Purpose: The rate of dislocation after revision total hip arthroplasty (RTHA) has varied from 8 to 28% in published series. Many causes are involved, but little work has been focused on the incidence of spinal disease in patients with postoperative dislocation.

Material and methods: We performed a prospective analysis of 267 patients who had undergone RTHA in search of risk factors of postoperative dislocation. Chi-square test or Student’s t test were used for the statistical analysis as appropriate. P < 0.05 was considered significant.

Results: We excluded 37 patients who had undergone first line THA with a restrained cup and who had not experienced dislocation. The 230 patients retained for analysis had undergone primary surgery with no intra or postoperative anti-dislocation measure. Among these, 31 (13.4% experienced dislocation). The cause was evident in ten cases (malposition, fracture of the greater trochanter, sciatic paralysis). For the other 21 patients, age, gender, types of surgery (uni or bipolary), surgical approach, size of the implant, and size of the femoral or acetabular defects were not found to influence the rate of postoperative dislocation. Conversely, a significant relationship was found between increased rate of dislocation and history of repeated dislocations (p < 0.001), prior surgery (p< 0.05), and association with spinal disease (p< 0.02). Characteristically, there was either radicular, or spinal disease, or both. Retrospective analysis of the radiograms showed that the measures of sacral incidence and inclination of lumbar lordosis were not predictive of dislocation. Inversely, the projection of a vertical line passing through the centre of rotation of the hips on L3 was different in patients who had experienced dislocation and those who had not (p< 0.02).

Discussion: This study confirms the role of a history of dislocation and prior surgery in the risk of postoperative dislocation. It also shows that associated spinal disease, which may results from radicular disease, as well as altered spinal static can have an influence. A prospective study is currently under way to distinguish these features.