Abstract
This report reviews the long-term results of treating acetabula with unusually severe problems, such as pelvic discontinuity or major column loss after failed total hip arthroplasty (THA) and reconstruction problems.
Loss of acetabular bone stock results from removal of bone during the original procedure, prosthetic failure, and osteolysis. In massive structural failure, the acetabular rim, quadrilateral plate, and associated columns become deficient. At worst, this may be combined with pelvic discontinuity and disruption of the ilium and ischium. Prosthetic protrusio may result from fixation loss and be associated with scarring of the femoral vessels, femoral nerve, ureter and bowel. A variety of implants has been used to in ace-tabular reconstruction. The results are often poor because of insufficient bone stock to support the implant.
In a consecutive series of 251 THA revisions done between 1988 and 1996, 17 patients were treated for major pelvic column loss, pelvic discontinuity or both.
In five patients, a posterolateral approach without trochanteric osteotomy was used. The extensile triradiate approach with ilioinguinal extension was used in 12 patients in whom severe prosthetic protrusio increased the risk of intrapelvic iatrogenic injury. A long anterior column pelvic plate was applied. A posteriorly placed AO 4.5-mm pelvic reconstruction plate with 10 to 12 holes was used in nine cases of pelvic discontinuity and in five cases of posterior column bone loss. This plate extended from the most inferior extent of the ischium across the wall of the posterior column to a point high on the ilium. Anterior column fixation was done in eight of nine cases of pelvic discontinuity and all three cases of anterior column deficiency. This called for an 8 to 12-hole 3.5-mm AO pelvic reconstruction plate that extended from the pubic symphysis across the pelvic rim. This spanned the anterior column defect, ranging from 4 cm to 8 cm, to the medial wall of the ilium.
Bulk allograft was used in 16 of the 17 patients. The patient in whom allograft was not used had pelvic discontinuity following pelvic irradiation. Whole pelvic acetabular transplants were used in seven with severe bone loss or following resection for chondrosarcoma and the other for pigmented or villonodular synovitis. Posterior segmental acetabular allograft was used in two cases of posterior column absence. Femoral heads were used in two posterior column defects, three pelvic discontinuities with anterior column defect, and two anterior column defects. Acetabular components were cemented in six of seven whole bulk ace-tabular transplants, six of nine pelvic discontinuities and two anterior column defects.
Cemented implants were classified as loose if there was a complete radiolucent line at the bone cement interface, measurable component migration or measurable change in position. Uncemented acetabular components were considered loose if component migration had occurred or screws had broken. Pelvic plates were considered loose if there was measurable migration or change in plate position or if fixation screws had backed out or broken.
Radiographic union was considered present when bridging callus or trabecular bone was visible across the discontinuity site. Junctional healing was considered probable when radiographs did not show obvious signs of failure. Grafts were considered unhealed if there was obvious displacement, bone gaps or hardware breakage.
Seven of the nine patients with pelvic discontinuity had late evidence of healing of the fracture and allograft consolidation. One underwent removal of the graft at three weeks after developing acute postoperative infection: early junctional healing of a whole bulk acetabular allograft required an osteotomy to break up the interface. Another patient, who underwent removal of the graft and implant at three months for chronic infection, had consolidation of a whole bulk ace-tabular allograft. One patient underwent revision of a pressfitted acetabular component at 60 months, and the pelvic discontinuity was solidly united. In a fourth patient, explored at 124 months for loosening of a cemented cup, there was near complete dissolution of the graft posterior acetabular wall and a loose posterior pelvic plate. In a patient with pelvic discontinuity after radiation therapy for uterine carcinoma, satisfactory healing of the pelvic discontinuity was confirmed at 32 months, when excisional arthroplasty for late chronic infection followed urinary sepsis.
Seven patients had major column loss with severe cavitary defects. Consolidation of the allograft was noted in all seven within the first 12 months of follow-up.
Revision (47%) was required for infection in three patients, implant loosening in four, and recurrent implant dislocation in one. The four loose cups were revised to a cemented all-polyethylene component. All four implants had been placed on less than 50% host bone. None of the four has required subsequent revision.
Dislocation postoperatively occurred in eight patients. In six, the extensile triradiate approach had been used. This approach led to dislocation in 50%. The main reasons for using the extensile triradiate approach were to avoid catastrophic injuries by direct exposure of vital structures and to allow stable anterior column plate fixation. In that no neurovascular injuries occurred and stable durable allograft consolidation and healing of pelvic discontinuity took place, these goals were largely met.
Three patients developed late sciatic palsy. In one, plaster immobilisation had possibly caused direct pressure over the fibular head and led to chronic peroneal palsy. The other two underwent additional exploration of the sciatic nerve for late entrapment caused by migration of screws from the posterior column plate. Two patients developed bladder infections postoperatively. Another developed superficial phlebitis of the lower leg.
Acetabular revision for loosening was necessary in three of seven cementless implants, while only two of 10 cemented implants failed. The acetabular component should be cemented into the allograft when more than 50% of the prosthetic interface is non-viable.
Virtually all graft material, including dense cortical grafts, may ultimately fail if used for implant fixation. Patients should be told about the inevitable risks. However, techniques used led to stable healing of the pelvic discontinuity in most cases. Long pelvic plates that securely stabilise the pelvis and allografts carefully opposed to host bone may explain the relative success in this series.
The abstracts were prepared by Professor M. B. E. Sweet. Correspondence should be addressed to him at The Department of Orthopaedic Surgery, Medical School, University of Witwatersrand, 7 York Road, Parktown, Johannesburg, 2193 South Africa
