Periprosthetic joint infection is one of the most dreaded and complex complications of total joint arthroplasty. Periprosthetic joint infection is now the major cause of failure following total knee arthroplasty1 and the third most common cause of failure following total hip arthroplasty2. It is estimated that the prevalence of periprosthetic joint infection may be on the rise3. A wide variety of pathogens are known to cause periprosthetic joint infection, with the majority of infections being caused by gram-positive bacteria, especially staphylococcal species4,5. The treatment of a confirmed periprosthetic joint infection often includes the need for surgical intervention, and two-stage exchange arthroplasty is the most common mode of surgical treatment in North America. Two-stage exchange arthroplasty relies on removal of all foreign material and insertion of an antibiotic-impregnated cement spacer for the purpose of delivering high doses of antibiotics locally in the interval of time between the resection arthroplasty and subsequent reimplantation.
Periprosthetic infection with fungi, although rare, represents a diagnostic and therapeutic challenge to which clear guidelines have not yet been established. It is not known if the protocol for treatment of a bacterial periprosthetic joint infection can also be applied in the same manner to fungal infections. Patients with fungal periprosthetic joint infection are believed to be a different type of host with decreased cellular immunity, mostly due to an underlying cause of immunosuppression, such as malignant disease, drug therapies (antineoplastic agents, corticosteroids, or immunosuppressive drugs), overuse or inappropriate use of antibiotics, and indwelling catheters (urinary or parenteral hyperalimentation). Other factors, such as diabetes, tuberculosis, intravenous drug use, and acquired immunosuppressive disease, are associated with an increased frequency of mycotic infection6. The lack of reliable antifungal medications for systemic and, in particular, local delivery poses a real challenge in pathogen-directed treatment. The literature contains few reports of fungal periprosthetic joint infection. In fact, our search of the entire English literature revealed a total of forty-six patients7-40. The vast majority of those infections were caused by Candida species. Previous case reports present a wide variety of treatment methods, both surgical and medical, as well as variable outcomes. Based on the very small number of patients in each report, it is difficult to draw firm conclusions regarding the outcome of treatment for this challenging problem. Further, the surgical treatment varies considerably for these patients, making interpretation of the presented data difficult and not directly comparable.
The purpose of this multicenter study was therefore to investigate the issues surrounding this rare condition with respect to patient characteristics and the currently implemented therapeutic strategies and their effectiveness. In particular, this study sought to determine the efficacy of two-stage exchange arthroplasty in the treatment of these complex infections.
We performed a review of joint arthroplasty databases in six centers to identify patients who were diagnosed with a periprosthetic fungal infection. Patients were diagnosed with fungal periprosthetic joint infection according to a positive preoperative aspiration culture and/or a positive intraoperative culture. We collected patient-specific information, including demographics, body mass index, smoking habits, comorbidities (especially immunity-impairing risk factors such as diabetes mellitus, corticosteroid therapy, malignant disease, and organ transplantation), and prolonged antibiotic treatment. Medical records were reviewed at each center to identify the clinical, laboratory, and operative characteristics of patients with infection as well as the treatment protocols that had been used. Patients were followed radiographically for at least two years or until recurrence of the infection. The average period of follow-up for patients who remained free of infection was forty-five months (range, twenty-four months to eleven years). Institutional review board approval was obtained in each center prior to initiating the study.
Source of Funding
The authors did not receive any outside funding or grants in support of their research for or preparation of this work.
From 1999 to 2006, a total of thirty-one patients with fungal infection about a total hip or knee prosthesis were identified. There were seventeen women and fourteen men with an average age of sixty-four years (range, thirty-six to eighty-six years) at the time of diagnosis, and they had an average body mass index of 29.1 kg/m2 (range, 17 to 48.6 kg/m2). Infection occurred after knee arthroplasty in seventeen patients and after hip arthroplasty in fourteen patients. Twenty-seven patients had one or more underlying systemic illnesses (Fig. 1), including cardiac disease in twelve patients, chronic liver disease in six patients, diabetes mellitus in six patients, rheumatoid arthritis in five patients, malignant disease in three patients, and chronic renal failure in two patients. Six patients were receiving systemic corticosteroid therapy at the time of presentation. Seven patients received a prolonged course of antibiotics prior to the development of the fungal infection. In four patients, no risk factors of impaired immunity could be identified. Two of those four patients developed the infection after multiple (more than five) revision surgeries. Another patient had a complicated history of multiple gunshot wounds to the extremity with the infected joint.
The interval from the index surgery to the diagnosis of infection averaged twenty-five months (range, one month to twelve years). Fungal periprosthetic infection occurred after a primary arthroplasty in eleven patients and after a revision arthroplasty in the remaining twenty patients. All patients presented with symptoms and signs of chronic infection, which included chronic pain and swelling. None of the patients had systemic symptoms including fever. Loosening of the implant was detected radiographically in seven patients, with extensive femoral osteolysis detected in one patient. The average erythrocyte sedimentation rate and C-reactive protein level were 54 mm/hr (range, 12 to 104 mm/hr) and 17.5 mg/L (range, 0.6 to 73.9 mg/L), respectively. In eleven patients, preoperative aspiration and examination of the joint fluid revealed an average white blood-cell count of 8761 cells/µL (8.76 × 109/L) (range, 440 to 26,700 cells/µL [0.4 to 26.7 × 109/L]) with an average neutrophil differential of 76% (range, 19% to 94%).
Candida species was the causative pathogen in the majority of the patients. Candida albicans was grown in cultures of specimens from twenty patients; Candida parapsilosis, in four patients; both fungi in three patients; and Candida glabrata, in one patient. Non-Candida species were isolated in three patients and included Aspergillus in one patient, Rhodotorula minuta in another patient, and Aureobasidium in the third patient. An intraoperative Gram stain did not reveal organisms in any of the patients. A coexisting bacterial infection was detected in five patients; the organisms included coagulase-negative Staphylococci in three patients, vancomycin-resistant enterococci in one patient, and Staphylococcus aureus in another patient.
Irrigation and débridement was the initial surgical treatment in seven patients, none of whom had resolution of the infection. Five of these patients required removal of the prosthesis to eradicate the infection. Due to the extent of bone loss, which prohibited further reconstruction, the infection was suppressed with oral fluconazole in the remaining two patients. Meanwhile, twenty-four patients were initially managed with resection of the components. Thus a total of twenty-nine patients required resection arthroplasty to control the infection (Fig. 2). Implants and cement, whenever present, were removed, and a thorough débridement of devitalized tissues was carried out in all patients. For all patients, a spacer was inserted that contained the antibiotics tobramycin and vancomycin; in five patients, the spacer also contained the antifungal medication amphotericin.
Of the twenty-nine patients who underwent resection arthroplasty and spacer insertion, delayed reimplantation was performed in nineteen at an average of seven months (range, two to fourteen months) after the resection arthroplasty. Multiple intraoperative cultures were obtained at the time of reimplantation, and these cultures were negative in all nineteen patients. Ten of the nineteen patients showed signs of persistent or recurrent infection after reimplantation. Four of these ten underwent further irrigation and débridement; the cultures that were obtained during that procedure did not demonstrate any fungal growth and these patients were free of infection at the time of the latest follow-up as of the time of writing. Two of the ten patients underwent above-the-knee amputation. Of the remaining four of these ten patients, one patient died due to uncontrolled sepsis that resulted in multiple organ system failure, another patient was managed on long-term suppressive antimicrobial therapy, and two patients refused additional surgery. As a result of uncontrolled infection, reimplantation was never performed in ten of the twenty-nine patients. Multiple operative débridement and spacer exchange procedures were performed in an attempt to control the infection. Nevertheless, the final outcome of these ten patients was above-the-knee amputation in three patients, knee arthrodesis in three patients, permanent resection arthroplasty of the hip in three patients, and hip disarticulation in one patient (Fig. 2).
Following the initial surgical treatment (resection arthroplasty or débridement), all thirty-one patients were managed with six weeks of intravenous antifungal agents. Fluconazole was used in twenty-three patients, with an initial dose of 800 mg/day that was gradually tapered to 400 mg/day. Three patients were treated with caspofungin (50 to 75 mg/day) in addition to fluconazole. Amphotericin B, 3 mg/kg/day, was used in five patients. Antifungal susceptibility testing showed a sensitivity to amphotericin with average minimal inhibitory concentration values ranging from 0.5 to 1 µg/mL. Two patients experienced side effects of amphotericin treatment, namely, an elevated serum creatinine level in one patient and rigors and nausea in another. Following reimplantation, oral fluconazole was routinely prescribed for six months with regular monitoring of liver function tests.
Note: The authors thank Mindy Tokarczyk, MT, ASCP, from the microbiology laboratory at Thomas Jefferson University Hospital, for her contribution to the scientific exhibit. They also thank Douglas Osmon, MD, from Mayo Clinic; Murat Pekmezci, MD, from University of Washington in St. Louis; Eric Hansen, MD, from University of California, San Francisco; and Nick Ting, BA, from Rush University Medical Center, for assistance in obtaining information about the patients.