BACKGROUND: Aseptic loosening has been identified a main failure mechanism of cemented total hip arthroplasty. It is very important to prevent loose by exploring its pathogenesis in studies of corpse specimens with bone cement prosthesis. OBJECTIVE: In this cadaveric retrieval study, we examined the hypotheses that debonding and stem migration were associated with mechanical failure of the cement mantle, and that cracking Of the mantle was indicative of aseptic loosening of cemented femoral stems. DESIGN, TIME AND SETTING: Observation experiment, The study was performed at Baylor college of Medicine and the Institute of Orthopedic Research, Houston, TX, USA between May and November 2003 PARTICIPANTS: A total of 18 femurs with cemented femoral stems were obtained from cadaveric donors in Department of Anatomy, Baylor College of Medicine at Houston, including 7 males, 11 females. METHODS: Eighteen cadaveric femurs with cemented femoral stems were cut from suitable area and immersed in a bath of Alcaim Blue pigment under vacuum for 24 hours to staining. Each femur was bilateral along the sagittal mid-plane and examined with stereomicroscopy to identify cracks within the cement mantle and recorded by Gruen zone. The widths of cracks present on the inner surface of the mantle were measured at three equidistant locations with an instrumented microscope system. The widths of gaps along the stem/cement interface were measured by reassembling its corresponding cement mantle, Any regions of burnishing of the inner surface of the cement mantle and the stem surface were recorded too. MAIN OUTCOME MEASURES: The following parameters were measured: location, width of cracks inner surface of the mantle; widths of gaps along the stern/cement interface; regions of burnishing of the inner surface of the cement mantle and the stem surface. RESULTS: Almost all components were debonded from the surrounding cement mantle. Slight burnishing of the inner surface of the cement mantle was evident in sixteen (89%) specimens. Cracks were present in fifteen (83%) of mantles, frequently corresponding to areas where the corners of the stem had come into contact with bone during implantation, (87%, P = 0.002). None of the stems had subsided by more than 0.5 mm. The average width of gaps between the stem and the cement mantle was (0.33 + 0.06) mm. There was a strong correlation between the minimum width of longitudinal cracks and the average stem/cement gap for each specimen (r~2=0.691, P= 0.001). CONCLUSION: Cement cracks are frequently observed in clinically successful cemented hip replacements. Stem migration occurs once cement fractures open. Cement cracks do not necessarily result in clinical failure of cemented stems. But, with the development of micromotion, the debris cause osteolysis and might affect implant stability and lead to harmful effects on the bone bed and subsequent loosing of the implant.