Twenty-three case series and five studies of implant registries were included. Eighteen studies had a quality score of 6 or less out of the maximum 10.
Acetabular fractures (four studies): Rates of acetabular fractures ranged from 2% to 5% .
Migration (six studies): Three studies reported rates of migrated cups of at least 10%.
Early loosening (five studies): One study reported three loose cups after an average period of 31 months. One study reported that two cups were loosened within two years after implantation. Three studies reported single cases of early loosening.
Aseptic loosening: Six studies reported failure rates higher than one (the NICE criteria). No significant difference was found in failure rates between cementless and cemented cups (rate ratio 0.6, 95% CI 0.14 to 2.60; six studies). The failure rate for the cup in case series without control group compared to the NICE criteria (failure rate/1) was 0.97 (95% CI 0.50 to 1.88; 16 studies).
National arthroplasty registries (four studies): In Finland results of press-fit porous coated cups after 10 years in rheumatoid arthritis were the same as those for cemented cups in patients aged below 55 years; uncemented cups performed better than cemented cups in patients aged at least 55 years. In Sweden cementless cups performed better than cemented cups in the general population. In Denmark the cumulative risk for revision for aseptic loosening after 14 years was slightly higher for rheumatoid arthritis than for osteoarthritis.
Femoral fractures (nine studies): No failure of implants was reported following conservative treatment or intra-operative direct fixation.
Subsidence (10 studies): In one study where subsidence rate was high (80%) only five out of 56 subsidised stems were found loose after a mean follow-up period of 14 years. Only one stem was found loose due to aseptic loosening in the remaining studies.
Early loosening: One episode was reported in one study.
Aseptic loosening (22 studies): Two studies reported increased failure rates. There was no significant difference in overall failure rates between cemented and uncemented stems (rate ratio 0.71, 95% CI 0.06 to 8.55; six studies).
The overall failure rate for the stem in case series without control (compared to the NICE criteria; failure rate/1) was 0.79 (95% CI 0.44 to 1.41).
National arthroplasty registries (five studies). In Finland cementless stems performed better than cemented stems across all age groups in patients with rheumatoid arthritis. In Norway and Sweden cementless stems performed better than cemented stems in the general population regardless of the diagnosis. In Denmark the cumulative risk for revision for aseptic loosening after 14 years was slightly higher for osteoarthritis than for rheumatoid arthritis.