Author(s): Parvizi Javad MD FRCS, Ketonis Constantinos BS, Antoci Valentin Jr. BS PhD, Davidson Helen M. BS, Ghanem Elie MD, Hickok Noreen J. PhD, Adams Christopher S. PhD, Shapiro Irvine M. PhD.; Philadelphia, PA
Title: Novel Rat Model for Periprosthetic Infection around a Long Bone Fracture
Purpose: Infection of implants used for treatment of long bone fracture, though rare, is an undesirable complication. Periprosthetic infection often retards healing of the fracture and may necessitate removal of the fixation implant.
Methods: To study the effects of periprosthetic infection on rat fracture healing, male Wistar rats (300and 350 g) were anesthetized with isoflurane gas, the animals were positioned supine and the intercondylar notch region of the knee was identified. Using a percutaneous stab incision, a 20 G followed by an 18 G needle was reamed into the femoral canal to 3 cm. The needle was left in the canal and used for delivery of bacterial inoculum (100 £gL suspension of 1 x 103 CFU). A titanium rod (1 x 20 mm) was press fit into the canal by retrograde insertion. The femur was then fractured with a dropped weight of 500 g from a height of 25 cm using a three point bending device. All fractures were verified radiographically and the rats were allowed to weightbear. The animals were examined daily for clinical signs of infection, including swelling and erythema of the knee and the thigh, loss of passive motion of the knee joint, and non-weightbearing on the affected extremity. Rats were sacrificed at 7, 14 and 21 days using CO2 and their femora radiographed. Periosteal elevation, cortical bone loss, abscess formation and overall bone health was further assessed using microCT. Next, the femur was carefully dissected, the implants extracted, rolled on a plate, and sonicated in 2 mL trypticase soy broth (TSB) to remove all bacteria into solution. Bacterial numbers were determined by serial dilution plating on TSB agar, and incubation for 24-48 h at 37oC. Colony forming units were counted and statistically analyzed for significant differences between infected and non-infected femurs. Further histology and microbiology was used to confirm infection status when deemed necessary.
Results: All rats were examined for development of periprosthetic infection and the specific features of fracture healing were closely studied. Infection was consistently induced in 100% of the rats with as little as 103 CFU in 100 uL of saline, as defined by clinical observation as soft tissue swelling, incapacitation, knee joint immobility, and guarding. Microbiology confirmed biofilm formation on the implants with expected variation between animals.
Discussion: This study describes an infected long bone fracture model. The cost-effectiveness of our model may provide an incentive for initiation of studies that may further elucidate the causes of PPI. The proposed rat peri-prosthetic infection model lends itself to more clinically relevant studies which can eventually help elucidate the clinical problem of PPI and possibly eradicate implant associated infection.