Method

First, we standardized the animal model to simulate implant contamination during the surgical procedure. We then implanted cobalt-chrome prostheses adapted to both knees of two New Zealand White rabbits, under standard aseptic measures and antibiotic prophylaxis with cefazolin. Prior to implantation, we immersed the prostheses in a 0.3 McFarland inoculum of S. aureus (ATCC 25923) for 30 seconds. In the first animal (control), the joint was directly closed after washing with saline. In the second animal (case), both prostheses were treated with electric current pulses for 30 seconds, washed with saline, and the joint was closed. After 72 hours, both animals were reoperated for the collection of periprosthetic tissue and bone samples, and prosthesis removal. In all samples, we performed quantitative cultures prior to vortexing and sonication, as well as prolonged cultures of the sonication broth. We confirmed the absence of contamination by identification with MALDI-TOF (VITEK-MS) and automated antibiotic susceptibility testing of the isolated colonies (VITEK-2).

Method

Using a 12-well culture plate designed for the application of electrical fields, strains of S. epidermidis were incubated at each well for 24 hours at 37ºC. After incubation, supernatant culture medium was removed, and each well was filled with 3ml of normal saline. Six different models were compared: a) control, b) low-pressure pulse lavage, c) high-pressure pulse lavage, d) pulsed electrical fields, e) low-pressure pulse lavage in combination with pulsed electrical fields, and f) high-pressure pulse lavage in combination with pulsed electrical fields. In all cases, exposure time was set to 25 seconds. In the electrical field models, 50 pulses were applied.

After exposure, each bottom electrode was scraped carefully to release adhered bacteria. Subsequently, different dilutions of biofilm removed were spread onto Müller Hinton agar plates and incubated for 24h at 37 ºC, and colony-forming units (CFU) per milliliters were counted. Bacterial counts were then compared to the control model.