Patients
Between August 1996 and July 1999, I performed fifty-two consecutive
primary bilateral total hip arthroplasties in fifty-two patients (104 hips)
who had bilateral osteonecrosis of the femoral head and no other joint
problems. All fifty-two patients (forty-eight men and four women) were
enrolled in the present study, which was approved by the institutional review
board at our institution. All patients provided informed consent. The average
age of the patients at the time of surgery was 44.2 years (range, thirty-one
to forty-eight years), and all were younger than fifty years of age. The
average weight of the patients was 68.9 kg (range, 62 to 92 kg), and their
average height was 168.3 cm (range, 158 to 184 cm). The mean duration of
follow-up was 7.1 years (range, five to eight years).
The bilateral total hip arthroplasties were performed during the same
anesthetic session, with one side treated immediately after the other.
Randomization of the use of either a zirconia or a cobalt-chromium femoral
head was determined from a sequential pool based on a table of randomized
numbers. Insertion of the zirconia and cobalt-chromium heads was assigned
alternatively to each side.
A cementless Duraloc acetabular shell (DePuy, Warsaw, Indiana) was used in
all hips; ninety-two hips received a Duraloc 100-series acetabular component
without screw holes, and twelve hips received a Duraloc 1200-series component
with multiple screw holes. All acetabular components were press-fit after
underreaming of the acetabulum by 2 mm. A 28-mm (inner diameter) ultra-high
molecular weight polyethylene liner (Enduron; DePuy, Warsaw, Indiana) made of
ram-extruded GUR 1050 polyethylene was used in all hips. The polyethylene was
irradiated (with between 2.4 and 4 Mrad) in a vacuum and was packaged in a
vacuum state.
A cementless Immediate Postoperative Stability (IPS) titanium-alloy femoral
component (DePuy, Leeds, United Kingdom) was used in all
hips8. The proximal,
metaphyseal portion of the stem (about one-third of the stem) is porous-coated
with sintered beads. The pore size is between 200 and 300 µm. The
porous-coated surface has a circumferential hydroxyapatite coating of 30 µm
in depth. The distal part of the stem is polished.
A 28-mm zirconia head was used in one hip and a 28-mm cobalt-chromium head
was used in the contralateral hip in all patients. The zirconia was yttria
tetragonal zirconium oxide polycrystal (Y-TZP) containing 3 mol% oxide for
stabilization (DePuy, Leeds, United Kingdom).
All operations were performed through a posterolateral approach. The
femoral component was inserted with a press-fit technique. The largest broach
that would fill the femoral canal and leave little cancellous bone remaining
was used.
The patients were allowed to stand on the second postoperative day, and
they progressed to full weight-bearing with crutches as tolerated. The average
time until full weight-bearing was ten days.
We performed a clinical and radiographic follow-up at six weeks, three and
six months, and one year after the operation and yearly thereafter. Harris hip
scores9 were
determined preoperatively and at each follow-up examination. Although several
components of the Harris hip score (pain, limp, use of support, ability to put
on socks and tie shoes, absence of deformity, and range of motion) were easily
differentiated between the two hips in the same patient, distance walked,
stair-climbing, sitting, and using public transportation were more difficult
to differentiate between the two hips. However, if the patients had
difficulties in these domains, they could always identify the hip that limited
their activities.
A supine anteroposterior radiograph of the pelvis with both hips in neutral
rotation and 0° of abduction was made for every patient. Consistent
patient positioning was ensured with the use of an x-ray frame. This frame is
constructed so that it can be placed at the end of a standard x-ray table.
Plastic polypropylene orthoses are secured to a plastic backboard through a
vertical slot. A wing nut allows adjustment for various limb lengths. Rotation
and abduction remain constant. Cross-table lateral radiographs were also made
of each hip.
The adequacy of the intramedullary fill by the stem was recorded as
satisfactory when the stem filled >80% of the proximal part of the canal in
the coronal plane and >70% in the sagittal plane, according to a previously
described method10.
The component was considered to be undersized if less of the canal was filled
in either or both planes.
Definite loosening of the femoral component was defined as progressive
axial subsidence of >3 mm or a varus or valgus
shift8. A femoral
component was considered to be possibly loose when there was a complete
radiolucent line surrounding the entire porous-coated surface on both the
anteroposterior and the lateral
radiograph8.
Anteversion of the acetabular component was measured on the true lateral
radiograph of the hip as the angle between a horizontal line and a second line
marking the plane of opening of the socket. To measure cup abduction, a line
that joined the inferior margins of the two acetabular teardrops was drawn on
the anteroposterior pelvic radiograph. The intersection of that line with a
line marking the plane of opening of the socket determined the angle of
abduction.
Definite loosening of the acetabular component was diagnosed when there was
a change in the position of the component (>2 mm vertically and/or medially
or laterally) or a continuous radiolucent line wider than 2 mm on both the
anteroposterior and the lateral
radiograph8. A
vertical change in the position of the cup was measured between the inferior
margin of the cup and the inferior margin of the ipsilateral teardrop, and a
horizontal change was measured between the Köhler line and the center of
the outer shell of the acetabular
component11 (see
Appendix).
The locations of areas of osteolysis in the acetabulum were recorded
according to the system of DeLee and
Charnley12, and
those in the femur were recorded according to the system of Gruen et
al.13. The length
and width of osteolytic lesions were measured, and the area was expressed in
square
centimeters3.
Proximal femoral bone resorption was graded
radiographically14,
with Grade 1 indicating atrophy or rounding off of the calcar; Grade 2, loss
of density in the calcar region with preservation of the medial cortical wall
to the level of the lesser trochanter; Grade 3, loss of density in the calcar
region with loss of the medial cortical wall to the level of the lesser
trochanter; and Grade 4, loss of density in the entire medial cortical wall
distal to the level of the lesser trochanter.
To differentiate osteolysis resulting from polyethylene or metallic debris
from stress-shielding-related bone resorption in the calcar region, three
Craig-needle biopsy specimens were obtained under fluoroscopic control from
the calcar region of four selected hips. Four patients who had had a total hip
prosthesis in place for at least five years (five, 5.6, six, and 7.5 years)
were randomly selected. The institutional review board approved this aspect of
the study, and these four patients provided specific informed consent for this
experimental procedure. Decalcified bone samples were embedded in paraffin wax
and were stained with Mayer hematoxylin and eosin. The samples were then
examined for evidence of polyethylene or metallic debris and/or macrophagic
resorption of bone (osteolysis) or diminished trabeculae without macrophagic
resorption of bone (stress-shielding).
Linear wear of the polyethylene liner was measured, with use of a software
program (AutoCAD, Release 13; Autodesk, Sausalito,
California)4, by one
observer who was blinded to the radiographic results. The observer made three
measurements on each radiograph, and the intraobserver error was ±0.038
mm. A ScanMaker 9600XL flat-bed imaging scanner (Microtek, Carson, California)
digitized the anteroposterior radiograph of the pelvis as two-dimensional
gray-scale arrays of twelve-bit (256 gray level) integers. The scanning
resolution was 600 psi (pixels per square inch). Wear, defined as penetration
of the head into the liner, was determined at annual intervals from
anteroposterior pelvic radiographs. The amount of penetration on radiographs
made six weeks postoperatively was considered to be the "zero
position."
Activity Level
The level of activity of the patients after the total hip arthroplasty was
assessed with the activity score of Tegner and
Lysholm15. This
activity grading scale, with which work and sports activities are graded
numerically, was used as a complement to the functional score. The patients
were given a score, according to the activities in which they engaged in daily
life, ranging from 0 points for a hip-related disability to 10 points for
participation in competitive sports at a national level.
Analysis of Retrieved Femoral Heads
Two zirconia femoral heads, retrieved at the time of revision procedures
due to aseptic loosening, were analyzed to determine the surface roughness (Ra
and Rpm) and the phases of crystalline structure of zirconia heads. The ages
of the two patients at the time of the revision surgery were forty-one and
forty-eight years. One patient was 160 cm tall and weighed 55 kg, and the
other was 170 cm tall and weighed 67 kg. The durations of implantation of
these two zirconia femoral heads were five and six years.
The parameter Ra was the mathematical average of all deviations (peaks and
valleys) from the mean line of the surface profile. The parameter Rpm (the
mean leveling depth) was defined as the distance between the mean line and a
line parallel to it that passed through the highest point. The intrinsic
errors of the measurement of Ra and Rpm were 0.0215 and 0.163 nm,
respectively.
The surface characteristics of the two explanted zirconia heads were
evaluated with use of interferometry (Wyko RST 500 interferometer; Wyko,
Tucson, Arizona) and environmental scanning electron microscopy. The
monoclinic content of the surface of the zirconia head was calculated with
x-ray diffraction. Interferometry, environmental scanning electron microscopy,
and x-ray diffraction studies were also used to compare the two explanted
zirconia heads with two identical zirconia heads that had never been
implanted. The two unused zirconia heads had been sterilized and packaged by
the same manufacturing company in a manner identical to that employed for the
implanted ones. Four unused cobalt-chromium heads that were identical to those
used in the study were also examined to determine their roughness values (Ra
and Rpm).
The interferometry measurements were undertaken at two different
magnifications with use of ×20 and ×40 lenses. The areas of
analyses were 125 by 125 µm and 64 by 64 µm, respectively. The results
are presented in terms of Ra and Rpm.
One investigator who had no knowledge of the clinical and experimental
results performed additional analysis using a CamScan 4 environmental scanning
electron microscope (University of Leeds, Leeds, United Kingdom), which is a
special type of scanning electron microscope that works under controlled
environmental conditions and requires no conductive coating on the specimen.
This microscope makes it possible to examine specimens in their natural state.
Secondary and back-scattered images were reviewed at various magnifications
(particularly at ×250) to assess the pits and scratches on the zirconia
head surfaces.
Statistical Analysis
The Student t test was used to determine possible correlations between the
rate of penetration and several specific variables, including age, gender,
activity level, abduction angle of the acetabular component, and head type.
Linear regression analysis was used to reveal any relationship between surface
roughness values and age, gender, weight, activity, duration in vivo, and head
type. The Kaplan-Meier curve
method16 was used
in the analysis of the failure-free rate. Also, Greenwood's
formula17 was used
to calculate confidence intervals of the failure-free rate at certain
time-points.
Clinical Results
Hip Score
The mean preoperative Harris hip score was 41 points (range, 8 to 48
points) in the zirconia head group and 43 points (range, 13 to 53 points) in
the cobalt-chromium head group. The mean hip score at the final follow-up
examination was 92 points (range, 63 to 100 points) in the zirconia head group
and 94 points (range, 84 to 100 points) in the cobalt-chromium head group.
Functional Outcome
At the latest follow-up examination, forty-five patients had no detectable
limp, five had a mild limp, and two had a severe limp due to aseptic loosening
of the femoral component. The ability to use stairs and public transportation,
to put on footwear, and to cut toenails was improved markedly after the
operation.
Activity Score
Many patients were quite active despite our admonitions to avoid activities
involving high impact after the total hip replacement. All but two patients
had an activity score of 5 or 6 points after the total hip replacement,
indicating participation in strenuous farm work (a score of 5 points) or
playing tennis (a score of 6 points). The two patients with a loose femoral
stem had a score of 1 point.
Radiographic Results
Stem Loosening
Two hips with a zirconia femoral head had aseptic loosening of the femoral
stem and were revised with a larger stem. In these two hips, the femoral stem
was undersized in both the coronal and the sagittal plane. The
histopathological examination of the retrieved tissues around the two loose
femoral stems revealed no evidence of infection or foreign-body reaction.
There was no subsidence or aseptic loosening of any stem in the
cobalt-chromium head group.
Position and Loosening of the Acetabular Component
The average lateral opening and anteversion of the acetabular component was
44.6° (range, 35° to 55°) and 21° (range, 16° to 24°),
respectively, in the zirconia head group and 45.4° (range, 38° to
52°) and 23° (range, 18° to 26°), respectively, in the
cobalt-chromium head group. All but two acetabular components were positioned
in satisfactory abduction (35° to 45°) and anteversion (15° to
25°). The abduction angles of the acetabular components of the two hips
that had loosening of the femoral component were 43° and 41°, and the
anteversion angles were 24° and 25°, respectively. No acetabular
component had aseptic loosening.
Bone-Remodeling
At the latest follow-up evaluation, forty-seven hips in each group had
Grade-2 bone loss and five hips in each group had Grade-3 bone loss in the
calcar region. No hip in either group had Grade-4 bone loss. There was no
evidence of polyethylene or metallic debris or macrophagic resorption of bone
in any of the Craig-needle biopsy specimens. No hip had distal femoral or
acetabular osteolysis. All hips in the cobalt-chromium head group and all but
two hips in the zirconia head group had radiographic evidence of bone growth
into the porous surfaces of the femoral stem (zones 1, 7, 8, and 14 as defined
by Gruen et
al.13).
Polyethylene Wear
The mean amount of polyethylene linear wear was 0.57 ± 0.035 mm in
the zirconia head group and 1.21 ± 0.042 mm in the cobalt-chromium head
group. The difference was significant (Student two-tailed t test, p = 0.004).
The mean polyethylene linear wear rate per year (and standard deviation) was
0.08 ± 0.009 mm in the zirconia head group and 0.17 ± 0.011 mm
in the cobalt-chromium head group. This difference was significant as well
(Student two-tailed t test, p = 0.004). The mean amount of volumetric
polyethylene wear was 350.8 mm3 (range, 112.15 to 638.06
mm3) in the zirconia head group and 744.7 mm3 (range,
381.28 to 1018.21 mm3) in the cobalt-chromium head group. This
difference was also significant (Student two-tailed t test, p = 0.004). On the
basis of the numbers available, there was no difference in the rates of
polyethylene wear between the cups with and those without screw holes (Student
two-tailed t test, p = 0.88). In both the cobalt-chromium and the zirconia
head group, there was a significant relationship between polyethylene wear and
the age of the patient (p = 0.028), male gender (p = 0.042), the activity of
the patient (p = 0.038), and the abduction angle of the acetabular component
(p = 0.047). With the numbers available, there was no significant correlation
between polyethylene wear and the patient's weight, Harris hip score,
anteversion of the cup, or range of motion of the hip.
Analysis of Retrieved Femoral Heads
The Ra values of the two explanted zirconia heads were 15.87 and 17.35 nm,
and the Rpm values were 153.86 and 156.18 nm. The surface roughness values for
the two zirconia heads that had not been implanted (the controls) were 5.31
and 5.48 nm of Ra and 65.27 and 66.35 nm of Rpm. The Ra values of the
unimplanted cobalt-chromium heads were between 25 and 50 nm, and the Rpm
values were between 262.6 nm and 525.2 nm
(Table I).
Evaluation of the two unused zirconia heads with use of environmental
scanning electron microscopy showed very smooth surfaces with some short and
long randomly distributed scratches (Fig.
1-A). The two retrieved zirconia heads revealed small pits and
voids on the surface (Fig.
1-B).
X-ray diffraction studies showed similar high peaks for the tetragonal
phase and only a small amount of monoclinic phase transformation on the pole
areas of the two unused and the two retrieved zirconia heads
(Figs. 2-A and 2-B).
Survivorship
Two femoral stems (4%) in the zirconia head group were revised because of
aseptic loosening. These two stems were undersized in both the coronal and the
sagittal plane. No acetabular component was revised in the zirconia head
group, and no femoral or acetabular component was revised in the
cobalt-chromium head group.
Several studies have demonstrated good clinical results and a low
wear rate of ultra-high molecular weight polyethylene cups used with a
zirconia
head5,18,
whereas others have shown no advantages with use of a zirconia
head3,4,6,19,20.
In a study of twenty-nine Charnley Elite total hip replacements with a 28-mm
zirconia ball and a cemented Hylamer cup in younger patients, Norton et al.
reported a five-year prosthetic survival rate of 32% and an average wear rate
of >0.6 mm/yr for the failed
cups20. Kim et al.
performed simultaneous bilateral total hip replacement using a Hylamer liner
and a cobalt-chromium ball on one side and a zirconia ball on the other side
in seventy
patients4. They
observed greater wear with the zirconia head (0.21 mm/yr) than with the
cobalt-chromium head (0.12 mm/yr). In contrast, Wroblewski et al. reported a
wear rate of 0.02 mm/yr in a study of more than 1000 total hip replacements
with a 22-mm zirconia head and a contemporary polyethylene
cup18. The low wear
rates seen by Wroblewski et al. and in the present study (0.08 mm/yr), in
which the zirconia heads articulated with a contemporary polyethylene instead
of Hylamer, suggest that the wear rates with zirconia heads might be even
lower when they articulate with newer, improved surfaces such as highly
cross-linked polyethylene.
One possible problem with the use of zirconia is in vivo phase
transformation7.
Zirconia ceramic has three phases of crystalline structure, which vary with
temperature. The monoclinic phase transforms into a tetragonal phase at
<1100°C, whereas the tetragonal phase transforms into the cubic phase
at 2370°C. The tetragonal phase is the most unstable, but zirconia has the
greatest mechanical strength in that phase. Transformation from the tetragonal
phase into the monoclinic phase is accompanied by a 3% increase in the volume
of the ceramic head. While this phase transformation plays an important role
in increasing the mechanical strength of the zirconia head, it may cause an
increase in the surface roughness and thereby may increase the polyethylene
wear rate.
Such phase transformation in zirconia heads may be clinically relevant
because it can be induced at relatively low temperatures (as low as body
temperature) in the presence of water and
pressure3,21.
Lu and McKellop reported that the surface temperature of the polyethylene
articulation with a zirconia head rose to 99°C in a hip
simulator22.
Haraguchi et al. reported surface deterioration of two zirconia ceramic
femoral heads associated with in vivo phase transformation after total hip
arthroplasty3. The
cause of the phase transformation of the zirconia head in their series was
unknown.
Our x-ray diffraction study of two unused (control) and two retrieved
zirconia heads demonstrated very little monoclinic phase content on the
surface of the explanted heads. This finding suggests that the two retrieved
zirconia heads underwent no deleterious volume expansion or delamination in
vivo.
In summary, this study of patients treated with bilateral total hip
arthroplasty showed that zirconia heads articulating with a contemporary
polyethylene produced less radiographic evidence of polyethylene wear than did
cobalt-chromium heads. While only two explanted zirconia heads were available
for comparison with unimplanted zirconia heads and with unimplanted
cobalt-chromium heads, the analysis of those components suggests that the
zirconia heads that were implanted were smoother than the cobalt-chromium
heads, and the radiographic wear data reflected those differences. Additional
retrieval and tribologic studies are needed to confirm or refute these
observations.
A figure demonstrating the measurement of the change in the position of the
acetabular component in the vertical and horizontal directions is 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).
Note: The author thanks Cath Hardaker at DePuy, Johnson and
Johnson International, Leeds, United Kingdom, for her evaluation of surface
characteristics of retrieved zirconia heads with use of interferometry and
environmental scanning electron microscopy.