Eleven RCTs and one quasi-RCT were included (1,490 knees).
In terms of study quality, 6 studies reported the randomisation method, six reported a blind outcome assessment, and three clearly reported intention-to-treat analysis.
Resurfacing was associated with a statistically significant reduction in the risk of reoperation compared with nonresurfacing; 2.3% versus 6.5% (RR 0.43, 95% CI: 0.27, 0.71, P=0.0008). The reduction in reoperation with resurfacing remained statistically significant when the meta-analysis was limited to the 3 highest quality studies (RR 0.45, 95% CI: 0.24, 0.82, P=0.01) and patients with osteoarthritis (RR 0.48, 95% CI: 0.27, 0.86, P=0.01). No statistically significant heterogeneity was detected.
Resurfacing was associated with a statistically significant reduction in the risk of significant anterior knee pain compared with nonresurfacing; 7.6% versus 22.3% (RR 0. 39, 95% CI: 0.20, 0.75, P=0.005), based on 7 studies. There was no statistically significant difference between treatments in anterior knee pain when the meta-analysis was limited to the 3 high-quality studies. Statistically significant heterogeneity was detected for both meta-analyses (P=0.01 and P=0.004).
There was no statistically significant difference between resurfacing and nonresurfacing in the IKS knee score; the WMD was 1.12 (95% CI: -2.13, 4.37), based on 4 studies. Statistically significant heterogeneity was detected (P=0.005). Following the exclusion of an outlier, which removed the significant heterogeneity, the meta-analysis showed a statistically significant effect in favour of resurfacing (WMD 2.67, 95% CI: 1.26, 4.08, P=0.0002). There was no statistically significant difference between resurfacing and nonresurfacing in the IKS function score (based on 3 studies), but statistically significant heterogeneity was detected (P=0.07).
Resurfacing was associated with a statistically significant reduction in the risk of significant pain during stair climbing compared with nonresurfacing; 12.7% versus 26.4% (RR 0.43, 95% CI: 0.22, 0.83, P=0.01), based on 2 studies.
There was no statistically significant difference between resurfacing and nonresurfacing in patient satisfaction (RR 0.71, 95% CI: 0.42, 1.19; based on 4 studies). No statistically significant heterogeneity was detected.