Abstract
Background: Stiffness is an uncommon but disabling problem after
total knee arthroplasty. The prevalence of stiffness after knee replacement
has not been well defined in the literature. In addition, the outcomes of
revision surgery for a stiff knee following arthroplasty have not been
evaluated in a large series of patients, to our knowledge. The purposes of
this study were to define the prevalence of stiffness after primary total knee
arthroplasty and to evaluate the efficacy of revision surgery for treatment of
the stiffness.
Methods: We defined a stiff knee as one having a flexion contracture
of =15° and/or <75° of flexion. Two separate groups were
evaluated. First, the results of 1000 consecutive primary total knee
replacements were reviewed to determine the prevalence of stiffness. Second,
the results of fifty-six revisions performed because of stiffness, sometimes
associated with pain or component loosening, after primary total knee
arthroplasty were evaluated.
Results: The prevalence of stiffness was 1.3%, at an average of
thirty-two months postoperatively. The patients with a stiff knee had had
significantly less preoperative extension and flexion than did those without a
stiff knee (p < 0.0001). There were no significant differences in age,
gender, implant design, diagnosis, or the need for lateral release between the
patients with and without stiffness. The second cohort, of knees revised
because of stiffness, were followed for an average of forty-three months. The
mean Knee Society score improved from 38.5 points preoperatively to 86.7
points at the time of follow-up; the mean Knee Society function score, from
40.0 to 58.4 points; and the mean Knee Society pain score, from 15.0 to 46.9
points. The mean flexion contracture decreased from 11.3° to 3.2°, the
mean flexion improved from 65.8° to 85.4°, and the mean arc of motion
improved from 54.6° to 82.2°. The arc of motion improved in 93% of the
knees, and flexion increased in 80%. Extension improved in 63%, and it
remained unchanged in 30%.
Conclusions: The prevalence of stiffness in our series of 1000
primary knee arthroplasties was 1.3%. Revision surgery was a satisfactory
treatment option for stiffness, as the Knee Society scores improved, the
flexion contractures diminished, and 93% of the knees had an increased arc of
motion. However, the results suggest that the benefits are modest.
Level of Evidence: Therapeutic study, Level IV (case
series [no, or historical, control group]). See Instructions to Authors for a
complete description of levels of evidence.
Stiffness after total knee arthroplasty, which has been variably referred
to as stiffness, arthrofibrosis, and limited motion, is a disabling
complication. The prevalence appears to be relatively low, but it has not been
clearly defined in the literature, to our
knowledge1-3.
Stiffness after total knee arthroplasty may be attributed to many factors,
including limited preoperative motion, a biological predisposition,
intraoperative technical problems, poor patient motivation, and inadequate
postoperative
rehabilitation1,4-9.
Sometimes the cause may not be identifiable.
Several treatments for stiffness have been reported with varying degrees of
success. Manipulation under anesthesia has been shown to be effective when it
is performed within six to twelve weeks after a primary total knee
arthroplasty10.
Other treatments include arthroscopic débridement of scar
tissue11,12,
release of the posterior cruciate
ligament13,14
(open or arthroscopic), and arthrolysis and exchange of the tibial
insert4,15.
We are aware of only a few studies of the outcomes of revision surgery for the
treatment of stiffness following total knee arthroplasty, and the sample sizes
were
small6,8,16.
The focus of the present study was to define and determine the prevalence
of stiffness after total knee arthroplasty. In addition, we sought to
determine the efficacy of revision surgery in patients with a stiff knee
following total knee arthroplasty.
The investigation was an open trial approved by the institutional review
board and performed at the University of Pennsylvania Medical Center,
Philadelphia, Pennsylvania.
We defined stiffness after total knee replacement as a flexion contracture
of =15° and/or <75° of flexion. These end points were selected
on the basis of gait analysis studies indicating an increased difficulty in
walking with increasing flexion contracture and that 67° of flexion is
required for normal
gait5.
Prevalence of Stiffness After Primary Total Knee Arthroplasty
In order to determine the prevalence of stiffness, we reviewed the data on
1000 primary total knee replacements (including nineteen bilateral procedures)
performed at our institution from 1997 to 2000. The arthroplasties were done
in 644 women and 337 men, with a mean age of fifty-six years (range,
twenty-six to seventy-three years) at the time of surgery. The diagnosis prior
to the total knee arthroplasty was osteoarthritis in 93% (932) of the knees,
rheumatoid arthritis in 5% (fifty-two), and osteonecrosis in 2% (sixteen). One
surgeon (P.A.L.) performed all of the arthroplasties with a consistent
technique17. A
posterior cruciate-substituting device was utilized in all knees. The average
duration of follow-up was thirty-four months (range, twenty-four months to 5.8
years). Preoperative, intraoperative, and follow-up data were analyzed. The
patients with stiffness following the arthroplasty were compared with those
who did not have stiffness at the most recent follow-up evaluation after the
arthroplasty.
Revision for Treatment of Stiffness
We reviewed the results of fifty-six of 545 revision arthroplasties that
had been performed by the senior surgeon (P.A.L.) from 1985 to 1999. These
fifty-six knees had stiffness prior to the revision surgery. Patients who had
undergone revision because of infection after primary total knee arthroplasty
were excluded from the study. The study group consisted of thirty-eight women
and fourteen men, with a mean age of sixty-nine years (range, thirty-six to
eighty-nine years) at the time of the index procedure. The diagnosis prior to
the primary total knee arthroplasty was osteoarthritis for 91% (fifty-one) of
the fifty-six knees and rheumatoid arthritis for 9% (five). The posterior
cruciate ligament, if present, was cut at the time of the revision surgery in
all patients, and a posterior cruciate-substituting prosthesis was used. All
knees were in 5° to 10° of valgus alignment prior to the revision
surgery. The range of motion prior to the primary total knee replacement was
analyzed for the nineteen knees in which the arthroplasty had been performed
at our institution.
Clinical data and Knee
Society18 scores
were analyzed at the preoperative visit before the revision, at the first
visit after the revision (at four to eight weeks postoperatively), and at the
final follow-up visit (at a mean of forty-three months [range, twenty-four
months to ten years]) for all fifty-six knees.
The femoral and tibial components were revised in all fifty-six knees. A
standard median parapatellar incision was used in all knees. Eight of the
knees required a quadriceps snip, one knee required a V-Y quadricepsplasty,
and four required a tibial tubercle osteotomy for exposure. A lateral
retinacular release was performed in six knees (11%). The intra-medullary
stems were cemented in all knees, and bone defects were filled with cement or
with impacted cancellous bone.
For stability, eight knees (14%) required the Press Fit Condylar TC3
prosthesis (DePuy [Johnson and Johnson], Warsaw, Indiana). Other implant
designs included the Press Fit Condylar Substituting prosthesis (DePuy
[Johnson and Johnson]) in fifteen knees, the Press Fit Condylar Sigma
prosthesis (DePuy [Johnson and Johnson]) in thirteen, the Scorpio prosthesis
(Stryker Howmedica Osteonics, Allendale, New Jersey) in seven, the Press Fit
Condylar Cruciate Retaining prosthesis (DePuy [Johnson and Johnson]) in eight,
the Insall-Burstein-II prosthesis (Zimmer, Warsaw, Indiana) in three, the
Genesis prosthesis (Smith and Nephew, Memphis, Tennessee) in one, and the
Posterior Stabilized prosthesis (DePuy [Johnson and Johnson]) in one.
Postoperatively, all patients were treated with a continuous-passive-motion
device set to move from 70° to 100° during the first night and then
from 0° to 100° for the duration of the hospital
stay17. Three knees
(5%) were placed in a cast in extension for ten days in order to prevent
recurrence of a large, fixed flexion contracture.
Statistical Methods
Prevalence of Stiffness After Primary Total Knee Arthroplasty
Differences in the range of motion between the preoperative and final
follow-up visits were analyzed in the series of 1000 primary arthroplasties.
Confidence intervals were determined for the prevalence of stiffness after
total knee arthroplasty. Two-sample t tests were performed to compare the mean
age and preoperative and postoperative ranges of motion between the patients
with postoperative knee stiffness and those without it. This method of
analysis assumes that the data are symmetrically distributed around the mean.
The data were examined for violations of this assumption. In addition,
Wilcoxon rank-sum tests, which require no parametric assumption, were
performed. Because of sparse cell counts, the Fisher exact test was used in
the analysis of gender, type of prosthesis, preoperative diagnosis, and
lateral release in the patients with knee stiffness.
Revision for Treatment of Stiffness
In the revision series, statistical analysis-of-covariance models were
performed to examine the mean change between the preoperative and
postoperative ranges of motion in relation to the institution at which the
surgery was performed, the preoperative diagnosis, the design of the revision
implant, and whether a lateral retinacular release, quadriceps snip, or tibial
tubercle osteotomy was necessary. The analysis-of-covariance model adjusted
for the baseline extension, flexion, or arc of motion value at the
preoperative visit. The nonparametric analogue to analysis of variance, the
Kruskal-Wallis test, was also used to compare knees according to the various
factors that may have affected outcome.
Statistical analysis also included a series of paired t tests to compare
the findings at the preoperative visit with those at the final follow-up visit
and to compare the findings at the first follow-up visit with those at the
final follow-up visit in the revision series. This method of analysis assumes
that the data are symmetrically distributed around the mean. The data were
examined for significant violations of this assumption. In addition, Wilcoxon
signed-rank tests, which require no parametric assumption, were performed. All
analyses were performed with use of SAS software (version 8.1; SAS Institute,
Cary, North Carolina) and S-PLUS software (version 6.0; MathSoft, Seattle,
Washington).
Prevalence of Stiffness After Primary Total Knee Arthroplasty
The prevalence of stiffness following the 1000 consecutive primary total
knee arthroplasties was 1.3% (thirteen knees). Of those thirteen knees, ten
had a flexion contracture of >10°, and the remaining three had a
contracture of 10° and also had severe limitation of flexion
(Fig. 1). The thirteen stiff
knees had had significantly less extension and flexion at the preoperative
visit than did the knees without stiffness (p < 0.0001)
(Table I). Preoperatively, the
stiff knees had a mean extension (and standard deviation) of 24.6°
± 10.1° and a mean flexion of 75.0° ± 23.7°, whereas
the non-stiff knees had a mean extension of 2.6° ± 3.4° and a
mean flexion of 108.4° ± 11.5°. Postoperatively, the stiff
knees had a mean extension of 16.9° ± 6.6° and a mean flexion
of 76.5° ± 22.5°, whereas the non-stiff knees had a mean
extension of 0.2° ± 0.98° and a mean flexion of 111.7°
± 9.9°.
A lateral retinacular release had been performed in 24% (232) of the 987
knees that were not stiff following the primary total knee arthroplasty. All
987 knees were in 5° to 10° of valgus alignment at the most recent
office visit. The Press Fit Condylar TC3 implant had been used in 1.6%
(sixteen) of the 987 knees; the Press Fit Condylar Sigma implant, in 40%
(395); the Scorpio implant, in 26% (253); the Legacy Posterior Stabilized
implant (Zimmer), in 3.5% (thirty-five); and the Press Fit Condylar
Substituting implant, in 29% (288).
Of the thirteen stiff knees, eleven had a diagnosis of osteoarthritis and
two had a diagnosis of rheumatoid arthritis prior to the primary total knee
arthroplasty. A lateral retinacular release had been performed in two knees.
All thirteen knees were in 5° to 10° of valgus alignment at the most
recent office visit. Two Press Fit Condylar TC3, five Press Fit Condylar
Sigma, two Scorpio, and four Press Fit Condylar Substituting prostheses had
been used in the thirteen knees. No significant differences were found between
the patients with knee stiffness and those without knee stiffness with regard
to age (p = 0.3371), gender (p = 0.1314), implant design (p = 0.0892),
preoperative diagnosis (p = 0.3095), or the need for a lateral retinacular
release (p = 0.7435).
All thirteen of the stiff knees were manipulated at least once within three
months after the surgery, and four were manipulated twice. Of the four knees
that were manipulated twice, one underwent arthroscopic débridement
because of persistent stiffness after the two manipulations, and the range of
motion improved from 30° to 80° preoperatively to 10° to 110°
postoperatively. Six knees eventually underwent revision because of continued
stiffness, and the improvements in the range of motion were from 30° to
90° preoperatively to 0° to 105° postoperatively, from 10° to
70° preoperatively to 5° to 100° postoperatively, from 10° to
70° preoperatively to 0° to 90° postoperatively, from 10° to
55° preoperatively to 0° to 95° postoperatively, from 15° to
35° preoperatively to 5° to 70° postoperatively, and from 15°
to 35° preoperatively to 0° to 85° postoperatively.
Revision Surgery for Treatment of Stiffness
Fifty-six knees in fifty-two patients underwent revision surgery for
stiffness at a mean of eighteen months (range, nine months to 5.2 years) after
the primary total knee arthroplasty. Prior to the revision surgery,
twenty-nine knees (52%) had undergone manipulation and one had been treated
with arthroscopic débridement.
There was significant improvement in the Knee Society functional and
clinical scores following the revisions (p < 0.001). The mean pain score
improved from 15.0 ± 11.1 to 46.9 ± 7.54 points; the mean knee
score, from 38.5 ± 18.8 to 86.7 ± 12.3 points; and the mean
overall function score, from 40.0 ± 19.1 to 58.4 ± 17.4
points.
There was also a significant improvement in extension, flexion, and the arc
of motion after the revision surgery (p < 0.001). Thirty-five knees (63%)
had a decreased flexion contracture, and forty-five knees (80%) had increased
flexion. Fifty-two knees (93%) had a gain in the arc of motion: 4% gained
5° to 9°; 23%, >10° to 19°; and 66%, =20°
(Fig. 2). From the preoperative
to the final follow-up visit, the mean flexion contracture decreased from
11.3° ± 11.9° to 3.2° ± 5.4°, the mean flexion
increased from 65.8° ± 24.0° to 85.4° ± 20.5°,
and the mean arc improved from 54.6° ± 20.5° to 82.2°
± 21.3°. The mean gains in extension, flexion, and the total arc
were 8.1°, 19.6°, and 27.6°, respectively (p < 0.001)
(Table II). No significant
differences in the ranges of motion were found between the first follow-up
visit and the final follow-up visit (see Appendix).
At the time of final follow-up, four knees (7%) had an increased flexion
contracture, seven knees (13%) had decreased flexion, and three knees (5%) had
a decreased arc of motion compared with the values at the preoperative visit.
Three of the four knees that had decreased extension also had a gain in
flexion and in the arc of motion; only one of the four knees had a decrease in
the arc of motion. Five of the seven knees that had decreased flexion also had
improved extension, and four had an increased arc of motion (see Appendix).
Seventeen (30%) of the knees that did not have an improvement in extension
already had had full extension prior to the surgery.
We attempted to define the reason for stiffness. Eighteen (32%) of the
knees had loosening of the femoral or tibial component, ten (18%) appeared to
have oversized components, three (5%) had heterotopic ossification, one was in
a patient with juvenile rheumatoid arthritis, and one knee had had a
preoperative tibial fracture. There was no clearly identified cause for the
stiffness of the remaining knees. With the numbers available, no significant
association was found between the mean change in the range of motion and the
institution at which the surgery was performed, the supposed preoperative
etiology, the design of the revision implant, or whether a lateral retinacular
release, quadriceps snip, or tibial tubercle osteotomy had been necessary (see
Appendix).
Four knees underwent manipulation under anesthesia within three months
after the revision. Three of those knees gained at least 10° in the arc of
motion, and only one did not undergo any change. Two knees were rerevised:
one, fifteen months after the first revision because of recurrent pain and
stiffness, and the other, ten years after the first revision because of
loosening and wear. The knee that was rerevised because of continued stiffness
gained 5° in the arc of motion.
This study showed that approximately 1.3% of 1000 knees have limited motion
after primary total knee replacement. There is ambiguity in the literature
concerning the definition of stiffness after total knee arthroplasty. Nicholls
and Dorr1 defined
stiffness as a flexion contracture of >20° or a total range of motion
of <45°. However, this definition fails to account for knees with
inadequate flexion.
Scranton3 defined
stiffness as flexion of <85°, and Christensen et
al.2 defined it as
an arc of motion of <70°. Neither definition includes patients who are
restricted by a flexion contracture alone. Therefore, we defined stiffness
after total knee replacement as a flexion contracture of =15° and/or
<75° of flexion.
Numerous types of treatment for stiffness after total knee arthroplasty,
including manipulation under
anesthesia10,13,19,20,
arthrolysis and exchange of the tibial
insert4,15,16,
arthroscopic release of the posterior cruciate
ligament13, removal
of the posterior cruciate
ligament16, and
revision
arthroplasty1-3,16,
have been evaluated and have shown variable results. Manipulation under
anesthesia is generally the initial option for patients with stiffness, and
some improvement can be
expected19,20.
Arthrolysis and exchange of the tibial insert generally has had poor
outcomes4.
Arthroscopic release of the posterior cruciate ligament for the treatment of
stiffness and pain can be of value if the retained posterior cruciate ligament
is
secure13,16.
Revision arthroplasty for the treatment of stiffness has had satisfactory
outcomes, although an optimal range of motion has not been
attained1-3,16.
As noted in the
literature21,22,
the preoperative range of motion is a good predictor of the final range of
motion. In our review of the results of 1000 primary total knee
arthroplasties, the thirteen knees that had stiffness postoperatively had had
limited motion preoperatively. Therefore, patients with loss of motion prior
to a primary total knee arthroplasty should be advised of the possibility of
diminished motion postoperatively.
Our evaluation of revision surgery for stiffness after total knee
arthroplasty demonstrated improvements in the range of motion as well as the
Knee Society functional and clinical scores. In addition, we found the range
of motion at the first follow-up visit after the revision surgery to be a
valuable indicator of long-term motion. Patients who have continued pain or
restriction of motion after surgery are unique and should be identified early
and watched closely by the surgeon.
The cause of stiffness after total knee arthroplasty remains unclear. Some
patients may be predisposed to it
biologically21-23.
Alternatively, there may be a series of technical errors that contribute to
postoperative loss of motion, including so-called overstuffing of the knee
with a too-large polyethylene spacer or using too-large a prosthesis,
decreasing the flexion space, or overtightening the patellar
retinaculum1,4-9.
One weakness of our study was that we did not stratify all of the knees
according to diagnosis. We were unable to do so because the reasons for
stiffness remain ambiguous or unidentified.
In conclusion, there is a relatively low risk (1.3%) of stiffness after
total knee arthroplasty. Revision surgery appears to be a reasonable option
for patients who have pain and stiffness after total knee replacement.
Nevertheless, the gains may be modest as the results do not approach those
usually achieved with a primary arthroplasty. Although the flexion
contractures were significantly reduced and 93% of the knees had a gain in the
arc of motion, the final mean arc of motion was only 82°.
Graphs and tables showing the changes in the range of motion between the
preoperative and postoperative evaluations and a table presenting an analysis
of various factors that might affect the range of motion gained after revision
surgery are available with the electronic versions of this article, on our web
site at
(go to the article citation and click on "Supplementary Material")
and on our quarterly CD-ROM (call our subscription department, at
781-449-9780, to order the CD-ROM).
Nicholls DW, Dorr LD. Revision
surgery for stiff total knee arthroplasty. J
Arthroplasty.1990;5 Suppl:
S73-7.5S73
1990
[CrossRef]
Christensen CP, Crawford JJ, Olin MD,
Vail TP. Revision of stiff total knee arthroplasty. J
Arthroplasty.2002;17:
409-15.17409
2002
[CrossRef]
Scranton PE Jr. Management of
knee pain and stiffness after total knee arthroplasty. J
Arthroplasty.2001;16:
428-35.16428
2001
[CrossRef]
Babis GC, Trousdale RT, Pagnano MW,
Morrey BF. Poor outcomes of isolated tibial insert exchange and
arthrolysis for the management of stiffness following total knee arthroplasty.
J Bone Joint Surg Am.2001;83:
1534-6.831534
2001
[PubMed]
Jordan LR, Olivo JL. Stiffness
and limitation of motion: prevention and management. In: Lotke PA, Garino JP,
editors. Revision total knee arthroplasty.
Philadelphia: Lippincott-Raven; 1999. p
461-7.461
1999
Kawamura H, Bourne RB. Factors
affecting range of flexion after total knee arthroplasty. J Orthop
Sci.2001;6:
248-52.6248
2001
[CrossRef]
Coutts RD, Engh GA, Mayor MB,
Whiteside LA, Wilde AH. The painful total knee replacement and the
influence of component design. Contemp Orthop.1994;28:
523-36,541-5.28523
1994
[PubMed]
Schurman DJ, Parker JN, Ornstein
D. Total condylar knee replacement. A study of factors influencing range
of motion as late as two years after arthroplasty. J Bone Joint
Surg Am.1985;67:
1006-14.671006
1985
Papagelopoulos PJ, Sim FH.
Limited range of motion after total knee arthroplasty: etiology, treatment,
and prognosis. Orthopedics.1997;20:
1061-7.201061
1997
[PubMed]
Brassard MF, Insall JN, Scuderi
GR. Complications of total knee arthroplasty. Insall JN, Scott WN,
editors. Surgery of the knee. 3rd ed. New York:
Churchill Livingstone; 2001. p
1801-44.1801
2001
Djian P, Christel P, Witvoet J.
[Arthroscopic release for knee joint stiffness after total knee arthroplasty].
Rev Chir Orthop Reparatrice Appar Mot.2002;88: 163-7.
French.88163
2002
[PubMed]
Court C, Gauliard C, Nordin JY.
[Technical aspects of arthroscopic arthrolysis after total knee replacement].
Rev Chir Orthop Reparatrice Appar Mot.1999;85: 404-10.
French.85404
1999
[PubMed]
Williams RJ 3rd, Westrich GH, Siegel
J, Windsor RE. Arthroscopic release of the posterior cruciate ligament for
stiff total knee arthroplasty. Clin Orthop.1996;331:
185-91.331185
1996
[PubMed][CrossRef]
Hirsch HS, Lotke PA, Morrison LD.
The posterior cruciate ligament in total knee surgery. Save, sacrifice, or
substitute? Clin Orthop.1994;309:
64-8.30964
1994
[PubMed]
Engh GA, Koralewicz LM, Pereles
TR. Clinical results of modular polyethylene insert exchange with
retention of total knee arthroplasty components. JBone Joint Surg
Am.2000;82:
516-23.82516
2000
Ries MD, Badalamente M.
Arthrofibrosis after total knee arthroplasty. Clin
Orthop.2000;380:
177-83.380177
2000
[CrossRef]
Lotke PA. Primary total knees:
standard principles and procedures. In: Lotke PA, Lonner JH, editors.
Knee arthroplasty. Philadelphia: Lippincott Williams
and Wilkins; 2003. p 49-72.49
2003
Insall JN, Dorr LD, Scott RD, Scott
WN. Rationale of the Knee Society clinical rating system. Clin
Orthop.1989;248:
13-4.24813
1989
Maloney WJ. The stiff total knee
arthroplasty: evaluation and management.
JArthroplasty.2002;17(4 Suppl 1):
71-3.1771
2002
[CrossRef]
Fox JL, Poss R. The role of
manipulation following total knee replacement. JBone Joint Surg
Am.1981;63:
357-62.63357
1981
Ritter MA, Stringer EA.
Predictive range of motion after total knee replacement. Clin
Orthop.1979;143:
115-9.143115
1979
Ritter MA, Campbell ED. Effect of
range of motion on the success of a total knee arthroplasty. J
Arthroplasty.1987;2:
95-7.295
1987
[CrossRef]
Kim JM, Moon MS. Squatting
following total knee arthroplasty. Clin Orthop.1995;313:
177-86.313177
1995
[PubMed]