Background. The past four decades, the problem of resistant bacteria has emerged. As a result of biofilm formation of (resistant) bacteria on the implant, more implant-associated infections (IAI) occurred. This has caused an increase in orthopaedic implant revisions, causing a high burden of disease. To overcome the rising problem of resistance, many studies have focused on new antibacterial agents, such as Ag, Cu, and Zn nanoparticles (NPs) incorporated on titanium (Ti6Al4V) implants. It is known they show antibacterial effects. It is, however, unknown what causes the antibacterial effects of these metals incorporated on titanium implants. This study aims to unravel the antibacterial mechanisms of titanium implants bearing Ag, Cu or Zn NPs behind the in vitro antibacterial effects against methicillin-resistant Staphylococcus aureus (MRSA). Methods. To obtain an implant surface bearing Ag, Cu or Zn NPs; porous Ti6Al4V implants and solid Ti6Al4Nb discs were treated by plasma electrolytic oxidation (PEO). The PEO electrolyte consisted of calcium acetate, calcium glycerophosphate, and Ag, Cu or Zn NPs. The surface morphology was visualized by scanning electron microscopy (SEM) and its chemical composition by energy dispersive X-ray spectroscopy (EDS). All implant groups contained either Ag, Cu or Zn NPs and were tested on its antibacterial leaching activity against MRSA by a zone of inhibition experiment. In addition, the antibacterial effects as a result of contact killing were examined by a direct contact assay. Porous and solid surfaces were compared to reveal the differences in their contact killing properties. Moreover, the porous implants were incubated for 2 h and 24 h. Furthermore, the generation of reactive oxygen species (ROS) of the implant with and without inoculation of bacteria was measured by electron paramagnetic resonance (EPR) for forty minutes. ROS generation after inoculation with bacteria was tested in two ways: (1) the implant was placed in a solution of bacteria PBS after which ROS generation was directly measured in 100 mM DMPO, and (2) the implant with bacteria in BHI was incubated for 2 h after which the implant was placed in 100 mM DMPO to examine ROS generation of the implant with adherent bacteria. Results. PEO processing resulted in four biofunctionalized groups: PT, PT+Ag, PT+Cu, and PT+Zn. The presence of Ag, Cu and Zn NPs was confirmed by SEM and EDS. The antibacterial leaching activity was only observed in PT+Ag. In addition, porous implants showed better contact killing properties than solid discs. All biofunctionalized groups were significantly different from a non-treated (NT) implant considering contact killing in 24 h. Moreover, ROS generation was observed in all biofunctionalized implants. However, solely PT+Cu was significantly different from a NT implant. Furthermore, the EPR results showed that bacteria generate ROS. In all biofunctionalized groups, however, ROS decreases when bacteria are added. In all groups, the ROS generation of adherent bacteria to the implants showed higher intensity than the ROS generation of bacteria in PBS in contact with the implant. Conclusion. Antibacterial surfaces incorporated with Ag NPs show most antibacterial leaching effects, attributed to the ion release of Ag. Furthermore, it is assumed that direct contact killing of Cu is a cause of ROS generation of an implant bearing Cu NPs.