Background: Modern total knee arthroplasty (TKA) femoral components are designed to provide a more optimal articular surface for the patella whether or not it has been resurfaced. Previous systematic reviews comparing outcomes of patellar resurfacing and no resurfacing combine both historic and modern designs.

Aims: The aim of this study was to investigate the effect of patellar resurfacing in modern "patellar friendly" implants on (1) incidence of anterior knee pain, (2) patient reported outcomes (3) complication rates, and (4) reoperation rates compared with unresurfaced patellae in primary TKA.

Methods: MEDline, PubMed and google scholar studies were evaluated using SIGN assessment tool and data analysis was conducted using Review Manager 5.2 on only randomised controlled trials. The search terms were: arthroplasty, replacement, knee (Mesh), TKA, prosthesis, patella, patellar resurfacing, patellar retaining.

Results: Thirty-two randomised controlled studies were identified that reported the type of TKA implant used: 11 used modern "patellar friendly" implants; and 21 older "patellar non-friendly" implants. Among "patellar friendly" TKAs there were no significant differences in anterior knee pain rates between resurfaced and unresurfaced groups. Patellar resurfacing with "patellar friendly" implants had significantly higher clinical (mean difference (MD) -0.77, p = 0.007) and functional (MD -1.87, p < 0.0001) knee society scores (KSS) than unresurfaced counterparts but these did not exceed the minimal clinically important difference (MCID). Resurfacing with "patellar friendly" implants was not associated with a significant (p = 0.59) difference in the Oxford knee score (OKS), in contrast when a "patellar non-friendly" implant was used there was a significant difference (MD 3.3, p = 0.005) in favour of resurfacing. There was an increased risk of reoperation for unresurfaced TKAs with "non-patellar friendly" implants (Odds ratio (OR) 1.68, 95% CI 1.03-2.74, p = 0.04), but not for unresurfaced patellae with "patellar friendly" implants (OR 1.17, CI 0.59-2.30).

Conclusions: Patellar resurfacing in combination with a modern patellar friendly implant was not associated with a lower rate of anterior knee pain, complications, or reoperations compared to not resurfacing, nor did it give a clinically significant improvement in knee specific function. In contrast patellar resurfacing in combination with a "non-friendly" TKA implant was associated with a significantly better OKS and lower reoperation rate. Implant design should be acknowledged when patellar resurfacing is being considered.