Antibiotics delivered by an implantable drug pump: A new application for treating osteomyelitis

Abstract

Osteomyelitis is usually treated with wide surgical debridement and prolonged intravenous antibiotics. The advent of the implantable drug pump has led us to evaluate this therapeutic modality for the treatment of osteomyelitis. We have previously shown that amikacin retains microbiologic activity in an implantable drug pump over a six-week period when incubated at 37°C. We have also demonstrated in rabbits that high local levels and low systemic levels of antibiotic can be achieved using an implantable drug pump as a delivery system and that this method can be used to sterilize infected wounds. This method was applied in the treatment of osteomyelitis in 14 patients whose duration of symptoms ranged from one month to 22 years. In 13 patients, active drainage had been present for more than six months, and all patients had undergone one or more attempts at eradication of their infection. Prior therapy included surgical debridement alone (one patient), one or more attempts at debridement and extended intravenous antibiotic therapy (13 patients), debridement and local flap procedures (four patients), and extended oral antibiotic therapy (two patients). The bones involved were the tibia (six patients), femur (four patients), and elbow, shoulder, hip, and radius (one patient each). Intra-operative cultures indicated that the infecting organism was Staphylococcus aureus alone (seven patients), S. aureus in combination with gram-negative bacteria (five patients), or gram-negative organisms alone (two patients). Debridement and pump implantation were performed in one stage. On an outpatient basis, the pumps were filled with amikacin in a concentration determined by each patient's serum level. Duration of therapy was based upon cessation of drainage and erythrocyte sedimentation rate. During therapy, serum amikacin levels ranged from less than 2.5 μg/ml to 8.2 μg/ml. The amikacin concentration in the wound drainage was always greater than the upper limits of what could be measured (greater than 55 μg/ml to greater than 5,000 μg/ml). There were no cases of ototoxicity and one case of minimal renal toxicity (posttreatment creatinine clearance of 66 ml per minute; normal equal to 70 ml per minute). Length of therapy ranged from 32 to 140 days (mean equal to 63 days). Length of hospitalization ranged from five to 52 days (mean equal to 24 days). Drainage in all patients stopped during therapy, but resumed in three patients after pump removal—in two patients, for brief periods of time, and in one patient, it continues. Twelve patients have posttreatment erythrocyte sedimentation rates less than or equal to 20 mm per hour. This is a potentially important method of treating patients with osteomyelitis. It offers the advantages of extremely high local and low systemic levels of antibiotic, a low incidence of side effects, a decreased length of hospitalization, and a high rate of suppression of infection.