Glycopeptide bone penetration in patients with septic pseudoarthrosis of the tibia

In the treatment of bone infections, a major determinant of the clinical response is the active drug concentration at the infected site. Because of the high prevalence of meticillin (methicillin)-resistant staphylococci and enterococci, glycopeptides are widely used for the treatment of bone and joint infections, but data on their penetration into human bone are lacking. The aim of our study was to measure vancomycin and teicoplanin concentrations in infected human bone under steady-state conditions and verify their relationship with inflammatory markers, patient demographic characteristics and pharmacodynamic microbiological markers.

Twenty-seven adult orthopaedic patients undergoing surgical debridement for septic pseudoarthrosis of the tibia and receiving either intravenous vancomycin (Vancocina) 1 g twice daily) or teicoplanin (Targosid) 10 mg/kg/day) were studied from January 2004 to January 2008. Plasma and bone specimens were simultaneously collected during surgery for pharmacokinetic and microbiological assays at a variable interval after antimicrobial administration. Bone samples were dissected into cortical and cancellous bone, cleaned of soft tissues, crushed and eluted into phosphate buffer. Necrotic samples and sequestra were not analysed.Plasma and bone antimicrobial concentrations were measured by a validated method of high-performance liquid chromatography with UV detection, and bone/plasma concentration ratios were calculated. Cortical and cancellous bone area under the concentration-time curve (AUC) over 24 hours (AUC(24)) values were measured by the linear-log trapezoidal rule, using WinNonlin) software, and were compared with the minimum inhibitory concentrations (MICs) of the infecting agents.

For vancomycin, the mean +/- SD concentrations were 2.66 +/- 1.2 mg/L in cortical bone and 11.53 +/- 7.8 mg/L in cancellous bone (corresponding to 20.67% and 89.39% of intraoperative plasma concentrations), and the mean +/- SD tissue AUC(24) values were 55.15 +/- 25.26 h . mg/L for cortical bone and 299.16 +/- 299.54 h . mg/L for cancellous bone. For teicoplanin, the mean +/- SD concentrations were 2.01 +/- 1.7 and 7.51 +/- 7.0 mg/L in cortical and cancellous bone, respectively (12.35% and 48.6% of intraoperative plasma concentrations), and the mean +/- SD teicoplanin tissue AUC(24) values were 34.08 +/- 23.6 h . mg/L and 155.17 +/- 132.8 h . mg/L for cortical bone and cancellous bone, respectively. The mean vancomycin AUC(24)/MIC ratios were 215.02 for plasma, 47.14 for cortical bone and 268.95 for cancellous bone. The mean teicoplanin AUC(24)/MIC ratios were 336.48, 36.27 and 197.21 for plasma, cortical bone and cancellous bone, respectively.

Bone penetration of both glycopeptides ranged from poor (<15%) to satisfactory (15-30%) in the cortical compartment, while it was far higher into the highly vascularized cancellous tissue. Vancomycin bone penetration was slightly higher than with teicoplanin, but the difference was not statistically significant. Higher bone concentrations were observed with higher inflammatory markers, possibly as a result of increased vascularization and vascular permeability under inflammatory conditions. Bone concentrations over the MIC and AUC/MIC ratios suggested that both glycopeptides achieve a satisfactory pharmacokinetic exposure in the cancellous bone, as far as Gram-positive pathogens are concerned. On the other hand, cortical bone exposure was suboptimal in most patients. Furthermore, as antimicrobial penetration may be affected by impaired blood supply, the role of radical surgical removal of purulent and necrotic tissues appears to be essential in order to shorten treatment duration and to reduce the risk of treatment failure.