Radiolucent lines and component stability in
knee arthroplasty
STANDARD VERSUS FLUOROSCOPICALLY-ASSISTED
RADIOGRAPHS
P. Vyskocil, C. Gerber, P. Bamert
From the General Hospital of Solothurn and the University of Zürich, Switzerland
he radiolucent lines and the stability of the
components of 66 knee arthroplasties were
assessed by six orthopaedic surgeons on conventional
anteroposterior and lateral radiographs and on
fluoroscopic views which had been taken on the same
day. The examiners were blinded as to the patients
and clinical results. The interpretation of the
radiographs was repeated after five months.
On fluoroscopically-assisted radiographs four of the
six examiners identified significantly more radiolucent
lines for the femoral component (p < 0.05) and one
significantly more for the tibial implant. Five
examiners rated more femoral components as
radiologically loose on fluoroscopically-assisted
radiographs (p = 0.0008 to 0.0154), but none did so for
the tibial components. The mean intra- and
interobserver kappa values were higher for
fluoroscopically-assisted radiographs for both
components.
We have shown that fluoroscopically-assisted
radiographs allow more reproducible, and therefore
reliable, detection of radiolucent lines in total knee
arthroplasty. Assessment of the stability of the
components is significantly influenced by the
radiological technique used. Conventional radiographs
are not adequate for evaluation of the stability of total
knee arthroplasty and should be replaced by
fluoroscopically-assisted films.
T
J Bone Joint Surg [Br] 1999;81-B:24-6.
Received 18 June 1998; Accepted after revision 17 August 1998
P. Vyskocil, MD, Orthopaedic Surgeon
C. Gerber, MD, Professor and Chairman
Department of Orthopaedics, University of Zürich, Balgrist 8008, Zürich,
Switzerland.
P. Bamert, MD, Head of Department
Department of Orthopaedics, General Hospital of Solothurn, Bürgerspital
4500, Solothurn, Switzerland.
Conventional radiographs remain in standard clinical use
although cadaver studies have shown that they are not
1
reliable in identifying interfaces reliably and in the assessment of radiological stability in knee arthroplasty. This is
because a tilt of the beam of only 2.3° to the plane of a
50 mm wide tibial component will obliterate a radiolucent
line 2 mm thick (Fig. 1). We have compared the use of
conventional and fluoroscopically-assisted radiographs of
knee prostheses to determine whether such standardised
radiographs would allow better reproducibility of their
interpretation and lead to a different appraisal of the frequency of total joint loosening.
Patients and Methods
We assessed 66 Freeman-Samuelson-Modular knee prostheses (Protek, Baar, Switzerland) which had been implanted between 1989 and 1994. Thirteen patients had bilateral
replacements. Fifteen patients were male and 38 were
female. The mean age at implantation was 74 years (55 to
90) and the mean follow-up was 41 months (12 to 78).
We took conventional, routine anteroposterior and lateral
radiographs, followed immediately by fluoroscopicallyassisted anteroposterior and lateral views, on a standard
fluoroscopic screening table with a moveable tube and an
image intensifer. The knee and the tube were moved until
the surfaces of the prostheses were parallel to the X-ray
beam. All the radiographs were evaluated by six orthopaedic surgeons who were blinded as to the patients and the
results. Conventional and fluoroscopically-assisted radiographs were assessed randomly by different panels for the
appearance of radiolucent lines in different parts of the
prosthesis (Fig. 2); each examiner determined whether he
thought that the components were radiologically stable or
loose. The interpretation of the original films was repeated
after five months.
We performed statistical analysis using the Wilcoxon test
for radiolucent lines and stability and the unweighted
kappa test for intra- and interobserver variability.
Correspondence should be sent to Dr P. Vyskocil at the Department of
Orthopaedics, General Hospital of Winterthur, Brauestrasse 15, 8401
Winterthur, Switzerland.
Results
©1999 British Editorial Society of Bone and Joint Surgery
0301-620X/99/19213 $2.00
Each examiner identified more radiolucent lines on fluoroscopically-assisted radiographs of both components (Table
24
THE JOURNAL OF BONE AND JOINT SURGERY
RADIOLUCENT LINES AND COMPONENT STABILITY IN KNEE ARTHROPLASTY
25
Fig. 1
A radiolucent line 2 mm thick is obscured on the radiograph if the central beam is tilted 2.3° to a tibial component 50 mm wide ( = invtan 2/50).
Table I. Radiolucent lines identified by the six examiners on conventional and fluoroscopically-assisted radiographs for the 66 femoral and
tibial components
Fig. 2
The different parts of the tibial (a) and femoral (b) component which were
assessed for radiolucent lines.
I). Four of the six examiners identified significantly more
radiolucent lines on these views of the femoral component
(p < 0.05) and one on those of the tibial component. Five
examiners rated loosening as higher for the femoral component on fluoroscopically-assisted radiographs (p = 0.0008
to 0.015). Conversely, on these no examiner assessed the
rate of loosening of the tibial components as higher. Intraobserver and interobserver reliability of evaluation of the
stability was better for the fluoroscopically-assisted radiographs for both components (Fig. 3). The mean intraobserver agreement on fluoroscopic radiographs was good
for the femoral component, but only moderate for the tibial
implant. On conventional radiographs intraobserver agreement for both components was only moderate. The mean
interobserver agreement on fluoroscopic radiographs was
moderate for the femoral and fair for the tibial component;
Fig. 3a
Examiner
Conventional
Fluoroscopicallyassisted
Difference
Femoral component
1
2
3
4
5
6
161
84
134
157
88
95
193
136
176
197
112
115
32
52
42
40
24
20
Tibial component
1
2
3
4
5
6
93
45
73
53
124
66
114
61
81
61
134
75
21
16
8
8
10
9
on conventional radiographs it was fair for the femoral and
poor for the tibial prosthesis with a kappa value of 0.17.
Discussion
1
Using a cadaver model, Mintz et al compared plain and
fluoroscopically-assisted radiographs in the assessment of
knee arthroplasty. They showed that the latter allowed
accurate measurement of lucent lines as small as 1 mm, but
plain radiographs were inadequate for their detection or
measurement. Fluoroscopically-assisted radiographs also
allowed measurement of the distance between the tibial
component and radiopaque markers in the proximal part of
Fig. 3b
The mean kappa values for intra- (a) and interobserver (b) variability for stability of the femoral and tibial components on conventional and
fluoroscopically-assisted radiographs.
VOL. 81-B, NO. 1, JANUARY 1999
26
P. VYSKOCIL, C. GERBER, P. BAMERT
the metaphysis which was reproducible to within 0.5 mm.
Plain radiographs could not achieve this. They therefore
recommended fluoroscopically-assisted radiography for the
detection of the presence and progression of radiolucent
lines in the tibial component. Our trigonometric calculations (Fig. 1) confirmed that a deviation of the central Xray beam of only 2.3° to the interface (the exact value
depending on the width of the prosthesis) is sufficient to
obscure a radiolucent line 2 mm wide. This agrees with the
2
findings of Magee and Weinstein who showed that a plain
radiograph with a divergence of the central beam of only 3°
from a plane parallel to the bone-implant interface will not
detect a lucent line of 2 mm beneath the tibial component.
3
Fehring and McAvoy reported that in 14 of 20 patients the
diagnosis of aseptic loosening could only be made by
fluoroscopically-assisted radiographs. They confirmed the
radiological interpretation at surgical revision, thereby
showing the clinical relevance of guided radiographs. None
the less, conventional radiographs remain the clinical standard although their accuracy for the detection of lucencies is
not well established and the reproducibility of the interpretation is uncertain.
Our results indicate that assessment of the femoral interface is significantly more sensitive to loosening zones and
more reproducible on fluoroscopically-assisted than on conventional radiographs. The lack of a significant difference
for the assessment of the stability of the tibial prostheses is,
however, surprising since, in comparison with the femoral
components, these were rated loose four times less often.
There were probably too few radiologically loose components to detect a significant difference. Since our tibial
prosthesis had three pegs, for the assessment of three of
seven parts of the component conventional radiographs
were adequate for the interpretation of stability. There was
better intraobserver reliability for the tibia when using
fluoroscopically-assisted imaging. Brand, Yoder and Ped4
ersen studied interobserver variability in interpreting
lucencies around total hip replacements. All three observers
agreed on the assessment of acetabular lucencies in only 26
hips (46%) and disagreed in 8 (13%). They found a significant interobserver variability in interpreting these
appearances. The problems with interpreting radiographs
which have not been taken in a standard manner and the
significance of the definition of ‘loosening’ have been
5
discussed by Brand, Pedersen and Yoder.
Our study shows that the same principles apply for total knee
arthroplasty. The routine use of fluoroscopically-assisted radiographs could improve sequential supervision of an individual
case and substantially improve the interobserver reliability.
No benefits in any form have been received or will be received from a
commercial party related directly or indirectly to the subject of this
article.
References
1. Mintz AD, Pilkington CAJ, Howie DW. A comparison of plain and
fluoroscopically guided radiographs in the assessment of arthroplasty
of the knee. J Bone Joint Surg [Am] 1989;71-A:1343-7.
2. Magee FP, Weinstein AM. The effect of position in the detection of
radiolucent lines beneath the tray. Trans Orthop Res Soc 1996;11:
357.
3. Fehring TK, McAvoy G. Fluoroscopic evaluation of the painful total
knee arthroplasty. Clin Orthop 1996;331:226-33.
4. Brand RA, Yoder SA, Pedersen DR. Interobserver variability in
interpreting radiographic lucencies about total hip reconstructions.
Clin Orthop 1985;192:237-9.
5. Brand RA, Pedersen DR, Yoder SA. How definition of “loosening”
affects the incidence of loose total hip reconstructions. Clin Orthop
1986;210:185-91.
THE JOURNAL OF BONE AND JOINT SURGERY