A Biofilm Approach to Detect Bacteria on Removed Spinal Implants

Study Design. This is a prospective study comparing the diagnosis of spinal implant infection by conventional peri-implant tissue culture with a technique which uses a combination of vortexing and bath sonication to dislodge bacteria growing as a biofilm on the surface of retrieved spinal implants. Objective. We hypothesized that the biofilm-sampling technique would be more sensitive than peri-implant tissue culture. Summary of Background Data. Culture of peri-implant tissue is inaccurate for the diagnosis of orthopedic device-related infection; cultures taken from the implant may be more sensitive. We have developed a technique which uses vortexing-bath sonication to sample bacterial biofilms on the surface of retrieved hip and knee implants, and shown that it is more sensitive than periprosthetic tissue culture for the microbiologic diagnosis of prosthetic knee, hip, and shoulder infection. Methods. We compared peri-implant tissue culture to the vortexing-bath sonication technique which samples bacterial biofilm on the surface of retrieved spinal implants, for the diagnosis of spinal implant infection. In addition, we compared detection of Staphylococcus and Propionibacterium acnes by rapid cycle real-time polymerase chain reaction with culture of sonicate fluid. Results. A total of 112 subjects were studied; 22 had spinal implant infection. The sensitivities of peri-implant tissue and sonicate fluid culture were 73% and 91% (P = 0.046), and the specificities were 93% and 97%, respectively. P. acnes and coagulase-negative staphylococci were the most frequent microorganisms detected among subjects with spinal implant infection, with P. acnes detected in 56 and 45%, and coagulase-negative staphylococci detected in 31 and 40% of peri-implant tissue and sonicate fluid cultures, respectively. Compared with the culture of sonicate fluid, polymerase chain reaction was 100 and 67% sensitive for the detection of culture-positive Staphylococcus and P. acnes spinal implant infection, respectively. Conclusion. Implant sonication followed by culture is more sensitive than peri-implant tissue culture for the microbiologic diagnosis of spinal implant infection.