with similar translation as well. They concluded that the
TB group in the anterior and posterior directions, but no sig- double-button augmentation failed to maintain normal
nificant difference was observed in the superior direction. anteroposterior translation after cyclic loading compared
Load-to-failure testing in the superior direction demon- with native AC joint ligamentous complex.45 By providing
strated greater displacement to failure in the TB group; a vertical link between the inferior surface of the clavicle
no significant differences were noted between the 2 con- and the superior aspect of the coracoid base, double-button
structs in terms of ultimate tensile load and stiffness. fixation may be the augmentation method that most accu-
To enhance the weak initial fixation that the CAL pro- rately reproduces the anatomic features of the native CC
vides in the WD reconstruction, numerous augmentation ligaments. However, since no repair of the AC complex is
techniques have been proposed to act as an additional attempted in the WD reconstruction, greater anteroposte-
and stronger link between the coracoid process and the dis- rior translations seem inevitable.13
tal clavicle during the biological healing of the CAL. The Motta et al30 highlighted the potential consequence of
most commonly reported method is a cerclage with suture an increased anteroposterior laxity of the TightRope device
or tape, but other augmentations such as screws and in acute cases. In 4 patients with joint hyperlaxity and
AJSM Vol. XX, No. X, XXXX AC Joint Triple-Bundle Repair Using Local Grafts 7

high-grade acute AC joint separations treated with a single closely the lines of action of the native CC ligaments,
double-button device, the authors described a nontrau- thus providing similar translational restraints to the distal
matic, painless, complete recurrence of the shoulder defor- clavicle as demonstrated herein. In a preliminary anatomic
mity without any signs of implant migration on study that we conducted in 12 paired cadaveric shoulders,
radiographs but with a trapezoid-shaped resorption of the SCT graft was found to be long enough in all specimens
the clavicular tunnel. The authors concluded that the to reconstruct both CC ligaments in such a fashion, with
lack of anteroposterior stability was responsible for a wind- a mean excess length of 39.9 6 5.7 mm (range, 32.2-
shield-wiper effect of the double-button sutures against the 47 mm) medially and 37.6 6 5 mm (range, 31-45.1 mm) lat-
sharp edges of the bone tunnels, which produced shear erally.22 These findings were confirmed during the present
abrasive forces accounting for these early failures. Facing study, since the reconstruction could be performed in all
similar complications with the use of a single device, Schei- cases. In addition, with the SCT restoring the CC complex,
bel et al33 switched to the double TightRope technique in the CAL may be used to reconstruct the AC ligamentous
acute cases, which seemed to prevent such shortcomings. complex, with its proximal end detached from the coracoid
In chronic cases, when double-button fixation was used tip to be transferred posteriorly to the distal clavicle. In the
to augment the CAL transfer, Boileau and colleagues6 did present TB technique, the resection of the distal end of the
not report any loss of reduction in their original study of clavicle is mandatory to perform the CAL transfer. Since
10 patients. One patient’s outcome was particularly note- Beitzel et al5 demonstrated that such resection led to
worthy, considering that a lateral migration of the coracoid increased horizontal translation, this may appear as a lim-
button was noted 6 months after surgery on follow-up radio- itation when compared with the TB reconstruction
graphs, which Boileau et al6 considered to be a coracoid described by Tauber and colleagues.38 However, no signif-
monocortical fracture. Since no loss of reduction was icant differences were found in the TB group between the
reported, this supports the hypothesis that satisfactory bio- intact state and after the reconstruction in the anterior
mechanical properties may be obtained after biological heal- and posterior directions (Table 1). This appears to confirm
ing of the CAL transfer. However, since the authors did not the conclusion from Beitzel et al,5 who stated that ‘‘if [a vio-
perform comparative dynamic imaging, the accuracy of lation of the AC capsule is] indicated in AC joint disloca-
their postoperative assessment of the AC joint’s horizontal tions, a reconstruction of the AC joint capsule should be
and vertical stability may be limited.36,37 More recently, considered.’’ In contrast to the original WD technique,
Kocaoglu and colleagues20 conducted a comparative study this new line of action of the transferred CAL seems to
involving the WD technique with a double-button augmen- achieve a satisfactory horizontal stabilization of the distal
tation only (ie, TightRope device; Arthrex) and a single- clavicle, as previously demonstrated by Shu et al.34 More-
bundle anatomic repair of the CC ligaments using palmaris over, as reported by Lafosse et al21 for the CAL and by
longus graft inserted in another double-button device (Graf- Kany et al18 for the SCT, both grafts may be harvested
tRope; Arthrex). Better clinical outcomes were reported arthroscopically to decrease the morbidity of the technique.
with the anatomic reconstruction, with a greater vertical However, Kany et al18 described a 3 cm–long graft harvest,
stability ascertained on anteroposterior comparative stress whereas a longer graft was needed in the present study (ie,
views; however, no axillary comparative stress views were 8 cm); such harvest may be technically challenging with
available, and thus it was not possible to compare the hori- the use of an arthroscope, and its feasibility remains to
zontal stability of the 2 techniques.37 be demonstrated.
First described by Vargas40 in 1942, the short head of Despite adequate biomechanical properties, 2 potential
the biceps is another local autograft that may be used to drawbacks of this technique need to be addressed. The contri-
stabilize chronic AC joint dislocation. In his original report, bution of the CAL to static glenohumeral stability has been
Vargas40 left the tendon proximally attached to the cora- well established, and sectioning it may alter kinematics of
coid tip, flipped it superiorly and posteriorly, and sutured the glenohumeral joint9; however, no functional consequen-
the tendon to itself after passing it through a single tunnel ces of such harvesting have been outlined, except in cases
drilled in the distal clavicle. Different modifications have with major alterations of the dynamic glenohumeral stabil-
hitherto been proposed, such as the intramedullary fixa- izers.12,44 Regarding the conjoint tendon, only the lateral
tion of the distal end of the tendon into the clavicle shaft, part needs to be harvested to obtain an adequate length to
as recommended by Jiang et al,17 or the arthroscopic reconstruct the CC ligaments without any significant dam-
approach described by Kany and colleagues.18 Kim et al19 age to the underlying neuromuscular structure. With a cau-
transferred both SCT and CAL distal ends to the conoid tious and sharp retrograde elevation of the superficial layer
and trapezoid tuberosities, respectively, advocating for an of the tendon, the biceps integrity is preserved since muscle
anatomic double-bundle reconstruction. fibers are retained on its deep layer; likewise, the musculocu-
With constructs similar to the WD reconstruction, none taneous nerve runs deep into the muscle belly, and such
of these previous techniques seem to accurately recon- a superficial dissection is thus safe. This recommendation
struct the anatomic lines of action of the native CC liga- has been confirmed clinically in several studies using this
ments; subsequently, unsatisfactory biomechanical graft, openly and arthroscopically; in fact, no author seems
outcomes could be expected.35 In fact, the CC ligaments to have encountered any substantial shortcoming such as
are tightened between the superior surface of the coracoid biceps impairment or musculocutaneous nerve injury.17-19,40
base and the inferior aspect of the distal clavicle.31 In our The present study and its findings should be considered in
study, the coracoid tunnel constrained the SCT to mimic light of its inherent limitations. This cadaveric model does
8 Le Hanneur et al The American Journal of Sports Medicine

not allow testing of the postoperative biological healing that Timothy Hewett, PhD, for their support in this study.
can be expected in patients, which subsequently limits the The authors dedicate this article to Pr Philippe Hardy,
relevance of our conclusions to the primary stability of the who passed away during the reviewing process, for his
reconstruction. Further, most of our specimens were issued active contribution to this work along with his continuing
from elderly donors with poor bone quality, which limited and relentless support and guidance despite his condition.
our experiments and their interpretability. For economic rea-
sons, bone density testing could not be conducted in our
specimens to accurately identify osteopenia, but experimen-
tal evidence of such condition was outlined. In a pilot study REFERENCES
of 2 pairs of shoulders, we encountered 3 coracoid fractures
during load-to-failure testing of the native ligaments; fur- 1. Barth J, Duparc F, Andrieu K, et al; French Society of Arthroscopy. Is
coracoclavicular stabilisation alone sufficient for the endoscopic
thermore, the coracoid-related failure rate was unusually
treatment of severe acromioclavicular joint dislocation (Rockwood
high compared with similar studies, especially in the WD types III, IV, and V)? Orthop Traumatol Surg Res. 2015;101
group.4,26,45 The protocol was consequently modified and (suppl 8):S297-S303.
the load to failure was performed only after reconstruction, 2. Barth J, Duparc F, Baverel L, et al. Prognostic factors to succeed in
which precluded any comparison with the ultimate tensile surgical treatment of chronic acromioclavicular dislocations. Orthop
load of the native AC joints. In addition, optical markers Traumatol Surg Res. 2015;101(suppl 8):S305-S311.
were not positioned directly on the bony structures but 3. Beitzel K, Cote MP, Apostolakos J, et al. Current concepts in the
treatment of acromioclavicular joint dislocations. Arthroscopy.
were placed on the resin blocks to avoid any additional weak-
2013;29(2):387-397.
ening of the specimens. Since the markers were positioned in 4. Beitzel K, Obopilwe E, Chowaniec DM, et al. Biomechanical compar-
similar fashion in all of our specimens, we considered that ison of arthroscopic repairs for acromioclavicular joint instability:
alterations of translational measurements at the AC joint suture button systems without biological augmentation. Am J Sports
space, if any, were similar in all specimens. Moreover, based Med. 2011;39(10):2218-2225.
on the report from Beitzel et al,4 measurement variations due 5. Beitzel K, Sablan N, Chowaniec DM, et al. Sequential resection of the
distal clavicle and its effects on horizontal acromioclavicular joint
to bending of bony structure may be considered negligible.
translation. Am J Sports Med. 2012;40(3):681-685.
The small sample size is another shortcoming of the present 6. Boileau P, Old J, Gastaud O, Brassart N, Roussane Y. All-arthroscopic
study. Comparisons of means for some parameters (ie, joint Weaver-Dunn-Chuinard procedure with double-button fixation for
displacements before and after reconstructions, ultimate ten- chronic acromioclavicular joint dislocation. Arthroscopy. 2010;26(2):
sile loads and subfailure stiffness of each construct) may be 149-160.
subject to statistical type II errors, putting their validity 7. Chamberlain CS, Crowley EM, Kobayashi H, Eliceiri KW, Vanderby R.
Quantification of collagen organization and extracellular matrix factors
into question. However, such a small sample was sufficient
within the healing ligament. Microsc Microanal. 2011;17(5):779-787.
to demonstrate that the TB reconstruction can provide ade- 8. Chamberlain CS, Crowley E, Vanderby R. The spatio-temporal
quate horizontal and vertical stability to the AC joint and dynamics of ligament healing. Wound Repair Regen. 2009;17(2):
that this construct was superior to the WD technique in 206-215.
terms of horizontal stability. Finally, translational testing 9. Curl LA, Warren RF. Glenohumeral joint stability: selective cutting
and load-to-failure testing as conducted herein are broadly studies on the static capsular restraints. Clin Orthop Relat Res.
used protocols, allowing for clarity and comparability of out- 1996;330:54-65.
10. Deshmukh AV, Wilson DR, Zilberfarb JL, Perlmutter GS. Stability of
comes. However, considering that the AC joint contributes to
acromioclavicular joint reconstruction: biomechanical testing of vari-
multiple shoulder motions and is thus subjected to forces in ous surgical techniques in a cadaveric model. Am J Sports Med.
variable directions, such testing hardly reflects the in vivo 2004;32(6):1492-1498.
conditions of this joint. 11. Epstein D, Day M, Rokito A. Current concepts in the surgical man-
agement of acromioclavicular joint injuries. Bull NYU Hosp Jt Dis.
2012;70(1):11-24.
12. Field LD, Dines DM, Zabinski SJ, Warren RF. Hemiarthroplasty of the
CONCLUSION shoulder for rotator cuff arthropathy. J Shoulder Elbow Surg. 1997;
6(1):18-23.
A triple-bundle anatomic reconstruction of the AC joint is fea- 13. Fukuda K, Craig EV, An KN, Cofield RH, Chao EY. Biomechanical study
of the ligamentous system of the acromioclavicular joint. J Bone Joint
sible by combining the SCT and the CAL to reconstruct the
Surg Am. 1986;68(3):434-440.
CC and AC ligaments, respectively. When compared with 14. Grutter PW, Petersen SA. Anatomical acromioclavicular ligament
native AC joints, the AC joints with TB repair demonstrated reconstruction: a biomechanical comparison of reconstructive tech-
satisfactory stability in all directions (ie, anterior, posterior, niques of the acromioclavicular joint. Am J Sports Med. 2005;33(11):
and superior). Furthermore, smaller horizontal translations 1723-1728.
were observed in comparison with the WD reconstruction 15. Harris RI, Wallace AL, Harper GD, Goldberg JA, Sonnabend DH,
Walsh WR. Structural properties of the intact and the reconstructed
augmented with a double-button device.
coracoclavicular ligament complex. Am J Sports Med. 2000;
28(1):103-108.
16. Jari R, Costic RS, Rodosky MW, Debski RE. Biomechanical function
of surgical procedures for acromioclavicular joint dislocations.
ACKNOWLEDGMENT Arthroscopy. 2004;20(3):237-245.
17. Jiang C, Wang M, Rong G. Proximally based conjoined tendon trans-
The authors thank Jim Postier for the drawings of the fer for coracoclavicular reconstruction in the treatment of acromio-
techniques. The authors thank Kristin Zhao, PhD, and clavicular dislocation. J Bone Joint Surg Am. 2007;89(11):2408-2412.
AJSM Vol. XX, No. X, XXXX AC Joint Triple-Bundle Repair Using Local Grafts 9

18. Kany J, Guinand R, Croutzet P. All arthroscopic augmented Vargas 31. Rios CG, Arciero RA, Mazzocca AD. Anatomy of the clavicle and cor-
procedure: an option after failed acromioclavicular joint dislocation acoid process for reconstruction of the coracoclavicular ligaments.
reconstruction. A technical note. Orthop Traumatol Surg Res. Am J Sports Med. 2007;35(5):811-817.
2016;102(5):669-672. 32. Rockwood CA. Injuries to the acromioclavicular Joint. In: Rockwood
19. Kim SH, Lee YH, Shin SH, Lee YH, Baek GH. Outcome of conjoined CA, Green DP, eds. Fractures in Adults. Vol 1, 2nd ed. Philadelphia,
tendon and coracoacromial ligament transfer for the treatment of PA: JB Lippincott; 1984:860-910.
chronic type V acromioclavicular joint separation. Injury. 2012;43(2): 33. Scheibel M, Droschel S, Gerhardt C, Kraus N. Arthroscopically assis-
213-218. ted stabilization of acute high-grade acromioclavicular joint separa-
20. Kocaoglu B, Ulku TK, Gereli A, Karahan M, Turkmen M. Palmaris lon- tions. Am J Sports Med. 2011;39(7):1507-1516.
gus tendon graft versus modified Weaver-Dunn procedure via 34. Shu B, Johnston T, Lindsey DF, McAdams TR. Biomechanical eval-
dynamic button system for acromioclavicular joint reconstruction in uation of a novel reverse coracoacromial ligament reconstruction
chronic cases. J Shoulder Elbow Surg. 2017;26(9):1546-1552. for acromioclavicular joint separation. Am J Sports Med. 2012;
21. Lafosse L, Baier GP, Leuzinger J. Arthroscopic treatment of acute 40(2):440-446.
and chronic acromioclavicular joint dislocation. Arthroscopy. 2005; 35. Sloan SM, Budoff JE, Hipp JA, Nguyen L. Coracoclavicular ligament
21(8):1017. reconstruction using the lateral half of the conjoined tendon. J Shoul-
22. Le Hanneur M, Delgrande D, Lafosse T, Werthel JD, Hardy P, Elhas- der Elbow Surg. 2004;13(2):186-190.
san B. Triple-bundle anatomical reconstruction using the coracoa- 36. Tauber M, Gordon K, Koller H, Fox M, Resch H. Semitendinosus ten-
cromial ligament and the short head of the biceps tendon to don graft versus a modified Weaver-Dunn procedure for acromiocla-
stabilize chronic acromioclavicular joint dislocations: a cadaver feasi- vicular joint reconstruction in chronic cases: a prospective
bility study. Orthop Traumatol Surg Res. 2018;104(1):27-32. comparative study. Am J Sports Med. 2009;37(1):181-190.
23. Lee DH, Ng J, Chung JW, Sonn CH, Lee KM, Han SB. Impact of 37. Tauber M, Koller H, Hitzl W, Resch H. Dynamic radiologic evaluation
chronicity of injury on the proportion of mesenchymal stromal cells of horizontal instability in acute acromioclavicular joint dislocations.
derived from anterior cruciate ligaments. Cytotherapy. 2014;16(5): Am J Sports Med. 2010;38(6):1188-1195.
586-598. 38. Tauber M, Valler D, Lichtenberg S, Magosch P, Moroder P, Haber-
24. Lee SJ, Nicholas SJ, Akizuki KH, McHugh MP, Kremenic IJ, Ben-Avi meyer P. Arthroscopic stabilization of chronic acromioclavicular joint
S. Reconstruction of the coracoclavicular ligaments with tendon dislocations: triple- versus single-bundle reconstruction. Am J Sports
grafts: a comparative biomechanical study. Am J Sports Med. 2003; Med. 2016;44(2):482-489.
31(5):648-655. 39. Tienen TG, Oyen JF, Eggen PJ. A modified technique of reconstruc-
25. Marsolais D, Côté CH, Frenette J. Neutrophils and macrophages tion for complete acromioclavicular dislocation: a prospective study.
accumulate sequentially following Achilles tendon injury. J Orthop Am J Sports Med. 2003;31(5):655-659.
Res. 2001;19(6):1203-1209. 40. Vargas L. Repair of complete acromioclavicular dislocation, utilizing
26. Mazzocca AD, Santangelo SA, Johnson ST, Rios CG, Dumonski ML, the short head of the biceps. J Bone Joint Surg Am. 1942;
Arciero RA. A biomechanical evaluation of an anatomical coracocla- 24(4):772-773.
vicular ligament reconstruction. Am J Sports Med. 2006;34(2):236- 41. Walz L, Salzmann GM, Fabbro T, Eichhorn S, Imhoff AB. The ana-
246. tomic reconstruction of acromioclavicular joint dislocations using 2
27. Metria Innovation Incorporation. MPT—Series 2. Metria Innovation Inc TightRope devices: a biomechanical study. Am J Sports Med. 2008;
web site. http://www.metriainnovation.com/mpt-series2. Accessed 36(12):2398-2406.
November 1, 2017. 42. Weaver JK, Dunn HK. Treatment of acromioclavicular injuries, espe-
28. Michlitsch MG, Adamson GJ, Pink M, Estess A, Shankwiler JA, Lee cially complete acromioclavicular separation. J Bone Joint Surg Am.
TQ. Biomechanical comparison of a modified Weaver-Dunn and 1972;54(6):1187-1194.
a free-tissue graft reconstruction of the acromioclavicular joint com- 43. Weinstein DM, McCann PD, McIlyeen SJ, Flatow EL, Bigliani LU. Sur-
plex. Am J Sports Med. 2010;38(6):1196-1203. gical treatment of complete acromioclavicular dislocations. Am J
29. Motamedi AR, Blevins FT, Willis MC, McNally TP, Shahinpoor M. Bio- Sports Med. 1995;23(3):324-331.
mechanics of the coracoclavicular ligament complex and augmenta- 44. Wiley AM. Superior humeral dislocation: a complication following
tions used in its repair and reconstruction. Am J Sports Med. decompression and debridement for rotator cuff tears. Clin Orthop
2000;28(3):380-384. Relat Res. 1991;263:135-141.
30. Motta P, Maderni A, Bruno L, Mariotti U. Suture rupture in acromio- 45. Zooker CC, Parks BG, White KL, Hinton RY. TightRope versus fiber
clavicular joint dislocations treated with flip buttons. Arthroscopy. mesh tape augmentation of acromioclavicular joint reconstruction:
2011;27(2):294-298. a biomechanical study. Am J Sports Med. 2010;38(6):1204-1208.

For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.