The difficulty in differentiating aseptic loosening from infection in painful total hip arthroplasty is well recognized. This prompted a review of the efficacy of the preoperative investigations used at the authors' institution. One hundred forty-four patients who underwent revision hip arthroplasty were reviewed. Seventy-two had sequential bone-gallium scan, and/or hip aspiration, and intraoperative Gram stain. These were compared to intraoperative culture as the gold standard. Twenty infected hips were detected on culture. For prediction of infection, the bone-gallium scan had a sensitivity of 38% and a specificity of 100%. Hip aspiration had a sensitivity of 57% and a specificity of 97%. The investigations combined gave a sensitivity of 64% and a specificity of 95%. Intraoperative Gram stain compared with subsequent culture yielded a sensitivity of 23% and a specificity of 100%. Uncemented hips were infected more frequently (47%) than cemented hips (9%), significant at P < .0001. It was concluded that bone-gallium imaging is not an effective method for investigating painful hip prostheses for sepsis and offers no additional advantage over hip aspiration. Intraoperative Gram stain also missed a large number of infections. Other modalities, such as indium-labeled-leukocyte imaging and capsular histologic examination, may be more efficacious. A significant difference in the number of infections found in cemented versus uncemented hips was shown, hence greater vigilance for infection is required when patients present with painful uncemented hip arthroplasties.