Long, cementless, femoral stem revisions are being used with increasing frequency. There is a relative lack of studies of late fractures after cementless implants, particularly in those patients who had a previous stem revision and are at higher risk for periprosthetic fracture after revision. In this paper, we review risk factors for periprosthetic fracture and revisions of long, cementless, locked stems and report implant survival compared with conventional, cemented, long-stem hip revision arthroplasties in such a group of patients. We used data recorded in our institution. Between 1996 and 2002, 85 cementless femoral stem Aura™ (distal locked stem) prostheses were implanted in 85 patients. Of the 85 revisions with long, cementless stems, 32 were performed after one stem revision, 20 after two stem revisions and 12 after three stem revisions. Between 2003 and 2010, 124 femoral revision stems were performed in 124 patients using an extensively long, titanium femoral stem (Ceraver Osteal™). Kaplan-Meier analysis was used to determine implant survival. The Cox regression model was used to study risk factors for reoperation and revision. The increase in stem length corresponded to a mean of 4.5 ± 2.1 femoral canal diameters and was not significantly different (p = 0.02) between the two groups. Cardiopulmonary distress, intraoperative or postoperative complications were not significantly different between the two groups. There was an overall increased risk of pain, periprosthetic fractures and revision for the cementless prostheses compared with the cemented stems. With regard to thigh pain at the last follow-up, most patients (95 %) reported no pain in the cemented group, while 15 % of the cementless group experienced thigh pain under stress and 6 % had incapacitating pain. In the cemented group, stem re-revision was not required in any hip; there was no periprosthetic fracture; five stems had radiological loosening but in the absence of pain were not revised. Among patients of the cementless group, 21 % sustained failure or revision of their interlocked stem with periprosthetic fractures observed in 15 cases; factors which contributed significantly to a higher risk of fracture included the number of previous revisions; the average time between surgery and failure was 3.2 years after one revision, 2.8 years after two revisions, and 1.6 years after three revisions. The long, cementless, locked stem showed more early complications compared with recementing of long-stem prosthesis. We therefore recommend the use of cemented long stems in patients with severe bone loss and previous revision.