A retrospective review of 1542 Charnley low-friction arthroplasties was carried out to compare the incidence of deep infection arising after plain and gentamicin-containing acrylic cement. There was no significant difference between the two series in primary operations, the infection rate being 1.72% and 1.65% respectively. However, in secondary operations (conversions and revisions) the gentamicin-containing cement gave significantly better results--a 0.81% infection rate as compared with 3.46%. All operations were performed in a clean air enclosure using total-body exhaust suits. Radiological diagnosis of deep infection was possible within one year of operation in all cases. Late haematogenous infection occurred in one case and was related to a skin lesion.

We examined 59 cemented high density polyethylene sockets removed at revision hip arthroplasty. Of these 19 showed areas of wear between the outside of the socket and the acetabular bone. This was associated with lack of acrylic cement in those areas and was also related to the depth of the wear on the articulating surface of the socket. It is suggested that, in some cases, changes at the bone-cement junction are secondary to socket loosening and abrasion against the bone of the acetabulum, rather than to particles migrating from the metal-polyethylene interface. It is therefore important that impingement of the neck of the femoral stem on the edge of the cup be avoided and that, when the socket is inserted, it is not in direct contact with the bone.

Twenty-two high-density polyethylene sockets from Charnley low-friction arthroplasties have been studied. Acrylic casts and shadowgraph techniques were used to measure both the real and radiographic rates of wear; these rates showed a significant correlation. In the sagittal plane, nine of the sockets had worn lateral to a line drawn vertically from the centre of curvature of the socket, 12 had worn medial to that line and only one was worn exactly in the line. In the coronal plane, nine sockets had worn in front of a similar vertical line, two behind that line and 11 had worn exactly in the line. Evidence of impingement of the neck of the stem onto the rim of the socket was found in 14 patients; this is considered to be one of the causes of socket loosening. The obvious solution is to reduce the diameter of the neck of the stem from 12.5 mm to 10 mm; provided that it is made of cold-formed, high nitrogen-content stainless steel, this narrower neck is strong enough not to fracture.

A method of performing a biplanar intracapsular trochanteric osteotomy with a Gigli saw was designed and tested prospectively in 431 cases of Charnley low friction arthroplasty. Three methods of trochanteric reattachment were tested, and a double cross-over wire with a compression spring was best; this method was successful in 222 out of 226 patients (98.2%), of which half were revision operations. Adduction seemed to be the main movement leading to trochanteric detachment.

In order to improve the fixation of the stem in primary Charnley low-friction arthroplasty, the medullary canal was routinely closed off with a cancellous bone block. A prospective study of 611 consecutive arthroplasties were carried out between 1 and 5 1/2 years (average 2 years and 9 months) after operation. There were two cases of radiological loosening of the stem, both of which could have been avoided. Five other patients showed demarcation of the bone-cement junction at the calcar; two of these are considered to be "at risk" of loosening. There was no evidence of radiological loosening of the stem in 604 cases and the method is recommended for routine use in primary total hip arthroplasty.

A retrospective study of 44 diabetic patients who, between them, had 62 Charnley low-friction arthroplasties, showed a superficial-infection rate of 9.7 per cent and a deep-infection rate of 5.6 per cent. All the operations were carried out in the Charnley clear-air enclosure and prophylactic antibiotics were not used. A statistically significant increase in the overall rate of infection was found in diabetic patients when compared with non-diabetic osteoarthritic patients (P less than 0.001) and rheumatoid patients (P less than 0.01). Hence it is suggested that in diabetic patients there may well be a place for prophylactic antibiotics in hip replacement surgery in addition to the use of the clean-air enclosure.

Twenty-one cases of the Charnley low-friction arthroplasty were revised because of recurrent or irreducible dislocation. Three main reasons for this revision, usually in combination, were identified: loss of the abductor mechanism due to trochanteric detachment; shortening of the limb due to high placement of the socket or low section of the femoral neck; and malorientation of the components. Sixteen patients had no further problems after revision. One patient had a single dislocation, four had more than one dislocation although they managed to cope with them and did not require a further revision.

A series of 400 patients (average age 57.7 years) with bilateral low-friction arthroplasties of the hip has been reviewed. Comparison was made with similar unilateral operations. The results showed that the bilateral arthroplasties can be carried out as a single operation without higher risk of local or general complications. However, this excludes the incidence of pulmonary embolism which is higher for the bilateral replacements.

Wear of high-density polyethylene on bone and cartilage has resulted in a large volume of plastic particles being shed into the two knees and two hips studied. The giant-cell foreign-body reaction of the synovium may not be sufficient to cope with the amount of debris presented and the destruction of the endosteal bone in one hip, caused by the wear particles and movement of the prosthesis, has made revision impossible. Articulation of high-density polyethylene against bone or cartilage either by design or by the failure of alignment of the component must be avoided.