Eighteen patients undergoing two-stage exchange arthroplasty for infected total hip or knee arthroplasty using gentamicin-loaded bone cement spacers (80g bone cement, 2 g gentamicin and 2 g clindamycin) were studied. The concentration of gentamicin eluted from the spacers was assessed on samples of blood, urine, and drainage fluid that were collected from each patient at set intervals during the 48 hours following the first-stage surgery. The hip and knee cement spacers showed similar curve of release over the first postoperative hours (early peak followed by slow release), but the mean gentamicin concentration in the drainage fluid was higher in patients with hip spacers compared to patients with knee spacers (30.61±19.47 mg/L vs 17.43±13,63 mg/L, p less than 0.05). In patients with hip spacers, the mean, maximum, and minimum concentration of gentamicin was higher with respect to the minimum inhibitory concentration (MIC) break point for Staphylococcus spp, Pseudomonas Aeruginosa and Enterobacteriaceae throughout the first postoperative 48 h. Conversely, in 25% of patients with a knee spacer a drug concentration below the MIC break point for Gram negative bacteria was found in the drainage fluid after 12 h. Gentamicin levels in the blood samples were negligible over the entire time interval and were steadily well below the renal toxicity reference. The highest urinary concentration of gentamicin was observed between 4 and 9 h postoperatively. Subsequently, it gradually declined until 48 h. Clinically, the rate of cure was 100% at a mean follow-up of 113 weeks (range 90-182). Gentamicin-loaded cement spacers offer the advantage of achieving early high concentrations of the antibiotic directly at the site of infection but especially in the knee a systemic antibiotic therapy must be given as a complement to the spacer implantation to eradicate periprosthetic joint infection (PJI).