It is generally accepted that a partial laceration of the
flexor tendon in a digit is best left unrepaired if the laceration involves £50% of
the tendon whereas a laceration involving 80% to 90% is
better treated with tendon repair and rehabilitation, as would be
done for a complete laceration1-5.
There is considerable controversy regarding this issue, however.
In a survey of hand surgeons, 30% of the respondents repaired
all partial lacerations whereas 45% repaired all lacerations
involving 50% of the tendon6.
The use of early mobilization techniques after repairs of complete
tendon lacerations has improved outcomes7-11,
and this clinical experience has been confirmed and extended in
a variety of animal models12-14.
However, the data regarding partial lacerations are far less abundant,
and rehabilitation after these injuries remains controversial. Horii
et al.15 showed that traditional
passive mobilization of flexor tendon repairs, with the wrist immobilized
in flexion, does not result in full tendon excursion because of
buckling within the tendon sheath. Active flexion programs can potentially
reduce the problem of buckling, but there is an increased risk of
tendon rupture16-18. Mobilization
with synergistic wrist and finger motion (wrist flexion with finger
extension and wrist extension with finger flexion) is a combination
therapy of active and passive motion15,19,20.
The motion force applied to the proximal portion of the flexor digitorum
profundus tendon is generated by extension of the wrist.
Rehabilitation employing wrist motion increases tendon excursion
compared with that during traditional passive motion rehabilitation
with the wrist flexed15,21. The
results of this therapy in vivo have only recently
been investigated, in a single report on complete lacerations in
a canine model22. In that study,
there were no differences in function among normal digits,
digits operated on and then treated with synergistic motion therapy,
and digits operated on and then treated with traditional passive
motion therapy. The group treated with traditional passive motion
therapy did have more adhesion formation than did the group treated
with synergistic motion. The results of this study may have been
affected by a high complication rate, which included tendon rupture
in 25% of the dogs and gap formation in 59% of
the tendons.
We believed that it was possible to isolate and study the effect of
a rehabilitation method while reducing the failure rate due to tendon
rupture by using a partial laceration model. In addition, such a
model allowed us to collect data about a challenging but seldom studied
injury, partial tendon laceration. The purpose of this study, therefore,
was to assess the effect of rehabilitation regimens involving either
synergistic motion or wrist fixation on adhesion formation after
repair of a partial tendon injury.
One hundred and thirty-two tendons from the second and fifth digits
of sixty-six mongrel dogs were used for this study. The dogs were
randomly divided into two postoperative therapy groups. The first
group was managed postoperatively with passive flexion and extension
of the digits with the wrist fixed in 45° of flexion (wrist fixation
group), and the second group was managed with synergistic motion
that combined passive digital flexion with passive wrist extension
and passive digital extension with passive wrist flexion (synergistic
motion group). The thirty-three dogs (sixty-six tendons) in each group
were further divided into three subgroups that were killed at one,
three, or six weeks after tendon repair. Thus, there were twenty-two
tendons from eleven dogs in each group. This study was approved
by our Institutional Animal Care and Use Committee.
Surgical Procedure and Postoperative Therapies
The dogs were anesthetized with pentobarbital. One randomly selected
forepaw was shaved, scrubbed with povidone-iodine, and sterilely
draped. A radial neurectomy proximal to the triceps innervation
was first performed to prevent postoperative weight-bearing on the
operatively treated limb. An elastic bandage was used to exsanguinate
the forelimb and act as a tourniquet for the procedure. The second
and fifth flexor digitorum profundus tendons were approached through
a midlateral incision in the paw between the proximal and distal annular
pulleys. According to the method of Dobyns et al.23,
the flexor digitorum profundus tendon was lacerated to 80% of
its transverse section at the level of the proximal interphalangeal
joint. At this level, there are two longitudinal collagen bundles,
with an intervening fibrous raphe. The 80% laceration involves
cutting all of one bundle, half of the other, and all of the raphe.
The method is reliable and reproducible23.
The tendons were subsequently repaired with a modified Kessler suture
of 5-0 Ticron suture (Davis and Geck, Wayne, New Jersey) with a
circumferential epitenon simple running suture of 6-0 nylon. After
tendon repair, the paw was closed in layers without closure of the
flexor sheath.
For the dogs in the wrist fixation group, a 3-mm threaded Kirschner wire
was used to fix the wrist in 45° of flexion, whereas the dogs in
the synergistic motion group were allowed free wrist movement. In
both groups, a dorsal aluminum splint was applied to the paw and
forearm to maintain 45° of wrist flexion and neutral position of
the digital joints between therapy sessions.
Rehabilitation began on the third postoperative day in each group.
In the wrist fixation group, passive motion of the operatively treated
digits from full flexion to extension was performed for ten repetitions
twice daily. In the synergistic motion group, the rehabilitation
protocol described above was also performed for ten repetitions
twice daily.
Assessment of Adhesion and Gap FormationGross
Evaluation
One of the two tendons that were operated on in each paw was randomly
assigned for adhesion grading. The adhesion formation was grossly
evaluated and graded by consensus by three of us (C.Z., T.M., and
P.C.), all orthopaedic surgeons. The flexor sheath was opened through
an area not included in the operation, away from the suture site,
with care taken to avoid interfering with any adhesions between
the tendon and sheath. With the use of loupe magnification, the
tendons were graded for adhesion formation at two sites: (1) between
the tendon and the flexor sheath, including the pulley and the synovial membrane,
and (2) between the tendon and the tendon bed, including the flexor
digitorum superficialis tendon and the surrounding soft tissues
of the phalanx. The rating scale at each site ranged from 0 (no
adhesion) to 4 (very severe) (Table I). Thus, the total of the scores
at the two sites ranged from 0 to 8. Any gap between the tendon
ends was measured with calipers.
Adhesion Breaking Strength
Adhesion breaking strength was measured in the other tendon of
each operatively treated paw. The digit was transected at the metacarpophalangeal
joint, leaving the digit intact. The flexor digitorum profundus
tendon was cut at the wrist level with the digit in neutral position,
and the flexor digitorum superficialis tendon was cut more distally,
so as not to interfere with the profundus stump. The distal end
of the flexor digitorum profundus tendon was detached from the base
of the distal phalanx. The proximal and distal vincula were carefully
transected without interfering with the repair site or any adhesions
that were present.
The specimen was mounted in the testing device (Fig. 1), which consists
of a load cell and a potentiometer attached to each end of the tendon.
The proximal part of the tendon was connected to an actuator while
a 0.3-N load was attached to the distal end of the tendon. This
load was sufficient to maintain tendon tension without moving the
tendon distally (the average gliding resistance of a tendon repaired
with the modified Kessler suture technique is >0.4 N in
vitro24). The actuator
was positioned at the preselected angle a, defined as the angle
in degrees formed between the horizontal plane and the proximal
cable extension. The pulley for the load was positioned at angle
b, defined as the angle in degrees formed between the horizontal
plane and the distal cable extension. On the basis of our experience
in previous studies25,26, the
arc of contact was set at 50° (α = 30° and β = 20°).
All specimens were kept moist throughout testing by immersion in
a saline solution bath, which was incorporated into the testing
jig. The actuator pulled the tendon proximally at a rate of 2.0
mm/sec. The excursion limit was set at 30 mm, which
was sufficient to rupture any adhesion that was present. The actuator
movement was then reversed. The tendon was pulled distally by the
weight attached to its distal end until the initial point of the
first test. Then, the test was repeated.
A displacement differential measured by the proximal and distal
potentiometers indicated the presence of an adhesion—i.e.,
if the potentiometer nearest to the actuator registered movement
while the other did not, this indicated that an intervening adhesion
was preventing tendon gliding. The adhesion strength was calculated
as the difference in maximum force between the first and second
tests (Fig. 2).
Following breakage of the adhesion, a second test confirmed synchronous
recording of movement at the two potentiometers and thus the absence
of additional intact adhesions.
Tendon Strength
Following the gross observation, the tendons were dissected from
the digits and were fixed in an MTS servohydraulic testing machine
(MTS Systems, Minneapolis, Minnesota) with use of clamps with interdigitating
grooves. The tendon gauge length was approximately 30 mm. Under
displacement control, the tendon was distracted at a rate of 20
mm/min until it ruptured completely. Tensile force and
displacement data were collected at a rate of 20 Hz. Throughout
testing, the tendons were kept moist by spraying with physiologic
saline solution.
Statistical Methods
The distribution of the values for adhesion grade was non-gaussian,
and thirty of the sixty values were zero. Therefore, exact
Wilcoxon tests were performed for the pairwise comparisons among
the three time-periods (one versus three weeks, one versus six weeks,
and three versus six weeks). This was done separately for the two
mobilization groups. Exact Wilcoxon tests were also used to assess
differences between the synergistic motion and fixation groups at
each of the three time-points.
The distribution of the values for adhesion breaking strength was
also non-gaussian. However, since the values were roughly continuous,
a two-factor analysis of variance was performed on the ranks of
the breaking strength values. The two factors of this model were
the group (synergistic motion or wrist fixation) and the week of
measurement (one, three, or six weeks). The model consisted of the
two main effects along with the interaction term. Contrasts for
the nine a priori comparisons of interest were
assessed.
Five tendons were excluded from the testing of the adhesion breaking
strength because they ruptured; one ruptured at three weeks and
one ruptured at six weeks in the synergistic motion group, and one
ruptured at one week and two ruptured at six weeks in the wrist
fixation group. Data for one tendon (a one-week specimen in the
wrist fixation group) were inadvertently not recorded because of
a technical error in the data collection. Gap formation was measured
only during the gross evaluation, as the gap might have been affected
by the adhesion breakage test. There was a 1 to 5-mm gap in ten
(30%) of the thirty-three tendons (four at one week, two
at three weeks, and four at six weeks) in the synergistic motion
group and in two (6%) of the thirty-three tendons in the
wrist fixation group (one at three weeks and one at six weeks).
There was no relationship between gap formation and adhesion
formation.
The median adhesion scores determined by gross observation in
the synergistic motion group were 0 (range, 0 to 2), 0 (range, 0
to 3), and 0 (range, 0 to 3) at one, three, and six weeks, respectively,
and there was no significant difference among the scores at the
different time-periods (p > 0.05). In the wrist fixation
group, the median adhesion scores were 0 (range, 0 to
5), 6 (range, 4 to 8), and 6 (range, 1 to 8) at one, three, and
six weeks, respectively. The adhesion score at one week was significantly
lower than those at three and six weeks (p < 0.001). There
was no significant difference between the three and six-week scores.
The adhesion scores in the wrist fixation group were significantly
higher than those in the synergistic motion group at three weeks
and six weeks (p < 0.001), but there was no significant
difference between the one-week scores of the two therapy groups
(p > 0.145) (Fig. 3).
The median adhesion breaking strength was 0.61 N (range, 0.14
to 1.58 N) at one week, 0.62 N (range, 0.07 to 4.35 N) at three
weeks, and 1.45 N (range, 0 to 7.68 N) at six weeks in the synergistic
motion group, with no significant difference among the values at
the three time-periods (p > 0.05). In the wrist fixation
group, the median adhesion breaking strength was 0.50 N (range,
0.08 to 2.24 N) at one week, 2.91 N (range, 0.60 to 13.71 N) at
three weeks, and 7.22 N (range, 0.05 to 18.33 N) at six weeks. The
strength at one week was significantly less than that at three weeks
or six weeks (p < 0.001), but there was no significant
difference between the three and six-week values (p = 0.478).
The adhesion breaking strength in the wrist fixation group was significantly
greater than that in the synergistic motion group at three weeks
and six weeks (p = 0.005 and 0.017, respectively), but
there was no significant difference between the values in the two
groups at one week (p = 0.815) (Fig. 4).
In the tendon strength tests, all of the tendons consistently failed
through the laceration site. In the synergistic motion group,
the mean maximum failure loads (and standard deviation) were 135 ±
59, 161 ± 54, and 183 ± 52 N at
one, three, and six weeks, respectively. The maximum load at six
weeks was significantly higher than that at one week (p < 0.05).
In the wrist fixation group, the mean failure load was 153 ±
49, 154 ± 45, and 171 ± 57 N at
one, three, and six weeks, respectively (p > 0.05) (Fig. 5).
It is generally acknowledged that postoperative rehabilitation is
an important factor influencing functional outcome after flexor
tendon repair. Postoperative therapy can be classified on the basis
of the force applied to the tendon and the tendon excursion19,27. According to this scheme, the
therapy in our wrist fixation group would be considered to be a
low-force/low-excursion method, whereas that in the synergistic
motion group would be considered a variable-force/high-excursion method.
It is also generally accepted that tendon excursion is important
to prevent adhesions, especially in the early stages after tendon repair,
although the exact amount of excursion needed is still controversial14,19,22,28-33. Loading applied to
the tendon must be great enough to overcome the tendon gliding resistance,
or the tendon will not move. Thus, the gliding resistance becomes
the minimum threshold for the force that must be applied to the
tendon if it is to move at all. In addition, force must be applied
to move the dead weight of the distal segment (phalanges and soft
tissue), to overcome friction within the joints, and to overcome any
soft-tissue stiffness, which may be increased by edema and scarring
after injury or surgery. These other forces can be overcome either
by additional tendon loading, as would occur with active digital
motion, or by passive manipulation. Although active motion protocols
are currently popular, excessive load can cause gap formation or
suture rupture, especially in the early stages after tendon repair.
Thus, rehabilitation that could predictably combine lower tendon
load with high excursion may be the ideal early therapy after tendon
repair.
For mobilization with synergistic wrist and finger motion, the force
applied to the tendon to induce flexion is generated by extension
of the wrist. The tendon is pulled proximally by wrist extension
and is pulled distally by finger extension. Therefore, the force
for this therapy is variable, depending primarily on the tendon
gliding resistance. The force necessary to overcome the other sources
of resistance to motion is applied externally by the passive digital
motion, and thus it does not increase the load perceived by the
tendon15. Although Lieber et al.19 reported that the force applied
to the tendon by synergistic wrist motion was similar to that applied
by passive digital motion with the wrist fixed in flexion, their
tests were performed on normal digits. After tendon repair, the
force on the tendon generated by synergistic motion must increase
as the resistance to tendon gliding increases. This increase in
tendon load may be the reason why gap formation was greater in the synergistic
motion group than it was in the fixation group in our study.
Recently, Silva et al.22 compared
these two postoperative therapies (synergistic wrist and digital
motion and digital motion with wrist fixation) in an in
vivo dog model. They reported that, at ten, twenty-one,
or forty-two days after surgery, there was more adhesion formation
in the fixation group than in the synergistic motion group. These observations
are similar to our own, but Silva et al. did not measure adhesion
breaking strength. A measurable repair-site gap was observed in
fifty-seven (69%) of eighty-three tendons in the study
by Silva et al. In our study, in which the laceration involved 80% of
the tendon, gap formation was found in 6% of the tendons
in the wrist fixation group. However, gap formation was found in
30% of the tendons in the synergistic motion group. This
observation tends to confirm that the synergistic method does indeed
apply higher tendon loads than the passive method does19. With a two-strand suture technique
in a complete laceration canine model, a force of approximately
15 N is needed for gap initiation and 25 N is needed to create a
2-mm gap34-37. We suspect that
the force applied to a repaired tendon in vivo with
synergistic wrist and finger motion might be greater than the 4
N that Lieber et al. reported in normal tendons27;
it may well approach the 15 to 25-N levels noted above.
In the current study, the increased gap formation in the synergistic
motion group did not lead to an increase in tendon rupture compared
with the prevalence in the wrist fixation group. This finding might
indicate that the intact portion of the tendon has sufficient strength
to withstand forces produced during early mobilization programs.
The average maximum strength in this study was over 150 N, which
was much higher than the force applied to the tendon during the
therapies27,38. This force is
also much greater than that in the complete laceration model with
the same suture technique in canines34,39,40.
Our partial laceration model was chosen after careful consideration.
We believe that the model is clinically relevant because partial
lacerations occur and because the surgical management of such injuries
is subject to debate1-3,5,6. The
other purpose of using partial laceration was to allow mobilization
immediately after tendon repair with the least risk of tendon rupture13. We believe that this model allowed
us to study the effect of the rehabilitation method itself in great
detail, with few confounding variables. Finally, by reducing the
risk of rupture, we also minimized the loss of data on our experimental
animals.
The main limitation of our study is that it addressed a special situation:
partial tendon laceration. We believe, however, that the findings
with regard to the rehabilitation methods and the relative adhesion
formation are consistent with data collected in studies of complete
lacerations2,13,22 and are likely
to be valid for that condition as well. In the adhesion breaking
test, the distal vinculum was surrounded by adhesions in several
cases; thus, these adhesions may have been disrupted during resection
of the vinculum. We do not think that this would have affected the
conclusions, however, as this situation occurred only when the adhesions
were very severe and extensive. Finally, the synergistic motion
was performed passively in this animal study. Active synergistic motion
by human patients may produce different results. However, the tendon
tension is changed by the different wrist positions regardless of
whether the motion is active or passive. Therefore, the trend of
the tendon excursion and tension should be similar regardless of
whether synergistic motion is active or passive.
The strengths of this study were that a relevant animal model of
tendon injury was used and the load needed to rupture adhesions
was measured directly. This study demonstrated that adhesion formation
is significantly less with a synergistic wrist and digital motion
therapy program than it is with traditional tendon rehabilitation.
Rehabilitation with an effective tendon-gliding program is a critical
element in minimizing adhesion formation. The gap formation noted
in the synergistic motion group is certainly a concern. We are optimistic, however,
that suture methods that combine high strength and low gliding resistance
may make it possible to achieve the benefits of postoperative synergistic
motion therapy while minimizing the risks of gap formation and tendon
rupture.