Study Design
Twenty-four dogs were used for this study, which was approved by our
Institutional Animal Care and Use Committee. The dogs were evenly divided into
three groups depending on whether they were killed at one, three, or six
weeks. The peroneus longus tendons of the two hindlegs (extrasynovial tendons)
were grafted into the second and fifth digits of one forepaw in each dog. One
tendon was treated with cd-HA (described below) prior to grafting, while the
other was immersed in saline solution as a control. The digits were randomly
assigned to the treatment and control groups. On the day on which the animal
was killed, the function of the operatively treated digits was assessed by
measuring the work of flexion and frictional force of the grafted tendon.
Following the mechanical evaluation, the grafts were stained to evaluate the
residual surface hyaluronic acid concentration with use of biotinylated
hyaluronic acid-binding-protein staining.
Surgical Procedures and Tendon Modification
The dogs were anesthetized with intravenous ketamine and diazepam. The
selected forelimb and both hindlimbs were shaved, scrubbed with povidone
iodine, and sterilely draped. The peroneus longus tendons were taken from the
musculotendinous junction to the ankle level, a length of approximately 10 to
15 cm. One of the two peroneus longus tendons was randomly selected for the
treatment group and was immersed for thirty seconds in cd-HA solution,
consisting of 1% sodium hyaluronate (95%; Acros, Geel, Belgium), 10% gelatin
(from porcine skin; Sigma Chemical, St. Louis, Missouri), 0.25%
1-ethy1-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) (Sigma),
and 0.25% N-hydroxysuccinimide (NHS) (Sigma) in 0.1-M NaCl (pH 6.0) and 0.9%
phosphate-buffered saline
solution41. When
the hyaluronic acid is mixed with the EDC, the reaction quickly forms a
gel-like mixture. The tendon should be put into this mixture as the gel begins
to form in order to achieve strong binding to the tendon surface. To keep it
hydrated until grafting, the treated tendon was wrapped with a smooth rubber
sheet (a section of the elastic bandage used to exsanguinate the limb) and a
towel moistened with saline solution was wrapped around the rubber sheet. The
control tendon was immersed in 0.9% NaCl solution and similarly wrapped.
After the peroneus longus tendons were harvested and treated, a proximal
radial neurectomy was performed through a lateral humeral incision on the
selected recipient forelimb, in order to denervate the triceps muscle and
prevent elbow extension and thus weight-bearing. An elastic bandage was then
used to exsanguinate the forelimb and to act as a tourniquet for the distal
part of the procedure.
The flexor tendons were approached through a midlateral incision at the
level of the distal interphalangeal joint. The flexor digitorum profundus
tendon was sharply divided at a level 5 mm proximal to the tendon-bone
insertion site. The distal portion of the flexor digitorum profundus tendon
was connected to the proximal end of the tendon graft by a towing suture
approximately 20 cm in length. A second incision was then made at the middle
metacarpal level to expose the flexor digitorum profundus tendon proximal to
the flexor sheath. By means of traction on the proximal part of the flexor
digitorum profundus in the proximal incision, the flexor digitorum profundus
tendon was carefully delivered from the sheath, with the towing suture left
within the sheath. Then 50 mm of the flexor digitorum profundus tendon was
excised with a transverse cut. The peroneus longus tendon graft was cut
transversely to approximately 80 mm in length and was pulled through the
flexor sheath by the towing suture, again by means of traction from the
proximal incision. Approximately 10 mm of the distal part of the tendon graft
was left to overlap the distal stump of the flexor digitorum profundus tendon
(5 mm was left to insert into the osseous tunnel and 5 mm, to overlap the
flexor digitorum profundus tendon stump), while 20 mm of the proximal graft
was used to overlap the proximal end of the flexor digitorum profundus tendon,
to create the proximal graft junction tendon weave. In this way, the length of
the flexor digitorum profundus tendon that was removed (approximately 50 mm)
and the interpositional length of the graft were kept the same. A
3-mm-diameter drill was then used to make a tunnel in the distal phalanx. A
20-gauge needle was driven through this tunnel, emerging dorsally through the
middle third of the nail. The distal end of the peroneus longus tendon was
sutured with a Pennington loop of 3-0 nylon suture. The suture was passed
through the needle and the needle was withdrawn, bringing the suture out
through the nail. The end of the graft was fed into the canal. The suture was
tied over a button on the nail. For additional reinforcement, the juncture of
the overlapping graft and the stump of the flexor digitorum profundus tendon
was sutured with a 4-0 Ethibond (Ethicon, Somerville, New Jersey).
The proximal repair of the graft with the flexor digitorum profundus
recipient was done with an interlace suture technique. The flexor digitorum
profundus tendon was slit at its end, and the graft was threaded into the
slit. Four corners of the interface of two tendons were sutured with 4-0
Ethibond suture (Fig. 2). The
graft was then threaded transversely in a different plane of the slit with the
same technique of suture reinforcement. The fish-mouth thus created was closed
with 4-0 Ethibond suture to embrace the graft. Finally, the skin was closed
with 3-0 Vicryl suture (polyglactin; Ethicon, Somerville, New Jersey).
Postoperative Care and Rehabilitation
Following recovery from the anesthesia, the dogs were fitted with a
custom-made sling-jacket to support the operatively treated forelimb on the
front of the chest with the elbow and wrist in flexion. The wounds were
checked and cleaned beginning at three days postoperatively. Five days after
the surgery, rehabilitation commenced with use of a synergistic motion
protocol43
consisting of twenty repeated motions twice daily, seven days per week.
Following the therapy, the sling-jacket was repositioned in order to keep the
limb on the front of the chest. This regimen continued until the animal was
killed at the designated time.
Measurement of Digital Work of Flexion
After the animal was killed with an overdose of pentobarbital, both
forepaws were amputated and the second and fifth digits of both the surgically
treated and the non-surgically treated paws were harvested. The grafted
tendons were carefully exposed at the proximal junction level. The proximal
graft-tendon junction was dissected free from the surrounding tissues. The
proximal end of the tendon graft was sutured to a cable that was connected to
a load transducer. The tendon graft within zone II, including the distal
attachment, which was kept enclosed by skin, was all preserved. A Kirschner
wire was inserted longitudinally through the metacarpal bone to fix the
metacarpophalangeal joint in extension. Two small metal beads (approximately
0.5 mm in diameter) were inserted into the middle phalanx and a short
Kirschner wire was inserted into the distal part of the nail to serve as
markers for the phalanges.
The prepared digit was mounted on the testing device by fixing the proximal
Kirschner wire to a custom jig. The testing device consisted of a testing
frame, actuator, linear potentiometer, and one load transducer. A 0.5-N weight
was attached to the extensor tendon to ensure full extension of the digit as a
starting position and to apply an initial tension to the graft
(Fig. 3). A 0.1-N preload was
applied to the proximal part of the graft before testing. The actuator pulled
the tendon proximally at a rate of 2 mm/sec, causing digital flexion. Data
from the linear potentiometer and the proximal load transducer were recorded
at 20 Hz. During the testing, digital motion (from extension to flexion) was
recorded simultaneously by digital fluoroscopy. The video images were
digitized with use of Analyze Software (Biomedical Imaging Resource, Mayo
Clinic, Rochester, Minnesota). Then, the range of motion of the proximal and
distal interphalangeal joints was determined by calibrating the digitized data
of the metal markers (a metal pin in the proximal phalanx, two metal beads in
the middle phalanx, and a metal pin in the distal phalanx)
(Fig. 4). The measurement of
work of flexion continued until the distal interphalangeal joint angle reached
40° of flexion, on the basis of a previously published
study44.
Work-of-flexion data were calculated from the tendon displacement versus
loading curve during digital flexion and then were normalized (divided) by the
total proximal interphalangeal and distal interphalangeal joint motion angle
at the point where the distal interphalangeal joint reached
40°44.
Measurement of Tendon Frictional Force
Following the measurement of work of flexion, the tendon graft was further
dissected, with the proximal pulley kept intact. The gliding resistance
between the tendon graft and the proximal pulley was measured with use of a
custom tendonpulley friction testing device, as previously
described45,46.
Quantification of Hyaluronic Acid on the Tendon Graft Surface
After measurement of the frictional force, a 10-mm segment of the central
portion of the tendon graft was excised and was stained with biotinylated
hyaluronic acid-binding protein. For the purpose of comparison, four peroneus
longus tendons treated with cd-HA and four untreated peroneus longus tendons
were used for time-0 staining. For this analysis, the graft segment was
incubated in 1% hydrogen peroxide for five minutes after a wash in
phosphate-buffered saline solution. Then the tendon was blocked with 1% bovine
serum albumin in phosphate-buffered saline solution for five minutes and was
incubated in biotinylated hyaluronic acid-binding protein (Calbiochem; EMD
Biosciences, San Diego, California) in 1% bovine serum albumin in
phosphate-buffered saline solution overnight at 4°C. An ABC Vectastain kit
(Vector Laboratories, Burlingame, California) was prepared according to the
manufacturer's instructions, and the tendon graft was incubated with the ABC
Vectastain for thirty minutes at room temperature. After washing with
phosphate-buffered saline solution for five minutes, the sample was incubated
with diaminobenzidine (DAB) for five minutes and then washed in distilled
water three times. The paired tendons (cd-HA and saline-solution groups) from
each dog were placed together on a white background and were photographed with
an Olympus C-7070 digital camera (Olympus America, Melville, New York), with
use of a constant background, lighting, distance, and camera settings. The
color images were converted to gray-scale images with use of Adobe Photoshop
6.0 (Adobe Systems, San Jose, California). The intensity of the gray scale was
analyzed with use of Scion Image Software (Scion, Frederick, Maryland).
Scanning Electron Microscopy
In order to better observe the tendon surface, two tendons from each of the
cd-HA and saline-solution-treated groups were examined with scanning electron
microscopy. The tendons were fixed in 1% glutaraldehyde and 4% formaldehyde in
0.1-M phosphate-buffered saline solution and were dehydrated through a graded
series of ethanol solutions in a critical point dryer. The specimens were then
attached to stubs, sputter-coated with a gold-palladium mixture, and viewed in
a Hitachi 4700 scanning electron microscope (Hitachi Scientific Instruments,
Pleasanton, California) at 5 kV.
Statistical Methods
The data obtained from the measurements of work of flexion and frictional
force were analyzed with use of two-factor (time and treatment) repeated
analysis of variance, followed by the Tukey Studentized range (HSD [honestly
significant difference]) post hoc test to compare control, cd-HA, and normal
digits at three time points (one, three, and six weeks). For the comparison of
the hyaluronic acid concentrations at different time points, the intensity
ratio between the cd-HA and saline-solution groups at each time point was
evaluated with one-way analysis of variance. In all cases, a level of p <
0.05 was considered to be significant.
None of the grafted digits showed infection or rupture. The normalized work
of flexion of the grafted digits in both the cd-HA-treated and the control
group were significantly higher than that of the normal digits at one, three,
and six weeks (p < 0.05). The normalized work of flexion of the
cd-HA-treated grafts was significantly lower than that of the
saline-solution-treated grafts at all three time points (p < 0.05). There
was no significant difference among the findings at one, three, and six weeks
in the normal digits. However, the normalized work of flexion in the
cd-HA-treated and saline-solution-treated groups at three weeks and six weeks
was significantly increased compared with the normalized work of flexion at
one week in the respective groups (p < 0.05). There was no significant
difference in the normalized work of flexion between the three and six-week
grafts in either the cd-HA or the saline-solution-treated group
(Fig. 5).
The gliding resistance of the saline-solution-treated grafts was
significantly higher than that of the normal flexor digitorum profundus
tendons as well as the cd-HA-treated tendon grafts at three and six weeks (p
< 0.05). There was no significant difference in the gliding resistance
between the normal flexor digitorum profundus tendons and the cd-HA-treated
tendon grafts at any time point. The gliding resistance of the
saline-solution-treated tendon grafts at three and six weeks was significantly
higher than that at one week (p < 0.05). There was no significant
difference in the gliding resistance between the three and six week
saline-solution-treated grafts (Fig.
6).
The intensity of the hyaluronic acid staining of the cd-HA-treated tendons
was significantly higher than that of the saline-solution-treated tendons
group at time-0 and at one week (p < 0.05). However, there was no
significant difference between these two groups at three or six weeks (p >
0.05) (Fig. 7). The intensity
ratio between the cd-HA and saline-solution-treated tendons at time-0 was
significantly higher than that at three or six weeks (p < 0.05). There was
no significant difference between the time-0 and one-week tendons.
As observed during dissection, adhesions occurred mainly at the proximal
and distal repair sites in all grafts. The cd-HA-treated grafts seemed to have
less adhesion along the graft itself than did the saline-solution-treated
grafts. The tendon graft surface within the flexor sheath was smooth and
shining in the cd-HA group but not in the control group, in which the surface
was rougher because of the greater number of adhesions. It was difficult to
separate the tendon graft from the surrounding soft tissue, especially in the
case of the six-week specimens, in the control group
(Fig. 8). Evaluation with
scanning electron microscopy confirmed that cd-HA treatment resulted in a
smooth surface, whereas the control tendons presented a rough surface
(Fig. 9). However, we did not
evaluate adhesions quantitatively because these observations were made after
the work-of-flexion test, which broke some of the adhesions before they could
be dissected and examined.
Extrasynovial tendons, such as the palmaris
longus47, the
plantaris48, and
the long-toe
extensors49, are
the most commonly available tendons for flexor tendon grafting. However,
experimental studies have demonstrated that extrasynovial tendon grafts are
associated with more adhesions to the surrounding tissue than are
intrasynovial tendon
grafts22,50.
Alternatives intended to improve the results of tendon grafting, such as
staged tendon
reconstruction19,51,52,
use of free vascularized tendon
grafts53, or use of
tendon graft
substitutes54,
remain controversial. The ideal tendon graft would be one that allows a
single-stage reconstruction with a simple surgical procedure, that can be
obtained from a readily available autologous source, and that has the
performance of the native intrasynovial tendon. The cd-HA that was developed
in this study provides the possibility of achieving this goal, although
additional studies are clearly needed and the formulation must be further
optimized.
In the current study, the tendon grafts treated with cd-HA exhibited
decreased digital work of flexion compared with the untreated grafts. While
the reason for this improvement cannot be definitely ascribed to the cd-HA
treatment, these in vivo findings are consistent with observations of improved
gliding ability after cd-HA treatment in in vitro
studies41,55.
It is possible that the cd-HA also serves as a mechanical barrier that
inhibits adhesion formation in vivo, as an antiadhesive effect of hyaluronic
acid has been
described41,55.
However, treatment of the grafts with cd-HA did not achieve normal function at
any time postoperatively. Compared with the saline-solution-treated grafts,
the cd-HA-treated tendon grafts had little adhesion within the flexor sheath,
but adhesions remained prominent at the proximal and distal repair sites.
These are areas where additional measures to reduce adhesions and improve
motion might be fruitful. The tendon graft surface was improved after cd-HA
treatment, as indicated by decreased frictional force and by the appearance on
scanning electron microscopy. This was true even though there was no
difference in the intensity of the hyaluronic acid staining between the cd-HA
and control groups at three or six weeks. This finding suggests to us that the
first three weeks after tendon grafting is a critical time with regard to
adhesion formation. A similar observation has been made in other
studies45,56.
The modification of hyaluronic acid with use of the carbodiimide (EDC)
reaction to create new biopolymers, crosslinking with other biological
molecules such as collagen, has been recently and widely investigated for
applications such as anti-adhesion
membranes40,57
and biodegradable
scaffolds58,59,
for tissue regeneration, or as timed-release drug-delivery
vehicles60,61.
However, to our knowledge, we are the first to report the use of cd-HA by
direct bonding to the tendon surface to improve flexor tendon grafting in
vivo. As the half-life of hyaluronic acid in the tissue is normally less than
three days62, the
exogenous hyaluronic acid concentration on the tendon surface should decrease
to the normal level as a result of hyaluronic acid turnover after that point.
However, our data showed that the intensity of the hyaluronic acid staining in
the cd-HA group was significantly higher than that in the saline-solution
group at one week, which suggests that this chemical modification of
hyaluronic acid might prolong its resident time somewhat. It is not clear,
however, whether this prolongation is caused by an increased resistance to
enzymatic digestion or an increased binding strength of hyaluronic acid to the
tendon surface, with delayed physical removal by abrasion.
The principal strength of this study is that it demonstrates that a
lubricant, hyaluronic acid, can be bound to the surface of a tendon graft and
persist in vivo for at least one week. This may have important clinical
implications. However, the study had several limitations as well. First, the
tendon graft surgery was performed under ideal conditions, which included a
normal digit, a normal flexor digitorum profundus tendon, and a normal flexion
sheath, that were quite different from the typical clinical situation of
previous injury, a damaged flexor sheath, and, often, prior failed tendon
surgery. Second, a follow-up period of longer than six weeks would be needed
to observe the full evolution of the healing period and the resulting changes
in mechanical and biological behavior. Third, we did not use a second control
group consisting of tendons treated with hyaluronic acid and/or gelatin but
without the cd modification. This would have required more animals. We based
the decision not to include such a control group in part on previous clinical
and animal studies of the use of unmodified hyaluronic acid with tendon
surgery26,34,35
as well as published reports in which there was no difference in gliding
resistance, after repetitive motion in vitro, between tendons treated with
unmodified hyaluronic acid and untreated
tendons41,55.
Finally, while the hyaluronic acid-binding-protein technique has been well
studied and accepted as a sensitive and specific hyaluronic acid staining
method63-68,
it is only semiquantitative as a result of a variety of factors including, but
not limited to, sample preparation (frozen or paraffin), dilution of the
antibody, staining techniques (for example, washing duration and frequency),
varying photographic conditions and lighting, and selection of the analytic
area. Using the ratio between the cd-HA and saline-solution groups may have
eliminated some errors in photographic and color density analysis as paired
tendons (one treated with cd-HA and one treated with saline solution) from the
same dog were photographed and analyzed together. However, any comparison
across time points was of necessity less precise.
In conclusion, we demonstrated that an extrasynovial tendon graft modified
with carbodiimide-derivatized hyaluronic acid improved graft lubrication and
resulted in lower digital work of flexion, decreased tendon gliding
resistance, and reduced adhesions in a canine in vivo model. We believe that
these findings have important clinical implications and that the bonding of
lubricants, cytokines, and other bioactive compounds to the tendon surface may
improve the results of tendon surgery in the future. ?