We report the results of 79 patients (81 hips) who underwent impaction grafting at revision hip replacement using the Exeter femoral stem. Their mean age was 64 years (31 to 83). According to the Endoklinik classification, 20 hips had a type 2 bone defect, 40 had type 3, and 21 had type 4. The mean follow-up for unrevised stems was 10.4 years (5 to 17).

There were 12 re-operations due to intra- and post-operative fractures, infection (one hip) and aseptic loosening (one hip). All re-operations affected type 3 (6 hips) and 4 (6 hips) bone defects. The survival rate for re-operation for any cause was 100% for type 2, 81.2% (95% confidence interval (CI) 67.1 to 95.3) for type 3, and 70.8% (95% CI 51.1 to 90.5) for type 4 defects at 14 years. The survival rate with further revision for aseptic loosening as the end point was 98.6% (95% CI 95.8 to 100). The final clinical score was higher for patients with type 2 bone defects than type 4 regarding pain, function and range of movement. Limp was most frequent in the type 4 group (p < 0.001). The mean subsidence of the stem was 2.3 mm (sd 3.7) for hips with a type 2 defect, 4.3 mm (sd 7.2) for type 3 and 9.6 mm (sd 10.8) for type 4 (p = 0.022).

The impacted bone grafting technique has good clinical results in femoral revision. However, major bone defects affect clinical outcome and also result in more operative complications.

Metal meshes are used in revision surgery of the hip to contain impacted bone grafts in cases with cortical or calcar defects in order to provide rotational stability to the stem. However, the viability of bone allografts under these metal meshes has been uncertain.

We describe the histological appearances of biopsies obtained from impacted bone allografts to the calcar contained by a metal mesh in two femoral reconstructions which needed further surgery at 24 and 33 months after the revision procedure.

A line of osteoid and viable new bone was observed on the surface of necrotic trabeculae. Active bone marrow between these trabeculae showed necrotic areas in some medullary spaces with reparative fibrous tissue and isolated reactive lymphocytes. This is interpreted as reparative changes after revascularisation of the cancellous allografts.

These pathological findings are similar to those reported in allografts contained by cortical host bone and support the hypothesis that incorporation of morcellised bone under metal meshes is not affected by these devices.

We determined the midterm survival, incidence of peri-prosthetic fracture and the enhancement of the width of the femur when combining struts and impacted bone allografts in 24 patients (25 hips) with severe femoral bone loss who underwent revision hip surgery. The pre-operative diagnosis was aseptic loosening in 16 hips, second-stage reconstruction in seven, peri-prosthetic fracture in one and stem fracture in one hip. A total of 14 hips presented with an Endoklinik grade 4 defect and 11 hips a grade 3 defect. The mean pre-operative Merle D’Aubigné and Postel score was 5.5 points (1 to 8).

The survivorship was 96% (95% confidence interval 72 to 98) at a mean of 54.5 months (36 to 109). The mean functional score was 17.3 points (16 to 18). One patient in which the strut did not completely bypass the femoral defect was further revised using a long cemented stem due to peri-prosthetic fracture at six months post-operatively. The mean subsidence of the stem was 1.6 mm (1 to 3). There was no evidence of osteolysis, resorption or radiolucencies during follow-up in any hip. Femoral width was enhanced by a mean of 41% (19% to 82%). A total of 24 hips had partial or complete bridging of the strut allografts.

This combined biological method was associated with a favourable survivorship, a low incidence of peri-prosthetic fracture and enhancement of the width of the femur in revision total hip replacement in patients with severe proximal femoral bone loss.

We reviewed the clinical and radiological results of 131 patients who underwent acetabular revision for aseptic loosening with impacted bone allograft and a cemented acetabular component. The mean follow-up was 51.7 months (24 to 156).

The mean post-operative Merle D’Aubigné and Postel scores were 5.7 points (4 to 6) for pain, 5.2 (3 to 6) for gait and 4.5 (2 to 6) for mobility. Radiological evaluation revealed migration greater than 5 mm in four acetabular components. Radiological failure matched clinical failure. Asymptomatic radiolucent lines were observed in 31 of 426 areas assessed (7%). Further revision was required in six patients (4.5%), this was due to infection in three and mechanical failure in three. The survival rate for the reconstruction was 95.8% (95% confidence interval 92.3 to 99.1) overall, and 98%, excluding revision due to sepsis.

Our study, from an independent centre, has reproduced the results of the originators of the method.

Bone allografts can store and release high levels of vancomycin. We present our results of a two-stage treatment for infected hip arthroplasty with acetabular and femoral impaction grafting using vancomycin-loaded allografts. We treated 29 patients (30 hips) by removal of the implants, meticulous debridement, parenteral antibiotic therapy and second-stage reconstruction using vancomycin-supplemented impacted bone allografts and a standard cemented Charnley femoral component. The mean follow-up was 32.4 months (24 to 60). Infection control was obtained in 29 cases (re-infection rate of 3.3%; 95% confidence interval 0.08 to 17) without evidence of progressive radiolucent lines, demarcation or graft resorption. One patient had a further infection ten months after revision caused by a different pathogen. Associated post-operative complications were one traumatic periprosthetic fracture at 14 months, a single dislocation in two hips and four displacements of the greater trochanter. Vancomycin-supplemented allografts restored bone stock and provided sound fixation with a low incidence of further infection.