Wear of cobalt chromium molybdenum alloy in a reciprocating ball-on-plate test was measured for a series of model synovial fluid samples, where the effect of protein and phospholipid content was examined. The protein content (albumin and γ-globulin) was varied to replicate a range of healthy and diseased SF pathologies. The results showed reduced wear was strongly correlated with increasing protein content. The effect of phospholipid addition on wear was more complex. Limited evidence suggested phospholipids reduced wear for a high albumin/γ-globulin ratio (A/G) but increased wear for low A/G ratios. Post-test examination showed thick (~μm) insoluble “gel-like” films were deposited in, and around, the wear scar. Micro Infrared Reflection Absorption Spectroscopy analysis indicated the films were predominately denatured β-sheet proteins although in some cases lipids were also present. Similar films were found in tests with human synovial fluid samples. Scanning Electron Microscopy imaging showed an aggregated fibril “rope” structure typical of non-native β-sheet proteins. The gel film is a protein-rich viscous phase which is entrained intermittently to form a lubrication film which contributes to surface protection and reduction of wear. We also suggest the formation of gel deposits is comparable to the “boosted” lubrication model of proposed by Professor Duncan Dowson for articular cartilage. In the boosted model high-viscosity, concentrated protein films are formed in depressions on the cartilage surface. The tests indicate the chemistry of human synovial fluid, particularly the protein content, could affect CoCrMo wear and therefore the risk of implant failure.