FormalPara Key Summary Points

Why carry out the study?

With increasing arthroplasties, more periprosthetic deep infections are being seen, but no standardized treatment protocols exist.

What was learned from the study?

Very few infections resolve without combined surgical and medical management.

Debridement and Girdlestone procedures underperform when compared to one-stage and two-stage revision.

Most cases are treated with 3–6 months of antifungal therapy. Lifelong fungal suppression may be used if surgery is not an option.

Introduction

The 2022 American Joint Replacement Registry reported 2,550,232 primary and revision hip and knee arthroplasties between 2012 and 2021 [1]. The number of knee and hip arthroplasties has risen significantly in the United States and is projected to continue to increase [2]. Concurrently, an increasing number of periprosthetic deep infections have been diagnosed, with a nearly two-fold increase between 1990 and 2004 [2]. Most infections are caused by Gram-positive bacteria, with Staphylococcus aureus and Staphylococcus epidermidis being the most commonly recovered isolates [3].

Fungal prosthetic joint infection represents less than 1% of infections but poses a diagnostic and therapeutic challenge to orthopedists and infectious disease physicians [4]. Often, these patients are in states of immunosuppression, possess other risk factors such as diabetes, or have had a prior revision arthroplasty [5]. The most effective treatment for fungal prosthetic joint infections remains unclear. Most cases are treated with two-stage revisions combined with systemic antifungal medications.

The most extensive studies of total hip and knee arthroplasty fungal infections have included 93 and 132 patients, respectively [6]. More data are required on the treatment and outcomes of patients affected by fungal prosthetic joint infection. The purpose of this study was to retrospectively compare the outcomes of fungal prosthetic joint infections involving different fungal isolates and modes of medical and surgical management in our institutions while evaluating available data on reported cases in the literature. The objectives were to assess the relationship between fungal species and outcome, fungal type and outcome, surgical repair and outcome, onset of infection from initial surgery and outcome, and monomicrobial and polymicrobial infections and outcome.

Methods

A retrospective record review was conducted from January 1, 2007, to December 31, 2018, of patients 18 years and older admitted to The University of Toledo Medical Center and the ProMedica Health System in Toledo, Ohio, with fungal prosthetic joint infections. This study was approved by both IRB boards and patient consent was waived due to the study involving retrospective data without active patient involvement. Inclusion criteria for the chart review included the presence of a prosthetic joint with at least two deep cultures for the same fungal organism or one positive bone culture for a fungal organism with or without a draining sinus tract. At least two deep cultures were selected to minimize the inclusion of patients with culture contamination. Standard laboratory techniques were used for the requested cultures. Patients were excluded if they were pregnant or if, on record review, the infection was not associated with the orthopedic implant. Demographic data, body mass index (BMI), smoking status, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), white blood cell count (WBC), presence of concurrent bacterial infection, surgical management, antimicrobial management, episodes of relapse, and treatment course were reviewed. Relapse was defined as requiring further surgical management or antimicrobial therapy after the initial treatment course.

A PubMed and Embase search of the English-language literature from January 1, 1980, to January 1, 2023, was additionally performed by reviewing pertinent references for cases meeting the inclusion criteria. Search terms included an extensive list of fungal species (see Table S1 in Appendix A for further information). Patients that met the following inclusion criteria were included: presence of prosthetic joint and positive deep culture for fungus (see Table S2 in Appendix A for further information). Cases were excluded if there were insufficient demographic data, including age, gender, site of the prosthesis, and culture results.

Data were compiled into an Excel spreadsheet, and descriptive statistics were collected. SPSS Statistics was used for statistical analysis to compare data. Crosstabulations were made for fungal species and outcome, fungal type (mold, yeast, dimorphic) and outcome, antifungal class and outcome, surgical repair and outcome, and time to onset of infection and outcome. Pearson Chi-square tests were performed for the above crosstabulations with a significance value of P < 0.05. ANOVA was used to test for significant differences in outcomes across age groups.

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of The University of Toledo and ProMedica Health Systems. Patient consent was waived, as this was a retrospective review and it was not feasible to contact all patients.

Results

Retrospective Record Review

Eighteen patients fit the inclusion criteria during this period in the retrospective record review from the University of Toledo and the ProMedica System in Toledo, Ohio. The average age at the time of infection was 61.9 years; 55.6% were female; seven patients were Caucasian, two were African American, and the race of the remaining nine patients was unknown. The average BMI was 32.1 kg/m2. Six patients had diabetes mellitus, and six were immunosuppressed for reasons including prior transplant on immunosuppressive medications, cancer on chemotherapy and/or radiation, and immunosuppressive medications for rheumatologic conditions. Fifty percent had an infected hip, and the other 50% had an infected knee. Twelve (66.7%) had at least one prior revision. The average ESR was 47.8 mm/h, CRP was 11.9 mg/dl, and WBC was 9.4 (× 109/l) (Table 1)

Table 1 Demographic data, comorbidities, and personal antecedent (including previous revisions and infections) of patients with prosthesis-related fungal infections
Full size table

Consistent with previously reported literature, 12 of 18 patients had a history of prior revision. Twelve patients presented with a painful joint, eight with drainage, and one with joint dehiscence. Fungal culture was positive for Candida species in all cases. Candida albicans was isolated in 65% of cases, Candida parapsilosis was isolated in 23%, and Candida glabrata and Candida dublinensis were each isolated in 6% of cases. Additionally, one patient with prosthetic joint infection of the knee was noted to have positive cultures for Trichophoron inkin and C. albicans. Thirteen of the 18 patients identified had a co-infecting bacterial organism identified on culture, with Staphylococcus aureus being the predominantly isolated organism. Other frequently isolated organisms identified were coagulase-negative Staphylococcus, Enterococcus species, and Serratia species (Table 2).

Table 2 Patient presentation, microbiological and prosthesis-related aspects of patients with prosthesis-related infections due to fungal species; signs and symptoms, and laboratorial tests
Full size table

Among the 18 affected patients, 15 had known outcomes. Debridement, antibiotics, and implant retention (DAIR), one-stage exchange, two-stage exchange, Girdlestone procedure, and amputation were used as treatment modalities (Fig. 1). Of the three patients that received DAIR, two were noted to have relapse of infection, and one patient died. Of the seven patients receiving two-stage exchange, three were noted with cure (defined as the absence of infection at > 1-year follow-up), three with relapse, and one with an unknown outcome. Of the other treatment modalities, one patient with a one-stage exchange, one patient with amputation, and one patient that received the Girdlestone procedure were cured. The remaining patient, who received a stage one revision as destination therapy, experienced a relapse of the fungal PJI.

Fig. 1
figure 1

Outcomes of fungal prosthetic joint infection

Full size image

Of the 15 patients in whom antifungal therapy was known, nine were treated with fluconazole, three were treated with micafungin, and two were treated with both modalities at some point during their fungal prosthetic joint infection treatment. Of the 14 patients in which the duration of therapy was known, eight received a 6-week course of therapy. Two received an 8-week course of therapy or longer, while three received a 2-week course of shorter. One received no antifungal therapy (Table 3).

Table 3 Initial management and antifungal treatment of patients with prosthesis-related fungal infections
Full size table

Review of the Literature

Demographics

A literature review revealed 271 cases reported between 1980 and 2023 that met inclusion criteria [7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117]. Patients’ age at the time of infection ranged from 31 to 89, with a mean of 66.7 ± 11.7. A total of 146 were female (54.3%); 106 (39.1%) of the reviewed cases reported comorbidities of diabetes mellitus (29.1%) or immunosuppression (17.0%). Patient race was not consistently reported.

The number of previous revisions or previous infections was not consistently reported. Previous revision history was reported in 49 cases, with reasoning for revision provided in 27 cases. Infection was cited as the reason in ten of these cases. One case had a previous fungal prosthetic joint infection caused by Candida species. The other cited reasoning for prior revisions included pain, aseptic loosening, and periprosthetic fractures.

Diagnostic Findings

Regarding the prosthetic joint, 43.5% of patients had an affected hip, and 53.8% had an affected knee. Other infected joints included five shoulders and one finger [16, 18, 53, 55, 62, 115]. On presentation, 53.1% of patients presented with a painful joint; 13.3% and 16.5% had dehiscence and drainage, respectively. Time to onset of infection was classified into: early (less than 3 months after), delayed (3–24 months after), and late (more than 24 months after). Time to onset was reported in 134 cases: 23 cases presented with early infection, 53 with delayed infection, and 58 with late infection.

The average CRP and ESR were 28.9 mg/l and 58.1 mm/h, respectively. The average WBC was 10.3 × 109/l. Among the fungal isolates, 13% were yeasts, 5% were molds, and a significant majority, 82%, were dimorphics. Also, 87% of cases were Candida species; 102 cases were C. albicans, 72 were C. parapsilosis, 25 were C. glabrata. Of the remainder, six cases were Histoplasma species, and 62 cases were other fungal species, including additional Candida species.

Sixty-nine (25.5%) cases presented with a bacterial co-infection. The bacterial species were predominantly coagulase-negative Staphylococcus species (35.2%), followed closely by Staphylococcus aureus (31.5%).

Treatment and Surgical Management

Most (69.4%) cases were treated with an azole. Figure 2 describes the antifungals used in the treatment of these cases. Fifty-three cases were treated with a combination of antifungals, with the combination of azoles and echinocandins being the most common. Treatment duration varied from < 6 weeks to > 1 year (Fig. 3). Eight patients were treated with lifelong antifungal suppression with fluconazole. The rationale for lifelong suppression included the presence of mold (2), immunosuppression (3), and refusal of surgery (3).

Fig. 2
figure 2

Counts of antifungal treatments used

Full size image
Fig. 3
figure 3

Treatment duration

Full size image

Surgical repair in combination with antifungal therapy was performed in 86.7% of cases. Different types of surgical management included debridement, one-stage exchange, two-stage exchange, Girdlestone procedure, and three-stage revision. The two-stage exchange was the most common (33%), followed by one-stage exchanges (25%) (Fig. 4).

Fig. 4
figure 4

Surgical approaches

Full size image

Combined Data from Retrospective Chart Review and Literature Review

Outcomes

Identified outcomes included complete recovery, reoccurrence not requiring repeat surgery, reoccurrence requiring repeat surgery, death secondary to the infection, and others (loss to follow-up, death from a secondary cause). Outcomes are as follows: 208 cases had recovery, 22 had a fungal prosthetic joint infection recurrence, 43 had recurrence requiring revision surgery, six died, and 15 had unreported or unknown outcomes. The observed outcomes are described in Fig. 5.

Fig. 5
figure 5

Outcomes in monomicrobial versus polymicrobial cases

Full size image

The continuous predictor variable of age was compared against the outcomes using ANOVA testing and did not identify a statistically significant difference (P = 0.570). When the outcomes for each surgical approach were compared using the Pearson Chi-square test, there was a statistically significant difference between the type of surgical approach and patient outcome (P = 0.022). An increase in negative outcomes (reoccurrence, need for revision, and death) in patients undergoing a Girdlestone procedure was noted. Sixty percent of patients who underwent the Girdlestone procedure had a negative outcome, compared to 15% and 18% for one and two-stage exchanges, respectively. There was no identified significance for the fungal species (P = 0.512), class of fungal infection (P = 0.344), and antifungal used (P = 0.728) when compared against outcomes. There was a marginal relationship between the time to onset of infection and outcome (P = 0.063).

Discussion

With increasing arthroplasties and periprosthetic deep infections, determining appropriate treatment and optimal surgical management is vital to preventing poor outcomes. Currently, no standardized guidelines exist for managing fungal prosthetic joint infection.

Patients most at risk for fungal prosthetic joint infections have comorbidities that delay wound healing and immune response, such as diabetes mellitus, immunosuppression, and kidney disease [5]. One previous study reported prior revision surgery and age > 65 as independent risk factors, though the current study did not reproduce those findings [118]. This may be because many case reports did not report demographic data or comorbidities for the patients presented.

Infections are classified based on time of development in relation to initial surgery: early (less than 3 months after), delayed (3–24 months after), and late (more than 24 months after) [119]. Presenting symptoms vary based on the time of presentation. In our study, over half of the patients were reported to present with an acutely painful joint (53.1%). Early infections often present with acute onset pain, effusion, erythema, and warmth. Delayed infections may present with more subtle signs, such as implant loosening and persistent pain. Inflammatory markers may aid diagnosis, but these are insufficient in and of themselves to determine the presence or absence of infection. Periprosthetic and intraoperative cultures are essential in determining the causative organism.

Very few of the presented cases resulted in a cure in the absence of surgery, demonstrating the importance of surgery in management. Surgical approaches include DAIR, one-stage exchange, stage one destination, two-stage exchange, Girdlestone procedure, three-stage revision, amputation, and arthrodesis. There was a significant difference in the type of surgical approach and patient outcome. Our study found that two-stage exchange was performed in a third of the presented cases, with 79% undergoing full recovery without any reported recurrence of infection. Previous reviews have reported two-stage exchange as the gold standard [5, 118]. Over half of the patients who underwent the Girdlestone procedure had a negative outcome. However, it must be noted that the Girdlestone procedure is often done in prior failed surgical reimplantation or in non-ambulatory patients, who are at higher risk for negative outcomes [120].

Antifungal usage is often determined by species resistance, hospital patterns of resistance, and patient comorbidities. Our study demonstrated azoles as the most common antifungal used, noted in 79% of reported cases, though 22% of cases were treated with a combination of antifungals. Treatment duration varied from < 6 weeks to > 1 year, with 3–6 months being the most common duration. For Candida joint infections, the IDSA currently recommends an azole such as fluconazole for 6 weeks or echinocandin for 2 weeks followed by an azole for at least 4 weeks as well as device removal for infection involving a prosthetic [121]. Chronic suppression with fluconazole is recommended if the device cannot be removed. This approach was used in 8 of the presented cases who either declined surgery or were poor surgical candidates. These patients did not have reported recurrence.

Our study is retrospective, so limitations of this study include a lack of some pertinent demographic, clinical, and diagnostic data. When considering polymicrobial and monomicrobial infections, a potential confounding variable is the receipt of broad-spectrum antimicrobials before and during surgical intervention. This variable could not be analyzed due to a lack of reporting in the literature, but it should be considered in future research. Additional confounding variables include the differences in the performance of laboratory personnel that may result in under and overreporting of fungal PJI and, additionally, orthopedic surgery practices may result in differences in the rate of fungal PJI among different centers.

This study identified a marginal relationship between the time to onset of infection and the patient outcome. However, this was limited by a low frequency of reporting of the time between initial prosthetic joint replacement and identification of infection. This relationship should continue to be explored in future research. As this study covers an extensive time frame from 1980 to 2023, it should be noted that medical practice and diagnostics have shifted accordingly. A recent study by Gonzalez et al. reviewed 225 cases of prosthetic joint infections emphasizing surgical approaches to treating fungal prosthetic joint infections [6]. The current study expands on this work by adding some case reports excluded in the former study. It includes a specific analysis of the microbiology and antifungal therapy with their relationship to the outcome.

Management of prosthetic fungal joint infections has not been standardized; it has been based on presented cases in the literature and physician experience. All studies discussed in this paper have been retrospective, as performing a randomized clinical trial with either medical or surgical intervention for fungal prosthetic joint infections may not be feasible. Current best practices would support aggressive surgical management, such as a two-stage exchange with prolonged (> 3 months) antifungal therapy directed at the specific pathogen.