Currently, there are three indications for the exchange of a
polyethylene liner: (1) polyethylene wear-induced synovitis, (2)
osteolysis, and (3) impending wear-through of the polyethylene liner.
As the longevity of hip replacement components has increased, the
number of asymptomatic patients who have potential wear-through
of the polyethylene liner at the time of clinical follow-up also has
increased. Because any revision surgery carries the potential for
complications, such as dislocation, it is not desirable to exchange
a polyethylene liner until it is absolutely necessary. Similarly,
because complete wear-through of the polyethylene liner also is
associated with complications, including the production of metal
debris from unintended metal-on-metal articulation, the ability of
a surgeon to correctly time the exchange of a polyethylene liner
is extremely important.
At our institution, we use computer-assisted radiographic techniques
to assess the thickness of the polyethylene liner in asymptomatic
patients who have gross wear of the polyethylene component. In previous
studies, we verified the accuracy of our digital technique with use
of clinical radiographs1,2. Given
this accuracy, we developed a protocol for the timing of polyethylene
liner exchange on the basis of computer-generated radiographic penetration
data. We are aware, however, that most orthopaedic surgeons do not
have access to computer-assisted radiographic techniques or to other
recently developed sophisticated systems. Typically, clinicians
assess femoral head penetration and polyethylene wear by manually
measuring radiographs; however, the accuracy of manual methods has
not been verified with use of clinical radiographs and is likely
to be inferior to that of contemporary computer-assisted techniques2,3. Thus, the usefulness of manual
methods for assessing component wear-through or for determining
the timing of polyethylene-liner exchange remains uncertain.
Recently, we developed a new manual technique whereby a clinician
can use a wear template to measure the minimum polyethylene thickness
on an anteroposterior pelvic radiograph. The template, developed
by the manufacturer of the acetabular component, depicts a cross-sectional
view of the cup and illustrates the thickness of the metal shell.
By placing the template over a follow-up radiograph, a surgeon can
simply measure the distance between the outer edge of the femoral
head and the inner edge of the metal shell to determine the thickness
of the polyethylene remaining in the liner.
In the present study, we examined the accuracy of this new manual
technique along with that of two other commonly used manual radiographic
methods. The purposes of the study were (1) to quantify the clinical accuracy
of two commonly used manual radiographic methods3,4;
(2) to quantify the clinical accuracy of the wear-template method
and to compare it with that of the previously described methods;
and (3) to determine the usefulness of the three methods for assessing
component wear-through on the basis of their accuracy.
Seventeen explanted polyethylene liners with corresponding prerevision
clinical radiographs were available for analysis. The liners were
removed, after a mean of 12.0 years (range, 5.9 to 14.9 years) in
situ, from seventeen patients who had a revision total
hip arthroplasty because of osteolysis or polyethylene wear. The
revision operation was a simple polyethylene liner exchange in sixteen
hips and a full cup revision in one hip. The acetabular components
included eleven Arthropor cups (Joint Medical Products, Stamford,
Connecticut), three Duraloc cups (DePuy [Johnson and Johnson],
Warsaw, Indiana), and three so-called Supercups (Joint Medical Products).
All cups had a hemispherical polyethylene liner.
A single observer directly measured the true polyethylene deformation
of the explanted liner with use of digital calipers (Digimatic Caliper;
Mitutoyo, Kanagawa, Japan). Because the liners were not marked when
they were removed at the time of revision, their exact orientation in
situ was not known. By probing the articulating surface
of the polyethylene liner with the calipers, the observer determined
the location of minimum thickness. The thickness of the polyethylene
liner at this location was then measured three times, and the average
value was recorded. In three cases in which the femoral head had actually
worn completely through the central portion of the liner, the minimum
thickness was recorded as zero. This direct measurement of minimum
polyethylene thickness was used for comparison with the radiographic measurements.
An observer who was blinded to the direct measurements then measured
the polyethylene deformation on anteroposterior pelvic radiographs
that had been made at a mean of 2.1 months (range, one day to 8.7
months) before the revision. Three manual techniques were used to
evaluate the radiographs. The first two techniques, the Livermore
method4 and the Dorr method3, have been extensively discussed
in the orthopaedic literature and, therefore, will only be described
briefly here. The third technique, a new method involving the use
of wear templates designed by the manufacturer, will be described
in greater detail.
With the Livermore method4,
the most recent clinical radiograph (in this case, the anteroposterior
pelvic radiograph made before the revision) was examined first.
Using a template of concentric circles and a compass, the observer
determined the location of the shortest radius from the center of
the femoral head to the outer surface of the acetabular cup (Fig. 1, A; line
OA). This line was defined as the line of maximum head penetration.
The observer then measured the distance along this line between
the surface of the femoral head and the outer surface of the acetabular
cup with a caliper (Fig. 1, A; line A"A).
Next, these lines were located on the initial postoperative anteroposterior
pelvic radiograph in the same position on the acetabular cup as
they had appeared on the prerevision radiograph, and the same measurements
were made. All measurements were corrected for magnification with
use of the known diameter of the femoral head. The amount of femoral
head penetration was then defined as the difference between the measurements
on the immediate postoperative and prerevision clinical radiographs.
With the Dorr method3, only
the most recent clinical radiograph (again, a prerevision anteroposterior
pelvic radiograph) was examined. A line was drawn from the superior
edge to the inferior edge of the metal acetabular cup. The observer measured
the distance from the superior margin of the acetabular component
to the edge of the femoral head (Fig. 1, B; line S"S)
as well as the distance from the inferior margin of the acetabular
component to the edge of the femoral head (line I"I). Linear
head penetration was defined as half of the difference between those
measurements. Again, all measurements were corrected for magnification
with use of the known diameter of the femoral head.
With both the Dorr and the Livermore technique, it was necessary
to contact the manufacturer of the acetabular component in order
to obtain the original thickness of the polyethylene for each cup
size and design. With this information, we calculated the minimum
polyethylene thickness as the difference between the original thickness
of the polyethylene liner (supplied by the manufacturer) and the
amount of femoral head penetration (determined radiographically).
The new technique involved the use of wear templates that were
developed for each component design and size. (Wear templates are
available for Joint Medical Products and DePuy acetabular components
and can be obtained by contacting DePuy, a Johnson and Johnson company;
other manufacturers’ acetabular component templates are
currently being developed.) These templates provided a cross-sectional
view of the cup that showed the thickness of the metal backing and
the original position of the femoral head (Fig. 2). Like preoperative
planning templates, the wear templates were created at 20% magnification
to match the approximate magnification of the radiograph. To assess polyethylene
thickness, the observer first drew a line on the prerevision anteroposterior
radiograph through the major axis of the ellipse that represented
the opening of the cup face (Fig. 3-A). The observer then placed the
appropriate template on top of the radiograph, using this line and
the outer surface of the metal shell as guides (Fig. 3-B). In cases
in which the femoral head was not completely visible, a template
of the femoral head was also placed over the radiograph to aid in
visualization of the edge of the head (Fig. 4-A). Once the templates were correctly
oriented, the observer measured the shortest distance between the edge
of the femoral head and the inside of the metal shell (Figs. 3-B [short
arrows] and 4-B [arrows]) three
times with a digital caliper and recorded the average value. This
distance, after correction for the 20% template magnification,
was defined as the minimum polyethylene thickness.
Statistical Methods
With use of linear regression analysis, the minimum polyethylene
thickness that had been calculated with each of the manual radiographic
techniques was compared with the minimum thickness that had been
determined by direct measurement of the explanted liner. Next, differences
between the radiographic and direct measurements were calculated
and examined with use of the statistical method of Bland and Altman5. This method, which is used to analyze
the mean difference between two measurements and the standard deviation
of the differences, was developed to assess agreement between two
methods of clinical measurement. A one-sample t test was used to
determine whether mean differences were significantly different
from zero. One-way analysis of variance with a post hoc Tukey test was
used to determine whether the mean differences (that is, the mean
error for each of the methods) differed among the three radiographic
techniques.
The minimum polyethylene thickness averaged 1.37 ±
1.06 mm with direct measurement, 2.91 ± 1.27 mm
with the Dorr radiographic technique, 1.44 ± 1.44
mm with the Livermore radiographic technique, and 1.33 ±
1.14 mm with the wear-template technique.
Linear regression analysis demonstrated a poor relationship between
the thickness measured directly and the thickness measured with
the Dorr technique (r2 = 0.22, Table I). The mean
difference between the radiographic measurements and the direct
measurements (that is, the error of the Dorr technique) was 1.54 ±
1.21 mm (Table II).
This value was significantly different from zero (p < 0.01),
indicating a bias for the radiographic measurement to significantly
overestimate the true thickness by an average of 1.54 mm.
A stronger relationship was found between the thickness measured
directly and the thickness measured with the Livermore technique
(r2 = 0.85, Table I). The mean
difference between the radiographic measurements and the direct
measurements (that is, the error of the Livermore technique) was
0.07 ± 0.62 mm (Table II). This value was not significantly
different from zero (p = 0.64), indicating that the radiographic
measurement did not consistently underestimate or overestimate the direct
measurement. With the assumption that these difference values were
derived from a normally distributed population, it can be expected
that 95% of all difference values for this radiographic
technique will fall within two standard deviations (1.24 mm) of
the mean value (0.07 mm). From this range, referred to as "the
limits of agreement,"5 it
can be inferred that, in 95% of all cases, the error of the
Livermore technique will fall between -1.17 and +1.31 mm.
The strongest relationship was found between the measurements
made with the wear templates and those made directly (r2 = 0.94, Table I). The mean
difference between the template measurements and the direct measurements
was -0.04 ± 0.28 mm (Table II), which was not significantly different
from zero (p = 0.58). The limits of agreement for the template
measurements were -0.60 to +0.52 mm.
One-way analysis of variance demonstrated that the average error
of the Dorr method was significantly larger than the average errors
of the Livermore and template methods (p < 0.01). However,
the average errors of the Livermore and template methods were not
significantly different from each other (p = 0.97).
Currently, computer-assisted radiographic methods are the most
accurate and most reproducible way to assess in vivo polyethylene
wear noninvasively, to track its progress over time, and to predict
when a worn polyethylene liner should be exchanged. However, these
techniques can be time-consuming and expensive because they require sophisticated
computer hardware and software. Additional personnel may also be
needed to digitally scan radiographs and to perform wear analyses.
Therefore, these techniques may not be practical for most orthopaedists
in the clinical setting.
Manual radiographic methods are far less accurate than contemporary
computer-assisted techniques6,7,
and, in some instances, can be laborious to perform. We acknowledge
several inherent weaknesses of the manual techniques that we studied.
First, the techniques do not correct for the position of the center
of the x-ray beam. Because the center of the x-ray beam on an anteroposterior
radiograph is not directly in line with the components, the radiographic
projections of the head and cup are elliptical. All of the manual
techniques assume that the radiographic projections are circular. Second,
as with any uncemented acetabular component, the presence of a rough
coating for biologic ingrowth makes detection of the edges of the
component more difficult. Finally, the accuracy of manual measurement
is highly dependent on the skill of the observer, and thus these
techniques are prone to high interobserver variability. All of these
factors decrease the accuracy of the manual radiographic techniques.
Despite these limitations, the manual methods continue to be used
because clinicians need an easy and economical way to assess polyethylene
liner deformation for individual patients.
In the current study, we compared the accuracy of two commonly
used manual radiographic methods with the accuracy of a new, template-based
method for the measurement of polyethylene wear. Several important
results were observed. First, we found a poor relationship between
the minimum polyethylene thickness that was determined by direct
measurement of the explanted liner and the thickness that was measured
radiographically with use of the Dorr technique. Although it is
quick and easy to perform, the Dorr technique demonstrated a significant
bias to overestimate the true polyethylene thickness. This result
was especially true for hips in which the femoral head had penetrated
mainly into the central portion of the acetabular cup. In such cases,
the Dorr technique was highly inaccurate. The error of the Dorr method
averaged 1.54 ± 1.21 mm. This finding implies that,
in the clinical setting, an average polyethylene liner with a calculated
thickness of 1.54 mm would actually be worn through. The negative
implications of such an error, including the metal-on-metal damage
to the cup and to the prosthetic femoral head as well as the potential
fracture of a ceramic head, are obvious. The inaccuracy of this
technique, combined with the fact that we had to contact the manufacturer
of the cup to obtain information on the original thickness of the
liner, made this the least useful of the techniques studied.
A stronger relationship was found between the thickness that
was determined by direct measurement of the explanted component
and the thickness that was measured radiographically with use of
the Livermore technique. Although the average error of the Livermore technique
(0.07 ± 0.62 mm) was significantly less than that
of the Dorr technique, the former method was much more laborious
and time-consuming. Unlike the Dorr technique, the small average
error did not indicate a bias for the radiographic measurement to
consistently underestimate or overestimate the true measurement;
however, the standard deviation (0.62 mm) implied a wide range of
error across the group. Furthermore, this method required multiple
measurements to be made on two different radiographs and also required
that we contact the manufacturer of the cup to obtain information
on the original thickness of the liner. Because of these difficulties,
the results of such an analysis cannot be demonstrated clearly or
easily to a patient, making this method impractical in the clinical setting.
The strongest relationship was found between the true polyethylene
thickness and the thickness that was measured radiographically with
use of the wear-template method. The average error of this technique
(—0.04 ± 0.28 mm) was significantly less
than that of the Dorr technique but was not significantly different from
that of the Livermore technique. Similar to the Livermore technique,
the average error for this method did not consistently underestimate
or overestimate the direct measurement. However, the standard deviation
for the template method (0.28 mm) was smaller than that for both
of the other manual methods, indicating less variation in the amount
of error. The limits of agreement for the template technique indicated
that for 95% of all hips the error would fall somewhere
between a 0.60-mm underestimation and a 0.52-mm overestimation of
the true polyethylene thickness. Although these limits may be greater
than one would find with use of the computer-assisted techniques,
this method is simple, quick, involves only one radiograph, and
requires no additional information from the manufacturer. The template of
the cup is superimposed on a radiograph, immediately demonstrating
the remaining thickness of the polyethylene liner to both the clinician
and the patient. Measurements can easily be performed in the office
with greater accuracy than has been noted in association with the
other manual techniques. Thus, we found this method to be the most
useful clinically.
Because we have confirmed the accuracy of the template method
with different types of hemispherical cups, we have found it useful
in our day-to-day practice. For orthopaedists who use a manual radiographic
method as a clinical tool to evaluate thinning polyethylene liners,
we advocate the wear-template method, with the understanding that
the measurements can be inaccurate by as much as one-half millimeter.
With knowledge of this level of inaccuracy, one can still assess
the minimum thickness of the polyethylene liner in patients who present
with eccentric wear.
In conclusion, the present study provides the clinician with
information about a new manual method for assessing polyethylene
liner thickness. This method is time-efficient and is more clinically
useful than other manual methods currently in use.
Graphs depicting the comparisons of the three radiographic techniques
with the direct measurement technique are available with the electronic
versions of this article, on our web site at www.jbjs.org (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).