|
INTRODUCTION
Revision
of loosened acetabular implants has generally given
disappointing results 1 3 7 9.
At the 1989 SOFCOT Symposium, Salvan 9
reported 22.9 % of recurrent loosenings, and 22.9 % of
potential loosenings, i.e.only 54.2 % of favourable results
among 210 cases of simple revision for acetabular implant
loosening.
The use
of a Müller-type reinforcement ring gave better
results 5 6,
and using this technique in 1993 Pascarel
6
reported only two failures in 141 cases at four years of
follow-up. However, the long-term results of these rings
wich cannot be incorporated remains an open question.
Remarkable
results at over ten years of follow-up have been recorded in
first-line surgery with impacted hemispheric acetabular
implants whose treated surface is treated surface allows an
osteointegration.. Can this technique be used in revisions?
Will grafting be necessary to restore bone stock?
To
address these questions, we studied a continuous series of
229 acetabular revisions by impacted implant with a mean
follow-up of six years and four months (four to nine
years).
MATERIALS AND METHODS
Between
January 1987 and September 1992, 229 acetabular implant
revisions by impacted implant were performed : 144 women and
85 men, 121 right and 108 left,mean age 66 years (18 to 90
years), 20 patients were over 80 years of age and 77 over 70
(30 %).
Acetabular
destruction was stage I in 47 cases, without substantial
loss of substance, stage II in 146 cases, with enlarged
cavity trough substance loss but without major deterioration
of the columns and lower edge, and stage III in 36 cases,
with marked acetabular deterioration.
Series A 1 :
From
1987 to 1989 we used the ATLAS 1 acetabular implant (split
hemispheric implant covered in a 1.5 mm layer of sanded
titanium, with 4.5 mm diameter screws) (91 cases). In each
case we reconstructed bone stock by impacting in the
acetabular cavity small cubes taken from femoral
heads stored in a bone bank. The implant was impacted in
this morsellized graft, and the screws transfixed the grafts
and were fixed in the healthy bone of the roof and columns.
Most of these acetabular implants were small : 50, 52 and 54
mm.
Series A 3 :
Since
1989 (138 cases), we have used the Atlas 3
hydroxyapatite-coated implant (2.5 mm titanium coating,
optional 6.5 mm screw, hydroxyapatite surface). A large
Atlas implant was impacted to fill the bone cavity and
ensure extensive contact between the surfaces of the implant
surface and the healthy acetabular bone. Little or no bone
grafting was done and the implants were large : mostly 56 ,
58 , 60 and up to 68 mm (Series A 3).
Follow-up at review was a mean six years and four mounths
(four to nine years).
Clinical
evaluation was done using the Merle d'Aubigné scoring
system. Radiographic assessment was done using topographic
classification into three zones: supraexternal, two central,
three inferointernal.
Survival was estimated by the actuarial method.
FAILURES AND PATIENTS LOST TO FOLLOW-UP
Series A 1 :
Of 91
patients with grafts, 25 died due to causes unrelated to the
implant and six were lost to follow-up. Among the remaining
60 patients (66 %), there were eight (13 %) acetabular
failures and nine femoral failures. The nine-year survival
rate was 88 % (table I). Forty-three patients were reviewed,
15 at nine years of follow-up, 20 at eight years, six at
seven years, and two at six years.
Series A 3 :
Of 138
patients, nine were lost to follow-up and 20 died. Among the
remaining 109 patients (80 %), there were three (3 %)
acetabular failures and five femoral revisions. The
eight-year survival rate was 97 % (table II). One hundred
and one patients were reviewed, 12 at four years of
follow-up, 38 at five years, 34 at six years, 14 at seven
years, and three at eight years.
Failures
The
eight failures in series A1 were due to resorption of the
grafts with secondary migration of the implant. As the cup
held by fixation screws could not follow this resorption of
the grafts, it was subject to painful micromotion with
rocking of the implant.
The
three failures of series A3 were related to an error in
indication : they were stage III cases with considerable
substance loss and disappearance of the lower edge.
Protrusion occured in the months following surgery. We
should have used a peripherally supported reinforcement
ring.
RESULTS
Clinical and
functional
Series A 1 - 43 revisions:
Results were excellent in 22 cases (51 %),
very good in 12 (28 %) and mediocre in 9 (21 %). Pain was
rare, and was absent (D6) in 38 patients, rare and slight
(D5) in two, and troublesome (D4) in two. Mobility was good
(M6) in 40 patients, moderate (M5) in two, and mediocre (M4)
in one. Function was generally less satisfactory with
frequent persistence of limping : no limping (F6) in 26
patients, slight limping (F5) in six, and clear limping and
use of a stick (F4) in 11.
Series A 3 - 101 revisions.
Results were excellent in 48 cases (48 %),
very good in 30 (30 %), good in 20 (20 %) and mediocre in 3
(2 %). Pain was absent (D6) in 71 patients, slight (D5) in
17, and troublesome (D4) in 13. Mobility was good (M6) in 95
patients, moderate (M5) in two, and mediocre (M4) in four.
Function was less satisfactory in this series: no limping
(F6) in 59 patients, slight limping (F5) in 31, and with
clear limping and use of a stick (F4) in 11.
Radiological results
Series A 1:
In the 43 cases, there was no reactive line 1
in 20 cases and a limited, nonprogressive reactive line in
zone 3 in 12 cases. Six reactive lines were complete in
zones one, two and three, five migrations of over 2mm were
seen, of which one with fracture of screws. The latter 11
cases, with extensive reactive lines or implant migration,
were obviously compromised and will probably progress to
failure, requiring surgical revision.
Series A 3:
In 85 of 101 cases (84 %) there was no
reactive line, and nonprogressive reactive lines were seen
in 16 cases in zone 3. No migration and no complete reactive
lines were seen in this series (table III). On the contrary,
the periprosthetic cavities were filled.
| 
