After approval by the scientific committee of our hospital, forty-three
patients with primary osteoarthritis of the thumb carpometacarpal were
enrolled in the study, which was conducted from November 1995 to June 1998.
None of the patients had responded to nonsurgical management, including
splinting, exercises, physical therapy, nonsteroidal antiinflammatory drugs,
and intra-articular injections with and without steroids. Nine patients had
bilateral involvement; thus, the study included a total of fifty-two thumbs.
After informed consent had been obtained from all patients, they were randomly
allocated for treatment with either trapezial excision with ligament
reconstruction alone (group I) or trapezial excision with ligament
reconstruction combined with tendon interposition (group II). The procedure to
be performed was determined by a computer-generated randomization list that
was made by an independent statistician. The validated software (Rancode Plus;
IDV Datenanalyse und Versuchsplanung, Gauting, Germany) randomly assigned
individual patient numbers to treatment groups with use of a block size of
four patients. Each patient included in the study was assigned to the next
free number on the randomization list and thus was allocated to treatment
group I or II. Patients who had rheumatoid or posttraumatic arthritis were
excluded from the study. Twelve patients were not included in the final
assessment because they had changed their mind about participating in the
study (three patients in group I and two patients in group II) or because they
had moved during the time of the investigation and could not be located
despite searches of the medical records department, the directory enquiry
services, and the internet (four patients in group I and three patients in
group II). The mean age of the remaining thirty-one patients (thirty-eight
thumbs) was fifty-nine years (range, forty-two to seventy-five years) at the
time of surgery. Of the thirty-one patients, six treated with ligament
reconstruction alone and one treated with ligament reconstruction and tendon
interposition had bilateral involvement. However, only the thumbs that had
been operated on first were used for the statistical analysis. Thus,
thirty-one thumbs in thirty-one patients were available for long-term clinical
assessment at an average of 48.2 months (range, thirty-two to sixty-four
months).
None of the affected thumbs had been operated on previously, and none of
the patients had a coexistent condition of the hand at the time of surgery.
Indications for surgery were severe pain, loss of strength, and loss of motion
at the base of the thumb that impaired functional activities of daily living.
The mean duration of symptoms before the operative intervention was 42.8
months (range, 15.0 to 120 months).
Group I (Ligament Reconstruction Only)
Fifteen patients (fifteen thumbs) were available for long-term follow-up at
an average of 46.2 months (range, thirty-two to sixty-four months after
ligament reconstruction alone). The assessment was performed with a
standardized outcome questionnaire, physical measurements, and radiographs.
There were thirteen women and two men, and their mean age at the time of
surgery was fifty-nine years (range, forty-two to seventy-five years). The
dominant hand was involved in nine patients and the nondominant hand, in six.
According to the radiographic criteria of Eaton and
Glickel12,
preoperatively two thumbs exhibited stage-II degenerative arthritis; eleven,
stage-III; and two, stage-IV with scaphotrapezial involvement
(Table I). At the time of
surgery, four patients were homemakers, one had retired, six had office jobs,
and four performed strenuous manual labor. Nine patients performed demanding
recreational tasks with their hands.
Group II (Ligament Reconstruction and Tendon Interposition)
Sixteen patients (sixteen thumbs) completed the outcome questionnaires and
were evaluated at an average of fifty months (range, thirty-five to sixty-two
months) after the combined ligament reconstruction and tendon interposition
procedure. Twelve patients were women and four were men; their mean age at the
time of surgery was 58.4 years (range, forty-two to seventy-eight years). The
dominant hand was involved in eleven patients and the nondominant hand, in
five. According to the radiographic criteria of Eaton and
Glickel12, three
thumbs had stage-II disease preoperatively; eleven, stage-III; and two,
stage-IV, including scaphotrapezial involvement
(Table I). At the time of
surgery, two patients were homemakers, three had retired, three had office
jobs, three performed light manual work, and five performed strenuous manual
labor. Ten patients performed demanding recreational activities with their
hands.
Operative Method
All of the procedures were performed by the senior author (G.K.-A.). The
surgical technique of Epping and
Noack13 was
slightly modified. The trapeziometacarpal joint was carefully exposed through
a dorsoradial hockey-stick-shaped incision over the base of the thumb. Care
was taken to avoid injury to the superficial branch of the radial sensory
nerve and the radial artery. The entire trapezium was excised and then
converted, with the use of rongeurs, to corticocancellous and cancellous bone
chips. In group I, the wrist was flexed and the radial half of the flexor
carpi radialis tendon was harvested in the depth of the trapezial void, with
its insertion left intact on the base of the index metacarpal. The distally
based strip of the flexor carpi radialis tendon was routed through an oblique
canal in the first metacarpal base, which had been created with a 3.2-mm burr.
The canal began dorsoradially at a distance of approximately 1 cm from the
metacarpal base and ran in a palmar-ulnar direction through the articular
surface of the metacarpal base (Fig.
1-A). With traction applied to the thumb, the tendon strip was
carefully tensioned and secured within the canal by gently impacting the
smooth corticocancellous and cancellous bone chips. This effectively prevented
the tendon from slipping out of the canal and stabilized the creation of the
suspensory ligament. The remaining tendon was sutured to the periosteum of the
first metacarpal base.
In group II, trapezial excision and ligament reconstruction were performed
in a similar fashion, except that the radial half of the flexor carpi radialis
tendon was harvested proximally through a second longitudinal incision in the
forearm. The tendon strip was drawn through the bone canal, blocked by bone
chips as it was in the other group, and sutured to the surrounding periosteum.
The remainder of the tendon was rolled up, placed into the trapezial void to
act as a spacer, and fixed with a 4-0 absorbable suture to the flexor carpi
radialis tendon in the depth of the arthroplasty space
(Fig. 1-B).
In both groups, the joint capsule was closed but Kirschner wires were not
inserted to stabilize the first ray. The thumb was immobilized in a spica cast
for three weeks. The cast was then replaced with an individually fitted thumb
spica splint that was worn constantly, except during bathing, for an
additional three weeks. Active and active-assisted range-of-motion and thenar
muscle-strengthening exercises were performed thereafter.
Objective and Subjective Assessment
Preoperative clinical and radiographic examinations and radiographic
follow-up immediately after the surgery were performed by independent
observers. The four-year follow-up assessment was performed by one of us
(E.F.), who had not been involved in either the surgical procedures or patient
care. Strength measurements were performed by independent ergotherapists at
each follow-up interval.
The overall long-term outcomes were assessed with use of the Buck-Gramcko
score21, a
standardized outcome measure that provides objective and subjective data, at
the time of final follow-up. Palmar and radial thumb abduction and tip-pinch
strength were scored objectively. For the subjective assessment, all patients
completed a self-administered questionnaire pertaining to pain, strength,
daily function, dexterity, cosmetic appearance, willingness to undergo the
surgery again under similar circumstances, and overall satisfaction with the
result of the procedure. Patients were instructed to grade daily function as
excellent when they had no difficulties with daily and strenuous or exertional
activities, good when they had occasionally noted minor problems with
tip-pinch or power, fair when they had moderate difficulties even with light
daily activities, and poor when function was completely impaired. Patients
graded the cosmetic result of the operation according to the appearance of the
scar and the position of the thumb. Objective and subjective values were
assigned a maximum score of 4 or 6 points each and were summed to calculate a
total score. A total score of 49 to 56 points (the maximum score) was rated as
an excellent surgical outcome; 40 to 48 points, good; 28 to 39 points, fair;
and <28 points, poor. The scores in each category and the total scores for
the overall surgical outcomes were analyzed at the time of final
follow-up.
Another questionnaire, in addition to the Buck-Gramcko score, was used to
determine the functional outcome. Two independent medical students asked the
patients to complete this questionnaire, which addressed their abilities to
perform ten functional tasks, including writing with a pen, brushing their
teeth, threading a needle, turning a key, opening a tight jar, using a knife
or scissors, buttoning and zipping clothes, picking up small objects (e.g.,
coins, paper clips, and thumbtacks), and card-playing at the time of the final
follow-up visit. The patients gave their responses according to a rating scale
of 1 to 5, with 1 expressing no difficulty and 5 indicating an inability to
perform the task. Furthermore, current and past employment status and activity
levels were determined.
All objective outcomes were determined preoperatively and postoperatively.
Grip and tip-pinch strength were measured with use of a Martin vigorimeter
(Martin Brothers, Tuttlingen, Germany), and the measurements were repeated
twice to calculate mean values for use in the statistical analysis. Radial and
palmar abduction of the carpometacarpal joint, the ability to oppose the thumb
tip to the palmar crease of the little finger, and the presence of
metacarpophalangeal joint hyperextension were determined.
Radiographic Evaluation
To calculate the degree of proximal migration of the thumb metacarpal,
standard posteroanterior and oblique radiographs with the hand at rest were
made for all thirty-one patients preoperatively and at each follow-up visit.
At the final follow-up examination, standard oblique radiographs were made of
the site of the arthroplasty with the hand at rest and with maximal pinch
stress. For each radiograph, the patient was asked to oppose the thumb tip to
the index finger. For the oblique stress radiograph, the position of the hand
remained unchanged, a new x-ray cassette was placed, and the patient exerted
maximum pinch effort. The distance between the distal articular surface of the
scaphoid and the proximal articular surface of the first metacarpal was
measured on all preoperative and postoperative radiographs
(Fig. 2). The index of the
height of the arthroplasty space was calculated by dividing the
scaphoid-metacarpal distance by the length of the first metacarpal, as
described by Yang and
Weiland22. The
postoperative indices were used to compare the degree of proximal migration of
the thumb metacarpal, both at rest and during stress, between the groups.
Statistical Analysis
Preoperative and postoperative statistical analyses were done by
independent statisticians. Between-group comparisons were conducted with use
of the Mann-Whitney U test or the chi-square test. Regression analyses were
performed to examine relationships between variables. A p value <0.05 was
considered significant.
Patient Demographics
Preoperatively, we found no significant differences between the two groups
with regard to age, gender ratio, duration of symptoms, follow-up period,
operating time, handedness, stage of disease, or pain status, with the numbers
available (Table I).
Furthermore, we detected no significant difference between the groups with
regard to mean tip-pinch or grip strength, palmar abduction, radial abduction,
or trapezial space prior to the operations
(Table II).
Objective and Subjective Outcomes
We detected a significant difference between the two groups with respect to
the mean total Buck-Gramcko scores (p = 0.036) at the most recent follow-up
assessment (Tables III and
IV). The mean total score for
the fifteen patients treated with ligament reconstruction alone (group I; 51.3
points, an excellent outcome) was superior to that for the sixteen patients
treated with ligament reconstruction and tendon interposition (group II; 44.6
points, a good outcome). Eleven patients in group I and four in group II had a
total score that was in the upper third of the distribution for the series;
none of the patients in group I and five in group II had score in the lower
third.
We detected significant differences between the groups with respect to the
Buck-Gramcko scores for palmar abduction (p = 0.02) and radial abduction (p =
0.008) at the time of final follow-up
(Table III). The mean degrees
of palmar abduction (p = 0.01) and radial abduction (p = 0.001) were
significantly greater in group I than in group II
(Table II). The values for
palmar abduction had an average increase of 34% in group I and 28% in group II
compared with the preoperative values. The values for radial abduction had an
average increase of 29% in group I and 15% in group II compared with the
preoperative values.
According to the Buck-Gramcko score and the objective strength
measurements, no significant differences could be detected between the groups
for tip-pinch or grip strength, with the numbers available (Tables
II and
III). At the final assessment,
the tip-pinch strengths had an average increase of 32% in group I and 9% in
group II compared with the preoperative values, whereas the grip strengths had
an average decrease of 20% in group I and 48% in group II.
We found no significant differences between the groups with regard to the
subjective parameters of pain, strength, daily function, dexterity, or the
overall assessment of the surgical outcome, with the numbers available
(Table IV). Pain relief was
similar for the two groups. Only two patients in group II reported constant
pain at the final follow-up visit. None of the patients in group I had a
decrease in subjective strength or dexterity compared with the preoperative
levels. However, in group II, three patients had a decrease in subjective
strength and four had a decrease in dexterity. The mean overall assessment of
the surgical result was good or excellent in both groups; only one patient in
group II was completely dissatisfied with the result of the operation, mainly
because of deficits in subjective strength, daily function, and dexterity.
Cosmetic appearance (p = 0.038) and willingness to undergo the surgery
again under similar circumstances (p = 0.038) were the only subjective
parameters that differed significantly between the groups, who had had similar
preoperative symptoms, at the final assessment
(Table IV). All patients in
group I gave both subjective items an excellent score. Despite a good or
excellent cosmetic result at the final evaluation, four patients in group II
were unwilling to undergo the surgery again, mainly because of impaired
strength, daily function, and dexterity compared with the preoperative
conditions. One of those patients was unwilling to undergo the surgery again
because of geographic and social reasons.
Although functional performance was slightly better in the group treated
with ligament reconstruction alone, we detected no significant differences in
the results of our functional-tasks questionnaire between the two groups, with
the numbers available (Fig. 3).
The groups also had similar results with regard to returning to their previous
occupation. Six of the ten patients in group I who had worked outside the home
preoperatively resumed working at their previous job, and four retired because
of their age. Nine of the eleven patients in group II who had worked outside
the home preoperatively returned to their previous work, and one retired
because of age. Only one patient treated with combined ligament reconstruction
and tendon interposition did not return to work because of constant pain and
thumb weakness. Eight of the nine patients in group I who had performed
demanding recreational tasks preoperatively were able to resume their previous
recreational activities. Three of the ten patients in group II who had
performed demanding recreational tasks preoperatively stated that they were
unable to resume their hobbies because of occasional or frequent pain after
strenuous activities, such as gardening, playing the clarinet, and sewing.
We found no apparent differences between group I and group II with respect
to the ranges of opposition (Table
II). The four patients in each group who had had mild
hyperextension of the metacarpophalangeal joint preoperatively exhibited no
evidence of progression.
Radiographic Outcomes
With the numbers available, we detected no significant differences between
the two groups with regard to the early postoperative or follow-up index of
the height of the arthroplasty space or the percentage decrease in the height
of the arthroplasty space, either at rest or with maximal pinch
(Table II). Postoperative
proximal migration of the thumb metacarpal at rest resulted in an average loss
of 42% of the initial postoperative height of the arthroplasty space in group
I and a 37% loss of the height in group II. Under stress, proximal metacarpal
migration was found to result in a 40% loss of height in group I and a 27%
loss of height in group II. An arthroplasty space was maintained in all of our
patients, and no patient had evidence of impingement of the metacarpal against
the scaphoid.
Effect of Arthroplasty Space on Pinch Strength and Pain
We did a regression analysis to determine whether the height of the
arthroplasty space affected tip-pinch strength. With the numbers available,
the analysis did not show a significant relationship between tip-pinch
strength and the decrease in the index of the arthroplasty space in either of
the two treatment groups at the most recent follow-up assessment (p = 0.4187).
Furthermore, no relationship was found between postoperative pain (with
exclusion of the patients in each group who were completely free of pain) and
the height of the arthroplasty space. There was also no significant
relationship between the height of the arthroplasty space and postoperative
pain or the objective and subjective outcomes achieved on the Buck-Gramcko
score.
Complications
Two patients in each treatment group had temporary paresthesias in the
distribution of the superficial sensory branch of the radial nerve. Severe
reflex sympathetic dystrophy occurred early after one combined ligament
reconstruction and tendon interposition procedure. Despite aggressive
conservative management and a supervised hand therapy program, it resulted in
permanent pain and impaired hand function at the final-follow-up
examination.
There is controversy in the literature regarding the effects of tendon
interposition, and the height of the arthroplasty space, on the objective and
subjective outcomes following ligament reconstruction to treat osteoarthritis
of the thumb carpometacarpal
joint20,22-25.
Some investigators have recommended ligament reconstruction and tendon
interposition2,17,18,
whereas others have found no correlation between the height of the
arthroplasty space and overall satisfaction, pain relief, or improvements in
thumb function or
strength22-24.
In the study by Mentzel et al., who used the Epping-Noack ligament
reconstruction, the mean total Buck-Gramcko score was 49.0 points (an
excellent outcome) after a mean forty-month follow-up, which is similar to the
mean score achieved in our ligament reconstruction group I (51.3
points)26. Mentzel
et al. reported a good or excellent final result in ten of eleven patients.
Hilty and Stober reported a good or excellent surgical outcome in fifty of
fifty-six patients after a mean duration of follow-up of thirty-two
months1. In our
study, the overall outcomes after the ligament reconstruction procedures were
good or excellent in all fifteen patients.
The rates of overall satisfaction and complete freedom from pain achieved
in our group II (ligament reconstruction and tendon interposition) compare
favorably with those of previous
studies19,23,27,28.
Only two authors have reported superior results, to our knowledge: Tomaino et
al. observed excellent relief of pain and satisfaction in twenty-one of
twenty-two patients at a mean of nine years
postoperatively2,
and Varitimidis et al. noted complete resolution of pain in fifty-five of
fifty-eight patients and overall satisfaction in all fifty-eight patients at a
mean of 42.5 months
postoperatively18.
In an analysis of the results in sixty-one patients, Wittemann et al.
suggested that the Epping-Noack ligament reconstruction is associated with a
discrepancy between excellent overall satisfaction and impaired hand function
during strenuous
tasks29. In
contrast to their study, we found the high degree of satisfaction in our group
to be consistent with the good-to-excellent scores on our functional-tasks
questionnaire. Our tendon-interposition group did have a slightly higher rate
of diminished thumb and hand function, but no significant differences could be
detected between our two groups with regard to their rating of ten functional
tasks and return to preoperative occupational levels. Unlike our findings and
those of Rayan and
Young27, previous
reports have shown an unrestricted ability to perform daily activities, such
as opening jars and automobile doors and turning
keys2,15,20.
In the present study, opening a tight jar remained the most difficult task,
followed by buttoning clothes, for both patient groups. Contrary to most
previous studies, which have demonstrated consistent postoperative
improvements in both pinch and grip
strengths2,15,18-20,22,23,30,
we found an increase in the mean tip-pinch strength and a decrease in the mean
grip strength after both procedures. Although we have no sound explanation for
the decrease in grip strength compared with the preoperative levels, it did
not seem to affect our patients' functional improvement.
In agreement with the findings of previous
studies22,23,26,30,
we noted no significant relationship between the height of the arthroplasty
space and tip-pinch strength or pain. Yang and Weiland observed a 21%
subsidence of the first metacarpal at rest and an additional 10.5% subsidence
during stress at a mean of thirty-two months after fifteen ligament
reconstruction and tendon interposition
arthroplasties22.
We found a greater percentage loss in arthroplasty space at rest and under
strain following both procedures than did authors of other studies in which
Kirschner wire stabilization was not
used18,22.
Nevertheless, our patients had satisfactory subjective outcomes and
improvements in tip-pinch strength, palmar abduction, radial abduction, and
opposition. Tomaino emphasized the importance of temporary pin fixation of the
thumb metacarpal to stabilize and protect the ligament reconstruction in the
early postoperative period and to preserve a trapezial space, thereby
restoring thumb
strength17.
Lanzetta and Foucher even noted decreased grip and tip-pinch strengths after
ligament reconstruction and tendon interposition in a subgroup of patients
without proximal metacarpal
migration31.
A weakness of our study is the small number of patients included in each
treatment group. Hence, additional prospective, randomized studies of larger
numbers of patients are needed.
In summary, our findings, together with those of Gerwin et
al.20, have shown
that tendon interposition does not affect the long-term objective and
subjective outcomes of ligament reconstruction for treatment of advanced
osteoarthritis of the thumb carpometacarpal joint. Furthermore, our results
suggest that the degree of proximal migration of the thumb metacarpal may have
no effect on postoperative thumb strength, function, and pain. We found both
procedures to have favorable long-term outcomes overall. We believe that
patient-oriented scoring systems may facilitate analysis of patient outcomes
and thus provide better assessment of the apparent benefits of different
treatment options.