Patients
The cases of all patients who had been treated with a triple arthrodesis
from May 1981 to May 1990 by the senior author (R.W.S.) in his private
practice were reviewed. Patients under the age of eighteen years at the time
of surgery or who had had a paralytic diagnosis as the cause of the foot and
ankle symptoms were excluded. Patients with a prior ankle arthrodesis were
excluded as well.
The patients were located through a review of old medical records, internet
searches, a review of death indices, and a credit-checking service. Contact
was made by telephone or
mail18. This
project and its protocol were approved by the institutional review board of
the Memorial Health Services Research Council, Long Beach Memorial Medical
Center, and all patients available for evaluation provided written
consent.
Fifty-two patients were eligible for the study and, of those, twelve had
died and thirteen had been lost to followup. The remaining twenty-seven
patients (thirty-one feet) completed an outcomes questionnaire. Twenty-two
patients (twenty-six feet) were available for the objective evaluation.
Seventeen patients (twenty feet) were evaluated by one of the authors (W.S.).
Five patients were evaluated in their local community by another orthopaedic
surgeon because they lived a great distance from the authors' facility. Four
surrogate examiners, fellowship-trained foot and ankle specialists, assisted
in the evaluation of the five patients. They were provided with
data-collecting worksheets, a brief history of the patients, the patient's
questionnaire, and a roll of colored tape to measure the course for the
functional tests.
The twenty-seven patients who returned their questionnaire were an average
of 47.5 years of age at the time of the surgery. There were nineteen women and
eight men, and the average duration of follow-up was fourteen years. The
diagnoses included systemic inflammatory arthritis in seven patients (nine
feet), six of whom had rheumatoid arthritis and one of whom had lupus;
posttraumatic arthrosis in eight patients (eight feet); primary osteoarthrosis
in five patients (five feet); tarsal coalition in three patients (three feet);
dysfunction of the posterior tibial tendon in four patients (four feet); and
post-clubfoot correction in one patient (two feet) (see Appendix).
Of the twenty-two patients who were available for examination, sixteen were
women and six were men. Their average age at the time of surgery was 45.4
years (range, nineteen to seventy-one years). The average duration of
follow-up was 14.0 years (range, eleven to eighteen years), and the average
age at the time of follow-up was 59.4 years (range, thirty-five to eighty-five
years). Eleven patients were less than sixty years of age at the time of
follow-up.
Surgery
All triple arthrodeses were performed by the senior author (R.W.S.) through
a lateral incision. In about one-third of the cases, a supplementary medial
incision at the talonavicular joint was used. In addition to the traditional
fusion of the talocalcaneal, calcaneocuboid, and talonavicular joints, the
space between the cuboid and navicular was also prepared for fusion as
described by Jahss et
al.19. Staples were
used for internal fixation in five procedures performed from 1981 to 1985, no
internal fixation was used in twenty procedures performed from 1983 to 1989,
and Steinmann pins were used in six procedures performed from 1989 to
1990.
Evaluation
Subjective Evaluation
The MODEMS (Musculoskeletal Outcomes Data Evaluation and Management System)
foot and ankle module from the American Academy of Orthopaedic Surgeons (AAOS)
was given to the patients to complete at home. This outcomes questionnaire was
developed by the AAOS and has been endorsed by several organizations,
including the American Orthopaedic Foot and Ankle Society (AOFAS). It consists
of a lower limb core scale, a global foot and ankle scale, and a shoe comfort
scale. Each scale has a maximum score of 100 points. The global foot and ankle
scale includes twenty questions assessing foot and ankle function, and the
shoe comfort scale includes five questions assessing the types of shoes that
can be worn comfortably. The expanded lower limb score includes nine questions
assessing pain, stiffness, swelling, function, influence of uneven surfaces,
and balance. We modified two questions, one of which referred to changes in
the responder's musculoskeletal condition and the other of which asked about
treatments subsequent to the surgery. On the original MODEMS questionnaire,
those questions asked the responder to compare the present state to the
"last time [the responder] completed [the] questionnaire." We
changed the time-frame to "before [the responder] had surgery."
The MODEMS questionnaire also includes the Medical Outcomes Study Short
Form-36
(SF-36)20,21,
which assesses overall function with a physical component score and a mental
component score.
The AAOS lower extremity scales that we used have been validated and shown
to be reliable and sensitive to change by Johanson et
al.22. Hunsaker et
al.23 determined
normative values for these scales from the general population.
Objective Evaluation
Gait, heel alignment, ankle range of motion, and plantar callosities were
assessed during the physical examination, and the results were recorded. These
parameters were chosen in part on the basis of previous long-term studies of
triple
arthrodesis6-8.
Gait was categorized by the degree of visible limp, which was noted to be
"none/mild," "moderate," or "severe." A
severe limp was defined as being notable on casual observation by someone who
was not a musculoskeletal professional. A moderate limp would be notable on
deliberate observation by someone who was not a musculoskeletal professional.
A mild limp would be notable only on deliberate observation by a
musculoskeletal professional. These categories of visible limp are conceptual
rather than derived from a tested protocol.
Heel alignment was noted as viewed from behind with the patient standing.
The active-assisted range of motion of the ankle was measured in the sagittal
plane with a goniometer, with use of the axis of the fibula and the axis of
the fifth metatarsal as landmarks while the knee was held in full
extension24. We
measured the motion between the foot and the leg with the subtalar complex
locked by the triple arthrodesis in the operatively treated foot and locked by
passive supination in the contralateral foot. The location and severity of any
tenderness or plantar callosities were noted as well.
Radiographs consisted of weight-bearing anteroposterior and lateral views
of the involved foot, a weight-bearing anteroposterior view of the ankle, and
a non-weight-bearing oblique view of the foot. Arthrosis in the ankle,
naviculocuneiform, and tarsometatarsal joints was classified as 0 (normal,
with no sign of arthritis), 1 (mild/moderate, with partial loss of the
cartilage space and/or osteophytes or subchondral cysts), or 2 (severe, with
complete loss of the cartilage space in at least one part of the joint, or a
fused joint).
Two functional tests were used: the 3-m up-and-go test and the six-minute
walk25. The 3-m
up-and-go test is used to assess agility. The patient starts in a sitting
position, stands up, walks to a marker 3 m away, turns around, returns to the
chair, and sits down. The time taken to perform the task is recorded. The
six-minute walk is a test of endurance. The patient walks for six minutes
along a protected course of known length, and the distance covered is
recorded. Both functional tests were conducted with the patient wearing
shoes.
We modified the normative data for the 3-m up-and-go and six-minute-walk
tests. This was necessary because Rikli and
Jones25 used 2.4 m
(8 ft) instead of 3 m (9.8 ft) for the up-and-go test when they derived the
normative values for patients older than sixty years of age, and Simmonds et
al.26 used five
minutes instead of six minutes for the walk test when they derived the
normative values for patients under sixty. The 6 m (19.7 ft) used for the
up-and-back cycle of the 3-m up-and-go test in our study is 23% greater than
the 16 ft (4.9 m) used by Rikli and
Jones25 for
patients older than sixty years of age, but we found that it took normal
subjects over sixty only 15% more time to complete the 3-m test than it took
them to complete the 8-ft test. For the six-minute walk by patients under
sixty, we added 20% to the normative values reported by Simmonds et al. for
their five-minute walk test.
Questionnaire
Overall Scores
The average MODEMS global foot and ankle score for the twenty-seven
patients was 68.8 points of a possible 100 points. To put this score in
context, a patient with mild pain, stiffness, swelling, and restriction of
activities would receive a score of approximately 78 points, whereas a patient
with moderate pain, stiffness, swelling, and restriction of activities would
receive a score of approximately 55 points. The global foot and ankle score
for the seven patients (nine feet) who had systemic inflammatory disease was
55.3 points. The score for the twenty patients (twenty-two feet) with
noninflammatory disease was 74.1 points.
Thirty-two questions in the MODEMS questionnaire deal with activity. Of
these, twelve are specific to the foot and ankle. Although there is no
specific pain subscore in the questionnaire, there are multiple questions
concerning pain as it relates to activity. Eleven (41%) of the twenty-seven
patients reported that they could perform moderate activity (moderate physical
work, jogging, or running) with mild or no pain in the foot and ankle, seven
patients had more than mild pain with moderate activity, and nine patients
could not perform moderate activity for reasons not related to the foot or
ankle. Eighteen patients (67%) reported that they could perform light activity
such as walking, housework, or yard work with mild or no pain. Five of the
nine patients who were unable to perform light activity were restricted for
reasons not related to the foot or ankle, such as severe low-back pain or
cardiopulmonary disease.
Twelve (44%) of the twenty-seven patients reported taking pain medications
(any kind, for any reason) at least once a day. Nine patients reported taking
no pain medications at all. Twenty patients used no support at all or mostly
walked without support or assistance, five patients used one or two canes or
crutches or a walker, and two patients used a wheelchair. Eight of the
twenty-two patients who were examined used custom orthoses. Twenty (74%) of
the twenty-seven patients reported moderate-to-severe difficulty with, or an
inability to negotiate, uneven surfaces. Four had mild difficulty, and three
reported no difficulty with walking on uneven surfaces.
Shoe Comfort
The shoe comfort scale is based on five additional questions in the foot
and ankle module of the MODEMS lower extremity instrument that ask what type
of shoes the subject is comfortable wearing (any shoe; most dress shoes,
sneakers, walking, or casual shoes; or prescription shoes). A score of 50
points indicates that the patient reported being comfortable wearing sneakers
or walking shoes or casual shoes, and a score of 25 points indicates that he
or she is comfortable only when wearing orthopaedic or prescription shoes. The
mean score for the shoe comfort scale was 47.9 points. Of the twenty-seven
patients, four reported that they were able to wear all types of shoes
comfortably and two stated that they could not wear sneakers, walking, or
casual shoes comfortably.
Treatment Expectations
The treatment expectation score is based on five questions about how much
improvement, with regard to relief of symptoms, comfort, stability, and
activity level, that the patients thought they had obtained after surgery. A
patient who answered "definitely yes" to all questions regarding
whether "my treatment met my expectations" would receive a score
of 100 points, whereas a patient who answered "not sure" to all
five questions would receive a score of 50 points. The mean treatment
expectation score was 68.9 points. One question specifically asked whether the
patients thought that the surgery had relieved the symptoms as they had
expected. Twenty-four (89%) of the twenty-seven patients answered
affirmatively to this question.
Satisfaction
The great majority (twenty-five; 93%) of the twenty-seven patients were
satisfied with the result of the surgery to the extent that they would seek
the same treatment if they "could go back in time and make the decisions
again." Sixteen (59%) of the twenty-seven patients answered
affirmatively when asked whether they would be satisfied to live with their
current symptoms for the rest of their lives. The responses were similar
between the patients with and without inflammatory arthritis.
Of the twenty-seven patients, fifteen (56%) (including three with
rheumatoid arthritis) thought that their general musculoskeletal condition was
much better or somewhat better than before the surgery. However, three thought
that it was the same, and eight thought that it was somewhat worse. One
patient answered "not applicable" because she had recently
sustained a tibial fracture in a fall and was wearing a cast. One patient
thought that the general musculoskeletal condition was much worse.
SF-36
The SF-36 questionnaire has two subscores: the physical component and the
mental component. Each has a 0 to 100-point scale, and the mean score for the
general United States adult population is 50 points. The mean physical
component score was 35.2 points for our twenty-seven patients, and the mean
mental component score was 54.8 points. The mean physical component score was
25.3 points for our patients with inflammatory arthritis and 38.6 points for
our patients without inflammatory arthritis (p = 0.004, Student t test). The
mental component scores for the two groups were very similar.
Complications
One patient had a deep wound infection that required antibiotic treatment
and local wound care for five months. Four patients had superficial wound
sloughing or delayed primary healing that spontaneously resolved without
antibiotic treatment. One other patient had a superficial pin-track infection
that was treated successfully with oral antibiotics. None of the patients
underwent revision surgery. Three patients had an ankle fusion subsequent to
the triple arthrodesis (at forty-seven, sixteen, and 173 months). One patient
with rheumatoid arthritis had a midfoot arthrodesis of the naviculocuneiform
and tarsometatarsal joints sixty-three months following the triple
arthrodesis.
Objective Evaluation and Physical Examination
Gait
Seven (32%) of the twenty-two patients who were examined walked with a
moderate or severe limp. Two of those patients had a severe limp; however, one
of them was recovering from a distal tibial fracture.
Clinical Alignment and Stability
Eleven hindfeet were in neutral alignment, three were in subtle varus,
seven were in subtle valgus, and five were in obvious valgus. All ankles were
grossly stable to varus and valgus stress on physical examination.
Sagittal Plane Range of Motion
Dorsiflexion was not compromised, but plantar flexion was. The mean
dorsiflexion was 1.6° (range, —19° to 10°) on the side of
the triple arthrodesis and 2.1° (range, —10° to 11°) on the
contralateral side in the patients who had had a unilateral operation. The
mean plantar flexion was 34.5° (range, 19° to 63°) on the side of
the triple arthrodesis and 47° (range, 32° to 70°) on the
contralateral side in the patients who had had a unilateral operation.
Plantar Callosities
Nine feet were noted to have plantar callosities; two were tender. There
was a suggestion of varus loading in five feet based on calluses at the base
of the fifth metatarsal.
Radiographs
Follow-up radiographs were available for twenty-two patients (twenty-six
feet). There were no nonunions. Twenty of the twenty-six feet exhibited some
degree of degenerative change in at least one of the adjacent joints at the
time of final followup (Fig. 1)
(see Appendix). In all but three of the feet in which there were degenerative
changes in the tarsometatarsal joint, there were also changes in the
naviculocuneiform joint. The prevalence of severe arthrosis was seven (27%) of
twenty-six in the ankle, seven of twenty-six in the naviculocuneiform joints,
and six of twenty-six in the tarsometatarsal joints
(Table I). Some feet had severe
arthrosis at more than one level.
There was poor correlation between ankle pain and radiographic signs of
ankle arthrosis. Twelve ankles were identified by the patient as the location
of continuing pain. Three of those ankles had radiographic evidence of
arthrosis, and two others had a healed fusion. Nine ankles exhibited
radiographic signs of arthrosis, but only three were painful. Correlation
between subjective reports of pain and radiographic findings in the midfoot
(naviculocuneiform and/or tarsometatarsal joints) seemed more consistent.
Eleven patients reported pain at the midfoot, and all eleven had radiographic
evidence of arthrosis. However, twelve feet that had radiographic signs of
midfoot arthrosis were not painful.
Functional Results (see Appendix)
Three-Meter Up-and-Go Test
All twenty-two patients tested were able to complete the 3-m up-and-go
test. Fifteen were able to complete the task in ten seconds or less, which is
the time limit required to categorize a geriatric patient as
"independent."16
Only four patients performed at or better than the age-adjusted mean. Of the
eighteen patients who performed worse than the age-adjusted mean, eight
performed at a level that was more than two standard deviations from the
mean.
Patients with systemic inflammatory disease performed more poorly on this
test, with a mean time (and standard deviation) of 15.1 ± 12.7 seconds
compared with 7.54 ± 2.25 seconds for the patients without inflammatory
arthritis (p = 0.003).
Six-Minute Walk
Two patients were unable to walk for the entire six minutes, one because of
cardiopulmonary reasons and the other because of pain and weakness. Only two
patients who completed the walk performed better than the age-controlled mean.
Of the eighteen patients who performed below the mean, six performed more than
two standard deviations below the mean. On the average, these patients
performed the six-minute walk test well below the standard for a normal
age-adjusted population.
The patients with systemic inflammatory disease walked a mean of 971
± 764 ft (296 ± 232.9 m), and those without inflammatory
arthritis walked a mean of 1436 ± 311 ft (437.7 ± 94.8 m). This
difference was not significant (p = 0.052).
Although the definition of arthritis varies somewhat, severe arthrosis of
the ankle and the naviculocuneiform joints was more common in our series than
it was in the studies by Saltzman et
al.6, Angus and
Cowell7, and Adelaar
et al.8. The
prevalence of severe ankle arthrosis was 27% (seven) of twenty-six in our
study, 9% (six) of sixty-seven in the study by Saltzman et al., and 15%
(twelve) of eighty in the study by Angus and Cowell. In the study by Adelaar
et al., the prevalence of severe ankle arthrosis was one of sixty and was
asymptomatic. Severe naviculocuneiform arthrosis occurred in 27% of our cases,
9% of the cases of Saltzman et al., 10% of those of Angus and Cowell, and 3%
of those of Adelaar et al. Perhaps the presence of paralysis in the other
three studies led to some sparing of the adjacent joints that did not occur in
our series.
We found correlation between radiographic signs of arthrosis and subjective
reports of pain to be inconsistent, especially in the ankle. Clinical
correlation was a little more consistent in the midfoot, as all patients who
reported pain in that location also had radiographic evidence of degenerative
disease there; however, not all patients with radiographic signs of midfoot
arthrosis had pain in the midfoot.
About one-quarter of our patients had systemic inflammatory disease, and
they had significantly worse global foot and ankle scores, SF-36 physical
component scores, and scores for their performance on the 3-m up-and-go test
compared with patients without systemic inflammatory disease. In contrast, in
the study by Pell et
al.16, in which the
duration of follow-up ranged from two to eleven years, patients with
rheumatoid arthritis scored higher on the American Orthopaedic Foot and Ankle
Society ankle-hindfoot scale than did those with posttraumatic conditions, but
the difference was insignificant.
Sagittal plane motion is characteristic of the ankle, but this study
reminds us of the component of plantar flexion that is due to the subtalar
complex, including the talonavicular joint. Plantar flexion averaged 27%
(12°) less than that on the contralateral side in our patients who had had
a unilateral triple arthrodesis, but there was no significant difference in
dorsiflexion between the two sides. The average arc of residual sagittal
motion was 36°, which is almost identical to the finding of Saltzman et
al.6.
During the period in which many of these procedures were performed,
surgeons paid less attention to clinical alignment than we do today. The most
common type of malalignment was hindfoot valgus. Although we are concerned
that fusion of the hindfoot in valgus may predispose a patient to lateral
degeneration in the ankle, we could not find evidence of this in our study
with the small numbers available. There were fusions in valgus with no ankle
degeneration, and there was severe ankle degeneration in some patients with
neutral alignment of the hindfoot.
Although it has been applied in the evaluation of low-back pain and the
results of total hip
replacement11,26,
functional testing, as far as we know, has not been used to analyze
postoperative function of the foot and ankle. Our patients performed two
easy-to-duplicate tests. One-third of them did not achieve the cutoff value
necessary for a classification of "independent" in the 3-m
up-and-go test, and one-quarter performed more than two standard deviations
below the mean in the six-minute walk test. These results are not encouraging
when compared with normal values, but, because of the retrospective design of
the study, we do not know if they represent an improvement from the
preoperative functional levels. Furthermore, these functional tests are
greatly affected by impairments other than those of the foot and ankle,
particularly in patients with rheumatoid arthritis.
Concerns about the structure of this study include the fact that thirteen
of the patients in the study cohort were not available for follow-up, despite
the use of various computer searches and death notice investigations. Our
patients proved to be quite mobile over the decade or more since their
surgical treatment. Forty-eight percent (thirteen) of the twenty-seven
patients whom we located lived outside of the Greater Los Angeles area at the
time of final follow-up. In addition, five patients who returned
questionnaires were not available for the protocol examination. Offers of
financial reimbursement did not entice them to return to our office or to
orthopaedists closer to them. The mean age of those five patients was greater
and their SF-36 scores were lower than those of the patients who were
examined, but the significance of those differences is not known.
The geographic dispersal of our patients raises the question of
interobserver reliability of multiple examiners, as five patients were
examined by an orthopaedist who was not an author of this paper. Although we
required that all examiners be orthopaedic surgeons with experience in foot
and ankle surgery and that all follow a specific protocol, uniformity of the
follow-up examinations could not be ensured.
We acknowledge the potential for error in the evaluation of the treatment
expectation score. This score depends on the patient's recollection of
symptoms prior to their treatment, which had been carried out more than ten
years prior to the time of follow-up in this study. Toolan et
al.27 found
recollections of preoperative perceptions to be inaccurate when they compared
them with the recorded preoperative perceptions of patients treated with foot
and ankle surgery.
It was necessary to do some extrapolating to relate our functional test
data to age-controlled means. The normal values for the timed walk reported by
Simmonds et al.26
were for patients under sixty years of age (average, thirty-five years of
age), and we extrapolated their five-minute-walk data to our six-minute-walk
results. The normal values for the up-and-go test reported by Rikli and
Jones25 were for
patients over sixty years of age, and we extrapolated their 8-ft-test data to
our 3-m-test results.
We used a somewhat uncommon outcome measurement tool. When our study was
initiated, we believed that the MODEMS outcome tools would be the method of
the future for orthopaedic clinical research. The questions in these tools
have been subjected to extensive peer review and have been shown to be valid,
highly reliable, and sensitive to change in patient status. The scoring scales
of these instruments correlate moderately to strongly with other measures of
pain and function, such as physician ratings, the Western Ontario and McMaster
Universities Osteoarthritis Index (WOMAC), and the
SF-3622. Also, the
lower extremity instruments incorporate the extensively used SF-36
questions.
Although the great majority of patients in this study said that they would
choose to have the surgery again under the same circumstances, the data showed
that a triple arthrodesis may lead to degenerative changes in adjacent joints.
These patients should have a strategy to prevent or retard such consequences.
Maintaining a reasonable body weight and being deliberate about the selection
of activities should be part of that strategy, and patients should be so
counseled by their orthopaedic surgeon. Knowledge of the adverse consequences
of a triple arthrodesis should remind orthopaedists to consider alternatives
to it whenever feasible.
In conclusion, the majority of patients were satisfied with the result of
the triple arthrodesis and had reasonable foot and ankle function as assessed
with the questionnaire and the objective examination. However, we learned that
patients treated with a triple arthrodesis will have limitations in their
selection of shoes and poor tolerance of walking on uneven surfaces. Patients
generally have mild gait disturbances but maintain a functional range of
motion in the sagittal plane. They are mildly to moderately impaired in their
agility and walking endurance. Finally, arthrosis in adjacent joints develops
in almost 30% of patients, and ankle arthrosis may lead to ankle arthrodesis
in some patients.