Thirty-three studies (n=unclear) were included in the review: 12 randomised controlled trials, 18 observational studies and three retrospective studies. A broad range of study quality details were reported using the PEDro questionnaire. A number of limitations were identified in many studies, including lack of blinding and intention-to-treat analysis.
Based on 16 studies (n=2,161 patients), the reviewers found a statistically significant difference in functional knee society score that favoured fixed-bearing total knee replacements (mean difference 1.93, 95% CI 1.13 to 2.73, I2=0%). Sixteen studies also found no statistically significant difference between groups in terms of clinical Knee Society score (I2=78%).
The reviewers found no statistically significant difference between groups in terms of Knee Society clinical or functional scores, Hospital for Special Surgery score, Oxford Knee score, Western Ontario and McMaster Universities Osteoarthritis Index score, and Functional Questionnaire of Hannover for Osteoarthritis. No statistically significant differences were identified in terms of pain, range of motion, patient satisfaction and patient preference.
Ten radiological outcomes were assessed and no statistically significant differences were identified for any outcome. No statistically significant differences were identified for any of 25 complication types.
The funnel plot did not provide evidence of publication bias.