Abstract
Background
Candida prosthetic joint infections (CPJI) are serious complications, for which optimal surgical management and antifungal therapy remain unclear. This systematic review and meta-analysis aimed at defining the outcome predictors of CPJI.
Methods
A systematic literature review was performed in PubMed, Medline, Embase, and Web of Science until July 2024. Articles (cohorts, case-series or case reports) reporting individual data of adult patients with CPJI were included. Data about underlying conditions, characteristics of infection, and outcomes were collected. Outcome predictors were assessed in univariate analysis. Significant variables were included in a multivariate model using logistic regression with a binomial link function. Multicollinearity among the independent variables was assessed using the variance inflation factor.
Results
A total of 385 CPJI (including 204 hip and 152 knee infections) from 110 publications were included. Polymicrobial infections accounted for 33% cases. Candida albicans (47.2%) was the predominant species followed by Candida parapsilosis (28.6%). In multivariate analysis, independent predictors of failure were co-infection with Staphylococcus aureus (odds ratio, 0.4; 95% confidence interval, 0.18–0.92; P = .032) and debridement/retention of the prosthesis (0.25; 0.11–0.55; P < .001), whereas first-line therapy with amphotericin B was associated with success (3.18; 1.25–9.87; P = .014). No difference according to the type of prosthesis exchange procedure (1, 2, or 3 stages) was found. Use of local antifungal therapy (eg, antifungal drug-impregnated spacers) had no significant impact on outcome.
Conclusions
This study confirms the importance of complete hardware removal in CPJI. Most importantly, it provides evidence supporting the use of amphotericin B as initial antifungal therapy.
Keywords: arthroplasty, candidiasis, hip, knee, prosthesis
This systematic review and meta-analysis of Candida prosthetic joint infections highlights the importance of complete prosthesis removal and first-line amphotericin B treatment for therapeutic success.
Prosthetic joint infections (PJI) are serious complications of total hip or knee arthroplasties, with an estimated incidence of 1%–2% [1–3]. Although bacteria (eg, Staphylococci) account for most PJI, fungi are encountered in about 1%–3% of cases (mainly in polymicrobial infections) with Candida spp. being the most frequent fungal pathogen (80%–90% of cases) [4–6]. These Candida PJI (CPJI) often affect patients with significant comorbidities, previous revisions of arthroplasty, and/or prolonged antibiotic exposure [4, 7, 8]. Moreover, the biofilm-forming capacity of Candida spp. and their limited therapeutic options make that these infections are difficult to treat, often requiring multiple-step exchange procedures and prolonged antifungal therapy [5, 7, 9, 10].
Given the rarity and heterogeneity of CPJI, robust evidence to support standardized approaches in the management of CPJI are lacking and guidelines only provide nonspecific or “experts opinion” recommendations [11–14]. While complete prosthesis removal is strongly recommended, the optimal surgical approach (eg, 1-stage, 2-stage, or 3-stage exchange procedure) remains debated. Similarly, the choice and duration of antifungal therapy are usually graded with a low quality of evidence.
Previous systematic reviews have described the epidemiology of CPJI without detailed analysis of outcome predictors [5, 7, 9]. A retrospective multicenter study of 269 CPJI has been recently published [15]. This study identified some parameters associated with success or failure, in particular regarding the surgical strategy, but did not find any association with the type of systematic antifungal therapy (consisting mainly of triazoles or echinocandins), or the use of local antifungal therapy which was rarely employed. Therefore, these results would deserve validation in a distinct dataset.
In the present study, we performed a systematic review of CPJI and analyzed the outcome predictors with a focus on the surgical management strategies and antifungal therapy.
MATERIAL AND METHODS
The study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses checklist (available at: https://www.prisma-statement.org/prisma-2020-checklist). The study protocol was registered in the international Prospective Register of Systematic Reviews under study number CRD42024526388.
Search Strategy
An exhaustive review of the literature was carried out using the PubMed Medline, Embase, and Web of Science databases. The literature review started in January 2024 and was last updated in July 2024. Full articles, as well as conference abstracts, were eligible. The search strategy is detailed in Supplementary material (Supplementary Appendix 1). After excluding duplicate entries from the different databases, a screening of the titles/abstracts was carried out by 2 authors to exclude articles that were not considered as appropriate for the study question (ie, reviews, opinion/guidelines, nonclinical studies, or nonhuman subjects). Articles from the same center or authors were also screened for exclusion of possible duplicate cases. The full texts of the remaining articles were assessed for eligibility according to the inclusion criteria. Any disagreement between the 2 reviewers was resolved by a third reviewer.
Eligibility Criteria
All articles reporting original data from individual human cases of CPJI were eligible, including clinical trials, retrospective and prospective cohort studies, case series, or case reports. There was no restriction related to the date/timing of study or language. Inclusion criteria were: (1) adult patients (≥18 years); (2) diagnosis of CPJI or Candida infection of other orthopedic prosthetic material, including mixed infections (eg, Candida and other microorganisms); (3) microbiological identification of Candida species (including yeast species that were later renamed or reassigned to other genera) by culture or nonculture method (polymerase chain reaction, metagenomics) from prosthetic material or periprosthetic tissue or joint fluid obtained by surgery or another sterile procedure (ie, joint puncture); and (4) presence of individual data regarding the patients underlying conditions, characteristics of infection, treatment, and outcome. PJI attributed to non-Candida yeasts (ie, never classified within the Candida genus according to the new or old taxonomy) were excluded. Publications with pooled data where individual data could be at least partially recovered were included. Pooled data that could not be individualized were considered for epidemiological description but were excluded from univariate and multivariate outcome analyses.
Data Collection
Data were extracted independently by 2 authors. All original data sources and database entries were reviewed by a third author. The collected data included patients characteristics (age, gender, immune status, immunosuppressive treatments, and relevant underlying diseases), infection characteristics (location and type of prosthetic material, timing from initial arthroplasty, type of Candida spp., co-infection with other microorganisms), surgical management of the infection (type of intervention, reinterventions), systemic antifungal therapy (initial and final antifungal drugs), local antifungal therapy (drug-impregnated spacers, intra-articular irrigation/injection of antifungal agents), and final outcome.
Risk of Bias Assessment
Considering that all included studies consisted of case reports or case-series, their quality was assessed using the National Heart, Lung, and Blood Institute quality assessment tool for case series, available at: https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools.
Outcome Analysis
The primary outcome was based on the first intervention performed with curative intent. Success was defined as completion of the initial surgical plan without any recurrence of the infection. Failure was defined as: (1) occurrence of an unexpected subsequent event leading to a change of the initial surgical plan (ie, need of a re-intervention that was not initially planned), (2) persistent signs of infection (ie, swelling, fistula, suffusion) after surgery, or (3) death of any cause occurring within 3 months from the intervention.
Statistical Analyses
Comparisons of proportions and continuous variables were performed using the Fisher's exact test and the Mann–Whitney test, respectively. Analyses were performed using GraphPad Prism 10.0 (GraphPad Software, Boston, MA). A P value < .05 was considered as statistically significant. To explore the associations between the outcome variable and potential predictors, a univariate analysis was conducted. All variables were summarized using the tbl_summary() function from the gtsummary package in R (version 2024.12.0 + 467). For categorical variables, proportions and percentages were reported, while continuous variables were described using means and standard deviations or medians and interquartile ranges, depending on their distribution. The analysis included stratified summaries based on the binary outcome variable. Associations between each independent variable and the outcome were assessed using appropriate statistical tests: chi-squared or Fisher's exact test for categorical variables, and t-tests or Wilcoxon rank-sum tests for continuous variables.
A multivariate analysis was performed to identify independent predictors of the outcome using logistic regression with a binomial link function. The generalized linear model framework was employed, and all significant variables in the univariate analysis were included as candidates in the multivariable model. Multicollinearity among the independent variables was assessed using the variance inflation factor (VIF). The VIF for each predictor was calculated using the vif() function from the car package in R. Predictors with a VIF >5 were considered to exhibit significant multicollinearity, and those with a VIF >10 were deemed to have severe multicollinearity. In cases of significant multicollinearity, nonessential variables were excluded from the multivariate model to improve model stability. Backward stepwise selection based on the Akaike Information Criterion was used to refine the model and identify the most parsimonious set of predictors. Adjusted odds ratios with 95% confidence intervals were reported to quantify the strength of associations. Model diagnostics, including goodness-of-fit tests, were conducted to assess the adequacy of the final model.
RESULTS
Characteristics of Articles
Following the systematic review process detailed in Supplementary material (Supplementary Figure 1), a total of 110 articles corresponding to 385 patients were included in the systematic review (Supplementary references). The selected articles consisted of 76 single case reports and 34 case series (ie, at least 2 cases; median 6; range, 2–29). These publications, dated from 1979 to 2024, originated from Europe (n = 44, 40%), North America (n = 34, 30.9%), South America (n = 1, 0.9%), and Asia (n = 31, 28.2%). A detailed list of the selected articles, as well as their quality assessment score, is provided in Supplementary material (Supplementary Table 1 and Supplementary Table 2, respectively).
Characteristics of Patients
The characteristics of patients and infections are described in Table 1. The median age was 68 years (range, 28–93), with a predominance of women (60.4%). Data on underlying conditions were present in 269 (69.9%) patients and a minority of them had an immunosuppressive condition (14.1%). Diabetes mellitus was reported in 27.1% of cases and use of intravenous drug in 1.9%. Most CPJI involved a knee (54.3%) or a hip (40.4%) prosthesis. The remaining cases (5.3%) included shoulder, elbow or ankle prostheses, megaprostheses (femur and/or tibia), and osteosynthetic material. Most infections (94.3%) were delayed (ie, occurring >4 weeks after prosthesis implantation). C albicans was the predominant pathogenic species (47.2%), followed by C parapsilosis (28.6%), C glabrata (8.1%), and C tropicalis (5.5%). Among rare Candida spp., C pelliculosa was the most frequent (2.4% of all cases). A previous or concomitant episode of candidemia was mentioned in only 7 cases. Polymicrobial infections (ie, Candida spp. and other microorganisms) were observed in 33% cases and this proportion was significantly higher in hip than in knee CPJI (39.5% vs 27.0%, P = .02). Staphylococcus aureus, coagulase-negative Staphylococci, other Gram-positive bacteria and Gram-negative bacteria were involved in 32.3%, 40.9%, 27.6%, and 33.9% of these polymicrobial infections, respectively.
Table 1.
Characteristics of Candida Prosthetic Joint Infections
| Patient Characteristics | |
| Age (y) (N = 348) | 68 (28–93) |
| Gender (female) (N = 333) | 201 (60.4) |
| Immunosuppressive condition (N = 269) | 38 (14.1)| |
| Diabetes mellitus (N = 269) | 73 (27.1) |
| Intravenous drug use (N = 269) | 5 (1.9) |
| Site of infection (N = 376)a | |
| Knee prosthesis | 204 (54.3) |
| Hip prosthesis | 152 (40.4) |
| Megaprosthesisb | 10 (2.7) |
| Shoulder prosthesis | 5 (1.3) |
| Elbow prosthesis | 1 (0.3) |
| Ankle prosthesis | 1 (0.3) |
| Osteosynthesis materialc | 3 (0.8) |
| Timing from initial implantation (N = 350) | |
| Early (<4 wk from implantation) | 20 (5.7) |
| Delayed (>4 wk from implantation) | 330 (94.3) |
| Type of infection (N = 385) | |
| Monomicrobial (only Candida spp.) | 258 (67.0) |
| Polymicrobial (Candida and other microbes) | 127 (33.0) |
| Pathogenic Candida spp. (N = 381)d | |
| C albicans | 180 (47.2) |
| C parapsilosis | 109 (28.6) |
| C glabrata | 31 (8.1) |
| C tropicalis | 21 (5.5) |
| Other Candida spp.e | 35 (9.2) |
| Mixed infectionf | 5 (1.3) |
Results are expressed as median (range) for continuous variables or number of cases (percentage) for proportions. N = total number of cases with available data.
aSite of infection could not be determined in 1 publication (N = 9 cases) because of pooled data.
bMegaprostheses involving total or partial femoral replacement ± proximal tibial replacement.
cSites of osteosynthesis material: hand, rachis, ankle.
d Candida species could not be determined in 1 publication (N = 4 cases) because of pooled data.
e C pelliculosa (n = 9), C lusitaniae (n = 5), C famata (n = 5), C guilliermondii (n = 4), C dubliniensis (n = 2), C freyschussii (n = 2), C boidinii (n = 1), C lipolytica (n = 1), C pseudotropicalis (n = 1), C utilis (n = 1), nonspecified Candida spp. (n = 4).
f C albicans and C parapsilosis (n = 2), C albicans and C glabrata (n = 2), C parapsilosis and C tropicalis (n = 1).
Management of CPJI
Data on surgical management and antifungal therapy are provided in Table 2. More than half of the patients (54.6%) were treated by a 2-stage exchange procedure. Other approaches included resection arthroplasty (12.8%), debridement with prosthesis retention (11.0%), 1-stage exchange procedure (10.7%), 3-stage exchange procedure (8.6%), and conservative treatment (2.3%).
Table 2.
Management of Candida Prosthetic Joint Infections
| Surgical Approach (N = 383) | |
| Conservative treatment | 9 (2.3) |
| Resection arthroplasty | 49 (12.8) |
| Debridement with prosthesis retention | 42 (11.0) |
| One-stage exchange procedure | 41 (10.7) |
| Two-stage exchange procedure | 209 (54.6) |
| Three-stage exchange procedure | 33 (8.6) |
| Initial systemic antifungal therapy (N = 353)a | |
| No antifungal therapy | 5 (1.4) |
| Fluconazole | 187 (53.0) |
| Other triazole | 19 (5.4) |
| Echinocandin | 59 (16.7) |
| Amphotericin B formulation | 46 (13.0) |
| Combined therapyb | 37 (10.5) |
| Final systemic antifungal therapy (N = 363) | |
| No antifungal therapy | 5 (1.4) |
| Fluconazole | 248 (68.3) |
| Other triazole | 30 (8.3) |
| Echinocandin | 42 (11.6) |
| Amphotericin B formulation | 17 (4.7) |
| Combined therapy | 21 (5.8) |
| Local antifungal agents (N = 383) | |
| None | 269 (70.2) |
| Amphotericin B-impregnated spacer | 64 (16.7) |
| Triazole-impregnated spacerc | 33 (8.6) |
| Intra-articular irrigation of amphotericin B | 5 (1.3) |
| Intra-articular irrigation of triazole3 | 12 (3.1) |
Results are expressed as number of cases (percentage) for proportions. N = total number of cases with available data.
aFirst antifungal drug administered for a duration of at least 5 d.
bAmphotericin B + flucytosine (n = 16), echinocandin + triazole (n = 10), amphotericin B + triazole (n = 6), triazole + flucytosine (n = 3), echinocandin + flucytosine (n = 2).
cFluconazole or voriconazole.
Complete data on antifungal therapy were available for 353 cases. No antifungal treatment was administered in 5 cases (1.4%). Fluconazole was the most frequent first-line agent (53%), followed by echinocandins (16.7%), amphotericin B formulations (13%), and other triazoles (5.4%). Combination first-line therapy was employed in 10.5% of cases. The last therapeutic line was fluconazole in most cases (68.3%). Local antifungal therapy was used in 29.8% of cases, primarily involving antifungal drug-impregnated spacers (amphotericin B or triazole).
Outcome of CPJI
Complete follow-up data were available for 376 (97.7%) cases. Success was achieved in 257 (68.4%) cases. Results of univariate and multivariate analyses of factors associated with outcome are presented in Table 3. In univariate analysis, female gender, knee prosthesis infection, C parapsilosis, rare Candida spp. (ie, other than C albicans, C parapsilosis, C glabrata, or C tropicalis), and initial antifungal therapy with an amphotericin B formulation were significantly associated with success. In contrast, hip prosthesis infection, co-infection with S aureus or Gram-negative bacteria, debridement with prosthesis retention, initial antifungal therapy with a triazole, and final antifungal therapy with an echinocandin were significantly associated with failure. In multivariate analysis, initial amphotericin B therapy was significantly associated with success, whereas co-infection with S aureus and debridement with prosthesis retention were significantly associated with failure. Overall, patients receiving an initial amphotericin B–containing regimen (either as monotherapy or in combination therapy) had a higher success rate compared to those receiving any other first-line treatment (83.3% vs 68.6%, P = .02).
Table 3.
Results of Univariate and Multivariate Analyses of Factors Associated With Success
| Characteristics | Failure (N = 119) | Success (N = 257) | P value | Multivariate Analysis OR (95% CI), P value |
|---|---|---|---|---|
| Underlying conditions | ||||
| Age >65 y | 65/99 (65.7) | 151/240 (62.9) | .6 | … |
| Gender female | 50/98 (51.0) | 145/226 (64.2) | .026 | … |
| Immunosuppressive disease | 10/66 (15.2) | 16/168 (9.5) | .2 | … |
| Diabetes mellitus | 21/66 (31.8) | 40/168 (23.8) | .2 | … |
| Characteristics of infection | ||||
| Knee | 43/116 (37.1) | 156/251 (62.2) | <.001 | 1.66 (0.91–3.04), .1 |
| Hip | 63/115 (54.8) | 85/251 (33.9) | <.001 | … |
| Other joint | 10/116 (8.6) | 10/251 (4.0) | .069 | … |
| Early infection | 6/100 (6.0) | 13/241 (5.4) | .8 | … |
| Delayed infection | 94/100 (94.0) | 228/241 (94.6) | .8 | … |
| C albicans | 55/102 (53.9) | 107/244 (43.9) | .087 | … |
| C parapsilosis | 23/102 (22.5) | 83/244 (34.0) | .035 | 1.83 (0.94–3.64), .074 |
| C glabrata | 12/102 (11.8) | 17/244 (7.0) | .14 | … |
| C tropicalis | 10/102 (9.8) | 12/244 (4.9) | .089 | … |
| Other Candida spp. | 4/102 (3.9) | 28/244 (11.5) | .027 | 2.83 (0.87–12.8), .087 |
| Polymicrobial infection | 40/103 (38.8) | 69/247 (27.9) | .045 | … |
| S aureus co-infection | 16/101 (15.8) | 19/244 (7.8) | .024 | 0.4 (0.18–0.92), .032 |
| CNS co-infection | 13/101 (12.9) | 30/244 (12.3) | .9 | … |
| Other gram-positive co-infection | 13/101 (12.9) | 21/244 (8.6) | .2 | … |
| Gram-negative co-infection | 16/101 (15.8) | 19/101 (7.8) | .024 | 0.48 (0.21–1.12), .088 |
| Surgical management | ||||
| Conservative approach | 2/119 (1.7) | 7/257 (2.7) | .7 | … |
| Resection arthroplasty | 12/119 (10.1) | 35/257 (13.6) | .3 | … |
| Debridement/retention | 23/119 (19.3) | 18/257 (7.0) | <.001 | 0.25 (0.11–0.55), <.001 |
| One-stage exchange | 10/119 (8.4) | 30/257 (11.7) | .3 | … |
| Two-stage exchange | 65/119 (54.6) | 141/257 (54.9) | >.9 | … |
| Three-stage exchange | 7/119 (5.9) | 26/257 (10.1) | .2 | … |
| Local antifungal agents | 35/119 (29.4) | 79/257 (30.7) | .8 | … |
| First-line antifungal therapy | ||||
| No antifungal therapy | 2/91 (2.2) | 3/230 (1.3) | .6 | … |
| Triazole | 64/91 (70.3) | 134/230 (58.3) | .045 | … |
| Echinocandin | 8/91 (8.8) | 29/230 (12.6) | .3 | … |
| Amphotericin B formulation | 5/91 (5.5) | 39/230 (17.0) | .007 | 3.18 (1.25–9.87), .014 |
| Combination therapy | 12/91 (13.2) | 25/230 (10.9) | .6 | … |
| Last-line (maintenance) antifungal therapy | ||||
| Triazole | 72/96 (75.0) | 196/235 (83.4) | .077 | … |
| Echinocandin | 11/96 (11.5) | 11/235 (4.7) | .025 | 0.43 (0.14–1.28), .13 |
| Amphotericin B formulation | 2/96 (2.1) | 13/235 (5.5) | .2 | … |
| Combination therapy | 9/96 (9.4) | 12/235 (5.1) | .15 | … |
Results are expressed as number of cases with condition/total number with available data (percentage).
Abbreviations: CI, 95% confidence interval; CNS, coagulase-negative Staphylococci; OR, odds ratio.
DISCUSSION
This study including 385 cases of CPJI represents the largest literature review on this topic. We identified 3 previous systematic reviews of CPJI (or more largely fungal PJI) published between 2017 and 2023, including 72, 225, and 350 cases of CPJI, respectively [5, 7, 9]. These epidemiological analyses provided only limited data about factors associated with success rates. A Spanish multicenter study retrospectively collected data about 43 CPJI between 2003 and 2015 in 16 centers [16]. A more recent multicenter cohort study reported data on 269 CPJIs from 2010 to 2021 in 7 European countries [15]. Compared to this later study, our work represents a larger dataset but suffers from the heterogeneity in the quality/completeness of data sources. The inclusion of case reports may have introduced a bias but resulted in a lower rate of mixed infections (33% in our study vs 51% in the cohort study of Dinh et al) [15].
Our results indicate that CPJI mainly occurred in patients without preexisting immunosuppressive conditions. Documentation of previous or concomitant candidemia was rare, possibly due to their underreporting in case-series. Among the causal Candida spp., we found a majority of C albicans and C parapsilosis. Although these species are common skin commensals, their respective proportions may also depend on local epidemiology. Surprisingly, C pelliculosa, a rare pathogenic species, accounted for 2.3% of cases (ie, 9 cases from 4 publications, all from Asia), ranking as the fifth cause of CPJI.
Most cases of CPJI were treated with prosthesis removal, and 2-stage exchange procedure was the preferred approach (>50% cases), as recommended by current guidelines [11, 13, 14]. Fluconazole was predominantly used as first-line and suppressive therapy, which is also in line with current recommendations [11, 14]. Local antifungal therapy, including antifungal drug-impregnated spacers or intra-articular injection/irrigation, was applied in about 30% cases, reflecting heterogenous in-house practices. We observed a relatively high success rate (ie, 73.7% vs 58% in the cohort of Dinh et al) [15], which may be overestimated due to the inclusion of case reports.
Our multivariate analysis of outcome determinants found some similar and distinct results compared to the cohort of Dinh et al [15]. The strong association of the debridement and implant retention approach with failure was demonstrated in both studies. We also failed to demonstrate the superiority of the 2-stage over the 1-stage exchange procedure. Our analysis included some cases with a 3-stage procedure, half of them originating from a single study [17]. While this approach is recommended by some experts [18], our result could not demonstrate its benefit but was statistically underpowered.
A unique and important finding of our study was the significant association of first-line amphotericin B therapy with success. Interestingly, Dinh et al did not observe this association, possibly due to the low number of amphotericin-B treated patients in their cohort [15]. Amphotericin B displays a better in vitro anti-biofilm activity compared to other antifungals, particularly triazoles [19, 20], which is consistent with our results. Indeed, we also observed an association between initial azole therapy and failure in univariate analyses. While echinocandins also display acceptable anti-biofilm activity, we did not find any benefit of its first-line use in CPJI. Similar observations were reported by Dinh et al [15]. Although C parapsilosis (less susceptible to echinocandins) was the cause of 29% of CPJI, only 9 of these cases received initial echinocandin therapy. The limited penetration of echinocandins in bone tissue might be a possible explanation for this result [21, 22]. Overall, in vivo drug efficacy does not necessarily correlate with in vitro data and may result from a conjunction of multiple parameters (eg, pharmacokinetic profile, planktonic and anti-biofilm activity, tissue penetration, protein binding, toxicity), which can impact outcome. The positive impact of first-line amphotericin B therapy should be confirmed in further retrospective or prospective studies because of the relatively low number of cases in our dataset. Of note, a majority of these cases (about 40%) originated from one study [23]. However, an extended analysis to all patients receiving a first-line amphotericin B containing regimen (as monotherapy or combined therapy) remained significant. The optimal duration of first-line amphotericin B therapy should also be investigated. We did not analyze the impact of the duration of antifungal therapy because this information was lacking in most case series and it is highly dependent on the type of surgical procedure. Overall, we observed a highly variable duration of antifungal therapy in case reports.
Regarding the use of local antifungal therapy, guidelines and experts opinions refrain from any recommendation because of the low level of evidence and lack of standardized approaches [11, 12, 14, 24]. Previous studies could not demonstrate the benefit of this approach but were limited by the low number of cases [9, 15]. Our analysis including a large dataset (114 cases with local antifungal therapy) did not find any impact on outcomes. The actual efficacy of antifungal drugs at the infection site may be influenced by multiple parameters, such as the type of antifungal agent and its dosage, the composition and porosity of the cement, as well as the drug/cement ratio [24]. Homogenous mixing of the antifungal drug powder with the cement may be difficult to achieve and mechanical properties of the cement may be affected by such procedures [25]. Optimized and standardized protocols would be needed to assess the potential benefit of this practice.
Co-infections with bacteria are frequent in CPJI. Not surprisingly, we found that S aureus co-infection was an independent predictor of failure in multivariate analysis, which was not the case with coagulase-negative Staphylococci.
Some other parameters displayed a significant association with outcome in univariate analyses and trends in the multivariate analysis. As reported earlier [15], C parapsilosis infections were associated with better outcomes compared to other CPJIs. Infections caused by rare Candida spp. demonstrated comparable trends. Hip CPJI, which were more frequently associated with polymicrobial infections, were associated with worse outcomes compared to knee CPJIs. Surprisingly, last-line echinocandin therapy was associated with worse outcome in univariate analysis. However, this observation may rather be a consequence than a cause. Indeed, patients with unfavorable outcomes would be more prone to receive prolonged intravenous therapy with a broad-spectrum antifungal drug, such as echinocandins, whereas those with favorable outcomes could be transitioned to maintenance therapy with an oral azole drug.
In conclusion, CPJI are rare but difficult-to-treat infections, for which current therapeutic approaches rely on low level of evidence and lack standardization. Despite some inherent limitations, such as the possible bias of individual case reports, the lack of data about predisposing conditions, antifungal susceptibility testing, duration of antifungal therapy, and the heterogeneity of therapeutic approaches, the present study represents the largest and most recent systematic review on this topic and complements the results of a recent multicenter cohort study on the same topic [15]. Taken together, these studies indicate that a complete removal of the infected prosthesis is crucial for optimal surgical management of CPJIs. However, the benefit of a multiple-stage versus 1-stage exchange procedure remains inconclusive. There are currently no clear recommendations regarding the type of antifungal therapy in CPJI [11, 12, 14]. Our results support the use of initial therapy with an amphotericin B lipid formulation as the first choice, whereas triazoles (mainly fluconazole) should be reserved for second-line or suppressive therapy. The role of echinocandins, also displaying acceptable anti-biofilm properties, should be further investigated. Despite analyzing a large dataset, our study could not demonstrate a benefit of local antifungal therapy, possibly due to the heterogeneity of such practices.
Supplementary Material
Notes
Author contributions. C.G.: study design, screening and selection process of publications, assessment of quality scores, data collection, statistical analyses, first draft of the manuscript. C.M.: study design, screening and selection process of publications, data collection, statistical analyses, review/editing of the manuscript. B.G.: statistical methodology, statistical analyses, review of results analyses, review/editing of the manuscript. S.S.: consultant for classification of surgical procedures and outcome assessment, review of results analyses, review/editing of the manuscript. T.F.: review of results analyses, review/editing of the manuscript. F.L.: study design, resolution of disagreements in selection process, data collection (review of all sources and data reporting), review of results analyses, first draft of the manuscript and finalization of the manuscript.
Financial support. No external funding was received for this study.
Contributor Information
Charles Gibert, Institute of Microbiology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland; Institut des Sciences Pharmaceutiques et Biologiques, Université Claude Bernard Lyon 1, Villeurbanne, France.
Camille Marchetti, Institute of Microbiology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland; Infectious Diseases Service, Department of Medicine, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland.
Benoît Guery, Infectious Diseases Service, Department of Medicine, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland.
Sylvain Steinmetz, Department of Orthopedics and Traumatology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland.
Tristan Ferry, Institut des Sciences Pharmaceutiques et Biologiques, Université Claude Bernard Lyon 1, Villeurbanne, France; Infectious and Tropical Diseases Service, Croix-Rousse Hospital, Hospices Civils de Lyon, Lyon, France; Centre de référence des IOA complexes de Lyon, CRIOAc Lyon, Lyon, France.
Frederic Lamoth, Institute of Microbiology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland; Infectious Diseases Service, Department of Medicine, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland.
Supplementary Data
Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
References
- 1. Marang-van de Mheen PJ, Bragan Turner E, Liew S, et al. Variation in prosthetic joint infection and treatment strategies during 4.5 years of follow-up after primary joint arthroplasty using administrative data of 41397 patients across Australian, European and United States hospitals. BMC Musculoskelet Disord 2017; 18:207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Springer BD, Cahue S, Etkin CD, Lewallen DG, McGrory BJ. Infection burden in total hip and knee arthroplasties: an international registry-based perspective. Arthroplast Today 2017; 3:137–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Tabaja H, Abu Saleh A, Osmon DR. Periprosthetic joint infection: what's new? Infect Dis Clin North Am 2024; 38:731–56. [DOI] [PubMed] [Google Scholar]
- 4. Gross CE, Della Valle CJ, Rex JC, Traven SA, Durante EC. Fungal periprosthetic joint infection: a review of demographics and management. J Arthroplasty 2021; 36:1758–64. [DOI] [PubMed] [Google Scholar]
- 5. Sambri A, Zunarelli R, Fiore M, et al. Epidemiology of fungal periprosthetic joint infection: a systematic review of the literature. Microorganisms 2022; 11:84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Tai DBG, Patel R, Abdel MP, Berbari EF, Tande AJ. Microbiology of hip and knee periprosthetic joint infections: a database study. Clin Microbiol Infect 2022; 28:255–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Cobo F, Rodriguez-Granger J, Lopez EM, et al. Candida-induced prosthetic joint infection. A literature review including 72 cases and a case report. Infect Dis (Lond) 2017; 49:81–94. [DOI] [PubMed] [Google Scholar]
- 8. Riaz T, Tande AJ, Steed LL, et al. Risk factors for fungal prosthetic joint infection. J Bone Jt Infect 2020; 5:76–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Gonzalez MR, Bedi ADS, Karczewski D, Lozano-Calderon SA. Treatment and outcomes of fungal prosthetic joint infections: a systematic review of 225 cases. J Arthroplasty 2023; 38:2464–71.e1. [DOI] [PubMed] [Google Scholar]
- 10. Karczewski D, Ren Y, Andronic O, et al. Candida periprosthetic joint infections—risk factors and outcome between albicans and non-albicans strains. Int Orthop 2022; 46:449–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Cornely OA, Sprute R, Bassetti M, et al. Global guideline for the diagnosis and management of candidiasis: an initiative of the ECMM in cooperation with ISHAM and ASM. Lancet Infect Dis 2025; 25:e280–93. [DOI] [PubMed] [Google Scholar]
- 12. Gamaletsou MN, Rammaert B, Brause B, et al. Osteoarticular mycoses. Clin Microbiol Rev 2022; 35:e0008619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Osmon DR, Berbari EF, Berendt AR, et al. Diagnosis and management of prosthetic joint infection: clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect Dis 2013; 56:e1–25. [DOI] [PubMed] [Google Scholar]
- 14. Pappas PG, Kauffman CA, Andes DR, et al. Clinical practice guideline for the management of candidiasis: 2016 update by the Infectious Diseases Society of America. Clin Infect Dis 2016; 62:e1–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Dinh A, McNally M, D’Anglejan E, et al. Prosthetic joint infections due to Candida species: a multicenter international study. Clin Infect Dis 2024; 80:347–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Escola-Verge L, Rodriguez-Pardo D, Lora-Tamayo J, et al. Candida periprosthetic joint infection: a rare and difficult-to-treat infection. J Infect 2018; 77:151–7. [DOI] [PubMed] [Google Scholar]
- 17. Baecker H, Frieler S, Gessmann J, Pauly S, Schildhauer TA, Hanusrichter Y. Three-stage revision arthroplasty for the treatment of fungal periprosthetic joint infection: outcome analysis of a novel treatment algorithm: a prospective study. Bone Jt Open 2021; 2:671–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Izakovicova P, Borens O, Trampuz A. Periprosthetic joint infection: current concepts and outlook. EFORT Open Rev 2019; 4:482–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Kraft L, Ribeiro VST, Goncalves GA, Suss PH, Tuon FF. Comparison of amphotericin B lipid complex, deoxycholate amphotericin B, fluconazole, and anidulafungin activity against Candida albicans biofilm isolated from breakthrough candidemia. Enferm Infecc Microbiol Clin (Engl Ed) 2023; 41:596–603. [DOI] [PubMed] [Google Scholar]
- 20. Maiolo EM, Furustrand Tafin U, Borens O, Trampuz A. Activities of fluconazole, caspofungin, anidulafungin, and amphotericin B on planktonic and biofilm Candida species determined by microcalorimetry. Antimicrob Agents Chemother 2014; 58:2709–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Cancela Costa A, Kritikos A, Glampedakis E, et al. Antifungal drug penetration in soft tissue abscesses: a comparative analysis. J Antimicrob Chemother 2024; 79:1668–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Felton T, Troke PF, Hope WW. Tissue penetration of antifungal agents. Clin Microbiol Rev 2014; 27:68–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Hwang B-H, Yoon J-Y, Nam C-H, et al. Fungal peri-prosthetic joint infection after primary total knee replacement. J Bone Joint Surg Br 2012; 94:656–9. [DOI] [PubMed] [Google Scholar]
- 24. Anagnostakos K, Becker SL, Sahan I. Antifungal-loaded acrylic bone cement in the treatment of periprosthetic hip and knee joint infections: a review. Antibiotics (Basel) 2022; 11:879. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Krampitz B, Steiner J, Trampuz A, Kuhn K-D. Voriconazole admixed with PMMA-impact on mechanical properties and efficacy. Antibiotics (Basel) 2023; 12:848. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.