Two-stage revision hip arthroplasty for infection using an antibiotic-loaded cement spacer has been used frequently with good results. However, spacer instability is also frequent. Proximal cementation of the spacer could avoid spacer dislocation. We retrospectively assessed 35 patients in whom a 2-stage revision hip arthroplasty for infection was carried out using an antibiotic-loaded cement spacer with gentamicin (Spacer-G) in which the spacer was proximally cemented in 16 patients. The mean follow-up was 32 months. We assessed spacer stability and infection elimination. There were 8 spacer dislocations (22.9%), 5 in hips without proximal cementation and 2 in hips with proximal cementation (p>0.05). There was no fracture in any hip. Reinfection occurred in 5 hips (14.3%), in 3 with the same microorganism, while 2 had a different microorganism. Our results indicate that the proximal cementation of the spacer prevents its dislocation. Infection was eliminated in 86% of the hips.