Post-operative surgical site infection following total joint arthroplasty occurs at rates between ~ 0.2–5 %, depending on the joint and the surgeon volume, as well as various patient risk factors. Given that an estimated over 700,000 knee and hip arthroplasties are performed in the US each year this translates to thousands of patients that are affected by this serious, costly and traumatic complication. In addition, it is now recognized that clinical culturing underestimates the infection rate and that a number of aseptic loosenings might actually have an infectious etiology. We have used a combination of non-culture based molecular methods to detect bacteria associated with hardware, antimicrobial impregnated cement, reactive tissue and pus collected during revision surgery in a total elbow arthroplasty (TEA) case and a total ankle revision (TAR) case. Confocal microscopy showed live cocci in biofilm cell clusters, and fluorescent in situ hybridization (FISH) demonstrated S. aureus biofilms. Reverse transcriptase (RT)-PCR, and multiplex PCR coupled with electrospray-ionization mass spectrometry (Ibis T5000) to identify S. aureus, S. epidermidis and genes for methicillin resistance. Together our complimentary techniques comprise compelling evidence that viable biofilm bacteria played an important role in the refractory infections in these cases.