The Atlas cup ⁶ is a press-fit acetabular component. It consists of a hemispherical shell made of a titanium alloy (TiAl6V4) with a split in the lower part to provide elasticity. The titanium shell is very thin (2.5 mm), so as to allow a maximal thickness of polyethylene (PE) to be inserted. The first cups of this pattern, used between July 1987 and December 1989, were of grit-blasted titanium alloy and did not have a hydroxyapatite (HA) coating. Since then, an HA-coated pattern has been used. In 1995, we performed a comparative study ⁷ of the porous titanium versus the HA-coated pattern and found the HA-coated cup to be superior. The present study reflects the follow-up period of more than ten years that has elapsed by now. The investigation described herein was a prospective study of a consecutive series of the first 151 HA-coated Atlas cups used, at primary arthroplasty, between 8 January 1990, and 27 September 1991.

All the acetabular components used in the study were HA-coated Atlas cups. The Atlas cup is made of TiAl6V4. It is hemisperical, and has a split that provides elasticity. The cup size chosen is 2 mm greater than the diameter of the final reamer. As the cup is impacted into place, the split closes; when the insert is placed in the cup, the split opens up again and produces a strong press-fit effect. ⁸

The stability of the insert is ensured by the circumferential clamping action of the cup around the peripheral cylindrical portion of the insert. This design counteracts any tendency of the insert to tilt ⁸ .

The thickness of the cup shell is limited to 2.5 mm, so as to allow a maximal thickness of PE to be inserted.

The femoral heads used were 22.2-mm diameter in 145 cases, and 26-mm diameter in six cases. Like the thin shell, these small heads allowed a greater thickness of PE to be used. The PE insert was always at least 10 mm thick.

On the femoral side, a PSM stem (HA-coated in the metaphyseal portion) was used in 147 cases; the remaining four cases were managed with a variety of cementless implants.

The study described herein was a prospective study of a consecutive series of 151 primary hip replacements performed at our center between 8 January 1990 and 27 September 1991. The mean patient age was 62 years (range, 25 to 82 years). Of the patients, 78 were women, and 73 men. There were 72 right and 79 left hips. In 129 cases, the indication for arthroplasty was primary osteoarthritis (OA) of the hip; eleven patients had developmental dysplasia of the hip (DDH); two had rheumatoid arthritis (RA); and two had osteonecrosis.

Twelve patients had died, and 16 were lost to follow-up. Fourteen patients were reoperated on: one for infection, three for dislocation, four for failure of the stem to obtain primary stability, four for metallosis caused by the deterioration of the titanium femoral head, and two for cup loosening. This left 109 patients for review in the present study, at a follow-up of eleven years post-arthroplasty in 61 cases, and of ten years in 48 cases. For the clinical assessment, the MDA scoring system was used. The radiographic assessment was concerned with the detection of lucent lines, cysts, and evidence of loosening, on the a.p. and the lateral films. Wear was measured on the A/P hip films using the Charnley method, ⁵ taking magnification into account. Cumulative survival out to eleven years could be calculated, with the Kaplan-Meier method, for the 151 hips that had undergone replacement.

B The one revision for deep infection occurred in the first year. The two instances of cup loosening were at six years and nine years, respectively. The first of these two cases was that of a female patient who had had several falls. At revision surgery, a deep notch was seen in the posterior part of the acetabular rim, which was interpreted as evidence of impingement. The other case had to be revised because the patient had fallen and sustained a fracture of the acetabular fossa, with loosening of the implant. This was, therefore, a case of intercurrent trauma necessitating revision.

In the 109 cases that were followed up, the MDA pain score was 6 in 95 cases, 5 in ten cases, and 4 in five cases. The MDA mobility score was 6 in 98 cases, 5 in seven cases (who had slight stiffness), and 4 in four cases (whose hips were stiff). The MDA function score was 6 in 93 cases (patients walking normally), 5 in seven cases (slight limp), and 4 in nine cases (limp). Seven of the last-mentioned nine patients were DDH cases.

Radiography showed a total absence of any periacetabular cysts. In the femur, there were 15 cysts, all of which were confined to the greater trochanter (Figure 3); there were no extensive cysts and no extensive lucent lines.

At the level of the cup, non-progressive lucent lines of less than 3 mm width were seen in Zone III in seven cases (6.2%). There were no extensive lucent lines.

Wear was 0 mm in eleven cases, 0.5 mm in 31 cases, 1 mm in 42 cases, 1.5 mm in 16 cases, 2 mm in seven cases, and 2.5 mm in one case. Mean wear was 0.90 mm. The mean annual wear rate in this series was 0.82 mm/year.

Mean cumulative survival was calculated, with the Kaplan-Meier method, for the 151 hips in the series. It was calculated out to eleven years, since 78 of the hips were still "alive" by the end of that time. The seven stem failures were not included in the analysis, since the study was concerned with cup survival. Also, following the policy of the Scandinavian registries, it was decided to apply the term 'revision' only to reoperations involving the exchange of at least one of the two major components in direct contact with the host bone (i.e. the cup and/or the stem), while the exchange of an acetabular insert or a femoral head would not be considered to be a 'revision'.

There were three failures involving the Atlas cup, one each in the first, the fifth, and the ninth year post-operatively. Two were revisions for 'intercurrent' reasons (one deep infection in the first year, the other as a result of an accidental fracture at nine years). Only in the case of the cup that needed revising at five years was failure of the implant fixation to blame. Thus, when analyzing cup failure as such, the rate was 1.98% (3/151 cups). This figure comprises the two failures not attributable to the implant as such, and one failure of implant fixation. The failure rate of the HA fixation of the Atlas cup at ten years was thus 0.66% (1/151 cups).

There were two instances of cup loosening. The ten-year survival, taking cup failure as the endpoint, was 99.18%. Four cases of metallosis were associated with the use of titanium femoral heads. ⁹ This type of deterioration remained rare (3.4%) over the period for which the patients were followed up. It is not impossible that further cases of metallosis will be encountered at a later stage of the in-service life of the implants.

The absence of macrophage-filled cysts around the cup provided evidence of the firm seating of the PE insert in the cup shell. The insert fixation system by circumferential clamping in the cup proved efficient in the clinical study, as it had done previously in a retrieval study. ¹⁰ In that study, no macroscopic or microscopic evidence of backside deterioration of the inserts had been found after more than ten years post-operatively. These results were certainly due, in some measure, to the absence of microfretting between the insert and the shell thanks to the stable seating of the insert, and to the great thickness of the PE. Another factor was the size of the femoral heads used. Most of the heads were 22-mm diameter ones, and this size is known to produce little wear debris and, consequently, little macrophage reaction. The findings in this study thus bear out Charnley's low-friction theory.

Lucent lines around the cup were rare (6.2%) and confined to Zone III. It is interesting to recall that, in a comparative study in 1995, ⁶ we found the non-HA-coated Atlas cups to be associated with a 40% rate of Zone III lucent lines.

Many authors ^{2 11 121317} have described the barrier to the migration of PE wear particles provided by the intimate contact, without any interposed fibrous tissue, between the host bone and a circumferential HA coating on the implant. It would appear that, in the case of the implants used in the present study, this barrier effect resulted in any femoral cysts being confined to the greater trochanter.

Linear PE wear was slight (0.08 mm/year). While Bloebaum et al ³ have voiced concern over possible third-body wear by dislodged HA particles becoming embedded in the PE bearing surface, this study showed that a good quality HA coating does not constitute such a hazard.

While Nashed¹⁶ , Bankson¹ and Hernandez¹⁴ with metal backed acetabular prostheses were refuted by the publications of Markel¹⁵ and of Callaghan⁴ found low wear rates. Wear is a multifactorial phenomenon, with PE quality, PE thickness, femoral head quality, patient's level of physical activity, component positioning, and other factors each playing a role. Under the conditions of use described in this study, there appears to be no cause for concern about accelerated PE wear.