We present the mid term results of 267, 3M Capital hip replacement performed in a single institution from 1991 to 1994. Patients were recalled back to clinic in April 1998 after the reported high failure rate of 3M hips. Fifty- nine hips were excluded from this study for various reasons.

Average follow-up was 68.8 months. Nine (4.2%) have been revised for aseptic loosening at the time of review and further 10(4.8%) of the stems are radiologically loose. There was no statistical significant difference between hips that failed, regarding grade of the operating surgeon, surgical approach: trochanteric osteotomy or Hardinge, or type of prosthesis used: monoblock (stainless steel) or modular (titanium). Acetabular wear rate, width of medial cement mantle or cancellous bone at level of neck at Gruen zone7, or stems with canal fill index less than 50%, 7 cm below the level of the collar also showed no statistical difference. Male patients had higher incidence of loosening (p=0.001) which was statistically significant. Both varus and valgus alignment of the stem had higher failure rate which was significant. We could no find any obvious reason for failure in 10 of the 18 patients. The stem was either in varus or there was an inadequate cement mantle in 8 of the failed hips on the initial postoperative radiograph. A feature of this study was high incidence of endosteolysis and debonding of prosthesis from cement in the failed cases.

The present series showed considerably lower revision and loosening rate of 3M stems compared to the published series, the reason for which is not clear. Only Palacos cement was used in this series, which may partly account for the lower failure rate. Surface finish of the stem leading to debonding of the prosthesis from cement along with different modulus of elasticity probably accounts for the higher rate. Technical failure is partly to blame for the higher failure rate.