Between July 1988 and December 1990, eighty-five patients underwent
a total of ninety-two total hip replacements in which a cementless,
anatomically designed femoral component with a circumferential proximal
porous coating (Anatomic Hip; Zimmer, Warsaw, Indiana) was coupled
with a cementless hemispheric porous-coated acetabular component (Harris-Galante
II; Zimmer). All operations were performed at a single institution
by one of the four senior authors (J.J.J., S.G., A.G.R., or J.O.G.).
The primary diagnosis was osteoarthritis in fifty-six hips (61%),
avascular necrosis in twenty (22%), rheumatoid arthritis
in eight (9%), posttraumatic arthritis in five (5%),
psoriatic arthritis in two (2%), and ankylosing spondy-litis
in one (1%). Of the eighty-five patients, six (seven hips)
died before eight years (the minimum duration of follow-up) had
elapsed and five (seven hips) were lost to follow-up. The remaining
seventy-eight hips in seventy-four patients (forty-six men and twenty-eight
women) were followed for a minimum of eight years. The average age
at the time of the arthroplasty was fifty-two years (range, thirty-one to
sixty-nine years). Four patients underwent staged bilateral hip
reconstruction (from two to eleven months apart).
The mean duration of follow-up was ten years (range, eight to eleven
years). Although the results for the fourteen hips that were followed
for less than eight years are reported and included in the survivorship
analysis, they are not included in the final, eight to eleven-year
results.
The indications for cementless total hip arthroplasty included an
age of less than seventy years and good bone quality (type A or
B according to the system of Dorr et al.14).
Hybrid total hip arthroplasty (with a cemented femoral component
and an uncemented acetabular component) was performed in younger
patients when an uncemented femoral component could not be used
because of poor femoral bone stock, sub-optimal proximal
femoral geometry, or an inability to comply with the weight-bearing
limitations necessary for postop-erative rehabilitation
after use of an uncemented femoral component. Furthermore, hybrid
total hip arthroplasty was performed in younger patients with a
lower-than-normal life expectancy because of cancer, severe inflammatory
disease, or cardiopulmonary disease. During the time of this study,
a total of 418 total hip replacements were implanted: 273 (65%) were
hybrid, 144 (34%) were cementless, and one (0.2%)
was cemented. Fifty-two of the 144 cementless arthroplasties were performed
with implants of other designs (that is, not the -design
evaluated in the present investigation) and thus were excluded from
the present study. The remaining group of eighty-five patients in
this study represent the younger and more active patients seen in
our practice at that time.
The Anatomic Hip is a curved collarless femoral component made
of Ti-6-Al-4-V alloy with a commercially pure circumferential titanium
fiber-metal porous coating diffusion bonded onto the proximal aspect
of the implant (Fig. 1-A). The modular femoral head is a
chromium-cobalt-alloy component mated to the femur by interference
fit with a conical taper. A 28-mm head was used in all patients.
The prosthetic design relies on maximum metaphyseal fill to provide
initial implant stability. Flexible reamers were used to ream the
medullary canal to the diameter of the desired stem, and this was followed
by rasping of the metaphysis and canal. Torque testing was performed
to confirm rotational stability by applying a 120 in-lb (13.5 Nm)
retroversion force with a torque wrench to the final broach. If
any motion of the broach was detected within the proximal part of
the femur, the next larger broach was used until this motion was
eliminated. In the event of an intraoperative, nondisplaced metaphyseal
fracture, the broach or prosthesis was removed, cerclage wires were
placed around the proximal part of the femur, and the final component
was implanted.
The acetabulum was reconstructed with use of a -porous-coated
hemispherical component (Harris-Galante II) with a similarly prepared
titanium fiber-metal ingrowth surface (Fig. 1-B). The acetabular component was
inserted in a line-to-line fashion; the diameter of the implant
matched the diameter of the last reamer used to prepare the acetabulum.
Two, three, or four (mean, three) supplemental 6.5-mm-diameter screws were
used in all patients. The polyethylene used at this time was machined
and sterilized with gamma irradiation in air.
The surgery was performed in a clean-air operating room with vertical
laminar flow. The operating team wore body exhaust systems. The
posterolateral approach was used in all patients. Prophylactic antibiotics,
a first-generation cephalosporin unless contraindicated, were given
preoperatively and continued for one to two days postoperatively.
Suction drains were typically removed twenty-four to forty-eight
hours postoperatively. All patients received a four-week course
of warfarin for prophylaxis against venous thromboembolism. All
patients walked with partial weight-bearing with two crutches for
six weeks and then with full weight-bearing with crutches for an additional
six weeks.
Clinical and Radiographic Analysis
Clinical and radiographic evaluation was performed preop-eratively
and postoperatively at six weeks, three months, six months, one
year, and annually thereafter. Clinical evaluation included physical
examination and determination of the -Harris hip score15. A result was considered excellent
when the Harris hip score was between 90 and 100 points, good when
it was between 80 and 89 points, fair when it was between 70 and
79 points, and poor when it was <70 points. The patients
were also questioned about thigh pain, and its severity was graded
with use of the same 44-point scale as is used for the Harris hip
pain score.
Standard radiographs included an anteroposterior view of the pelvis
and anteroposterior and lateral views of the proximal part of the
femur. Two nonbiased observers who had not been involved in the
implantation evaluated all radiographs. The six-week postoperative
radiograph served as the baseline for identifying subsequent subsidence,
remodeling of the femur, interfacial radiolucencies, osteolysis,
and component loosening.
On the anteroposterior16 and
lateral17 radiographs, the femoral
component was divided into seven zones. Radiolucencies at the bone-prosthesis
interface were recorded as £1 mm, >1 mm but £2
mm, or >2 mm in width. Radiolucencies adjacent to the porous-coated
portion of the stem were recorded separately. The extent and location
of any osteolysis was recorded and was compared with that seen on previous
radiographs. Subsidence was determined by measuring the distance
between two lines drawn perpendicular to a longitudinal line through
the center of the femoral canal18.
The first perpendicular line was drawn at the level of the tip of
the greater trochanter, and the second line was drawn at the junction
of the lateral base of the Morse taper and the shoulder of the prosthesis.
Any change of >2 mm in the distance between these two lines
represented vertical subsidence of the femoral component18.
Change in the position of the prosthesis into varus or valgus was
determined by drawing a line through the center of the prosthesis
and another through the center of the proximal part of the femur.
A change of >2°, into varus or valgus, on comparable radiographs
represented a change in stem orien-tation18. Remodeling changes in the proximal
and distal parts of the femur including rounding of the calcar,
cortical hypertrophy, and formation of an intramedullary bone pedestal
were evaluated. Cortical hypertrophy was recorded for each Gruen
zone16. The presence of a pedestal
was recorded as partial or complete. Heterotopic ossification was
graded according to the method of Brooker et al.19.
The radiographs were also used to evaluate the status of the acetabular
component with regard to stability, migration, radiolucency around
screws, and periacetabular osteolysis. We classified the periacetabular
osteolytic lesions as small (<1 cm in diameter), large
(1 cm in diameter), or combined (two or more lesions in different
size categories). We also categorized the lesions according to their
location (peripheral, retroacetabular, or ischial).
Wear of the acetabular component was measured with a digitized
analysis, as described by Martell et al.18.
Kaplan-Meier survivorship analysis20 was
performed on the entire cohort of ninety-two hips.
Clinical Results
The average Harris hip score increased from 51 points (range, 24
to 76 points) preoperatively to 94 points (range, 48 to 100 points)
at the time of final follow-up. Of the seventy-eight hips that were
followed for more than eight years, fifty-six (72%) had
an excellent result; eleven (14%), a good result; seven
(9%), a fair result; and four (5%), a poor result.
Of the four poor results, only one was related to the total hip
arthroplasty; this patient reported moderate pain and a moderate limp
from a 2-cm leg-length discrepancy. The other three poor results
were related to contralateral hip osteoarthritis, spinal stenosis,
or an unstable total knee prosthesis in a patient with rheumatoid
arthritis.
Patients were questioned about pain around the hip and thigh separately.
At the time of final follow-up, fifty-nine hips (76%) were
pain-free, eleven (14%) were slightly painful without compromising
activity, six (8%) were mildly painful after activities,
two (3%) were moderately painful and required concessions,
and none were markedly painful or caused disabling pain. Seven hips
(9%) were associated with mild-to-severe- thigh
pain. Thigh pain was related to a larger stem size (p = 0.06,
analysis of variance). Five hips (6%) were associated with
mild thigh pain; one (1%), moderate thigh pain; and one
(1%), disabling thigh pain. At the time of final follow-up,
sixty hips (77%) were in patients with no limp; fourteen
(18%), in those with a slight limp; three (4%),
in those with a moderate limp; and one (1%), in a patient
with a severe limp. Sixty-six hips (85%) were in patients
who required no cane; six (8%), in those who required a
cane only for long walks; four (5%), in those who required
a cane full-time; and two (3%), in those who required a
walker.
In the entire cohort of ninety-two hips (eighty-five patients), four
(4%) had an intraoperative complication: a nondisplaced metaphyseal
fracture that occurred with implantation of the broach or the femoral
component. In all four hips, two cerclage wires were placed and
the implant was stable intraoperatively. All four femoral components
subsequently had stable ingrowth. Postoperative complications included
thromboembolic events in two patients. Bilateral lower-extremity
thrombi developed two weeks postoperatively in one patient, and
a symptomatic pulmonary embolus developed three weeks postoperatively
in the other. There were four postoperative dislocations, all of
which were initially treated with closed reduction and six weeks
of bracing. Two of the patients had recurrent dislocations and required
a reoperation: one of them had an acetabular revision at nine years,
and the other had an exchange of the femoral head and the acetabular
liner at seven years after the arthroplasty. No femoral component
was revised.
In the entire cohort of ninety-two hips, there were nine reoperations
(10%) in eight patients. Seven acetabular reoperations (including
two acetabular revisions) and two femoral reoperations were required.
There were no femoral revisions. One patient required two reoperations:
an exchange of the polyethylene liner because of excessive wear
eight years after the arthroplasty and, one year later, revision
to a constrained acetabular component because of recurrent dislocation.
One patient fell, five years after the arthroplasty, and sustained
a periprosthetic acetabular fracture with loosening of the acetabular
component, which required revision six months after the fracture.
One patient had recurrent (four) dislocations -requiring
exchange of the liner and head at seven years after the arthroplasty.
One patient had a pathologic fracture of the greater trochanter
through an osteolytic lesion and required open reduction and internal
fixation, bone-grafting of the lesion, and exchange of the liner
and head at twelve years. Three additional patients required exchange
of the liner and head with grafting of periacetabular and/or
trochanteric areas of osteolysis at nine years (one patient) and
ten years (two patients) after the arthroplasty. One patient required
removal of a symptomatic cerclage wire from the proximal part of
the femur.
The fourteen hips in the eleven patients who had less than eight
years of clinical follow-up were followed for a mean of fifty-four
months (range, one to eighty months). The mean Harris hip score
in this group increased from 50 points (range, 33 to 72 points)
preoperatively to 88 points (range, 48 to 100 points) at the most
recent follow-up evaluation. The result was excellent for ten hips,
good for one, fair for one, and poor for two. At the time of final
follow-up, ten hips were pain-free, two were slightly painful without
compromising activity, none were mildly painful after activities,
and two were moderately painful and required concessions. At the
time of final follow-up, nine hips were in patients with no limp;
two, in patients with a slight limp; and three, in patients with
a moderate limp. Ten hips were in patients who did not require use of
a cane, one was in a patient who used a cane only for long walks,
and three were in patients who used a cane full-time. There was
only one reoperation in this group, an acetabular revision, which
was reported in the previous paragraph.
Survivorship Analysis
Kaplan-Meier survivorship analysis was performed with use of
revision or loosening of any component as the end point. The ten-year
rate of survival was 96.4% ± 2.1% for the
total hip prosthesis, 100% for the femoral component, and 96.4% ±
2.1% for the acetabular component. With any reoperation
as the end point, the ten-year survival rate was 91.9% ±
3.6%.
Radiographic Results
Of the seventy-eight hips in the seventy-two patients with a minimum
of eight years of clinical follow-up, seventy-three had complete
radiographic follow-up (mean, ten years; range, eight to eleven
years). No femoral component had subsided, and one stem had a change
of >2° into varus. This femoral component was undersized
and had been implanted in slight varus. It had migrated into varus
and was surrounded by a nearly complete radiolucency at thirty-six
months. At the most recent follow-up examination, at 113 months,
the femoral component had osseous growth into the proximal porous surfaces,
as seen on anteroposterior and lateral radiographs, without subsidence
or additional varus shift, and the Harris hip score was 100 points.
In the overall series, forty-four (60%) of the stems were
in neutral alignment, ten (14%) were in valgus alignment
(mean, 1.5°), and sixteen (22%) were in varus alignment
(mean, 3.1°) at the time of the last follow-up. The radiographs
of three hips were inadequate for measurement of varus-valgus position.
Qualitative changes in the proximal and distal parts of the femur
were evaluated to assess the remodeling characteristics. Seventy-one
hips (97%) had calcar rounding, which was partial in fifty-five
hips (75%) and complete in sixteen (22%). Cortical
hypertrophy was present in fifty-four hips (74%), with
the most common locations being distal zones 3, 4, and 5 (affected
in 19%, 23%, and 16% of the hips, respectively).
An incomplete bone pedestal was present in fifty-nine hips (81%),
and a complete pedestal was present in two (3%). Neither
patient with a complete pedestal had other findings of loosening.
Radiolucent lines were present in one or more zones in sixty-eight
(93%) of the seventy-three hips at the last follow-up evaluation.
All radiolucencies were £1 mm in width and were not progressive.
On the anteroposterior radiographs, twenty-three hips (32%)
had a radiolucency in zone 1; ten (14%), in zone 2; twenty-four
(33%), in zone 3; forty-seven (64%), in zone 4;
twenty-one (29%), in zone 5; twenty-three (32%),
in zone 6; and thirty-one (42%), in zone 7. On the -lateral
radiographs, twenty-two (30%) of the hips had a radiolucency
in zone 1; seventeen (23%), in zone 2; twenty-nine (40%),
in zone 3; forty-four (60%), in zone 4; thirty (41%),
in zone 5; sixteen (22%), in zone 6; and twenty-seven (37%),
in zone 7. These radiolucencies were adjacent to the nonporous portion
of the stem, were nonprogressive, and were £1 mm in width.
Two hips showed a 1-mm radiolucency adjacent to the porous ingrowth
surface (zone 1) on the anteroposterior radiograph. One of these
hips also had a lucency of <1 mm in zone 7 on the lateral
radiograph. In both cases, there was evidence of bone growth into
the other areas and no evidence of loosening. The last Harris hip
scores of these two patients were 97 and 100 points.
Osteolysis was adjacent to four femoral components (5%); three
of the osteolytic lesions were in zone 1A, and one was in zone 7A.
These four lesions were proximal to the lesser trochanter, and three
of the four did not extend into the region of the porous surface.
The fourth lesion was seen, on the anteroposterior radiograph, to
incorporate the region of the porous surface, in zone 1, and it
resulted in a pathologic fracture of the greater trochanter at 123
months. There was no osteolysis distal to the lesser trochanter
in any hip.
Acetabular radiolucency was present in zone A1 in twenty-one
(29%) of the hips, in zone 2A in fourteen (19%),
in zone B1 in six (8%), in zone B2 in twelve (16%),
and in zone C in twenty-three (32%). All radiolucencies
were <1 mm in width and were nonprogressive. No acetabular
components were surrounded by a complete radiolucent line. Two acetabular components
migrated >2 mm. One of the patients with acetabular migration
was asymptomatic and did not require a reoperation. The other patient,
in whom the migration was a result of periprosthetic fracture, had
an acetabular revision. Radiolucencies were identified around screws
in seven acetabular components. No screw broke.
Osteolysis was identified adjacent to twelve (16%) of
the acetabular components (Figs. 2-A and 2-B). The acetabular osteolytic lesions
were large (1 cm in diameter) in six hips, small in five, and combined
in one.
The mean rate of wear of the acetabular component was 0.16 mm/yr
(range, 0.00 to 0.47 mm/yr). An association between increased
wear and younger age was identified (p < 0.001). There
was no association between wear and diagnosis. Acetabular osteolysis
was more common in patients with a higher wear rate and more total
wear (p < 0.001, analysis of variance). Hips without acetabular
osteolysis had an average wear rate of 0.015 mm/yr, whereas
hips with acetabular osteolysis had an average wear rate of 0.27
mm/yr (p < 0.005, analysis of variance). There
were four ischial lesions (all large), seven retroacetabular lesions
(four small and three large), and three peripheral lesions (all
small) in these twelve hips. Peripheral osteolysis was seen in younger
patients, at an average age of thirty-six years at the time of the
index arthroplasty, whereas retroacetabular and ischial osteolysis
was seen in older patients, at an average age of fifty-four years
at the time of the index arthroplasty (p = 0.003, analysis
of variance).
Of the seventy-three hips with more than eight years of radiographic
follow-up, fifty-six (77%) were found to have heterotopic
ossification. According to the method of Brooker et al.19, twenty hips (27%) had
grade-I heterotopic ossification; thirty (41%), grade-II;
six (8%), grade-III; and none, grade-IV. No patient was
functionally limited by or required a reoperation secondary to heterotopic
ossification.
The fourteen hips in the eleven patients with less than eight years
of follow-up were followed radiographically for a mean of fifty-four
months (range, one to eighty-one months). Thirteen of the fourteen
hips had radiolucency in at least one zone of the femoral component.
All radiolucencies were <1 mm in width and were not progressive.
There was no cortical hypertrophy in any hip at this short follow-up
interval. No patients in this group had subsidence of >2
mm or a change in varus or valgus position of >2°. One
patient had a femoral osteolytic lesion in zone 7A (proximal to
the lesser trochanter) seen initially at sixty-seven months. Acetabular
lucencies, all <1 mm, were seen in seven hips. There was
no lucency around any screw. One patient had acetabular migration
of >2 mm, which was initially recognized at sixty-nine
months. The patient remained asymptomatic at the time of final follow-up,
at eighty-nine months, and did not require revision. Two hips in one
patient had osteolysis adjacent to the acetabular component. These
were included in the group of acetabular components associated with
osteolysis described above. One lesion was peripheral, and the other
was small (<1 cm).
The purpose of this prospective study was to evaluate the results
at an average of ten years after a second-generation cementless
total hip arthroplasty incorporating an anatomically designed femoral
component with a circumferential proximal porous coating and a hemispheric
acetabular component. At an average of ten years, 86% of
the hips demonstrated a good or excellent clinical result. Furthermore,
there was no aseptic femoral loosening and no femoral osteolysis distal
to the lesser trochanter up to eleven years after the arthroplasty.
Polyethylene wear and periacetabular osteolysis remain a concern
in this relatively young and active patient population.
As a result of the disappointing results associated with early proximally
coated cementless femoral components such as the porous-coated anatomic
implant (PCA; Howmedica, Rutherford, New Jersey)21,22 and
the Harris-Galante-I implant (Zimmer, Warsaw, Indiana)18, the second-generation proximally
coated femoral stems were developed. The Harris-Galante-I femoral
component had a relatively small porous surface area that was noncircumferential,
resulting in a susceptibility to distal osteolysis23,24. The clinical results in the
present series, in which a second-generation proximally coated femoral
component with a circumferential porous surface was used, compare
favorably with those in previously published reports on -cemented
femoral components25-29 and on
other suc-cessful cementless designs10,11,30.
While there are only limited intermediate-term data on these second-generation
cementless proximally porous-coated femoral components, our findings
of no loosening or distal osteolysis compare favorably with those
available in the literature10,30,31.
Our results compare favorably with the reported success rates of
extensively porous-coated femoral components as well. Engh et al.
reviewed the results of 174 consecutive arthroplasties with an extensively
coated femoral component, followed for at least ten years, and found
a 2% rate of femoral revision, a 2% rate of radiographic
loosening, and a 97% rate of survival of the femoral component
at twelve years11. They identified
osteolysis in 39% of the hips. Kim et al. recently reported
on fifty-two hips treated with an extensively porous-coated stem
and identified a 2% femoral revision rate at a mean of
eleven years32. They found a 55% rate
of femoral osteolysis at this time, with the osteolysis distal to
zones 1 and 7 in 17% (nine) of the hips.
The results in our series of femoral components with a proximal
circumferential coating also compare favorably with those in series
of femoral components implanted with modern cementing techniques29,33,34.
In our series, seven hips (9%) were associated with
mild-to-severe thigh pain, which was mild in association with five hips
and moderate and disabling in association with one each. In other
series of proximally and fully coated femoral components, the prevalence
of thigh pain has been variable, ranging from 1.5% to 27%10,11,22,35,36. In a recent study of
the same prosthesis as was used in our study, Ragab et al. identified
a 5% prevalence of thigh pain, not associated with stem
size, at a mean of seven years30.
In our study, thigh pain was associated with a larger stem size
(analysis of variance, p = 0.06). In a study of the porous-coated
anatomic prosthesis, Bourne et al. found a 27% prevalence
of thigh pain at five years22.
In our series, the prevalence of reported thigh pain did not decrease
with increases in the duration of follow-up. Xenos et al., also
in a study of the porous-coated anatomic prosthesis, found a peak
prevalence of thigh pain of 23%, which decreased to 12% by
ten years35. Vresilovic et al.
found a prevalence of thigh pain of 12% in patients with
a wedge-shaped cementless stem, and the prevalence correlated positively
with the size of the stem36. Lastly,
Engh et al. identified an 8% rate of thigh pain in their study
of 174 extensively coated femoral stems followed for a minimum of
ten years11.
Fixation of cementless femoral components was classified by Engh
et al. as stable with osseous ingrowth, stable with fibrous ingrowth,
and unstable37. This classification
system is dependent on a cylindrical implant and a radiograph made
tangential to the porous surface. However, it is difficult to use
the system for anatomically designed stems with limited ingrowth
surfaces. Incomplete sclerotic and radiolucent lines adjacent to
the prosthesis do not correlate well with implant stability. Most stems
in our series had a radiolucent line in at least one zone; however,
these lines were generally seen adjacent to the nonporous surface
of the implant and were not associated with other changes indicative
of stem loosening. We also identified an incomplete pedestal in
fifty-nine hips (81%) and a complete pedestal in two (3%).
In none of these hips was there other evidence of aseptic loosening,
such as subsidence or complete radiolucency, and most stems had
clear evidence of osteointegration as seen by trabecularization
into the pads. Ragab et al., in their review of the same stem as
was used in our series, found a partial pedestal in twenty-eight
hips (29%), all of which had a radiographically stable
implant30. In the case of proximally
porous-coated femoral components, therefore, radiographic subsidence
or a complete radiolucent line is an indicator of loosening, whereas
partial radiolucencies that are not adjacent to the ingrowth surface and
pedestal formation are not. Lastly, trabecularization into ingrowth
pads is an indication of osteointegration.
Particulate debris from polyethylene wear and resultant osteolysis
remain the primary factors limiting the longevity of hip prostheses38. While a tight seal at the bone-implant
interface and stable fixation without motion potentially inhibit
or slow migration of the debris to the distal femoral zones, periacetabular
osteolysis has become relatively common at intermediate follow-up intervals4,10,11,22,30,35,39,40. The femoral
osteolysis seen in our series (in 5% of the hips) was localized
to Gruen zones 1 and 7, proximal to the lesser trochanter. No hip
had osteolysis distal to the lesser trochanter. However, we identified
an 8% rate of small periacetabular osteolytic lesions and
an 8% rate of large lesions. Despite this high rate of
osteolysis, acetabular ingrowth and fixation remained excellent,
with only one revision for aseptic loosening (following a periprosthetic
acetabular fracture).
We showed that the presence of these osteolytic lesions is directly
related to polyethylene wear, with acetabular osteolysis more likely
to develop in patients with a higher wear rate. Furthermore, the
location of these osteolytic lesions—that is, peripheral
compared with retroacetabular—was related to age, with
peripheral osteolysis developing in younger patients. Perhaps this
is ultimately related to bone quality, as patients with poor bone
quality are susceptible to retroacetabular osteolysis. Given these
findings, screening radiographs at annual intervals and early intervention
are warranted to prevent the difficult situation that results from
advanced osteolysis and loosening. In addition, as plain radiographs
often underestimate the extent of osteolysis, especially adjacent
to the hemispherical acetabular component, we have found computed
tomography to be of assistance in characterizing lesions found on
plain radiographs.
The outcome of cementless total hip arthroplasty depends on many
factors, including component design, patient selection, and surgical
technique. Our results at a mean of ten years suggest that the use
of a second-generation, circumferentially porous-coated femoral
stem provides reproducible ingrowth and excellent clinical results.
However, polyethylene wear and its sequelae remain matters of concern
in this relatively young and active group of patients.