Nine studies were included in the review: 2 prospective cohort studies, 6 retrospective cohort studies and 1 cross-sectional study. The total number of patients was 411.
The short-term clinical results appeared to be comparable to the alternative treatment of arthrodesis, but problems with study quality limited the conclusions that could be drawn.
None of the studies addressed radiologic loosening, a recognised precursor of clinical loosening or polyethylene wear. None of the studies systematically assessed the subsidence of fusion cages into vertebral bodies.
Operated segments appeared to move with an average range of 5 to 12 degrees (6 studies), but the mobility of the motion segment was frequently lost as a number of arthroplasties eventually resulted in a fusion of the two vertebrae. This was as high as 26% in one case series.
Adjacent segment deterioration was hard to assess because the duration of follow-up was too short. In one study, reoperation was performed at the adjacent segment in 11 of the 50 patients within 2 years and there was a further 13 reoperations at the arthroplasty level during the same period (a total of 24 reoperations in 50 patients).
The complication rate showed wide variation and was described in various ways. Complications were observed in 3 to 50% of the patients (9 studies). No infections were reported, but there were 6 venous injuries, 2 arterial injuries and 6 thrombotic complications. Two studies described a type of implant no longer used and gave information on implant failures.