Rifampin, bound in high concentrations to prosthetic grafts, has been
proposed for the treatment of vascular graft infections. The optimum a
ntibiotic concentration and duration of treatment for infected grafts
is not known. This study compared the in vitro and in vivo efficacy of
varying concentrations of rifampin against three different strains of
slime producing Staphylococcus epidermidis (RP62A, KC2, and KB1) boun
d to knitted Dacron at high and low concentrations at 10(4) and 10(7)
CFU/cm(2) of prosthetic. Time kill experiments were performed at 4, 18
, and 42 hr, in which each Dacron bound bacterial strain was exposed i
n vitro to 4x, 64x, 100x, and 1,000x minimum inhibitory concentration
(MIG) of rifampin. In vivo,the Dacron bound laboratory strain RP62A wa
s implanted subcutaneously into the backs of male Swiss-Webster mice a
nd exposed to 4x, 100x, and 1,000x the MIC of rifampin for similar tim
e periods. In addition, systemic vancomycin (10 mg/kg) was assessed fo
r synergy and prevention of rifampin resistance. In vitro,all concentr
ations of rifampin showed near total killing (<1 log) of all bacterial
strains at low initial concentrations (10(4) CFU/cm(2)) but not high
(10(7) CFU/cm(2)) to 42 hr. Importantly, resistance was shown to devel
op in all three strains of S. epidermidis with high initial bacterial
biofilm concentrations. In vivo, rifampin concentrations between 4x MI
C and 100x MIC achieved a balance between optimal killing and preventi
on of resistance. Systemic vancomycin slightly improved bacterial clea
rance but did not alter the development of rifampin resistance at high
local concentrations. Caution is advised with the use of antibiotic b
onded grafts because resistance may develop, even with the addition of
systemic antibiotics.