Aims

To investigate the bone penetration of intravenous antibiotic prophylaxis with flucloxacillin and gentamicin during hip and knee arthroplasty, and their efficacy against Staphylococcus (S.) aureus and S. epidermidis.

Joint fluids obtained for diagnostic purposes from 25 patients were assayed for the presence of gentamicin. All of the patients had presented with failing or painful joints at periods up to 10 years following primary hip or knee arthroplasty using gentamicin-impregnated cement. Gentamicin was detected in the joint fluids from 9 of 15 patients with knee prostheses and 4 of 10 with hip prostheses. Gentamicin concentrations ranged from 0.06mg/L to 0.85 mg/L with no significant differences in concentration between patients with hip or knee prostheses, or type of prosthesis, and no identifiable relationship was found gentamicin concentration and the time after primary arthroplasty. Although the majority of the gentamicin concentrations were found to be below the levels required to inhibit susceptible pathogens, we conclude that gentamicin release around failing implants may lead to false negative cultures in some patients and provide selective pressure for the emergence of resistance where infection is present in others.

Twelve patients undergoing total hip replacements were given 600mg linezolid as a 20min intravenous infusion along with conventional prophylaxis of 1gm cefamandole immediately before surgery. Routine total hip arthroplasty was performed and at timed intervals during surgery, samples of bone, fat, muscle and blood were collected for assay by HPLC analysis. Samples of haematoma fluid that formed around the operation site and further blood samples were also collected at timed intervals following the operation for assay. The penetration of linezolid into bone was rapid with mean levels of 9.1mg/L (95% CI: 7.7–10.6mg/L) achieved at 10min after the infusion, decreasing to 6.3mg/L (95% CI: 3.9–8.6mg/L) at 30min. Correcting for the simultaneous blood concentrations gave values for bone penetration of 51% at 10min, 60% at 20min and 47% at 30min. although the penetration of linezolid into fat was also rapid, mean levels and degree of penetration were approximately 60% of those seen in bone at 10min: 4.5mg/L (95%CI:3–6.1mg/L; penetration 27%) 20min: 5.2mg/L (95% CI:4–6.4mg/L; penetration 37%) and 30min:4.1mg/L (95% CI:3.3–4.8mg/L; penetration 31%). For muscle, the corresponding values were 10min: 10.4mg/L (95%CI:8.1–12.7mg/L; penetration 58%), 20min 13.4mg/L (95%:10.2–16.5mg/L; penetration 94%) and 30min 12mg/L (95% CI:9.2–14.8mg/L; penetration 93%). Mean concentration of linezolid in the haematoma around the operation site were 8.2mg/L at 6–8h and 5.6mg/L at 8–10h after the infusion and 7mg/L at 2–4h following a second 600mg infusion given 12h postoperatively.

We conclude that linezolid exhibits rapid penetration in bone, fat and muscle of patients undergoing hip arthroplasty to achieve levels in excess of the MIC for sensitive organisms (MIC of < _ 4mg/L); with therapeutic levels maintained in the drainage which surrounds the operation site for more than 16h. This pharmaco-kinetic profile is similar to those of agents currently used for the treatment of bone and associated soft tissue infections and suggests a role for linezolid in the management of such patients

Bone cement containing gentamicin may release antibiotic when fractured during revision operations. Tissue samples taken during surgery may be contaminated by gentamicin and give inaccurate microbiological assessment.

We studied five patients in whom cement containing gentamicin had been used in the primary procedure. During revision hip replacement, samples of joint fluid, tissues and cement were taken both before and after disruption of the cement.

With the exception of one sample of joint fluid, low concentrations of gentamicin were recorded in the samples taken before the cement was disrupted, but after disruption the specimens contained gentamicin at concentrations high enough to inhibit or prevent growth of sensitive organisms. The cement contained very high levels up to ten years after insertion.

Our findings suggest that no reliance can be placed on the microbiological assessment of specimens taken once cement splitting has started and that specimens should therefore be taken as early as possible.