Infected total hip arthroplasty (THA) is catastrophic, but it is treatable with a high degree of success. Two-stage revision with an antibiotic-loaded cement spacer is the most widely accepted method of treatment, and considered by some to be the best method; however, single-stage treatment currently is used widely, and is gaining acceptance. Although antibiotic-loaded cement is considered to be important for antibiotic delivery after surgery, cementless revision is equally successful with one- or two-stage procedures. Delivery of antibiotics with depot methods, such as cement or bone graft impregnated with antibiotics, is considered to be effective, but the antibiotic levels rapidly deteriorate after the first three days, leaving the cement itself vulnerable to colonization by resistant organisms. Nephrotoxicity is not common, but it does occur, and necessitates removal of the cement. This can be catastrophic if the implants are fixed with antibiotic-impregnated cement. Success rates of THA revision for infection can be as high as 98%, but this rate is dependent on the organism. Failure rates of 20% are the norm for resistant organisms such as methicillin-resistant Staphylococcus aureus, The cost of this failure rate is huge. Failure probably is due to the low concentration of antibiotics in the operative site. Antibiotic infusion into the operative site achieves concentrations that are hundreds of times higher than can be achieved with any other technique and has the additional advantage of being able to be discontinued in the case of renal or auditory damage. Limited personal experience suggests that the failure rate of revision total hip with resistant organisms is significantly lower with intra-articular delivery than with other currently available methods. Between January 2002 and July 2013, 9 patients (9 hips) presented with late-onset acute infections in cementless THA with bone-ingrown implants. These patients were all more than 2 years from their original surgery and had acute symptoms of infection for 4 to 9 days. Two were the author's patients and 7 were referred from another institution. None had symptoms until the onset of their infection, and none had postoperative wound complications, fever, or prolonged pain suggestive of a chronic process. All were treated with debridement and head/liner exchange, followed by catheter infusion of intraarticular antibiotics. All remained free of signs of infection at a mean follow-up of 74 months (range, 62–121 months). This sequential series of successful treatment of late-onset infection of osteointegrated total hip replacement suggests that this is a highly effective method. It has the advantages of being a single-stage procedure, and of avoiding the catastrophic surgical procedure of removing fully osteointegrated total hip replacements

Infection is one of the most devastating complications following total joint arthroplasty. Treatment is difficult, often requiring multiple surgical procedures, prolonged hospitalization, and long-term intravenous (IV) antibiotic therapy. Failure rates are high for resistant organisms and mixed-flora infections, and antibiotic-loaded cement spacers deliver antibiotics for only a few days and can harbor resistant bacteria on the surface. We have adopted a direct-exchange method with antibiotics infused directly into the joint using Hickman catheters to achieve extremely high levels of intraarticular (IA) antibiotics for six weeks. Hickman catheters have a fibrous cuff that allows soft-tissue ingrowth and seals the surface of the tube to prevent contamination of the joint by tracking along the catheter. Two catheters are inserted to ensure that at least one will be functional for six weeks. The safety and efficacy of this protocol was evaluated in patients undergoing primary or revision TKA by measuring joint and serum levels of vancomycin following IV administration (as a prophylactic) and IA administration (as a treatment for infected TKA), and comparing the levels with each method. Therapeutic levels of vancomycin were present in the knee following IV or IA administration, but much higher levels were possible with IA administration (avg. of 6.8 and 9,242 µg/mL). Vancomycin achieved therapeutic levels in the synovial fluid of the knee with IV administration, but clearance from the knee was rapid, suggesting that the synovial fluid concentration may be sub-therapeutic for hours before the next IV dose is given. In contrast, IA delivery of vancomycin resulted in peak levels that were many orders of magnitude higher, and trough levels remained therapeutic for 24 hours in both the joint space and in the serum (minimum trough levels of 8.4 and 4.2 µg/mL, respectively). The elimination constant (half-life) of IA-administered vancomycin was 3.1 hours. This protocol was used in 18 knees (18 patients) with methicillin-resistant Staphylococcus aureus treated between January 2001 and January 2007 with one-stage revision that included débridement, uncemented revision of total knee components, and IA infusion of 500 mg vancomycin via Hickman catheter once or twice daily for 6 weeks. No IV antibiotics were used after the first 24 hours. Serum vancomycin levels were monitored to maintain levels between 3 and 10 µg/mL. Mean serum vancomycin peak concentration was 6 ± 2 µg/mL and the mean serum vancomycin trough concentration was 3 ± 1 µg/mL at 2 weeks post-operative. Knee synovial fluid peak and trough vancomycin levels were measured in two knees. Synovial fluid peak concentrations were 10,233 µg/mL and 20,167 µg/mL and trough concentrations were 724 µg/mL and 543 µg/mL, respectively. Minimum follow-up was 27 months (range, 27–75 months). Mean followup was 62 months, (range, 27–96 months). At 2-year follow-up, mean Knee Society score was 83 ± 9. No radiographic evidence of implant migration has occurred. One knee reinfected with MRSA and was reoperated at 5 months. A necrotic bone segment was found, the knee was debrided and revised, and the antibiotic infusion protocol was readministered. The knee remained free of infection at 42 months post-operatively. Directly infusing antibiotics into the infected area maintains a high local concentration level while minimizing systemic toxicity. This method avoids the use of antibiotic-loaded cement and the potential for growth of antibiotic-resistant strains of bacteria. These findings support single-stage revision in cases treated with cementless revision and IA antibiotics

Bactericidal levels of antibiotics are difficult to achieve in infected total joint arthroplasty when intravenous antibiotics or antibiotic-loaded cement spacers are used, but intra-articular (IA) delivery of antibiotics has been effective in several studies. This paper describes a protocol for IA delivery of antibiotics in infected knee arthroplasty, and summarises the results of a pharmacokinetic study and two clinical follow-up studies of especially difficult groups: methicillin-resistant Staphylococcus aureus and failed two-stage revision. In the pharmacokinetic study, the mean synovial vancomycin peak level was 9242 (3956 to 32 150; sd 7608 μg/mL) among the 11 patients studied. Serum trough level ranged from 4.2 to 25.2 μg/mL (mean, 12.3 μg/mL; average of 9.6% of the joint trough value), which exceeded minimal inhibitory concentration. The success rate exceeded 95% in the two clinical groups. IA delivery of antibiotics is shown to be safe and effective, and is now the first option for treatment of infected total joint arthroplasty in our institution.

Cite this article: Bone Joint J 2016;98-B(1 Suppl A):31–6.

Infection is one of the most devastating complications following total joint arthroplasty. Treatment is difficult, often requiring multiple surgical procedures, prolonged hospitalisation, and long-term intravenous (IV) antibiotic therapy. Failure rates are high for resistant organisms and mixed-flora infections, and antibiotic-loaded cement spacers deliver antibiotics for only a few days and can harbor resistant bacteria on the surface. We have adopted a direct-exchange method with antibiotics infused directly into the joint using Hickman catheters to achieve extremely high levels of intraarticular (IA) antibiotics for six weeks. Hickman catheters have a fibrous cuff that allows soft-tissue ingrowth and seals the surface of the tube to prevent contamination of the joint by tracking along the catheter. Two catheters are inserted to ensure that at least one will be functional for six weeks. The safety and efficacy of this protocol was evaluated in patients undergoing primary or revision TKA by measuring joint and serum levels of vancomycin following IV administration (as a prophylactic) and IA administration (as a treatment for infected TKA), and comparing the levels with each method. Therapeutic levels of vancomycin were present in the knee following IV or IA administration, but much higher levels were possible with IA administration (avg. of 6.8 and 9,242 µg/mL). Vancomycin achieved therapeutic levels in the synovial fluid of the knee with IV administration, but clearance from the knee was rapid, suggesting that the synovial fluid concentration may be sub-therapeutic for hours before the next IV dose is given. In contrast, IA delivery of vancomycin resulted in peak levels that were many orders of magnitude higher, and trough levels remained therapeutic for 24 hours in both the joint space and in the serum (minimum trough levels of 8.4 and 4.2 µg/mL, respectively). The elimination constant (half-life) of IA-administered vancomycin was 3.1 hours. This protocol was used in 18 knees (18 patients) with methicillin-resistant Staphylococcus aureus treated between January 2001 and January 2007 with one-stage revision that included debridement, uncemented revision of total knee components, and IA infusion of 500 mg vancomycin via Hickman catheter once or twice daily for 6 weeks. No IV antibiotics were used after the first 24 hours. Serum vancomycin levels were monitored to maintain levels between 3 and 10 µg/mL. Mean serum vancomycin peak concentration was 6±2 µg/mL and the mean serum vancomycin trough concentration was 3±1 µg/mL at 2 weeks postoperative. Knee synovial fluid peak and trough vancomycin levels were measured in two knees. Synovial fluid peak concentrations were 10,233 µg/mL and 20,167 µg/mL and trough concentrations were 724 µg/mL and 543µg/mL, respectively. Minimum follow-up was 27 months (range, 27–75 months). Mean followup was 62 months, (range, 27–96 months). At 2-year follow-up, mean Knee Society score was 83±9. No radiographic evidence of implant migration has occurred. One knee reinfected with MRSA and was reoperated at 5 months. A necrotic bone segment was found, the knee was debrided and revised, and the antibiotic infusion protocol was readministered. The knee remained free of infection at 42 months postoperatively. Directly infusing antibiotics into the infected area maintains a high local concentration level while minimising systemic toxicity. This method avoids the use of antibiotic-loaded cement and the potential for growth of antibiotic-resistant strains of bacteria. These findings support single-stage revision in cases treated with cementless revision and IA antibiotics

Introduction. Resistant organisms are difficult to eradicate in infected total knee arthroplasty, and treatment of methicillin-resistant Staphylococcus aureus (MRSA) is especially challenging. Whereas most surgeons use antibiotic-impregnated cement during revision to treat infection, the delivery of the drug in adequate doses is limited in penetration and duration. This study presents the 2- to 8-year prospective results of one-stage revision and intraarticular antibiotic infusion protocol used to treat MRSA. Methods. Eighteen knees (18 patients) with methicillin-resistant Staphylococcus aureus were treated between January 2001 and January 2007 with one-stage revision protocol that included débridement, uncemented revision of total knee components, and intraarticular infusion of 500 mg vancomycin via Hickman catheter once or twice daily for 6 weeks. (Figure 1) No intravenous antibiotics were used after the first 24 hours. Serum vancomycin levels were monitored to maintain levels between 3 and 10 μg/mL. The mean serum vancomycin peak concentration was 6 ± 2 μg/mL and the mean serum vancomycin trough concentration was 3 ± 1 μg/mL at 2 weeks postoperative. Results. Knee synovial fluid peak and trough vancomycin levels were measured in two knees. Synovial fluid peak concentrations were 10,233 μg/mL and 20,167 μg/mL and trough concentrations were 724 μg/mL and 543μg/mL, respectively. Minimum followup was 27 months (range, 27-75 months). Mean followup was 62 months, (range, 27–96 months). At 2-year followup, mean Knee Society score was 83 ± 9. No radiographic evidence of implant migration has occurred. One knee became reinfected with methicillin-resistant Staphylococcus aureus and was reoperated at 5 months. A necrotic bone segment was found, the knee was debrided and revised, and the antibiotic infusion protocol was readministered. The knee remained free of infection at 42 months postoperatively. Conclusions. One-stage revision with uncemented components and 6 weeks intraarticular vancomycin administration safely and effectively treated MRSA-infected TKA with no apparent complications