Abstract
Background: Ankylosing spondylitis is a seronegative
spondyloarthropathy that primarily affects the sacroiliac joints,
spine, hips, and, less commonly, the knee joints. The purpose of
this study was to evaluate the results in a consecutive group of
patients with ankylosing spondylitis who underwent total knee arthroplasty.
Methods: The results of thirty total knee arthroplasties
in twenty patients with ankylosing spondylitis were reviewed. There were
seventeen men and three women, with an average age of fifty-five
years (range, twenty-eight to sixty-seven years)
at the time of the arthroplasty. The diagnosis of ankylosing spondylitis
was established preoperatively with use of the New York criteria.
All patients received a cemented condylar-type implant.
The average duration of follow-up was 11.2 years (range,
three to sixteen years).
Results: The average Knee Society pain score improved
from 14 points preoperatively to 76.3 points at the time of the
latest follow-up. The improvement in the average Knee Society function
score was less impressive, with an increase from 16.3 points preoperatively
to 58.7 points at the time of the latest follow-up. The average
arc of motion was 84.8° prior to the arthroplasty and 86.7° at the
time of the final follow-up. Six knees (20%) had heterotopic
bone formation. Three knees required manipulation under anesthesia
because of poor motion postoperatively. There was one revision,
due to loosening of a patellar component. All other components were radiographically
stable at the time of the latest follow-up.
Conclusions: Total knee arthroplasty with cement
in patients with ankylosing spondylitis provided excellent pain
relief and durable fixation at an average of 11.2 years postoperatively.
However, patients with ankylosing spondylitis are at increased risk
for the development of stiffness and heterotopic bone formation.
Ankylosing spondylitis is a type of seronegative arthropathy with
an insidious onset, typically at a young age, that invariably affects
amphiarthrodial joints, such as the sacroiliac joints. Although
involvement of peripheral diarthrodial joints, including the knee,
is less common, it has been reported in up to 70% of patients
(in series ranging from twenty to ninety-eight patients)1-4. On occasion, a patient with ankylosing
spondylitis presents with hip, knee, and shoulder involvement before
the sacroiliitis is evident2,5.
There are various reports in the literature regarding the outcome
of total hip arthroplasty in patients with ankylosing spondylitis1,6-14. According to these studies,
total hip arthroplasty, despite a relatively high rate of complications,
provides excellent, durable, and predictable pain relief8,10,11.
There is very little available information on the results of
total knee arthroplasty in patients with ankylosing spondylitis. Some
series have included a few total knee arthroplasties in patients
with ankylosing spondylitis together with a much greater number
of total knee arthroplasties in patients with osteoarthritis or
rheumatoid arthritis15,16. It
is difficult to ascertain the clinical outcomes in patients with
ankylosing spondylitis from the data in those reports. Finsterbush
et al., in a report on the results of joint procedures in twenty-three
patients with ankylosing spondylitis, noted excellent pain relief
following the six knee arthroplasties performed in their series15. The functional results of the total
knee arthroplasties were, however, less predictable and were noted
to deteriorate with time in all six patients.
The purpose of the present study was to evaluate the midterm clinical
and radiographic results of total knee arthroplasty in patients
with ankylosing spondylitis.
We used our computerized database to identify all patients with
a diagnosis of ankylosing spondylitis who underwent condylar total
knee arthroplasty with cement between 1982 and 1997. During this
time-period, of 13,821 primary total knee arthroplasties performed
for various diagnoses, thirty-four were done in twenty-three
patients with ankylosing spondylitis. Three patients (four knees)
were excluded: one because of inadequate follow-up, and
the other two because of a mixed diagnosis of ankylosing spondylitis
and rheumatoid disease. All patients were examined by a rheumatologist,
and all met the New York diagnostic criteria for "definite" ankylosing
spondylitis17. HLA-B27
assay was positive in sixteen patients (twenty-five knees).
Chest expansion measured at the fourth intercostal space was found
to be limited to <2.5 cm in all patients. Extraskeletal
manifestations of ankylosing spondylitis included iritis in three
patients, apical pulmonary fibrosis in one patient, and aortitis
in one patient. In addition, seventeen patients reported some degree
of fatigue, weight loss, and anorexia. The characteristic radiographic
changes of subchondral sclerosis with complete obliteration of the
sacroiliac joint and squaring of the anterior vertebral borders
were present in all patients. Ossification of the anterior longitudinal
ligaments leading to the so-called bamboo spine was also
seen in eighteen patients. Six patients underwent surgery on other
joints because of deformity and pain: an ipsilateral total hip arthroplasty
was performed in three patients; a contralateral total hip arthroplasty,
in two patients; and a shoulder arthroplasty, in one patient.
There were seventeen men and three women, with an average age
of fifty-five years (range, twenty-eight to sixty-seven years)
at the time of surgery. The average weight of the patients was 76.4
kg (range, 50 to 105 kg), and the average height was 166.6 cm (range,
150 to 179 cm). All patients had insertion of a cemented condylar
prosthesis. Ten of the patients had bilateral knee replacement,
with the two procedures performed simultaneously in five and in
a staged fashion in five. A press-fit condylar implant (DePuy, a
Johnson and Johnson company, Warsaw, Indiana) was used in seventeen
knees; a Genesis implant (Richards, Memphis, Tennessee), in seven;
and a cruciate condylar implant (Howmedica, Rutherford, New Jersey),
in six. The patella was resurfaced with a cemented all-polyethylene
component in all patients. Five knees had had surgery prior to the
knee arthroplasty: two knees had had a medial meniscectomy; one,
a proximal tibial osteotomy; one, a synovectomy; and one, an arthroscopic débridement.
The average duration of the ankylosing spondylitis prior to the
knee arthroplasty was twenty-five years (range, one to
fifty-seven years), and the average duration of symptomatic
knee involvement prior to the knee arthroplasty was twelve years
(range, three to twenty-nine years).
Clinical and radiographic data in the total joint registry were collected
prospectively. Patients were contacted at two months, one year,
two years, five years, and every five years thereafter on a routine
basis. Knee scores were calculated with use of the Knee Society18 knee-scoring system, which consists
of a score for pain and a score for function, each with a maximum
of 100 points. Knee scores were assessed before surgery, at two
years, and at the latest follow-up examination.
Institutional approval and the consent of all patients were obtained
prior to the review, and no patient refused to participate in the
study. The duration of clinical follow-up averaged 11.2 years (range,
three to sixteen years), and the duration of radiographic follow-up
averaged 8.6 years (range, two to twenty-two years). All
patients were followed for a minimum of two years or until failure
of the prosthesis or death. There were six deaths during the follow-up
period.
Serial anteroposterior and lateral radiographs of the involved joint
were reviewed to assess the position of the prosthesis and the presence
of loosening and/or wear during the follow-up
period. Radiographs were also scrutinized for the presence of heterotopic
ossification. Definitely loose components were defined as those
with a complete lucent line on any radiograph or femoral or tibial
subsidence of 2 mm19.
Perioperative antibiotics and antithromboembolic medications were
used in all patients. All wore compressive stockings, and nineteen
used sequential compression devices. One patient received ibuprofen
for prophylaxis against heterotopic ossification because heterotopic
ossification had developed in the hip following a previous hip arthroplasty.
Heterotopic ossification did not develop in the knee of this patient
following the total knee arthroplasty.
Surgical Data
The operations were performed with use of general anesthesia in
eleven patients (eighteen knees) and with use of epidural anesthesia
in nine patients (twelve knees). Fiberoptic intubation was necessary
in seven patients because of a lack of adequate neck extension.
A tourniquet was used during the arthroplasty in twelve patients
(fourteen knees).
A lateral release was performed from within out, with sparing of
the lateral superior genicular artery, in seven patients (nine knees).
In addition, six patients (eight knees) with preoperative flexion
contracture required extensive soft-tissue release posteriorly and
a mild elevation (2 to 6 mm) of the joint line. A continuous-passive-motion
machine was used by all patients. The patients were allowed to walk
with full weight-bearing and use of bimanual support on the first
postoperative day. Additional physical therapy in the hospital and
after discharge consisted of range-of-motion exercises, muscle-strengthening,
and gait-training.
Statistical Analysis
The changes in the knee-pain and function scores were evaluated
with the Wilcoxon signed-rank test. Comparison of continuous measures
between groups was performed with use of the two-sample t test.
Significance was attributed to observations with less than a 0.05
likelihood of occurring by chance.
Knee Scores
The average preoperative Knee Society pain and function scores
were 14 points (range, 0 to 37 points) and 16.3 points (range, 0
to 40 points), respectively. These scores improved significantly,
to an average pain score of 87.5 points (range, 75 to 100 points)
and an average function score of 80 points (range, 30 to 90 points)
at two years postoperatively (p < 0.0001). At the time
of the final follow-up, these scores had deteriorated slightly,
to an average of 76.3 points (range, 45 to 100 points) for pain
and an average of 58.7 points (range, 0 to 85 points) for function.
Range of Motion
The average arc of motion was 84.8° (range, 25° to 130°) preoperatively,
which improved to 94.3° (range, 50° to 125°) at two years postoperatively
(p < 0.0002). The average arc of motion had decreased to
86.7° (range, 40° to 125°) at the time of the latest follow-up (no
significant difference from the preoperative range of motion). Prior
to the index surgery, eleven knees had an extensor lag of 10°, including
one knee with an extensor lag of 25° and another with an extensor
lag of 32°. At the two-year follow-up evaluation, three
knees had an extensor lag of 10°. Over the years, the number of
knees with an extensor lag of 10° increased to ten. On the average, the
patients with heterotopic ossification had a significantly smaller
arc of motion than the patients without heterotopic ossification,
both preoperatively (p < 0.05) and at the time of the latest
follow-up (p < 0.007).
Pain
Prior to the total knee arthroplasty, two knees, both of which had
been replaced to treat marked flexion contracture and deformity,
were mildly painful; fifteen were moderately painful; and thirteen
were severely painful. At two years, the pain was rated as none
in twenty-four knees, mild in four knees, and moderate
in two knees. At the time of the final follow-up, pain was absent
in twenty-three knees, mild in five knees, moderate in
one knee, and severe in one knee.
Walking
Prior to the operation, two patients (two knees) could not walk,
six patients (ten knees) required a walker, and twelve patients
(eighteen knees) required a cane or crutches full-time. All patients
could walk at two years after the total knee arthroplasty. At the
time of the latest follow-up, there was a slight deterioration in
the walking status: two patients (two knees) could not walk, four
patients (seven knees) required a walker, three patients (five knees)
needed crutches or a cane full-time, eight patients (eleven
knees) required a cane or crutches part-time, and three
patients (five knees) were able to walk without any aids.
Radiographic Findings
A complete set of postoperative radiographs was available for all
thirty knees. The average tibiofemoral varus-valgus angulation
was 4.8° of valgus. The patella was located centrally in all thirty
knees. All nonrevised total knee arthroplasty components were well
fixed at the time of the latest follow-up. There was evidence of
nonprogressive radiolucent lines of £2 cm in two zones
adjacent to two tibial components and in one zone adjacent to one
femoral component.
Reoperations
Three patients (three knees) required manipulation of the knee in
an attempt to improve the range of motion. Compared with the values
before the manipulation, the maximum flexion at the time of the
latest follow-up was improved by 20° in one knee, 17° in one knee,
and 11° in the third knee. One knee required revision at ten years
because of aseptic loosening of the patellar component. There had
been radiographic evidence of avascular necrosis involving the patella
prior to the revision surgery in this patient. During the revision
surgery, the patellar component was confirmed to be grossly loose.
The tibial and femoral components were well fixed and did not require revision.
Complications
There was no deep venous thrombosis or deep infection in this
study. A superficial wound infection, however, developed in one
patient; it was treated successfully with oral antibiotics. A wound
hematoma developed in one patient, and it resolved following treatment
with a compression sleeve. Heterotopic ossification was noted on
the latest radiographs of six knees (four patients, two of whom
had bilateral total knee arthroplasty) (Figs. 1-A and 1-B). Only one of the
six patients had moderate symptoms (pain) at the time of the latest
follow-up. Heterotopic ossification developed following manipulation
of two knees. Three of the six knees with heterotopic ossification
had been operated on previously. The average erythrocyte sedimentation
rate was not significantly different between the patients in whom
heterotopic ossification developed and those in whom it did not.
In spite of the retrospective nature of this study and the relatively
small number of patients, we were able to make some important observations.
We found that total knee arthroplasty provides excellent and predictable
pain relief in patients with ankylosing spondylitis. Although the
walking and functional abilities of these patients also improve,
the gain in the range of motion is less optimal. We believe that
the inferior range of motion in these patients is related to three
factors: the nature of the disease in that it causes soft-tissue
contracture and joint ankylosis, the poor preoperative range of
motion, and the relatively high rate of heterotopic ossification.
Three patients in our series required manipulation of the knee because
of a poor postoperative range of motion. The rate of manipulation
of the knee following total knee arthroplasties performed by the
senior author (R.T.T.) for other diagnoses was approximately 3% (thirteen
manipulations after the most recent 500 primary knee arthroplasties).
The rate of postoperative knee manipulation in patients with ankylosing
spondylitis is therefore high when compared with that of patients
with osteoarthritis or rheumatoid arthritis.
The occurrence of heterotopic ossification in our patients with ankylosing
spondylitis is similar to that reported in patients who have undergone
total hip arthroplasty7,8,10,12,14.
Some degree of heterotopic ossification developed in 20% (four)
of our twenty patients over the years following the total knee arthroplasty.
Although heterotopic ossification was obviously responsible for
functional disability in only one of our patients, it seems intuitive
that heterotopic ossification had some deleterious effect on the
range of motion in the remainder of the patients.
We observed that the occurrence of heterotopic ossification was
more frequent in knees that had been operated on previously and
was more likely to develop when there was heterotopic ossification
in the contralateral knee. Heterotopic ossification developed following
manipulation in two knees. It is difficult to ascertain whether
the poor range of motion of the two knees requiring manipulation
was the cause or the consequence of heterotopic ossification. It
is plausible, however, that manipulation is another predisposing
factor for heterotopic ossification. Four patients (six knees) in
our study received ibuprofen perioperatively. Heterotopic ossification did
not develop in any of these patients, despite the fact that one
of them had had severe heterotopic ossification in the hip. It seems
that, after total knee arthroplasty, there is a major risk of clinically
important heterotopic ossification developing in patients who have
ankylosing spondylitis. Prophylaxis to prevent heterotopic ossification
should be considered in this patient subgroup, especially in patients
with previous ipsilateral knee surgery, severe preoperative stiffness,
or a history of heterotopic ossification elsewhere. In agreement
with the findings of a previous study12,
we also found that the erythrocyte sedimentation rate was not predictive
of the development of heterotopic ossification.
Patients with ankylosing spondylitis are known to have a compromised
vital capacity because of the involvement of costovertebral joints.
One might expect these patients to be at a greater risk for intraoperative
or postoperative pulmonary complications. However, we were surprised
to discover a low rate of pulmonary complications in our patients
with ankylosing spondylitis. Postoperative atelectasis developed
in only two patients, and it resolved with minimal intervention.
No other intraoperative or postoperative pulmonary complications
were observed. It is important to note that the awareness of a potentially
higher complication rate in patients with ankylosing spondylitis
may have encouraged the use of specific measures to minimize the
risk of these complications. For example, because of severe cervical
spine ankylosis, fiberoptic intubation was used in seven patients.
Also, compared with the general population, most patients required higher
average airway pressures for adequate ventilation during anesthesia.
In conclusion, total knee arthroplasty provided excellent pain relief
and a variable degree of improvement in function in patients with
ankylosing spondylitis. The result of the arthroplasty was surprisingly
durable in this group of young patients with severe spinal and joint
deformities. These patients had a higher rate of complications,
including development of heterotopic ossification and stiffness,
and at a mean of 11.2 years postoperatively they had minimal improvement
in the range of motion compared with the preoperative status. Total
knee arthroplasty can be performed successfully, with marked relief
of pain and improvement in function, in this highly debilitated
subset of patients.
Brinker MR, Rosenberg AG, Kull L,Cox DD. Primary noncemented total hip arthroplasty in patients
with ankylosing spondylitis. Clinical and radiographic results at
an average follow-up period of 6 years. J Arthroplasty,1996;11: 802-12. 11802
1996
[PubMed]
Forouzesh S,Bluestone R. The clinical spectrum of ankylosing spondylitis. Clin Orthop,1979;143: 53-8. 14353
1979
[PubMed]
Resnick D, Dwosh IL, Goergen TG, Shapiro RF, Utsinger PD, Wiesner KB,Bryan BL. Clinical and radiographic abnormalities in ankylosing spondylitis:
a comparison of men and women. Radiology,1976;119: 293-7. 119293
1976
[PubMed]
Resnick D. Patterns of peripheral joint disease in ankylosing spondylitis . Radiology,1974;110: 523-32.. 110523
1974
[PubMed]
West HF. Aetiology in ankylosing spondylitis. Ann Rheum Dis,1949;8: 143. 8143
1949
[PubMed]
Toni A, Baldini N, Sudanese A, Tigani D,Giunti A. Total hip arthroplasty in patients with ankylosing spondylitis with
a more than two year follow-up. Acta Orthop Belg,1987;53: 63-6. 5363
1987
[PubMed]
Baldursson H, Brattstrom H,Olsson T. Total hip replacement in ankylosing spondylitis. Acta Orthop Scand,1977;48: 499-507. 48499
1977
[PubMed]
Bisla RS, Ranawat CS,Inglis AE. Total hip replacement in patients with ankylosing spondylitis with
involvement of the hip. J Bone Joint Surg Am,1976;58: 233-8. 58233
1976
[PubMed]
DeLee J, Ferrari A,Charnley J. Ectopic bone formation following low friction arthroplasty
of the hip. Clin Orthop,1976;121: 53-9. 12153
1976
[PubMed]
Kilgus DJ, Namba RS, Gorek JE, Cracchiolo A 3rd,Amstutz HC. Total hip replacement for patients who have ankylosing spondylitis.
The importance of the formation of hetertopic bone and of the durability
of fixation of cemented components. J Bone Joint Surg Am,1990;72: 834-9. 72834
1990
[PubMed]
Shih LY, Chen TH, Lo WH,Yang DJ. Total hip arthroplasty in patients with ankylosing spondylitis: longterm
followup. J Rheumatol,1995;22: 1704-9. 221704
1995
[PubMed]
Sundaram NA,Murphy JC. Heterotopic bone formation following total hip arthroplasty
in ankylosing spondylitis. Clin Orthop,1986;207: 223-6. 207223
1986
[PubMed]
Walker LG,Sledge CB. Total hip arthroplasty in ankylosing spondylitis. Clin Orthop,1991;262: 198-204. 262198
1991
[PubMed]
Williams E, Taylor AR, Arden GP,Edwards DH. Arthroplasty of the hip in ankylosing spondylitis. J Bone Joint Surg Br,1977;59: 393-7. 59393
1977
[PubMed]
Finsterbush A, Amir D, Vatashki E,Husseini N. Joint surgery in severe ankylosing spondylitis. Acta Orthop Scand,1988;59: 491-6. 59491
1988
[PubMed]
Lu H, Mow CS,Lin J. Total knee arthroplasty in the presence of severe flexion
contracture: a report of 37 cases. J Arthroplasty,1999;14: 775-80. 14775
1999
[PubMed]
Moll JM,Wright V. New York clinical criteria for ankylosing spondylitis.
A statistical evaluation. Ann Rheum Dis,1973;32: 354-63. 32354
1973
[PubMed]
Insall JN, Dorr LD, Scott RD,Scott WN. Rationale of the Knee Society clinical rating system. Clin Orthop,1989;248: 13-4. 24813
1989
[PubMed]
Ewald FL. The Knee Society total knee arthroplasty roentgenographic evaluation
and scoring system. Clin Orthop,1989;248: 9-12. 2489
1989
[PubMed]