Mechanical failure due to aseptic loosening is the most common long-term
complication of total hip arthroplasty with
cement1-5
and is the most common indication for revision
surgery6. One
strategy for avoiding such loosening is the use of porous-coated implants that
are designed to achieve stable biological fixation without cement. Cementless
fixation allows for a direct biological interface between implant and bone
that has the ability to remodel over time to maintain mechanical stability.
Few large, long-term studies of so-called first-generation cementless hip
prostheses have been reported. The Anatomic Medullary Locking prosthesis (AML;
DePuy, Warsaw, Indiana), the Porous-Coated Anatomic prosthesis (PCA;
Howmedica, Rutherford, New Jersey), and the Harris-Galante Porous prosthesis
(HGP; Zimmer, Warsaw, Indiana) have been studied most
extensively7-11.
The results associated with these first-generation stems have been mixed as
numerous groups have reported a high rate of clinical complications such as
limp and thigh
pain9,11-13.
Although the adequacy of bone ingrowth and the stability of the biological
interface have been demonstrated in both canine models and human
subjects7,14-17,
early enthusiasm for the first generation of porous-coated femoral prostheses
was tempered by concerns about proximal osteolysis around fully porous-coated
stems (such as the AML), distal osteolysis around noncircumferentially
proximally porous-coated stems (such as the HGP), and bone resorption related
to
stress-shielding9-13,18-20.
Efforts to improve the clinical outcomes and survival rates associated with
cementless femoral prostheses have focused on modifications of prosthetic
design and construction. The most important design changes have included
limiting the porous-coated surface to only the proximal portion of the femoral
component in order to enhance proximal stress transfer, altering the implant
geometry in order to improve initial metaphyseal and diaphyseal fit and fill,
and using materials with a lower elastic modulus in order to minimize
stress-shielding.
The purpose of the present study was to assess the intermediate-term
implant survival rates, clinical outcomes, and radiographic results for
patients who had been followed for five to ten years after undergoing a
primary total hip arthroplasty that was performed with use of the HG Multilock
femoral prosthesis (Zimmer, Warsaw, Indiana).
Materials
The specific design elements of the HG Multilock prosthesis include a
proximal circumferential fiber-metal porous coating consisting of a titanium
wire mesh that is diffusion-bonded to the substrate and that stands 0.5 mm
proud on all sides. The porous surface area is trapezoidal, is enlarged
relative to that of first-generation components, and extends into the
metaphyseal-diaphyseal junction of the stem
(Fig. 1). The stem is composed
of surface-hardened Ti-6Al-4V alloy. The distal part of the stem is smooth,
straight, and cylindrical and has flutes that increase in depth as the stem
diameter increases to attenuate bending and torsional stiffness.
Study Population
Between February 1991 and March 1994, 123 consecutive primary total hip
arthroplasties were performed in 101 patients with use of the HG Multilock
cementless femoral component and the Harris-Galante II porous-coated
acetabular component (Zimmer). Institutional review board approval for the
present study was obtained from the University of Pittsburgh. All patients who
were less than seventy-five years old and who presented for total hip
replacement during this period were included in the study. The study group
included sixty-four men and thirty-seven women. The mean age of the patients
at the time of the index procedure was 53.8 years (range, twenty-seven to
seventy-five years), and the mean weight was 84.8 kg (range, 50.8 to 128.8
kg). Sixty-seven right hips and fifty-six left hips were involved. The
preoperative diagnosis was osteoarthritis for sixty-five hips (53%),
osteonecrosis for forty-seven (38%), rheumatoid arthritis for six (5%), and
other conditions for five (4%); the other conditions included anky-losing
spondylitis (two hips), posttraumatic osteoarthritis (one), synovial
osteochondromatosis (one), and fusion (one).
Four patients (five hips) died during the study period because of problems
that were unrelated to the operation; none had had a revision or a reoperation
before death. Twenty-five patients (thirty hips) were interviewed by telephone
more than five years after surgery; none of these patients had had additional
surgery, and all reported that they were extremely satisfied with the outcome
and would not return for new radiographs. The remaining seventy-two patients
(eighty-eight hips) had complete radiographic and clinical follow-up.
Operative Technique
A direct lateral modified Hardinge approach was used in all
cases21. The
femoral canal was reamed in a line-to-line fashion with the distal diameter of
the prosthesis to be used, and no channels were cut into the endosteum. As a
result, the stems were locked only in the proximal metaphyseal region. The
fully seated femoral rasp was torque-tested by applying a retroversion force
of 120 in-lb to determine broach motion. If >100 µm of motion was
detected, or if motion was visible, the next size of broach was inserted and
micromotion was reassessed. This technique resulted in a larger-sized rasp and
implant for sixteen hips (13%). The most common size for women was 13 mm
(range, 11 to 17 mm), and the most common size for men was 15 mm (range, 11 to
18 mm).
Physical therapy was begun on the first postoperative day. A standardized
rehabilitation protocol consisting of gait training allowed the patient to
bear partial (20%) weight on the involved limb for the initial six weeks
postoperatively. Progressive weight-bearing as tolerated was initiated at six
weeks. All patients were fully weight-bearing by ten weeks.
Clinical Outcome
The mean duration of follow-up was seventy-eight months (range, sixty to
117 months). Patients were assessed at six weeks; at three, six, and twelve
months; and annually thereafter. All complications related to the hip
arthroplasty, including those requiring revision of the femoral or acetabular
component, were recorded. With revision of the femoral or acetabular component
as the clinical end point, survival statistics were calculated and plotted
according to the method of Kaplan and
Meier22-24.
Preoperative and postoperative Harris hip scores were calculated for each
patient25. All
patients were specifically asked about thigh pain, and those who had pain were
instructed to grade the pain on a visual analog scale from 1 to 10. Consistent
with previous
reports9, mild pain
received a score of 1 to 4; moderate pain received a score of 5, 6, or 7; and
severe pain received a score of 8, 9, or 10.
Radiographic Assessment
Anteroposterior and lateral radiographs of the involved hip (or hips) were
made on the fourth postoperative day and at six weeks. Additional radiographs
were made at three, six, and twelve months and annually thereafter. The
follow-up radiographs were made at an outside institution for thirteen
patients. All radiographs were reviewed by two independent observers (R.K.S.
and J.D.B.). A standard set of measurements for both the femoral component and
the acetabulum was made with use of a caliper that had an accuracy of
±0.02 mm and a goniometer that had an accuracy of ±1°. The
locations of radiographic findings in the femur were recorded with use of the
zones described by Gruen et
al.26 and Johnston
et al.27. When
these zones were related to the particular structure of the HG Multilock
prosthesis, zones 1 and 2 were divided by the junction of the porous and
smooth areas of the implant laterally, zone 4 included the distal smooth tip,
zones 3 and 5 divided the remaining smooth portion of the stem laterally and
medially, and zones 6 and 7 encompassed the proximal-medial area of the
implant. Specific radiographic signs related to implant stability included
radiolucent lines, endosteal hypertrophy (so-called spot welds), implant
subsidence, proximal bone resorption, and rounding of the proximal part of the
medial aspect of the femur (calcar atrophy).
Femoral stem fill was determined by measuring the ratio between the width
of the femoral component and the width of the femoral canal on both the
anteroposterior and lateral radiographs as described by Callaghan et
al.28. Diaphyseal
fill was measured at the midpoint of the smooth fluted stem, and metaphyseal
fill was measured at the proximal border of the lesser trochanter. Canal fill
was defined as contact (or within 1 mm of contact) of the implant with the
medial and lateral endosteal cortices at the same
level28.
Stem position (varus, valgus, or neutral alignment) was established as the
angle formed by the intersection of a line drawn through the longitudinal axis
of the proximal part of the femoral canal and a line drawn through the
longitudinal axis of the stem in accord with the method described by Martell
et al.10. Vertical
subsidence was determined by evaluating the change in distance from the
superolateral edge of the porous coating to the tip of the greater trochanter
on the anteroposterior radiograph. Instability of the femoral stem was defined
as subsidence of >2 mm or a change in the stem angle of >2°, as
described by Campbell et
al.29. Initial
contact between the osteotomy site and the femoral collar was judged to be
present if there was =2 mm of distance between the proximally porous-coated
region of the prosthesis and the calcar at the level of the proximal border of
the lesser trochanter. Cortical remodeling was evaluated by assessing the
proximal femoral osteotomy site for hypertrophy (densification) or rounding
(atrophy) as described by Hedley et
al.30 and by
assessing the region of the fluted stem for periosteal hypertrophy. Formation
of intramedullary bone at the distal tip was classified as none, slight (as
indicated by a radiopaque halo), or complete (as indicated by a pedestal)
according to the criteria described by Engh et
al.31.
Statistical Analysis
The association with femoral osteolysis was evaluated with use of the
Student t test for continuous variables and with use of the Fisher exact test
for categorical variables. Kaplan-Meier survivorship analysis was performed to
determine the cumulative survival rate for the uncemented total hip prosthesis
with revision as the end point and with 95% confidence intervals calculated
with use of the Greenwood
formula32. The Cox
regression model33
was used to identify multivariate predictors of revision with the following
variables as candidates: age, weight, gender, side of involvement, laterality,
diagnosis, femoral osteolysis, head size, polyethylene thickness, cup
inclination, the presence of cerclage wires, the rate of polyethylene wear,
walking distance, and the presence of thigh pain. Hazard ratios were derived
for significant variables. Statistical analysis was conducted with use of SPSS
software (version 11.0; SPSS Science, Chicago, Illinois). A two-tailed p value
of <0.05 was considered significant.
Clinical Results
Harris hip scores were collected prospectively for all hips at the time of
the index operation and for 118 hips at the time of the final follow-up. The
mean Harris hip score was 49 points (range, 21 to 70 points) preoperatively
and 95 points (range, 76 to 100 points) at the time of the latest examination.
The average duration of follow-up was seventy-eight months (range, sixty to
117 months). A good or excellent score (90 to 100 points) was recorded for 117
hips (99%). The remaining hip had a fair score.
No thigh pain was reported in association with 109 (89%) of the 123 hips.
Mild thigh pain (corresponding with a visual analog pain score of 1 to 4) was
described in association with twelve hips (10%) at eighteen months; this pain
typically occurred only with prolonged activity, was not disabling, and did
not require medication. Moderate pain (corresponding with a visual analog pain
score of 5) was described in association with one hip at eighteen months; the
pain was related to activity and was not disabling. At the time of the
five-year follow-up, these minor complaints persisted intermittently in twelve
hips. The thigh pain progressed in only one patient. At the time of the
seven-year follow-up, this patient reported severe pain (corresponding with a
visual analog pain score of 10) and the femoral component subsequently was
revised. At the time of the revision procedure, there was no evidence of bone
ingrowth. The primary diagnosis for this patient had been synovial
osteochondromatosis. Recurrence of the disease was seen at the time of
revision surgery.
Radiographic Assessment of the Femoral Component
Stable osseous ingrowth occurred in eighty-four (95%) of eighty-eight hips,
and stable fibrous fixation occurred in three hips (3%). One stem was
loose.
Two specific patterns of bone ingrowth were seen, with some hips having
endosteal spot welds that were indicative of a circumscribed area of stress
transfer and others having a diffuse area of ingrowth with or without a
circumscribed spot weld (Figs. 2-A and
2-B). Seventy-two hips (82%) had some degree of stress-shielding
in the proximal metaphysis, but only two hips had cortical resorption.
Table I presents the location
of spot welds according to the zones used for the radiographic analysis.
Table II presents additional
data regarding the initial and final positions of the stem.
Osteolysis
Femoral osteolysis was seen in thirty-two hips (38%). The presence of these
endosteal erosions was restricted to the proximal zones. No distal osteolysis
was noted.
Larger head size (32 mm compared with 28 mm) and male gender were the only
predictors of femoral osteolysis. With the numbers available, there was no
significant correlation between femoral osteolysis and age, weight, side of
involvement, bilateral involvement, diagnosis, polyethylene thickness, wear
rate, cup inclination, or thigh pain (see Appendix).
Heterotopic Ossification
Heterotopic ossification was seen in fifty-seven (46%) of the 123 hips
within eighteen months after the index procedure. This finding was localized
to the region of the abductor musculature. According to the classification
devised by Brooker et
al.34, thirty-seven
hips were categorized as class I, thirteen were categorized as class II, six
were categorized as class III, and one was categorized as class IV. At the
time of the most recent follow-up, the location and amounts of heterotopic
ossification remained unchanged.
Revisions
There were a total of five revisions involving four cups and one stem;
specifically, two polyethylene liners were exchanged because of accelerated
wear and pelvic osteolysis, two cups were revised because of recurrent
dislocations, and one stem was revised because of loosening. In all four of
the hips that had acetabular revision, the stem was stable with osseous
ingrowth.
With revision of either component for any reason as the end point, the
cumulative survival rate was 97% (95% confidence interval, 95% to 99%) at
seven years and 78% (95% confidence interval, 66% to 90%) at ten years
(Fig. 3).
With revision of the femoral stem as the end point, the cumulative survival
rate was 100% at seven years and 97% (95% confidence interval, 94% to 100%) at
ten years (Fig. 4).
Complications
Two patients sustained a nondisplaced fracture of the anterior aspect of
the femoral neck during insertion of the femoral component. The fractures were
longitudinal and extended <3 cm distally. Both fractures were treated with
autogenous bone-grafting and cerclage wiring. The fractures united primarily,
and the most recent Harris hip scores for these patients were 97 and 95.
Two patients had recurrent postoperative dislocations of the hip. One of
these patients required polyethylene liner exchange, and the other required
revision and repositioning of the acetabular component. The femoral component
was stable in both patients. A third patient sustained a traumatic dislocation
as the result of a motor-vehicle accident. This patient required open
reduction and internal fixation of a fracture of the greater trochanter.
Three patients had perioperative proximal deep venous thromboses and were
treated with a standard anticoagulation regimen, and one patient had
development of a postoperative hematoma that resolved without sequelae.
A table showing the variables that were assessed with regard to the
association with the presence of femoral osteolysis and a section describing
the methods of evaluating the acetabular components in this series as well as
the results of the radiographic assessment 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).
Note: The authors thank Dr. Harry E. Rubash for his
contributions to the preparation of this manuscript. Dr. Rubash initiated the
study and provided tremendous insight and guidance during its completion.
Also, Drs. James Fenwick and Timothy Conlan contributed to the early
evaluation of patients and collection of data. Dr. J. Dennis Bobyn contributed
to the radiographic evaluation.